Practice Guidelines for Central Venous Access a Report by the American Society of Anesthesiologists Task Force on Central Venous Access

Home , Giacomini

Practice Guidelines for Central Venous Access A Report by the American Society of Anesthesiologists Task Force on Central Venous Access

RACTICE Guidelines are systematically developed rec- • What other guideline statements are available on this topic? ommendations that assist the practitioner and patient P X Several major organizations have produced practice guide- in making decisions about health care. These recommenda- lines on central venous access128–132 tions may be adopted, modified, or rejected according to • Why was this Guideline developed? clinical needs and constraints, and are not intended to re- X The ASA has created this new Practice Guideline to provide place local institutional policies. In addition, Practice Guide- updated recommendations on some issues and new recommendations on issues that have not been previously ad- lines developed by the American Society of Anesthesiologists dressed by other guidelines. This was based on a rigorous (ASA) are not intended as standards or absolute require- evaluation of recent scientific literature as well as findings ments, and their use cannot guarantee any specific outcome. from surveys of expert consultants and randomly selected Practice Guidelines are subject to revision as warranted by ASA members • the evolution of medical knowledge, technology, and prac- How does this statement differ from existing guidelines? X The ASA Guidelines differ in areas such as insertion site tice. They provide basic recommendations that are sup- selection (e.g., upper body site) guidance for catheter place- ported by a synthesis and analysis of the current literature, ment (e.g., use of real-time ultrasound) and verification of expert and practitioner opinion, open forum commentary, venous location of the catheter and clinical feasibility data. • Why does this statement differ from existing guidelines? X The ASA Guidelines differ from existing guidelines because it addresses the use of bundled techniques, use of an as- Methodology sistant during catheter placement, and management of ar- A. Definition of Central Venous Access terial injury For these Guidelines, central venous access is defined as placement of a catheter such that the catheter is inserted into internal jugular veins, subclavian veins, iliac veins, and com- a venous great vessel. The venous great vessels include the mon femoral veins.* Excluded are catheters that terminate in superior vena cava, inferior vena cava, brachiocephalic veins, a systemic artery.

Developed by the American Society of Anesthesiologists Task B. Purposes of the Guidelines Force on Central Venous Access: Stephen M. Rupp, M.D., Seattle, The purposes of these Guidelines are to (1) provide guid- Washington (Chair); Jeffrey L. Apfelbaum, M.D., Chicago, Illinois; ance regarding placement and management of central ve- Casey Blitt, M.D., Tucson, Arizona; Robert A. Caplan, M.D., Seattle, Washington; Richard T. Connis, Ph.D., Woodinville, Washington; nous catheters, (2) reduce infectious, mechanical, throm- Karen B. Domino, M.D., M.P.H., Seattle, Washington; Lee A. Fleisher, botic, and other adverse outcomes associated with central M.D., Philadelphia, Pennsylvania; Stuart Grant, M.D., Durham, North venous catheterization, and (3) improve management of Carolina; Jonathan B. Mark, M.D., Durham, North Carolina; Jeffrey P. Morray, M.D., Paradise Valley, Arizona; David G. Nickinovich, Ph.D., arterial trauma or injury arising from central venous cath- Bellevue, Washington; and Avery Tung, M.D., Wilmette, Illinois. eterization. Received from the American Society of Anesthesiologists, Park Ridge, Illinois. Submitted for publication October 20, 2011. Accepted C. Focus for publication October 20, 2011. Supported by the American Society of Anesthesiologists and developed under the direction of the Committee These Guidelines apply to patients undergoing elective cen- on Standards and Practice Parameters, Jeffrey L. Apfelbaum, M.D. tral venous access procedures performed by anesthesiologists (Chair). Approved by the ASA House of Delegates on October 19, or health care professionals under the direction/supervision 2011. Endorsed by the Society of Cardiovascular Anesthesiologists, October 4, 2010; the Society of Critical Care Anesthesiologists March 16, of anesthesiologists. The Guidelines do not address (1) clin- 2011; the Society of Pediatric Anesthesia March 29, 2011. A complete ical indications for placement of central venous catheters, (2) list of references used to develop these updated Guidelines, arranged emergency placement of central venous catheters, (3) pa- alphabetically by author, is available as Supplemental Digital Content 1, http://links.lww.com/ALN/A783. tients with peripherally inserted central catheters, (4) place- Address correspondence to the American Society of Anesthesi- ment and residence of a pulmonary artery catheter, (5) inser- ologists: 520 North Northwest Highway, Park Ridge, Illinois 60068- tion of tunneled central lines (e.g., permacaths, portacaths, 2573. These Practice Guidelines, as well as all ASA Practice Param- eters, may be obtained at no cost through the Journal Web site, www.anesthesiology.org. Supplemental digital content is available for this article. Direct * This description of the venous great vessels is consistent with URL citations appear in the printed text and are available in the venous subset for central lines defined by the National Health- both the HTML and PDF versions of this article. Links to the care Safety Network (NHSN). digital files are provided in the HTML text of this article on the Copyright © 2012, the American Society of Anesthesiologists, Inc. Lippincott Journal’s Web site (www.anesthesiology.org). Williams & Wilkins. Anesthesiology 2012; 116:539–73

Anesthesiology, V 116 • No 3 539 March 2012 Practice Guidelines

Hickman®, Quinton®, (6) methods of detection or treat- Scientific Evidence ment of infectious complications associated with central ve- Study findings from published scientific literature were ag- nous catheterization, or (7) diagnosis and management of gregated and are reported in summary form by evidence cat- central venous catheter-associated trauma or injury (e.g., egory, as described in the following paragraphs. All literature pneumothorax or air embolism), with the exception of ca- (e.g., randomized controlled trials, observational studies, case rotid arterial injury. reports) relevant to each topic was considered when evaluating the findings. However, for reporting purposes in this D. Application document, only the highest level of evidence (i.e., level 1, 2, These Guidelines are intended for use by anesthesiologists or 3 within category A, B, or C, as identified in the following and individuals who are under the supervision of an anes- paragraphs) is included in the summary. thesiologist. They also may serve as a resource for other physicians (e.g., surgeons, radiologists), nurses, or health Category A: Supportive Literature care providers who manage patients with central venous Randomized controlled trials report statistically significant catheters. (P Ͻ 0.01) differences between clinical interventions for a specified clinical outcome. E. Task Force Members and Consultants Level 1: The literature contains multiple randomized con- The ASA appointed a Task Force of 12 members, including trolled trials, and aggregated findings are supported anesthesiologists in both private and academic practice from by meta-analysis.‡ various geographic areas of the United States and two con- Level 2: The literature contains multiple randomized consulting methodologists from the ASA Committee on Stan- trolled trials, but the number of studies is insuffi- dards and Practice Parameters. cient to conduct a viable meta-analysis for the pur- The Task Force developed the Guidelines by means of a pose of these Guidelines. seven-step process. First, they reached consensus on the cri- Level 3: The literature contains a single randomized con- teria for evidence. Second, original published research stud- trolled trial. ies from peer-reviewed journals relevant to central venous access were reviewed and evaluated. Third, expert consul- Category B: Suggestive Literature tants were asked to (1) participate in opinion surveys on the Information from observational studies permits inference of effectiveness of various central venous access recommenda- beneficial or harmful relationships among clinical interventions and (2) review and comment on a draft of the Guide- tions and clinical outcomes. lines. Fourth, opinions about the Guideline recommenda- Level 1: The literature contains observational comparisons tions were solicited from a sample of active members of the (e.g., cohort, case-control research designs) of clin- ASA. Opinions on selected topics related to pediatric pa- ical interventions or conditions and indicates statis- tients were solicited from a sample of active members of the tically significant differences between clinical inter- Society for Pediatric Anesthesia (SPA). Fifth, the Task Force ventions for a specified clinical outcome. held open forums at three major national meetings† to solicit Level 2: The literature contains noncomparative observa- input on its draft recommendations. Sixth, the consultants tional studies with associative (e.g., relative risk, were surveyed to assess their opinions on the feasibility of correlation) or descriptive statistics. implementing the Guidelines. Seventh, all available informa- Level 3: The literature contains case reports. tion was used to build consensus within the Task Force to finalize the Guidelines. A summary of recommendations Category C: Equivocal Literature may be found in appendix 1. The literature cannot determine whether there are beneficial or harmful relationships among clinical interventions and F. Availability and Strength of Evidence clinical outcomes. Preparation of these Guidelines followed a rigorous meth- Level 1: Meta-analysis did not find significant differences odologic process. Evidence was obtained from two principal (P Ͼ 0.01) among groups or conditions. sources: scientific evidence and opinion-based evidence. Level 2: The number of studies is insufficient to conduct meta-analysis, and (1) randomized controlled trials † Society for Pediatric Anesthesia Winter Meeting, April 17, 2010, San Antonio, Texas; Society of Cardiovascular Anesthesia 32nd have not found significant differences among Annual Meeting, April 25, 2010, New Orleans, Louisiana, and Inter- groups or conditions or (2) randomized controlled national Anesthesia Research Society Annual Meeting, May 22, 2011, trials report inconsistent findings. Vancouver, British Columbia, Canada. Level 3: Observational studies report inconsistent findings ‡ All meta-analyses are conducted by the ASA methodology group. Meta-analyses from other sources are reviewed but not or do not permit inference of beneficial or harmful included as evidence in this document. relationships.

Anesthesiology 2012; 116:539–73 540 Practice Guidelines SPECIAL ARTICLES

Category D: Insufficient Evidence from Literature Disagree. Median score of 2 (at least 50% of responses are 2 The lack of scientific evidence in the literature is described by or 1 and 2). the following terms: Strongly Disagree. Median score of 1 (at least 50% of responses are 1). Inadequate: The available literature cannot be used to assess relationships among clinical interventions and clinical outcomes. The literature either does not Category C: Informal Opinion meet the criteria for content as defined in the “Fo- Open-forum testimony, Internet-based comments, letters, cus” of the Guidelines or does not permit a clear and editorials are all informally evaluated and discussed dur- interpretation of findings due to methodologic con- ing the development of Guideline recommendations. When cerns (e.g., confounding in study design or imple- warranted, the Task Force may add educational information mentation). or cautionary notes based on this information. Silent: No identified studies address the specified relation- Guidelines ships among interventions and outcomes. I. Resource Preparation Opinion-based Evidence Resource preparation includes (1) assessing the physical environment where central venous catheterization is planned to de- All opinion-based evidence relevant to each topic (e.g., survey data, termine the feasibility of using aseptic techniques, (2) availabil- open-forum testimony, Internet-based comments, letters, editori- ity of a standardized equipment set, (3) use of an assistant for als) is considered in the development of these Guidelines. However, central venous catheterization, and (4) use of a checklist or pro- only the findings obtained from formal surveys are reported. tocol for central venous catheter placement and maintenance. Opinion surveys were developed by the Task Force to address each clinical intervention identified in the docu- The literature is insufficient to specifically evaluate the ment. Identical surveys were distributed to expert consul- effect of the physical environment for aseptic catheter inser- tants and ASA members, and a survey addressing selected tion, availability of a standardized equipment set, or the use pediatric issues was distributed to SPA members. of an assistant on outcomes associated with central venous catheterization (Category D evidence). An observational study reports that the implementation of a trauma intensive care Category A: Expert Opinion unit multidisciplinary checklist is associated with reduced Survey responses from Task Force-appointed expert consultants catheter-related infection rates (Category B2 evidence).1 Ob- are reported in summary form in the text, with a complete servational studies report reduced catheter-related blood- listing of consultant survey responses reported in appendix 5. stream infection rates when intensive care unit-wide bundled protocols are implemented (Category B2 evidence).2–7 These Category B: Membership Opinion studies do not permit the assessment of the effect of any Survey responses from active ASA and SPA members are re- single component of a checklist or bundled protocol on out- ported in summary form in the text, with a complete listing of come. The Task Force notes that the use of checklists in other ASA and SPA member survey responses reported in appendix 5. specialties or professions has been effective in reducing the error rate for a complex series of activities.8,9 Survey responses are recorded using a 5-point scale and summarized based on median values.§ The consultants and ASA members strongly agree that central venous catheterization should be performed in a location Strongly Agree. Median score of 5 (at least 50% of the that permits the use of aseptic techniques. The consultants and responses are 5). ASA members strongly agree that a standardized equipment set Agree. Median score of 4 (at least 50% of the responses are should be available for central venous access. The consultants 4 or 4 and 5). and ASA members agree that a trained assistant should be used Equivocal. Median score of 3 (at least 50% of the responses during the placement of a central venous catheter. The ASA are 3, or no other response category or com- members agree and the consultants strongly agree that a check- bination of similar categories contain at list or protocol should be used for the placement and mainte- least 50% of the responses). nance of central venous catheters.

§ When an equal number of categorically distinct responses are Recommendations for Resource Preparation. Central ve- obtained, the median value is determined by calculating the arith- nous catheterization should be performed in an environ- metic mean of the two middle values. Ties are calculated by a ment that permits use of aseptic techniques. A standard- predetermined formula. ized equipment set should be available for central venous Refer to appendix 2 for an example of a list of standardized equipment for adult patients. access. A checklist or protocol should be used for place- # Refer to appendix 3 for an example of a checklist or protocol. ment and maintenance of central venous catheters.# An ** Refer to appendix 4 for an example of a list of duties per- assistant should be used during placement of a central formed by an assistant. venous catheter.**

Anesthesiology 2012; 116:539–73 541 Practice Guidelines Practice Guidelines

II. Prevention of Infectious Complications 100%), caps (100% and 94.7%), and masks covering both Interventions intended to prevent infectious complica- the mouth and nose (100% and 98.1%). tions associated with central venous access include, but are not limited to (1) intravenous antibiotic prophylaxis, (2) Selection of Antiseptic Solution aseptic techniques (i.e., practitioner aseptic preparation and patient skin preparation), (3) selection of coated or Chlorhexidine solutions: A randomized controlled trial com- impregnated catheters, (4) selection of catheter insertion paring chlorhexidine (2% aqueous solution without alcohol) site, (5) catheter fixation method, (6) insertion site dress- with 10% povidone iodine (without alcohol) for skin prep- ings, (7) catheter maintenance procedures, and (8) aseptic aration reports equivocal findings regarding catheter colonization (P ϭ 0.013) and catheter-related bacteremia (P ϭ techniques using an existing central venous catheter for 13 injection or aspiration. 0.28) (Category C2 evidence). The literature is insufficient to evaluate chlorhexidine with alcohol compared with povi- Intravenous Antibiotic Prophylaxis. Randomized con- done-iodine with alcohol (Category D evidence). The litera- trolled trials indicate that catheter-related infections and ture is insufficient to evaluate the safety of antiseptic solu- sepsis are reduced when prophylactic intravenous antibi- tions containing chlorhexidine in neonates, infants and otics are administered to high-risk immunosuppressed children (Category D evidence). cancer patients or neonates. (Category A2 evidence).10,11 Solutions containing alcohol: Comparative studies are in- The literature is insufficient to evaluate outcomes associ- sufficient to evaluate the efficacy of chlorhexidine with alco- ated with the routine use of intravenous antibiotics (Cat- hol in comparison with chlorhexidine without alcohol for egory D evidence). skin preparation during central venous catheterization (Cat- The consultants and ASA members agree that intrave- egory D evidence). A randomized controlled trial of povidone- nous antibiotic prophylaxis may be administered on a iodine with alcohol indicates that catheter tip colonization is case-by-case basis for immunocompromised patients or reduced when compared with povidone-iodine alone (Cate- high-risk neonates. The consultants and ASA members gory A3 evidence); equivocal findings are reported for cathe- agree that intravenous antibiotic prophylaxis should not ter-related infection (P ϭ 0.04) and clinical signs of infection be administered routinely. (P ϭ 0.09) (Category C2 evidence).14 Recommendations for Intravenous Antibiotic Prophylaxis. The consultants and ASA members strongly agree that For immunocompromised patients and high-risk neonates, chlorhexidine with alcohol should be used for skin prep- administer intravenous antibiotic prophylaxis on a case-by- aration. SPA members are equivocal regarding whether case basis. Intravenous antibiotic prophylaxis should not be chlorhexidine-containing solutions should be used for administered routinely. skin preparation in neonates (younger than 44 gestational weeks); they agree with the use of chlorhexidine in infants (younger than 2 yr) and strongly agree with its use in Aseptic Preparation and Selection of Antiseptic Solution children (2–16 yr). Aseptic preparation of practitioner, staff, and patients: A randomized controlled trial comparing maximal barrier precautions (i.e., mask, cap, gloves, gown, large full-body drape) Recommendations for Aseptic Preparation and Selection with a control group (i.e., gloves and small drape) reported of Antiseptic Solution equivocal findings for reduced colonization (P ϭ 0.03) and In preparation for the placement of central venous catheters, catheter-related septicemia (P ϭ 0.06) (Category C2 evi- use aseptic techniques (e.g., hand washing) and maximal bar- 12 dence). The literature is insufficient to evaluate the efficacy rier precautions (e.g., sterile gowns, sterile gloves, caps, masks of specific aseptic activities (e.g., hand washing) or barrier covering both mouth and nose, and full-body patient precautions (e.g., sterile full-body drapes, sterile gown, drapes). A chlorhexidine-containing solution should be used gloves, mask, cap) (Category D evidence). Observational stud- for skin preparation in adults, infants, and children; for ne- ies report hand washing, sterile full-body drapes, sterile onates, the use of a chlorhexidine-containing solution for gloves, caps, and masks as elements of care “bundles” that skin preparation should be based on clinical judgment and result in reduced catheter-related bloodstream infections institutional protocol. If there is a contraindication to chlo- (Category B2 evidence).2–7 However, the degree to which each rhexidine, povidone-iodine or alcohol may be used. Unless particular element contributed to improved outcomes could contraindicated, skin preparation solutions should contain not be determined. alcohol. Most consultants and ASA members indicated that the Catheters Containing Antimicrobial Agents. Meta-analysis following aseptic techniques should be used in preparation of randomized controlled trials15–19 comparing antibiotic- for the placement of central venous catheters: hand washing coated with uncoated catheters indicates that antibiotic- (100% and 96%); sterile full-body drapes (87.3% and coated catheters reduce catheter colonization (Category A1 73.8%); sterile gowns (100% and 87.8%), gloves (100% and evidence). Meta-analysis of randomized controlled trials20–24

Anesthesiology 2012; 116:539–73 542 Practice Guidelines SPECIAL ARTICLES

comparing silver-impregnated catheters with uncoated cath- Recommendations for Selection of Catheter Insertion Site. eters report equivocal findings for catheter-related blood- Catheter insertion site selection should be based on clin- stream infection (Category C1 evidence); randomized con- ical need. An insertion site should be selected that is not trolled trials were equivocal regarding catheter colonization contaminated or potentially contaminated (e.g., burned or (P ϭ 0.16–0.82) (Category C2 evidence).20–22,24 Meta-anal- infected skin, inguinal area, adjacent to tracheostomy or yses of randomized controlled trials25–36 demonstrate that open surgical wound). In adults, selection of an upper catheters coated with chlorhexidine and silver sulfadiazine body insertion site should be considered to minimize the reduce catheter colonization (Category A1 evidence); equivo- risk of infection. cal findings are reported for catheter-related bloodstream in- Catheter Fixation. The literature is insufficient to evaluate fection (i.e., catheter colonization and corresponding posi- whether catheter fixation with sutures, staples or tape is as- tive blood culture) (Category C1 evidence).25–27,29–35,37,38 sociated with a higher risk for catheter-related infections Cases of anaphylactic shock are reported after placement of a (Category D evidence). catheter coated with chlorhexidine and silver sulfadiazine Most consultants and ASA members indicate that use of 39–41 (Category B3 evidence). sutures is the preferred catheter fixation technique to mini- Consultants and ASA members agree that catheters coated mize catheter-related infection. with antibiotics or a combination of chlorhexidine and silver Recommendations for Catheter Fixation. The use of su- sulfadiazine may be used in selected patients based on infectious tures, staples, or tape for catheter fixation should be deter- risk, cost, and anticipated duration of catheter use. mined on a local or institutional basis. Recommendations for Use of Catheters Containing Anti- Insertion Site Dressings. The literature is insufficient to microbial Agents. Catheters coated with antibiotics or a evaluate the efficacy of transparent bio-occlusive dressings combination of chlorhexidine and silver sulfadiazine should to reduce the risk of infection (Category D evidence). Ran- be used for selected patients based on infectious risk, cost, domized controlled trials are equivocal (P ϭ 0.04–0.96) and anticipated duration of catheter use. The Task Force regarding catheter tip colonization50,51 and inconsistent notes that catheters containing antimicrobial agents are not a (P ϭ 0.004–0.96) regarding catheter-related blood- substitute for additional infection precautions. stream infection50,52 when chlorhexidine sponge dressings Selection of Catheter Insertion Site. A randomized con- are compared with standard polyurethane dressings (Cate- trolled trial comparing the subclavian and femoral insertion gory C2 evidence). A randomized controlled trial is also equiv- sites report higher levels of catheter colonization with the ocal regarding catheter tip colonization for silver-impreg- femoral site (Category A3 evidence); equivocal findings are nated transparent dressings compared with standard reported for catheter-related sepsis (P ϭ 0.07) (Category C2 dressings (P Ͼ 0.05) (Category C2 evidence).53 A randomized evidence).42 A randomized controlled trial comparing the in- controlled trial reports a greater frequency of severe localized ternal jugular insertion site with the femoral site reports no contact dermatitis when neonates receive chlorhexidine-im- difference in catheter colonization (P ϭ 0.79) or catheter pregnated dressings compared with povidone-iodine im- related bloodstream infections (P ϭ 0.42) (Category C2 evi- pregnated dressings (Category A3 evidence).54 dence).43 Prospective nonrandomized comparative studies The ASA members agree and the consultants strongly are equivocal (i.e., inconsistent) regarding catheter-related agree that transparent bio-occlusive dressings should be used colonization44–46 and catheter related bloodstream infec- to protect the site of central venous catheter insertion from tion46–48 when the internal jugular site is compared with the infection. The consultants and ASA members agree that subclavian site (Category C3 evidence). A nonrandomized dressings containing chlorhexidine may be used to reduce the comparative study of burn patients reports that catheter col- risk of catheter-related infection. SPA members are equivocal onization and bacteremia occur more frequently the closer regarding whether dressings containing chlorhexidine may the catheter insertion site is to the burn wound (Category B1 be used for skin preparation in neonates (younger than 44 evidence).49 gestational weeks); they agree that the use of dressings con- Most consultants indicate that the subclavian insertion taining chlorhexidine may be used in infants (younger than 2 site is preferred to minimize catheter-related risk of infec- yr) and children (2–16 yr). tion. Most ASA members indicate that the internal jugular Recommendations for Insertion Site Dressings. Transpar- insertion site is preferred to minimize catheter-related ent bio-occlusive dressings should be used to protect the risk of infection. The consultants and ASA members agree site of central venous catheter insertion from infection. that femoral catheterization should be avoided when pos- Unless contraindicated, dressings containing chlorhexi- sible to minimize the risk of infection. The consultants dine may be used in adults, infants, and children. For and ASA members strongly agree that an insertion site neonates, the use of transparent or sponge dressings con- should be selected that is not contaminated or potentially taining chlorhexidine should be based on clinical judg- contaminated. ment and institutional protocol.

Anesthesiology 2012; 116:539–73 543 Practice Guidelines Practice Guidelines

Catheter Maintenance. Catheter maintenance consists of (1) connectors with standard caps indicate decreased levels of determining the optimal duration of catheterization, (2) con- microbial contamination of stopcock entry ports with 63,64 ducting catheter site inspections, (3) periodically changing needleless connectors (Category A2 evidence); no differ- catheters, and (4) changing catheters using a guidewire in- ences in catheter-related bloodstream infection are reported 65,66 stead of selecting a new insertion site. (P ϭ 0.3–0.9) (Category C2 evidence). Nonrandomized comparative studies indicate that longer The consultants and ASA members strongly agree that catheterizations are associated with higher rates of catheter catheter access ports should be wiped with an appropriate colonization, infection, and sepsis (Category B2 evi- antiseptic before each access. The consultants and ASA mem- dence).45,55 The literature is insufficient to evaluate whether bers agree that needleless ports may be used on a case-by-case specified time intervals between catheter site inspections are basis. The consultants and ASA members strongly agree that associated with a higher risk for catheter-related infection central venous catheter stopcocks should be capped when not (Category D evidence). Randomized controlled trials report in use. equivocal findings (P ϭ 0.54–0.63) regarding differences in Recommendations for Aseptic Techniques Using an Ex- catheter tip colonizations when catheters are changed at 3- isting Central Line. Catheter access ports should be wiped versus 7-day intervals (Category C2 evidence).56,57 Meta-anal- with an appropriate antiseptic before each access when using ysis of randomized controlled trials58–62 report equivocal an existing central venous catheter for injection or aspiration. findings for catheter tip colonization when guidewires are Central venous catheter stopcocks or access ports should be used to change catheters compared with the use of new in- capped when not in use. Needleless catheter access ports may sertion sites (Category C1 evidence). be used on a case-by-case basis. The ASA members agree and the consultants strongly agree that the duration of catheterization should be based on III. Prevention of Mechanical Trauma or Injury clinical need. The consultants and ASA members strongly Interventions intended to prevent mechanical trauma or agree that (1) the clinical need for keeping the catheter in injury associated with central venous access include, but place should be assessed daily; (2) catheters should be are not limited to (1) selection of catheter insertion site, promptly removed when deemed no longer clinically neces- (2) positioning the patient for needle insertion and cath- sary; (3) the catheter site should be inspected daily for signs of eter placement, (3) needle insertion and catheter place- infection and changed when infection is suspected; and (4) ment, and (4) monitoring for needle, guidewire, and cath- when catheter infection is suspected, replacing the catheter eter placement. using a new insertion site is preferable to changing the cath- 1. Selection of Catheter Insertion Site. A randomized con- eter over a guidewire. trolled trial comparing the subclavian and femoral insertion Recommendations for Catheter Maintenance. The dura- sites reports that the femoral site had a higher frequency of tion of catheterization should be based on clinical need. The thrombotic complications in adult patients (Category A3 ev- clinical need for keeping the catheter in place should be as- idence).42 A randomized controlled trial comparing the in- sessed daily. Catheters should be removed promptly when no ternal jugular insertion site with the femoral site reports longer deemed clinically necessary. The catheter insertion equivocal findings for arterial puncture (P ϭ 0.35), deep site should be inspected daily for signs of infection, and the venous thrombosis (P ϭ 0.62) or hematoma formation (P ϭ catheter should be changed or removed when catheter inser- 0.47) (Category C2 evidence).43 A randomized controlled trial tion site infection is suspected. When a catheter related in- comparing the internal jugular insertion site with the subcla- fection is suspected, replacing the catheter using a new inser- vian site reports equivocal findings for successful veni- tion site is preferable to changing the catheter over a puncture (P ϭ 0.03) (Category C2 evidence).67 Nonran- guidewire. domized comparative studies report equivocal findings for arterial puncture, pneumothorax, hematoma, hemotho- Aseptic Techniques Using an Existing Central Venous rax, or arrhythmia when the internal jugular insertion site Catheter for Injection or Aspiration is compared with the subclavian insertion site (Category C3 evidence).68–70 Aseptic techniques using an existing central venous catheter for injection or aspiration consist of (1) wiping the port with Most consultants and ASA members indicate that the an appropriate antiseptic, (2) capping stopcocks or access internal jugular insertion site is preferred to minimize ports, and (3) use of needleless catheter connectors or access catheter cannulation-related risk of injury or trauma. ports. Most consultants and ASA members also indicate that the The literature is insufficient to evaluate whether wiping internal jugular insertion site is preferred to minimize ports or capping stopcocks when using an existing central catheter-related risk of thromboembolic injury or trauma. venous catheter for injection or aspiration is associated with a Recommendations for Catheter Insertion Site Selection. reduced risk for catheter-related infections (Category D evi- Catheter insertion site selection should be based on dence). Randomized controlled trials comparing needleless clinical need and practitioner judgment, experience, and

Anesthesiology 2012; 116:539–73 544 Practice Guidelines SPECIAL ARTICLES

skill. In adults, selection of an upper body insertion site tion, and the skill and experience of the operator. The should be considered to minimize the risk of thrombotic consultants and ASA members agree that the selection of a complications. modified Seldinger technique versus a Seldinger technique 2. Positioning the Patient for Needle Insertion and Cath- should be based on the clinical situation and the skill and eter Placement. Nonrandomized studies comparing the Tren- experience of the operator. The consultants and ASA delenburg (i.e., head down) position with the normal supine members agree that the number of insertion attempts position indicates that the right internal jugular vein increases in should be based on clinical judgment. The ASA members diameter and cross-sectional area to a greater extent when adult agree and the consultants strongly agree that the decision patients are placed in the Trendelenburg position (Category B2 to place two central catheters in a single vein should be evidence).71–76 One nonrandomized study comparing the Tren- made on a case-by-case basis. delenburg position with the normal supine position in pediatric Recommendations for Needle Insertion, Wire Placement, patients reports an increase in right internal jugular vein diam- and Catheter Placement. Selection of catheter size (i.e., eter only for patients older than 6 yr (Category B2 evidence).77 outside diameter) and type should be based on the clinical The consultants and ASA members strongly agree that, situation and skill/experience of the operator. Selection of when clinically appropriate and feasible, central vascular ac- the smallest size catheter appropriate for the clinical situ- cess in the neck or chest should be performed with the patient ation should be considered. Selection of a thin-wall needle in the Trendelenburg position. (i.e., Seldinger) technique versus a catheter-over-the-needle (i.e., modified Seldinger) technique should be based on the clinical situation and the skill/experience of the Recommendations for Positioning the Patient for Needle operator. The decision to use a thin-wall needle technique Insertion and Catheter Placement or a catheter-over-the-needle technique should be based at When clinically appropriate and feasible, central venous ac- least in part on the method used to confirm that the wire cess in the neck or chest should be performed with the patient resides in the vein before a dilator or large-bore catheter is in the Trendelenburg position. threaded (fig. 1). The Task Force notes that the catheter- over-the-needle technique may provide more stable ve- 3. Needle Insertion, Wire Placement, and Catheter Place- nous access if manometry is used for venous confirmation. ment. Needle insertion, wire placement, and catheter place- The number of insertion attempts should be based on ment includes (1) selection of catheter size and type, (2) use of a clinical judgment. The decision to place two catheters in a wire-through-thin-wall needle technique (i.e., Seldinger tech- single vein should be made on a case-by-case basis. nique) versus a catheter-over-the-needle-then-wire-through- 4. Guidance and Verification of Needle, Wire, and Catheter the-catheter technique (i.e., modified Seldinger technique), (3) Placement. Guidance for needle, wire, and catheter placement limiting the number of insertion attempts, and (4) introducing includes ultrasound imaging for the purpose of prepuncture two catheters in the same central vein. vessel localization (i.e., static ultrasound) and ultrasound for Case reports describe severe injury (e.g., hemorrhage, he- vessel localization and guiding the needle to its intended venous matoma, pseudoaneurysm, arteriovenous fistula, arterial dis- location (i.e., real time or dynamic ultrasound). Verification of section, neurologic injury including stroke, and severe or needle, wire, or catheter location includes any one or more of the lethal airway obstruction) when there is unintentional ar- following methods: (1) ultrasound, (2) manometry, (3) pressure terial cannulation with large bore catheters (Category B3 waveform analysis, (4) venous blood gas, (5) fluoroscopy, (6) 78–88 evidence). The literature is insufficient to evaluate continuous electrocardiography, (7) transesophageal echocardi- whether the risk of injury or trauma is associated with the ography, and (8) chest radiography. use of a thin-wall needle technique versus a catheter-over- the needle technique (Category D evidence). The literature is insufficient to evaluate whether the risk of injury or Guidance trauma is related to the number of insertion attempts Static Ultrasound. Randomized controlled trials comparing (Category D evidence). One nonrandomized comparative static ultrasound with the anatomic landmark approach for lo- study reports a higher frequency of dysrhythmia when two cating the internal jugular vein report a higher first insertion central venous catheters are placed in the same vein (right attempt success rate for static ultrasound (Category A3 evi- internal jugular) compared with placement of one cathe- dence);90 findings are equivocal regarding overall successful can- ter in the vein (Category B2 evidence); no differences in nulation rates (P ϭ 0.025–0.57) (Category C2 evidence).90–92 In carotid artery puncture (P ϭ 0.65) or hematoma (P ϭ addition, the literature is equivocal regarding subclavian vein 0.48) were noted (Category C3 evidence).89 access (P ϭ 0.84) (Category C2 evidence) 93 and insufficient for The consultants agree and the ASA members strongly femoral vein access (Category D evidence). agree that the selection of catheter type (i.e., gauge, The consultants and ASA members agree that static ultra- length, number of lumens) and composition (e.g., poly- sound imaging should be used in elective situations for pre- urethane, Teflon) should be based on the clinical situa- puncture identification of anatomy and vessel localization

Anesthesiology 2012; 116:539–73 545 Practice Guidelines Practice Guidelines

Fig. 1. Algorithm for central venous insertion and verification. This algorithm compares the thin-wall needle (i.e., Seldinger) technique versus the catheter-over-the needle (i.e., Modified-Seldinger) technique in critical safety steps to prevent unintentional arterial placement of a dilator or largebore catheter. The variation between the two techniques reflects mitigation steps for the risk that the thin-wall needle in the Seldinger technique could move out of the vein and into the wall of an artery between the manometry step and the threading of the wire step. ECG ϭ electrocardiography; TEE ϭ transesophageal echocardiography. when the internal jugular vein is selected for cannulation; Real-time Ultrasound. Meta-analysis of randomized con- they are equivocal regarding whether static ultrasound imag- trolled trials94–104 indicates that, compared with the ana- ing should be used when the subclavian vein is selected. The tomic landmark approach, real-time ultrasound guided ve- consultants agree and the ASA members are equivocal re- nipuncture of the internal jugular vein has a higher garding the use of static ultrasound imaging when the fem- first insertion attempt success rate, reduced access time, oral vein is selected. higher overall successful cannulation rate, and decreased

Anesthesiology 2012; 116:539–73 546 Practice Guidelines SPECIAL ARTICLES

rates of arterial puncture (Category A1 evidence). identifying the position of the catheter tip (Category B2 evi- Randomized controlled trials report fewer number of dence). Randomized controlled trials indicate that continu- insertion attempts with real-time ultrasound guided ous electrocardiography is effective in identifying proper venipuncture of the internal jugular vein (Category A2 catheter tip placement compared with not using electrocar- evidence).97,99,103,104 diography (Category A2 evidence).115,126,127 For the subclavian vein, randomized controlled trials report The consultants and ASA members strongly agree that fewer insertion attempts with real-time ultrasound guided veni- before insertion of a dilator or large- bore catheter over a puncture (Category A2 evidence),105,106 and one randomized wire, venous access should be confirmed for the catheter or clinical trial indicates a higher success rate and reduced access thin-wall needle that accesses the vein. The Task Force be- time, with fewer arterial punctures and hematomas compared lieves that blood color or absence of pulsatile flow should not with the anatomic landmark approach (Category A3 evi- be relied upon to confirm venous access. The consultants dence).106 agree and ASA members are equivocal that venous access For the femoral vein, a randomized controlled trial re- should be confirmed for the wire that subsequently resides in ports a higher first-attempt success rate and fewer needle the vein after traveling through a catheter or thin-wall needle passes with real-time ultrasound guided venipuncture com- before insertion of a dilator or large-bore catheter over a wire. pared with the anatomic landmark approach in pediatric The consultants and ASA members agree that, when feasible, patients (Category A3 evidence).107 both the location of the catheter or thin-wall needle and wire The consultants agree and the ASA members are equivocal that, should be confirmed. when available, real time ultrasound should be used for The consultants and ASA members agree that a chest guidance during venous access when either the internal radiograph should be performed to confirm the location of jugular or femoral veins are selected for cannulation. The the catheter tip as soon after catheterization as clinically ap- consultants and ASA members are equivocal regarding the propriate. They also agree that, for central venous catheters use of real time ultrasound when the subclavian vein is placed in the operating room, a confirmatory chest radio- selected. graph may be performed in the early postoperative period. The ASA members agree and the consultants strongly agree that, if a chest radiograph is deferred to the postoperative Verification period, pressure waveform analysis, blood gas analysis, ultra- Confirming that the Catheter or Thin-wall Needle Resides sound, or fluoroscopy should be used to confirm venous in the Vein. A retrospective observational study reports that positioning of the catheter before use. manometry can detect arterial punctures not identified by blood flow and color (Category B2 evidence).108 The literature is insuf- Recommendations for Guidance and Verification of ficient to address ultrasound, pressure-waveform analysis, blood Needle, Wire, and Catheter Placement gas analysis, blood color, or the absence of pulsatile flow as The following steps are recommended for prevention of me- effective methods of confirming catheter or thin-wall needle chanical trauma during needle, wire, and catheter placement venous access (Category D evidence). in elective situations: Confirming Venous Residence of the Wire. An observational ● Use static ultrasound imaging before prepping and study indicates that ultrasound can be used to confirm venous draping for prepuncture identification of anatomy to placement of the wire before dilation or final catheterization 109 determine vessel localization and patency when the in- (Category B2 evidence). Case reports indicate that transesoph- ternal jugular vein is selected for cannulation. Static ageal echocardiography was used to identify guidewire position 110–112 ultrasound may be used when the subclavian or femoral (Category B3 evidence). The literature is insufficient to vein is selected. evaluate the efficacy of continuous electrocardiography in con- ● Use real time ultrasound guidance for vessel localization firming venous residence of the wire (Category D evidence), al- and venipuncture when the internal jugular vein is selected though narrow complex electrocardiographic ectopy is recog- for cannulation (see fig. 1). Real-time ultrasound may be nized by the Task Force as an indicator of venous location of the used when the subclavian or femoral vein is selected. The wire. The literature is insufficient to address fluoroscopy as an Task Force recognizes that this approach may not be fea- effective method to confirm venous residence of the wire (Cat- sible in emergency circumstances or in the presence of egory D evidence); the Task Force believes that fluoroscopy may other clinical constraints. be used. Confirming Residence of the Catheter in the Venous Sys- ● After insertion of a catheter that went over the needle or a tem. Studies with observational findings indicate that fluo- thin-wall needle, confirm venous access.†† Methods for confirming that the catheter or thin-wall needle resides in roscopy113,115 115–125 and chest radiography are useful in the vein include, but are not limited to, ultrasound, ma- †† For neonates, infants, and children, confirmation of venous nometry, pressure-waveform analysis, or venous blood gas placement may take place after the wire is threaded. measurement. Blood color or absence of pulsatile flow

Anesthesiology 2012; 116:539–73 547 Practice Guidelines Practice Guidelines

should not be relied upon for confirming that the catheter nonsurgically, as follows: 54.9% (for neonates), 43.8% (for in- or thin-wall needle resides in the vein. fants), and 30.0% (for children). SPA members indicating that ● When using the thin-wall needle technique, confirm the catheter may be nonsurgically removed without consulta- venous residence of the wire after the wire is threaded. tion is as follows: 45.1% (for neonates), 56.2% (for infants), and When using the catheter-over-the-needle technique, 70.0% (for children). The Task Force agrees that the anesthesi- confirmation that the wire resides in the vein may not be ologist and surgeon should confer regarding the relative risks needed (1) when the catheter enters the vein easily and and benefits of proceeding with elective surgery after an arterial manometry or pressure waveform measurement pro- vessel has sustained unintended injury by a dilator or large-bore vides unambiguous confirmation of venous location of catheter. the catheter; and (2) when the wire passes through the Recommendations for Management of Arterial Trauma or catheter and enters the vein without difficulty. If there is Injury Arising from Central Venous Access. When unin- any uncertainty that the catheter or wire resides in the tended cannulation of an arterial vessel with a dilator or vein, confirm venous residence of the wire after the wire large-bore catheter occurs, the dilator or catheter should is threaded. Insertion of a dilator or large-bore catheter be left in place and a general surgeon, a vascular surgeon, may then proceed. Methods for confirming that the wire or an interventional radiologist should be immediately resides in the vein include, but are not limited to, ultra- consulted regarding surgical or nonsurgical catheter re- sound (identification of the wire in the vein) or trans- moval for adults. For neonates, infants, and children the esophageal echocardiography (identification of the wire decision to leave the catheter in place and obtain consul- in the superior vena cava or right atrium), continuous tation or to remove the catheter nonsurgically should be electrocardiography (identification of narrow-complex based on practitioner judgment and experience. After the ectopy), or fluoroscopy. injury has been evaluated and a treatment plan has been ● After final catheterization and before use, confirm resi- executed, the anesthesiologist and surgeon should confer dence of the catheter in the venous system as soon as regarding relative risks and benefits of proceeding with the clinically appropriate. Methods for confirming that the elective surgery versus deferring surgery to allow for a pe- catheter is still in the venous system after catheterization riod of patient observation. and before use include manometry or pressure waveform measurement. Appendix 1: Summary of ● Confirm the final position of the catheter tip as soon as clinically appropriate. Methods for confirming the position of Recommendations the catheter tip include chest radiography, fluoroscopy, or Resource Preparation continuous electrocardiography. For central venous catheters placed in the operating room, perform the chest radiograph ● Central venous catheterization should be performed in an envi- no later than the early postoperative period to confirm the ronment that permits use of aseptic techniques. position of the catheter tip. ● A standardized equipment set should be available for central venous access. IV. Management of Arterial Trauma or Injury Arising ● A checklist or protocol should be used for placement and maintenance of central venous catheters. from Central Venous Catheterization ● An assistant should be used during placement of a central venous Case reports of adult patients with arterial puncture by a catheter. large bore catheter/vessel dilator during attempted central venous catheterization indicate severe complications (e.g., Prevention of Infectious Complications cerebral infarction, arteriovenous fistula, hemothorax) af- • For immunocompromised patients and high-risk neonates, ter immediate catheter removal; no such complications administer intravenous antibiotic prophylaxis on a case-by- were reported for adult patients whose catheters were left case basis. in place before surgical consultation and repair (Category 80,86 Intravenous antibiotic prophylaxis should not be adminis- B3 evidence). tered routinely. The consultants and ASA members agree that, when unin- • In preparation for the placement of central venous catheters, use tended cannulation of an arterial vessel with a large-bore cathe- aseptic techniques (e.g., hand washing) and maximal barrier pre- ter occurs, the catheter should be left in place and a general cautions (e.g., sterile gowns, sterile gloves, caps, masks covering surgeon or vascular surgeon should be consulted. When unin- both mouth and nose, and full-body patient drapes). tended cannulation of an arterial vessel with a large-bore cathe- • A chlorhexidine-containing solution should be used for skin ter occurs, the SPA members indicate that the catheter should be preparation in adults, infants, and children. left in place and a general surgeon, vascular surgeon, or inter- For neonates, the use of a chlorhexidine-containing solution ventional radiologist should be immediately consulted before for skin preparation should be based on clinical judgment and deciding on whether to remove the catheter, either surgically or institutional protocol.

Anesthesiology 2012; 116:539–73 548 Practice Guidelines SPECIAL ARTICLES

If there is a contraindication to chlorhexidine, povidone-io- In adults, selection of an upper body insertion site should dine or alcohol may be used as alternatives. be considered to minimize the risk of thrombotic Unless contraindicated, skin preparation solutions should complications. contain alcohol. • When clinically appropriate and feasible, central venous access in • If there is a contraindication to chlorhexidine, povidone-iodine the neck or chest should be performed with the patient in the or alcohol may be used. Unless contraindicated, skin preparation Trendelenburg position. solutions should contain alcohol. • Selection of catheter size (i.e., outside diameter) and type • Catheters coated with antibiotics or a combination of chlo- should be based on the clinical situation and skill/experience rhexidine and silver sulfadiazine should be used for selected of the operator. patients based on infectious risk, cost, and anticipated dura- Selection of the smallest size catheter appropriate for the tion of catheter use. clinical situation should be considered. Catheters containing antimicrobial agents are not a substi- • Selection of a thin-wall needle (a wire-through-thin-wall-needle, tute for additional infection precautions. or Seldinger) technique versus a catheter-over-the-needle (a catheter-over-the-needle-then-wire-through-the-catheter, or Modi- • Catheter insertion site selection should be based on clinical fied Seldinger) technique should be based on the clinical situation need. and the skill/experience of the operator. An insertion site should be selected that is not contami- The decision to use a thin-wall needle technique or a cath- nated or potentially contaminated (e.g., burned or infected eter-over-the-needle technique should be based at least in skin, inguinal area, adjacent to tracheostomy or open sur- part on the method used to confirm that the wire resides in gical wound). the vein before a dilator or large-bore catheter is In adults, selection of an upper body insertion site should threaded. be considered to minimize the risk of infection. The catheter-over-the-needle technique may provide • The use of sutures, staples, or tape for catheter fixation should be more stable venous access if manometry is used for venous determined on a local or institutional basis. confirmation. • Transparent bio-occlusive dressings should be used to protect • The number of insertion attempts should be based on clinical the site of central venous catheter insertion from infection. judgment. Unless contraindicated, dressings containing chlorhexidine • The decision to place two catheters in a single vein should be may be used in adults, infants, and children. made on a case-by-case basis. For neonates, the use of transparent or sponge dressings • Use static ultrasound imaging in elective situations before prep- containing chlorhexidine should be based on clinical judg- ping and draping for prepuncture identification of anatomy to ment and institutional protocol. determine vessel localization and patency when the internal jug- • The duration of catheterization should be based on clinical ular vein is selected for cannulation. need. Static ultrasound may be used when the subclavian or femoral The clinical need for keeping the catheter in place should be vein is selected. assessed daily. • Use real-time ultrasound guidance for vessel localization and Catheters should be removed promptly when no longer venipuncture when the internal jugular vein is selected for deemed clinically necessary. cannulation. • The catheter insertion site should be inspected daily for signs of Real-time ultrasound may be used when the subclavian or infection. femoral vein is selected. Real-time ultrasound may not be feasible in emergency The catheter should be changed or removed when catheter circumstances or in the presence of other clinical insertion site infection is suspected. constraints. • When a catheter-related infection is suspected, replacing the • After insertion of a catheter that went over the needle or a catheter using a new insertion site is preferable to changing the thin-wall needle, confirm venous access.†† catheter over a guidewire. Methods for confirming that the catheter or thin-wall nee- • Catheter access ports should be wiped with an appropriate anti- dle resides in the vein include, but are not limited to: ultra- septic before each access when using an existing central venous sound, manometry, pressure-waveform analysis, or venous catheter for injection or aspiration. blood gas measurement. • Central venous catheter stopcocks or access ports should be Blood color or absence of pulsatile flow should not be relied capped when not in use. upon for confirming that the catheter or thin-wall needle • Needleless catheter access ports may be used on a case-by-case resides in the vein. basis. • When using the thin-wall needle technique, confirm venous residence of the wire after the wire is threaded. • When using the catheter-over-the-needle technique, confir- Prevention of Mechanical Trauma or Injury mation that the wire resides in the vein may not be needed (1) • Catheter insertion site selection should be based on clinical need when the catheter enters the vein easily and manometry or and practitioner judgment, experience, and skill. pressure waveform measurement provides unambiguous con-

Anesthesiology 2012; 116:539–73 549 Practice Guidelines Practice Guidelines

firmation of venous location of the catheter, and (2) when the Appendix 2. Example of a Standardized Equipment wire passes through the catheter and enters the vein without Cart for Central Venous Catheterization for Adult difficulty. Patients If there is any uncertainty that the catheter or wire resides in the Item Description Quantity vein, confirm venous residence of the wire after the wire is threaded. Insertion of a dilator or large-bore catheter may then First Drawer proceed. Bottles Alcohol-based Hand Cleanser 2 Methods for confirming that the wire resides in the vein Transparent bio-occlusive dressings with catheter 2 include, but are not limited to surface ultrasound (identifi- stabilizer devices cation of the wire in the vein) or transesophageal echocar- Transducer kit: NaCL 0.9% 500 ml bag; single- 1 diography (identification of the wire in the superior vena line transducer, pressure bag cava or right atrium), continuous electrocardiography Needle Holder, Webster Disposable 5 inch 1 (identification of narrow-complex ectopy), or fluoroscopy. Scissors, 4 1/2 inchSterile 1 • After final catheterization and before use, confirm residence of Vascular Access Tray(Chloraprep, Sponges, 1 the catheter in the venous system as soon as clinically Labels) Disposable pen with sterile labels 4 appropriate. Sterile tubing, arterial line pressure-rated (for 2 Methods for confirming that the catheter is still in the manometry) venous system after catheterization and before use include Intravenous connector with needleless valve 4 waveform manometry or pressure measurement. Second Drawer • Confirm the final position of the catheter tip as soon as clin- ϫ ically appropriate. Ultrasound Probe Cover, Sterile 3 96 2 Applicator, chloraprep 10.5 ml 3 Methods for confirming the position of the catheter tip Surgical hair clipper blade 3 include chest radiography, fluoroscopy, or continuous Solution, NaCl bacteriostatic 30 ml 2 electrocardiography. Third Drawer • For central venous catheters placed in the operating room, perform the chest radiograph no later than the early postoperative Cap, Nurses Bouffant 3 Surgeon hats 6 period to confirm the position of the catheter tip. Goggles 2 Mask, surgical fluidshield 2 Management of Arterial Trauma or Injury Arising from Gloves, sterile sizes 6.0–8.0 (2 each size) 10 Packs, sterile gowns 2 Central Venous Catheterization Fourth Drawer • When unintended cannulation of an arterial vessel with a dilator or large-bore catheter occurs, the dilator or catheter should be left Drape, Total Body (with Femoral Window) 1 in place and a general surgeon, a vascular surgeon, or an interven- Sheet, central line total body (no window) 1 tional radiologist should be immediately consulted regarding sur- Fifth Drawer gical or nonsurgical catheter removal for adults. Dressing, Sterile Sponge Packages 4 For neonates, infants, and children, the decision to leave the Catheter kit, central venous pressure single 1 catheter in place and obtain consultation or to remove the lumen14 gauge catheter nonsurgically should be based on practitioner judg- Catheter kits, central venous pressure two 2 ment and experience. lumens 16 cm 7 French • After the injury has been evaluated and a treatment plan Sixth Drawer has been executed, the anesthesiologist and surgeon should Triple Lumen Centravel Venous Catheter Sets, 2 confer regarding relative risks and benefits of proceeding with 7 French Antimicrobial Impregnated the elective surgery versus deferring surgery for a period of Introducer catheter sets, 9 French with sideport 2 patient observation.

Anesthesiology 2012; 116:539–73 550 Practice Guidelines SPECIAL ARTICLES

Appendix 3. Example of a Central Venous Catheterization Checklist

Central Line Insertion Standard Work & Safety (Bundle) Checklist for OR and CCU

Date: ______Start Time: ______End Time: ______Procedure Operator: ______Person Completing Form: ______Catheter Type: Central Venous PA/Swan-Ganz

French Size of catheter: ______. Catheter lot number: ______

Number of Lumens: 1 2 3 4 Insertion Site: Jugular Upper Arm Subclavian Femoral Side of Body: Left Right Bilateral Clinical Setting: Elective Emergent

1. Consent form complete and in chart Exception: Emergent procedure

2. Patient’s Allergy Assessed (especially to Lidocaine or Heparin)

3. Patient’s Latex Allergy Assessed (modify supplies)

4. Hand Hygiene: Operator and Assistant cleanse hands (ASK, if not witnessed) 5. Optimal Catheter Site Selection: In adults, Consider Upper Body Site Check / explain why femoral site used: ______OR Anatomy – distorted, prior surgery/rad. Scar Chest wall infection or burn Exception(s) Coagulopathy COPD severe/ lung disease checked to left Emergency / CPR Pediatric

6. Pre-procedure Ultrasound Check of internal jugular location and patency if IJ 7. Skin Prep Performed (Skin Antisepsis): Chloraprep 10.5 ml applicator used Dry technique (normal, unbroken skin): 30 second scrub + 30 second dry DRY time WET Wet technique (abnormal or broken skin): 2 minute scrub + 1 minute dry time

8. MAXIMUM Sterile Barriers: Operator wearing hat, mask, sterile gloves, and sterile gown Others in room, (except patient) wearing mask Patient’s body covered by sterile drape 9. Procedural “Time out” performed: Patient ID X 2 Procedure to be performed has been announced Insertion site marked Patient positioned correctly for procedure (Supine or Trendelenburg) Assembled equipment/ supplies including venous confirmation method verified Labels on all medication & syringes are verified

(continued)

Anesthesiology 2012; 116:539–73 551 Practice Guidelines Practice Guidelines

Appendix 3. Continued

10. Ultrasound Guidance Used for Elective Internal Jugular insertions (sterile Used for IJ probe cover in place) Not used (Other site used) 11. Confirmation of Venous Placement of Access Needle or Catheter: (do not Manometry rely on blood color or presence/absence of pulsatility) Ultrasound Transducer Blood Gas 12. Confirmation of Venous Placement of the Wire: Access catheter easily in vein & confirmed (catheter-over needle technique) Not Needed

Access via thin-wall needle (confirmation of wire recommended) Ultrasound or ambiguous catheter or wire placement when using catheter-over-the-needle TEE technique Fluoroscopy ECG

13. Confirmation of Final Catheter in Venous System Prior to Use: Manometry Transducer 14. Final steps: Verify guidewire not retained Type and Dosage (ml / units) of Flush: ______ Catheter Caps Placed on Lumens Tip position confirmation: Fluoroscopy Chest radiograph ordered Catheter Secured / Sutured in place

15. Transparent Bio-occlusive dressing applied

16. Sterile Technique Maintained when applying dressing

17. Dressing Dated

18. Confirm Final Location of Catheter Tip CXR Fluoroscopy Continuous ECG 19. After tip location confirmed, “Approved for use” Written on Dressing

20. Central line (maintenance) Order Placed

Comments:

Tip location:

Anesthesiology 2012; 116:539–73 552 Practice Guidelines SPECIAL ARTICLES

Appendix 4. Example Duties Performed by an Silver-impregnated catheters versus no coating Assistant for Central Venous Catheterization Chlorhexidine combined with silver sulfadiazine catheter coating versus no coating Reads prompts on checklist to ensure that no safety Selection of catheter insertion site step is forgotten or missed. Completes checklist as Internal jugular task is completed Subclavian Verbally alerts anesthesiologist if a potential error or mistake is about to be made. Femoral Gathers equipment/supplies or brings standardized Selecting a potentially uncontaminated insertion site supply cart. Catheter fixation Brings the ultrasound machine, positions it, turns it on, Suture, staple, or tape makes adjustments as needed. Insertion site dressings Provides moderate sedation (if registered nurse) if Clear plastic, chlorhexidine, gauze and tape, cyanoacrylate, needed. antimicrobial dressings, patch, antibiotic ointment Participates in “time-out” before procedure. Catheter maintenance Washes hands and wears mask, cap, and nonsterile Long-term versus short-term catheterization gloves (scrubs or cover gown required if in the sterile Frequency of insertion site inspection for signs of infection envelope). Attends to patient requests if patient awake during Changing catheters procedure. Specified time intervals Assists with patient positioning. Specified time interval versus no specified time interval (i.e.,as Assists with draping. needed) Assists with sterile field setup; drops sterile items into One specified time interval versus another specified time interval field as needed. Changing a catheter over a wire versus a new site Assists with sterile ultrasound sleeve application to Aseptic techniques using an existing central line for injection or ultrasound probe. aspiration Assists with attachment of intravenous lines or Wiping ports with alcohol pressure lines if needed. Capping stopcocks Assists with application of a sterile bandage at the end Needleless connectors or access ports of the procedure. Assists with clean-up of patient, equipment, and Prevention of Mechanical Trauma or Injury supply cart; returns items to their proper location. Selection of catheter insertion site Internal jugular Appendix 5: Methods and Analyses Subclavian Femoral State of the Literature Trendelenburg versus supine position For these Guidelines, a literature review was used in combination Needle insertion and catheter placement with opinions obtained from expert consultants and other sources Selection of catheter type (e.g., double lumen, triple lumen, (e.g., ASA members, SPA members, open forums, Internet post- Cordis) ings). Both the literature review and opinion data were based on Selection of a large-bore catheter evidence linkages, or statements regarding potential relationships Placement of two catheters in the same vein between clinical interventions and outcomes. The interventions Use of a Seldinger technique versus a modified Seldinger listed below were examined to assess their effect on a variety of technique outcomes related to central venous catheterization. Limiting number of insertion attempts Resource Preparation Guidance of needle, wire and catheter placement Selection of a Sterile Environment Static ultrasound versus no ultrasound (i.e., anatomic Availability of a standardized equipment set landmarks) Use of a checklist or protocol for placement and maintenance Real-time ultrasound guidance versus no ultrasound Use of an assistant for placement Verification of placement Manometry versus direct pressure measurement (via pressure Prevention of Infectious Complications transducer) Intravenous antibiotic prophylaxis Continuous electrocardiogram Aseptic techniques Fluoroscopy Aseptic preparation Venous blood gas Hand washing, sterile full-body drapes, sterile gown, gloves, Transesophageal echocardiography mask, cap Chest radiography Skin preparation Chlorhexidine versus povidone-iodine Management of Trauma or Injury Arising from Central Venous Aseptic preparation with versus without alcohol Catheterization Selection of catheter coatings or impregnation Not removing versus removing central venous catheter on Antibiotic-coated catheters versus no coating evidence of arterial puncture.

Anesthesiology 2012; 116:539–73 553 Practice Guidelines Practice Guidelines

For the literature review, potentially relevant clinical studies were 0.70, Var (Sav) ϭ 0.016; (3) linkage assignment, Sav ϭ 0.94, Var identified via electronic and manual searches of the literature. The (Sav) ϭ 0.002; (4) literature database inclusion, Sav ϭ 0.65, Var electronic and manual searches covered a 44-yr period from 1968 (Sav) ϭ 0.034. These values represent moderate to high levels of through 2011. More than 2,000 citations were initially identified, agreement. yielding a total of 671 nonoverlapping articles that addressed topics related to the evidence linkages. After review of the articles, 383 Consensus-based Evidence studies did not provide direct evidence, and were subsequently Consensus was obtained from multiple sources, including (1) sur- eliminated. A total of 288 articles contained direct linkage-related vey opinion from consultants who were selected based on their evidence. A complete bibliography used to develop these Guide- knowledge or expertise in central venous access, (2) survey opinions lines, organized by section, is available as Supplemental Digital solicited from active members of the ASA and SPA, (3) testimony Content 2, http://links.lww.com/ALN/A784. from attendees of publicly-held open forums at two national anes- Initially, each pertinent outcome reported in a study was classi- thesia meetings, (4) Internet commentary, and (5) task force opin- fied as supporting an evidence linkage, refuting a linkage, or equiv- ion and interpretation. The survey rate of return was 41.0% (n ϭ 55 ocal. The results were then summarized to obtain a directional of 134) for the consultants (table 2), 530 surveys were received from assessment for each evidence linkage before conducting formal active ASA members (table 3), and 251 surveys were received from meta-analyses. Literature pertaining to five evidence linkages con- active SPA members (table 4). tained enough studies with well-defined experimental designs and An additional survey was sent to the expert consultants asking statistical information sufficient for meta-analyses (table 1). These them to indicate which, if any, of the evidence linkages would linkages were (1) antimicrobial catheters, (2) silver sulfadiazine change their clinical practices if the Guidelines were instituted. The catheter coatings, (3) chlorhexidine and silver sulfadiazine catheter rate of return was 16% (n ϭ 22 of 134). The percentage of respond- coatings, (4) changing a catheter over a wire versus a new site, and ing consultants expecting no change associated with each linkage (5) ultrasound guidance for venipuncture. were as follows: (1) availability of a standardized equipment set ϭ General variance-based effect-size estimates or combined prob- 91.8%, (2) use of a trained assistant ϭ 83.7%, (3) use of a checklist or ability tests were obtained for continuous outcome measures, and protocol for placement and maintenance ϭ 75.5%, (4) use of bundles Mantel-Haenszel odds-ratios were obtained for dichotomous out- that include a checklist or protocol ϭ 87.8%, (5) intravenous antibiotic come measures. Two combined probability tests were employed as prophylaxis ϭ 93.9%, (6) aseptic preparation (e.g., hand washing, caps, follows: (1) the Fisher combined test, producing chi-square values masks) ϭ 98.0%, (8) skin preparation ϭ 98.0%, (9) selection of cath- based on logarithmic transformations of the reported P values from eters with antibiotic or antiseptic coatings/impregnation ϭ 89.8%, the independent studies, and (2) the Stouffer combined test, pro- (10) selection of catheter insertion site for prevention of infection ϭ viding weighted representation of the studies by weighting each of 100%, (11) catheter fixation methods ϭ 89.8%, (12) insertion site the standard normal deviates by the size of the sample. An odds- dressings ϭ 100%, (13) catheter maintenance ϭ 100%, (14) aseptic ratio procedure based on the Mantel-Haenszel method for combin- techniques using an existing central line for injection or aspiration ϭ ing study results using 2 ϫ 2 tables was used with outcome fre- 95.9%, (15) selection of catheter insertion site for prevention of me- quency information. An acceptable significance level was set at P Ͻ chanical trauma or injury ϭ 100%, (16) Trendelenburg versus supine 0.01 (one-tailed). Tests for heterogeneity of the independent stud- patient positioning for neck or chest venous access ϭ 100%, (17) ies were conducted to assure consistency among the study results. needle insertion and catheter placement ϭ 100%, (18) guidance DerSimonian-Laird random-effects odds ratios were obtained of needle, wire, and catheter placement ϭ 89.8%, (19) verification of when significant heterogeneity was found (P Ͻ 0.01). To control needle puncture and placement ϭ 98.0%, (20) management of trauma for potential publishing bias, a “fail-safe n ” value was calculated. or injury ϭ 100%. No search for unpublished studies was conducted, and no reliability Fifty-seven percent of the respondents indicated that the Guide- tests for locating research results were done. To be accepted as lines would have no effect on the amount of time spent on a typical significant findings, Mantel-Haenszel odds ratios must agree with case, and 43% indicated that there would be an increase of the combined test results whenever both types of data are assessed. In amount of time spent on a typical case with the implementation of the absence of Mantel-Haenszel odds-ratios, findings from both the these Guidelines. Seventy-four percent indicated that new equip- Fisher and weighted Stouffer combined tests must agree with each ment, supplies, or training would not be needed to implement the other to be acceptable as significant. Guidelines, and 78% indicated that implementation of the Guide- Interobserver agreement among Task Force members and two lines would not require changes in practice that would affect costs. methodologists was established by interrater reliability testing. Agreement levels using a kappa (␬) statistic for two-rater agreement Combined Sources of Evidence pairs were as follows: (1) type of study design, ␬ ϭ 0.70–1.00l; (2) Evidence for these Guidelines was formally collected from multiple type of analysis, ␬ ϭ 0.60–0.84; (3) evidence linkage assignment, sources, including randomized controlled trials, observational liter- ␬ ϭ 0.91–1.00; and (4) literature inclusion for database, ␬ ϭ 0.65– ature, surveys of expert consultants, and randomly selected samples 1.00. Three-rater chance-corrected agreement values were (1) study of ASA and SPA members. This information is summarized in table design, Sav ϭ 0.80, Var (Sav) ϭ 0.006; (2) type of analysis, Sav ϭ 5, with a brief description of each corresponding recommendation.

Anesthesiology 2012; 116:539–73 554 Practice Guidelines SPECIAL ARTICLES

Table 1. Meta-analysis Summary

Heterogeneity Weighted Fisher P Stouffer P Effect Odds Confidence P Effect Evidence Linkages N Chi-square Value Zc Value Size Ratio Interval Values Size Antibiotic-coated catheters vs. no coating Catheter colonization 5 0.35 0.23–0.55 ns Silver sulfadiazine catheter coating vs. no coating Catheter-related 5 0.70 0.45–1.10 ns bloodstream infection Chlorhexidine ϩ silver sulfadiazine catheter coating vs. no coating Catheter colonization 12 0.43 0.34–0.54 ns Catheter-related 12 0.70 0.47–1.03 ns bloodstream infection Changing a catheter over a wire vs. a new site Catheter colonization 5 1.18 0.66–2.09 ns Real-time ultrasound guidance vs.no ultrasound* Successful insertion/ 11 7.15† 1.33–18.27 0.005 cannulation First attempt success 5 3.24 1.93–5.45 ns Time to insertion 6 70.67 0.001 Ϫ7.15 0.001 Ϫ0.23 ns ns Arterial puncture 10 0.24 0.15–0.38 ns

* Findings represent studies addressing internal jugular access. † Random-effects odds ratio. ns ϭ P Ͼ 0.01.

Table 2. Consultant Survey Responses*

Percent Responding to Each Item Strongly Strongly N Agree Agree Equivocal Disagree Disagree I. Resource preparation 1. Central venous catheterization should be 54 92.6* 7.4 0.0 0.0 0.0 performed in a location that permits the use of aseptic techniques 2. A standardized equipment set should be 55 78.2* 16.4 5.4 0.0 0.0 available for central venous access 3. A trained assistant should be present during 54 33.3 29.6* 16.7 18.4 1.9 placement of a central venous catheter 4. A checklist or protocol should be used for the 54 59.3* 20.4 9.3 9.3 1.8 placement and maintenance of central venous catheters II. Prevention of infectious complications 5. Intravenous antibiotic prophylaxis should not be 55 43.6 32.7* 12.7 7.3 3.6 administered routinely 6. For immunocompromised patients and high-risk 55 23.6 36.4* 27.3 10.9 1.8 neonates, intravenous antibiotic prophylaxis may be administered on a case-by-case basis 7. The practitioner should use the following aseptic techniques in preparation for the placement of central venous catheters (check all that apply) 55 Percentage Hand washing 100.0 Sterile full-body drapes 87.3 Sterile gowns 100.0 Gloves 100.0 Caps 100.0 Masks covering both mouth and nose 100.0 (continued)

Anesthesiology 2012; 116:539–73 555 Practice Guidelines Practice Guidelines

Table 2. Continued

Percent Responding to Each Item Strongly Strongly N Agree Agree Equivocal Disagree Disagree 8. Chlorhexidine with alcohol should be used for 55 72.7* 27.3 0.0 0.0 0.0 skin preparation 9. Catheters coated with antibiotics or a 55 38.2 45.5* 16.3 0.0 0.0 combination of chlorhexidine and silver sulfadiazine may be used in selected patients based on infectious risk, cost, and anticipated duration of catheter use 10. Please indicate your preferred central venous catheter insertion site to minimize catheter- related risk of infection (check one) 55 Percentage Internal jugular 41.8 Subclavian 52.7 Femoral 0.0 No preference 5.5 11. Femoral catheterization should be avoided when 54 37.0 53.7* 3.7 3.7 1.9 possible to minimize the risk of infection 12. An insertion site should be selected that is not 53 71.7* 24.5 7.8 0.0 0.0 contaminated or potentially contaminated (e.g., burned or infected skin, inguinal area, adjacent to tracheostomy or open surgical wound) 13. Please indicate your preferred catheter fixation technique to minimize catheter-related risk of infection (check one) 54 Percentage Sutures 70.4 Staples 3.7 Tape 5.5 No preference 20.4 14. Transparent bio-occlusive dressings should be 55 52.7* 41.8 3.6 1.8 0.0 used to protect the site of central venous catheter insertion from infection 15. Dressings containing chlorhexidine may be used 55 20.0 34.6* 45.4 0.0 0.0 to reduce the risk of catheter-related infection 16. The duration of catheterization should be based 55 61.8* 30.9 0.0 7.3 0.0 on clinical need 17. The clinical need for keeping a catheter in place 53 90.6* 9.4 0.0 0.0 0.0 should be assessed daily 18. Catheters should be promptly removed when 54 88.9* 11.1 0.0 0.0 0.0 deemed no longer clinically necessary 19. The catheter site should be inspected daily for 54 88.9* 11.1 0.0 0.0 0.0 signs of infection 20. The catheter should be changed or removed 55 74.6* 20.0 3.6 1.8 0.0 when infection is suspected 21. When catheter-related infection is suspected, 55 70.9* 27.3 1.8 0.0 0.0 replacing the catheter using a new insertion site is preferable to changing the catheter over a guidewire 22. Catheter access ports should be wiped with an 55 69.1* 21.8 7.3 1.8 0.0 appropriate antiseptic before each access 23. Needleless catheter access ports may be used 55 30.9 47.3* 12.7 3.6 5.5 on a case-by-case basis 24. Central venous catheter stopcocks should be 54 81.5* 18.5 0.0 0.0 0.0 capped when not in use (continued)

Anesthesiology 2012; 116:539–73 556 Practice Guidelines SPECIAL ARTICLES

Table 2. Continued

Percent Responding to Each Item Strongly Strongly N Agree Agree Equivocal Disagree Disagree III. Prevention of mechanical trauma or injury 25. Please indicate your preferred central venous catheter insertion site to minimize catheter cannulation-related risk of injury or trauma (check one) 55 Percentage Internal jugular 81.8 Subclavian 9.1 Femoral 3.6 No preference 5.6 26. Please indicate your preferred central venous catheter insertion site to minimize catheter- related risk of thromboembolic injury or trauma (check one) 55 Percentage Internal jugular 76.4 Subclavian 7.3 Femoral 0.0 No preference 16.3 27. When clinically appropriate and feasible, central 54 51.9* 33.3 9.6 5.6 0.0 venous access in the neck or chest should be performed in the Trendelenburg position 28. Selection of catheter type (i.e., gauge, length, 55 49.1 38.2* 9.1 3.6 0.0 number of lumens) and composition (e.g., polyurethane, Teflon) should be based on the clinical situation and skill/experience of the operator 29. Selection of a modified Seldinger technique vs. 55 36.4 49.1* 5.4 7.3 1.8 a Seldinger technique should be based on the clinical situation and the skill/experience of the operator 30. The number of insertion attempts should be 55 45.5 32.7* 3.6 16.4 1.8 based on clinical judgment 31. The decision to place two catheters in a single 55 55.6* 40.0 3.6 1.8 0.0 vein should be made on a case-by-case basis 32. Ultrasound imaging (i.e., static) should be used 53 49.1 26.4* 11.3 9.4 3.8 in elective situations for pre-puncture identification of anatomy and vessel localization when the internal jugular vein is selected for cannulation 33. Ultrasound imaging (i.e., static) should be used in 55 12.7 18.2 32.7* 25.5 10.9 elective situations for pre-puncture identification of anatomy and vessel localization when the subclavian vein is selected for cannulation 34. Ultrasound imaging (i.e., static) should be used 55 18.2 32.7* 21.8 23.6 3.6 in elective situations for pre-puncture identification of anatomy and vessel localization when the femoral vein is selected for cannulation 35. When available, real-time ultrasound should be 54 44.4 33.3* 13.0 9.3 0.0 used for guidance during venous access when the internal jugular vein is selected for cannulation 36. When available, real-time ultrasound should be 53 11.3 17.0 37.7* 28.3 5.7 used for guidance during venous access when the subclavian vein is selected for cannulation (continued)

Anesthesiology 2012; 116:539–73 557 Practice Guidelines Practice Guidelines

Table 2. Continued

Percent Responding to Each Item Strongly Strongly N Agree Agree Equivocal Disagree Disagree 37. When available, real-time ultrasound should be 54 14.8 35.2* 33.3 14.8 1.9 used for guidance during venous access when the femoral vein is selected for cannulation 38. Before insertion of a dilator or large bore 54 57.4* 25.9 7.4 9.3 0.0 catheter over a wire, venous access should be confirmed for the catheter or thin-wall needle that accesses the vein 39. Before insertion of a dilator or large bore 55 29.1 29.1* 25.5 12.7 3.6 catheter over a wire, venous access should be confirmed for the wire that subsequently resides in the vein after traveling through a catheter or thin-wall needle 40. When feasible, both the location of the catheter 55 25.4 38.2* 18.2 15.6 3.6 or thin-wall needle and wire should be confirmed 41. A chest radiograph should be performed to 55 30.9 41.8* 9.1 14.5 3.6 confirm the location of the catheter tip as soon after catheterization as clinically appropriate 42. For central venous catheters placed in the 55 47.3 50.9* 0.0 1.8 0.0 operating room, a confirmatory chest radiograph may be performed in the early postoperative period 43. If a chest radiograph will be deferred to the 55 56.4* 30.9 5.4 7.3 0.0 postoperative period, pressure/waveform analysis, blood gas analysis, ultrasound or fluoroscopy should be used to confirm venous positioning of the catheter before use IV. Management of arterial trauma or injury arising from central venous 44. When unintended cannulation of an arterial 55 45.4 36.4* 7.3 9.1 1.8 vessel with a large bore catheter occurs, the catheter should be left in place and a general or vascular surgeon should be consulted

*Nϭ number of consultants who responded to each item. An asterisk next to a percentage score indicates the median.

Anesthesiology 2012; 116:539–73 558 Practice Guidelines SPECIAL ARTICLES

Table 3. ASA Member Survey Responses*

Percent Responding to Each Item Strongly Strongly N Agree Agree Equivocal Disagree Disagree I. Resource preparation 1. Central venous catheterization should be 529 78.1* 19.1 2.1 0.8 0.0 performed in a location that permits the use of aseptic techniques 2. A standardized equipment set should be 530 64.5* 30.0 4.2 0.9 0.4 available for central venous access 3. A trained assistant should be present during 526 24.1 35.6* 24.0 13.1 3.2 placement of a central venous catheter 4. A checklist or protocol should be used for 528 35.6 37.5* 16.3 8.9 1.7 The placement and maintenance of central venous catheters II. Prevention of infectious complications 5. Intravenous antibiotic prophylaxis should 526 29.7 44.5* 16.9 7.0 1.9 not be administered routinely 6. For immunocompromised patients and 523 25.0 54.1* 15.9 4.2 0.8 high-risk neonates, intravenous antibiotic prophylaxis may be administered on a case-by-case basis 7. The practitioner should use the following aseptic techniques in preparation for the placement of central venous catheters (check all that apply) 524 Percentage Hand washing 96.0 Sterile full-body drapes 73.8 Sterile gowns 87.8 Gloves 100.0 Caps 94.7 Masks covering both mouth and nose 98.1 8. Chlorhexidine with alcohol should be used 522 57.3* 34.1 7.8 0.8 0.0 for skin preparation 9. Catheters coated with antibiotics or a 526 24.3 54.8* 19.2 1.7 0.0 combination of chlorhexidine and silver sulfadiazine may be used in selected patients based on infectious risk, cost, and anticipated duration of catheter use 10. Please indicate your preferred central venous catheter insertion site to minimize catheter- related risk of infection (check one) 524 Percentage Internal jugular 51.3 Subclavian 44.3 Femoral 0.0 No preference 4.4 11. Femoral catheterization should be avoided 525 33.9 49.7* 9.3 4.7 2.3 when possible to minimize the risk of infection 12. An insertion site should be selected that is 523 58.9* 37.9 2.5 0.7 0.0 not contaminated or potentially contaminated (e.g., burned or infected skin, inguinal area, adjacent to tracheostomy or open surgical wound) (continued)

Anesthesiology 2012; 116:539–73 559 Practice Guidelines Practice Guidelines

Table 3. Continued

Percent Responding to Each Item Strongly Strongly N Agree Agree Equivocal Disagree Disagree 13. Please indicate your preferred catheter fixation technique to minimize catheter- related risk of infection (check one) 524 Percentage Sutures 80.2 Staples 5.7 Tape 3.6 No preference 10.5 14. Transparent bio-occlusive dressings 522 46.9 44.4* 6.5 1.3 0.8 should be used to protect the site of central venous catheter insertion from infection 15. Dressings containing chlorhexidine may be 525 18.7 37.9* 41.3 1.9 0.2 used to reduce the risk of catheter-related infection 16. The duration of catheterization should be 523 49.5 44.5* 3.1 2.5 0.4 based on clinical need 17. The clinical need for keeping a catheter in 523 65.8* 32.5 1.3 0.4 0.0 place should be assessed daily 18. Catheters should be promptly removed 521 78.7* 20.9 0.4 0.0 0.0 when deemed no longer clinically necessary 19. The catheter site should be inspected 521 79.1* 19.6 1.1 0.2 0.0 daily for signs of infection 20. The catheter should be changed or 524 72.7* 24.4 2.5 0.2 0.2 removed when infection is suspected 21. When catheter-related infection is 525 64.8* 30.7 3.8 0.8 0.0 suspected, replacing the catheter using a new insertion site is preferable to changing the catheter over a guidewire 22. Catheter access ports should be wiped 522 64.6* 31.0 3.4 1.0 0.0 with an appropriate antiseptic before each access 23. Needleless catheter access ports may be 522 33.9 51.3* 12.3 1.7 0.8 used on a case-by-case basis 24. Central venous catheter stopcocks should 527 70.6* 26.2 2.6 0.6 0.0 be capped when not in use III. Prevention of mechanical trauma or injury 25. Please indicate your preferred central venous catheter insertion site to minimize catheter cannulation-related risk of injury or trauma (check one) 525 Percentage Internal jugular 79.4 Subclavian 10.7 Femoral 2.7 No preference 7.2 26. Please indicate your preferred central venous catheter insertion site to minimize catheter-related risk of thromboembolic injury or trauma (check one) 525 Percentage Internal jugular 67.6 Subclavian 12.8 Femoral 1.9 No preference 17.7 (continued)

Anesthesiology 2012; 116:539–73 560 Practice Guidelines SPECIAL ARTICLES

Table 3. Continued

Percent Responding to Each Item Strongly Strongly N Agree Agree Equivocal Disagree Disagree 27. When clinically appropriate and feasible, 528 57.0* 37.7 3.0 1.9 0.4 central venous access in the neck or chest should be performed in the Trendelenburg position 28. Selection of catheter type (i.e., gauge, 530 52.1* 38.1 6.2 3.4 0.0 length, number of lumens) and composition (e.g., polyurethane, Teflon) should be based on the clinical situation and skill/experience of the operator 29. Selection of a modified Seldinger 531 47.8 36.9* 9.8 4.7 0.8 technique vs. a Seldinger technique should be based on the clinical situation and the skill/experience of the operator 30. The number of insertion attempts should 528 47.3 43.6* 4.2 3.8 1.1 be based on clinical judgment 31. The decision to place two catheters in a 527 45.9 36.2* 12.1 4.4 1.3 single vein should be made on a case-by- case basis 32. Ultrasound imaging (i.e., static) should be 526 28.9 25.1* 21.3 18.8 5.9 used in elective situations for pre-puncture identification of anatomy and vessel localization when the internal jugular vein is selected for cannulation 33. Ultrasound imaging (i.e., static) should be 528 9.7 14.2 41.5* 26.5 8.1 used in elective situations for pre-puncture identification of anatomy and vessel localization when the subclavian vein is selected for cannulation 34. Ultrasound imaging (i.e., static) should be 527 11.9 29.8 30.6* 21.4 6.3 used in elective situations for pre-puncture identification of anatomy and vessel localization when the femoral vein is selected for cannulation 35. When available, real time ultrasound 525 24.0 24.2 23.2* 21.5 7.1 should be used for guidance during venous access when the internal jugular vein is selected for cannulation 36. When available, real time ultrasound 530 8.1 13.4 42.1* 27.9 8.5 should be used for guidance during venous access when the subclavian vein is selected for cannulation 37. When available, real-time ultrasound 528 13.5 23.5 31.4* 25.0 6.6 should be used for guidance during venous access when the femoral vein is selected for cannulation 38. Before insertion of a dilator or large bore 524 52.9* 32.1 8.4 6.3 0.4 catheter over a wire, venous access should be confirmed for the catheter or thin-wall needle that accesses the vein 39. Before insertion of a dilator or large bore 524 24.0 25.4 25.6* 22.9 2.1 catheter over a wire, venous access should be confirmed for the wire that subsequently resides in the vein after traveling through a catheter or thin-wall needle (continued)

Anesthesiology 2012; 116:539–73 561 Practice Guidelines Practice Guidelines

Table 3. Continued

Percent Responding to Each Item Strongly Strongly N Agree Agree Equivocal Disagree Disagree 40. When feasible, both the location of the 526 23.8 32.5* 22.1 19.4 2.3 catheter or thin-wall needle and wire should be confirmed 41. A chest radiograph should be performed 525 39.8 45.5* 7.1 7.0 0.6 to confirm the location of the catheter tip as soon following catheterization as clinically appropriate 42. For central venous catheters placed in the 524 46.8 48.1* 2.5 1.9 0.8 operating room, a confirmatory chest radiograph may be performed in the early postoperative period 43. If a chest radiograph will be deferred to 527 33.0 35.3* 12.7 16.7 2.3 the postoperative period, pressure/waveform analysis, blood gas analysis, ultrasound or fluoroscopy should be used to confirm venous positioning of the catheter before use IV. Management of arterial trauma or injury arising from central venous 44. When unintended cannulation of an 526 28.5 35.6* 16.3 17.9 1.7 arterial vessel with a large bore catheter occurs, the catheter should be left in place and a general or vascular surgeon should be consulted

* Number of ASA members who responded to each item. An asterisk next to a percentage score indicates the median.

Anesthesiology 2012; 116:539–73 562 Practice Guidelines SPECIAL ARTICLES

Table 4. SPA Member Survey Responses*

Percent Responding to Each Item Strongly Strongly N Agree Agree Equivocal Disagree Disagree

1. A chlorhexidine-containing solution should 250 17.2 26.0 31.6* 17.2 8.0 be used for skin preparation in neonates† 2. A chlorhexidine-containing solution should 248 46.0 40.3* 11.3 2.4 0.0 be used for skin preparation in infants‡ 3. A chlorhexidine-containing solution should 249 62.7* 30.9 5.2 1.2 0.0 be used for skin preparation in children§ 4. Dressings containing chlorhexidine may be 243 7.0 14.0 52.2* 20.2 6.6 used in neonates 5. Dressings containing chlorhexidine may be 249 22.5 36.6* 35.3 4.8 0.8 used in infants 6. Dressings containing chlorhexidine may be 249 38.6 35.3* 24.5 1.2 0.4 used in children 7. When unintended cannulation of an arterial vessel with a large bore catheter occurs in neonates (check one) 244 Percentage The catheter should be left in place5 54.9 The catheter may be nonsurgically 45.1 removed 8. When unintended cannulation of an arterial vessel with a large-bore catheter occurs in infants (check one) 249 Percentage The catheter should be left in place 43.8 The catheter may be nonsurgically 56.2 removed 9. When unintended cannulation of an arterial vessel with a large bore catheter occurs in children (check one) 244 Percentage The catheter should be left in place 30.0 The catheter may be nonsurgically 70.0 removed

* Number of SPA members who responded to each item. An asterisk beside a percentage score indicates the median response. † Younger than 44 gestational weeks. ‡ Younger than 2 yr. § 2–16 yr of age. The complete wording of the response category is: The catheter should be left in place and a general surgeon, vascular surgeon, or interventional radiologist should be immediately consulted before deciding on whether to remove the catheter, either surgically or nonsurgically. # The complete wording of the response category is: The catheter may be nonsurgically removed without consulting a general surgeon, vascular surgeon, or interventional radiologist.

Anesthesiology 2012; 116:539–73 563 Practice Guidelines Practice Guidelines

Table 5. Evidence Summary*

Evidence Consultant ASA Member SPA Member Guideline Interventions Category1 Survey2 Survey2 Survey2 Recommendation

I. Resource preparation Catheterization in environment D Strongly agree Strongly agree Should be performed that permits use of aseptic techniques Standardized equipment set D Strongly agree Strongly agree Should be available An assistant D Agree (trained) Agree (trained) Should be used A checklist or protocol B23 Strongly agree Agree Should be used II. Prevention of infectious complications Intravenous antibiotic prophylaxis Prophylactic intravenous D Agree Agree Should not be routinely antibiotics should administered not be administered routinely Prophylactic intravenous A24 Agree Agree Administer on a case-by- antibiotics should be case basis administered to immunocompromised patients and high- risk neonates Aseptic techniques and barrier precautions: Maximal barrier vs. gloves C25,6 and small drape only ЉBundledЉ elements: B23 hand-washing, sterile full body drapes, sterile, gloves, caps, and masks Specific activities: Hand washing D 100% agreement 96% agreement Use Sterile full-body drape D 87% agreement 74% agreement Use Sterile gown D 100% agreement 88% agreement Use Sterile gloves D 100% agreement 100% agreement Use Caps D 100% agreement 95% agreement Use Masks covering both D 100% agreement 98% agreement Use mouth and nose Skin preparation: Solutions containing chlorhexidine: Chlorhexidine with D Strongly agree Strongly agree Should be used for adults, alcohol (patient age infants and children not specified) Antiseptic solutions containing chlorhexidine for: Neonates D Equivocal Should be based on clinical judgment and Institutional protocol Infants D Agree Should be used Children D Strongly agree Should be used Solutions containing alcohol: Chlorhexidine without C25,7 alcohol vs. povidone-iodine without alcohol Chlorhexidine with D alcohol vs. Povidone-iodine with alcohol Skin preparation solutions with vs. without alcohol: Chlorhexidine D (continued)

Anesthesiology 2012; 116:539–73 564 Practice Guidelines SPECIAL ARTICLES

Table 5. Continued

Evidence Consultant ASA Member SPA Member Guideline Interventions Category1 Survey2 Survey2 Survey2 Recommendation

Povidone-iodine A35/C28 Skin preparation solutions Use unless contraindicated containing alcohol Catheters containing antimicrobial agents: Antibiotic-coated A15 Agree (selected Agree (selected Should be used for catheters pts) pts) selected patients Silver-impregnated C13/C25 No recommendation catheters Chlorhexidine and silver A15/B39/C13 Agree (selected Agree (selected Should be used for sulfadiazine coated pts) pts) selected patients catheters Selection of catheter insertion site: Internal jugular vs. C23,5/C33,5 Majority prefer Majority prefer Site selection should be subclavian subclavian site internal jugular based on clinical need to site minimize risk of catheter- related infection Subclavian vs. femoral A35/C24 Agree (avoid Agree (avoid Site selection should be femoral) femoral) based on clinical need. In adults, upper body site should be considered to minimize risk of infection Catheter fixation: Risk of catheter-related D Majority prefer Majority prefer Should be determined on a infections with suture suture local or institutional basis suture, staple, tape Catheter insertion site dressings: Transparent bio-occlusive D Strongly agree Strongly agree Should be used Chlorhexidine sponge C23,5 Agree Agree May be used unless dressings (patient contraindicated age not specified) Chlorhexidine- A310 Should be based on clinical impregnated judgment and institutional transparent protocol dressings for neonates Chlorhexidine sponge dressings For neonates Equivocal Should be based on clinical judgment and institutional protocol For infants Agree May be used, unless contraindicated For children Agree May be used, unless contraindicated Silver-impregnated C25 No recommendation transparent dressings Catheter maintenance: Duration of catheterization B24,5 related to higher colonization/infection rates Duration of catheterization Strongly agree Agree Duration should be based should be based on on clinical need clinical need Specific time intervals D between insertion site inspections Catheter change interval C25 3-days vs. 7-days Daily assessment of Strongly agree Strongly agree Clinical need for keeping clinical need for catheter in place should continuing be assessed daily catheterization (continued)

Anesthesiology 2012; 116:539–73 565 Practice Guidelines Practice Guidelines

Table 5. Continued

Evidence Consultant ASA Member SPA Member Guideline Interventions Category1 Survey2 Survey2 Survey2 Recommendation

Conduct daily catheter Strongly agree Strongly agree Catheter insertion site site inspections should be inspected daily for signs of infection Change or remove Strongly agree Strongly agree Catheter should be catheter when changed or removed infection is when Catheter insertion suspected site infection is suspected When catheter-related C15 Strongly agree Strongly agree When catheter-related infection is (Suspected (Suspected infection is suspected, suspected, replace infection) infection) replacing the catheter catheter using new using a new insertion site insertion site vs. is preferred catheter change over a guidewire Promptly remove catheter Strongly agree Strongly agree Promptly remove catheter when deemed no when deemed no longer longer clinically clinically necessary necessary Aseptic techniques using an existing central venous catheter: Wipe port with an D Strongly agree Strongly agree Catheter access ports appropriate should be wiped with an antiseptic before appropriate antiseptic access before each access Cap stopcocks or access Strongly agree Strongly agree Central venous catheter ports when not in stopcocks or access use ports should be capped when not in use Needleless catheter connectors/access ports vs. standard caps Needleless catheter A211/C23 Agree Agree Needless catheter access connectors/ports vs. (case-by case (case-by case ports may be used on a standard caps basis) basis) case-by-case basis III. Prevention of mechanical trauma or injury Selection of catheter insertion site: Internal jugular vs. C213,14,15,16/C317 subclavian Subclavian vs. femoral A312 Preferred catheter Majority prefer Majority prefer Insertion site selection insertion site internal jugular internal jugular should be based on clinical need and practitioner judgment, experience and skill. In adults, selection of an upper body insertion site should be considered to minimize the risk of thromboembolic injury or trauma Positioning the patient for needle insertion and catheter placement: Trendelenburg vs. normal B218 Strongly agree Strongly agree When clinically appropriate supine and feasible, central venous access in the neck or chest should be performed with the patient in the Trendelenburg position (continued)

Anesthesiology 2012; 116:539–73 566 Practice Guidelines SPECIAL ARTICLES

Table 5. Continued

Evidence Consultant ASA Member SPA Member Guideline Interventions Category1 Survey2 Survey2 Survey2 Recommendation

Needle insertion, wire and catheter placement: Selection of catheter size Strongly agree Strongly agree Should be based on the and type clinical situation and the skill and experience of the practitioner; selection of the smallest size catheter appropriate for the clinical situation should be considered Large-bore catheters B319 Select the smallest size associated with catheter appropriate for unintentional arterial the clinical situation cannulation Modified Seldinger vs. D Agree Agree Should be based on the Seldinger technique clinical situation and the skill and experience of the operator; the decision to use a catheter-over-the- needle (modified Seldinger) technique or a thin-wall needle (Seldinger) technique should be based at least in part on the method used to confirm that the wire resides in the vein before a dilator or large-bore catheter is threaded Limiting the number of D Agree Agree Should be based on clinical insertion attempts judgment Introducing two catheters B220/C313,15 Strongly agree Agree Should be decided on a in the same central (case-by-case) (case-by-case) case-by-case basis vein Guidance of needle placement in elective situations: Static ultrasound for preprocedural vessel localization vs. landmark approach: Internal jugular vein A321/C222 Agree Agree Use access (elective (elective situations) situations) Subclavian vein access C222 Equivocal Equivocal May be used (elective (elective situations) situations) Femoral vein access D Agree Equivocal May be used (elective (elective situations) situations) Real-time ultrasound for guiding needle vs. landmark approach: Internal jugular vein A113,21,22,23/A224 Agree Equivocal Use access (when available) (when available) Subclavian vein access A224/A313,15,16,23 Equivocal Equivocal May be used (when available) (when available) Femoral vein access A321,24 Agree Equivocal May be used (when available) (when available) (continued)

Anesthesiology 2012; 116:539–73 567 Practice Guidelines Practice Guidelines

Table 5. Continued

Evidence Consultant ASA Member SPA Member Guideline Interventions Category1 Survey2 Survey2 Survey2 Recommendation

Verification of venous access: Confirm that catheter or Strongly agree Strongly agree Confirm venous access thin-wall needle is in after insertion of catheter a vein that went over the needle or a thin-wall needle Ultrasound D An identified method Manometry B213 An identified method Pressure waveform D An identified method analysis Venous blood gas D An identified method Absence of pulsatility, D Should not be relied upon blood color to confirm venous access (based on Task Force opinion) Confirm venous residence Agree Equivocal When using the thin-wall of the wire needle technique, confirm venous residence of the wire after the wire is threaded Ultrasound B225 An identified method Transesophageal B325 An identified method ultrasound Continuous D An identified method (based electrocardiography on Task Force opinion) Fluoroscopy D An identified method (based on Task Force opinion) Confirm both the location Agree Agree Confirm if there is any of the catheter or (when feasible) (when feasible) uncertainty that the thin-wall needle and catheter or wire resides wire in the vein Verification of catheter placement: Confirmation of final Confirm the final position of position of tip of the catheter tip as soon as catheter clinically appropriate (based on Task Force opinion) Fluoroscopy B226 Strongly agree Agree An identified method Chest radiograph B226 Agree Agree An identified method Continuous A226 An identified method electrocardiography Unintended cannulation of an arterial vessel with a large bore catheter: Leave catheter in place B327 Agree Agree For adults, the catheter (patient age not should be left in place specified) and a general surgeon, a vascular surgeon, or an interventional radiologist should be immediately consulted For neonates Majority prefer Should be based on clinical leaving in place judgment For infants Majority prefer Should be based on clinical nonsurgical judgment removal (continued)

Anesthesiology 2012; 116:539–73 568 Practice Guidelines SPECIAL ARTICLES

Table 5. Continued

Evidence Consultant ASA Member SPA Member Guideline Interventions Category1 Survey2 Survey2 Survey2 Recommendation

For children Majority prefer Should be based on clinical Nonsurgical removal judgment

* Categories of evidence for literature: Category A: Supportive Literature. Randomized controlled trials report statistically significant (P Ͻ 0.01) differences between clinical interventions for a specified clinical outcome. Level 1: The literature contains multiple randomized controlled trials, and aggregated findings are supported by meta-analysis. † Level 2: The literature contains multiple randomized controlled trials, but the number of studies is insufficient to conduct a viable meta-analysis for the purpose of these Guidelines. Level 3: The literature contains a single randomized controlled trial. Category B: Suggestive Literature. Information from observational studies permits inference of beneficial or harmful relationships among clinical interventions and clinical outcomes. Level 1: The literature contains observational comparisons (e.g., cohort, case-control research designs) of clinical interventions or conditions and indicates statistically significant differences between clinical interventions for a specified clinical outcome. Level 2: The literature contains noncomparative observational studies with associative (e.g., relative risk, correlation) or descriptive statistics. Level 3: The literature contains case reports. Category C: Equivocal Literature. The literature cannot determine whether there are beneficial or harmful relationships among clinical interventions and clinical outcomes. Level 1: Meta-analysis did not find significant differences (P Ͼ 0.01) among groups or conditions. Level 2: The number of studies is insufficient to conduct meta-analysis, and (1) randomized controlled trials have not found significant differences among groups or conditions or (2) randomized controlled trials report inconsistent findings. Level 3: Observational studies report inconsistent findings or do not permit inference of beneficial or harmful relationships. Category D: Insufficient Evidence from Literature. The lack of scientific evidence in the literature is described by the following terms. Inadequate: The available literature cannot be used to assess relationships among clinical interventions and clinical outcomes. The literature either does not meet the criteria for content as defined in the “Focus” of the Guidelines or does not permit a clear interpretation of findings due to methodological concerns (e.g., confounding in study design or implementation). Silent: No identified studies address the specified relationships among interventions and outcomes. 1 All meta-analyses are conducted by the ASA methodology group. Meta-analyses from other sources are reviewed but not included as evidence in this document. 2 Survey data recorded on a 5-point scale: strongly agree - agree - equivocal - disagree - strongly disagree; reported findings represent the median survey response. 3 Catheter-related bloodstream infection. 4 Catheter-related infection and sepsis. 5 Catheter colonization. 6 Catheter-related septicemia. 7 Catheter-related bacteremia. 8 Catheter-related infection and clinical signs of infection. 9 Anaphylactic shock. 10 Localized contact dermatitis. 11 Microbial contamination of stopcock entry ports. 12 Thrombotic complications. 13 Arterial puncture. 14 Deep vein thrombosis. 15 Hematoma. 16 Successful venipuncture. 17 Pneumothorax, hemothorax, or arrhythmia. 18 Diameter and cross sectional area of right internal jugular vein for patients older than 6 yr. 19 Severe injury (e.g., hemorrhage, hematoma, pseudoaneurysm, arteriovenous fistula, arterial dissection, neurologic injury including stroke, and severe or lethal airway obstruction) may occur. 20 Dysrhythmia. 21 First insertion attempt success rate. 22 Overall successful cannulation rate. 23 Access time. 24 Number of insertion attempts. 25 Confirmation of venous placement of wire. 26 Identifying the position of the catheter tip. 27 Fewer severe complications in adult patients.

References‡‡ ated bloodstream infection rates decline after bundles and checklists. Pediatrics 2011; 127:436–44 1. Chua C, Wisniewski T, Ramos A, Schlepp M, Fildes JJ, Kuhls DA: Multidisciplinary trauma intensive care unit check- 7. Warren DK, Cosgrove SE, Diekema DJ, Zuccotti G, Climo list: Impact on infection rates. J Trauma Nurs 2010; 17: MW, Bolon MK, Tokars JI, Noskin GA, Wong ES, Sepkowitz 163–6 KA, Herwaldt LA, Perl TM, Solomon SL, Fraser VJ, Preven- sion Epicenter Program: A multicenter intervention to pre- 2. Berenholtz SM, Pronovost PJ, Lipsett PA, Hobson D, Earsing vent catheter-associated bloodstream infections. Infect K, Farley JE, Milanovich S, Garrett-Mayer E, Winters BD, Control Hosp Epidemiol 2006; 27:662–9 Rubin HR, Dorman T, Perl TM: Eliminating catheter-related bloodstream infections in the intensive care unit. Crit Care 8. Boorman D: Today’s electronic checklists reduce likelihood Med 2004; 32:2014–20 of crew errors and help prevent mishaps. ICAO J 2001: 17–20-36 3. Higuera F, Rosenthal VD, Duarte P, Ruiz J, Franco G, Safdar N: The effect of process control on the incidence of central venous 9. Karl R: Briefings, checklists, geese, and surgical safety. Ann catheter-associated bloodstream infections and mortality in inten- Surg Oncol 2010; 17:8–11 sive care units in Mexico. Crit Care Med 2005; 33:2022–7 10. Spafford PS, Sinkin RA, Cox C, Reubens L, Powell KR: 4. Miller MR, Griswold M, Harris JM 2nd, Yenokyan G, Huskins Prevention of central venous catheter-related coagulase- WC, Moss M, Rice TB, Ridling D, Campbell D, Margolis P, negative staphylococcal sepsis in neonates. J Pediatr 1994; Muething S, Brilli RJ: Decreasing PICU catheter-associated 125:259–63 bloodstream infections: NACHRI’s quality transformation 11. Vassilomanolakis M, Plataniotis G, Koumakis G, Hajichras- efforts. Pediatrics 2010; 125:206–13 tou H, Skouteri H, Dova H, Efremidis AP: Central venous 5. Pronovost P, Needham D, Berenholtz S, Sinopoli D, Chu H, catheter-related infections after bone marrow transplanta- Cosgrove S, Sexton B, Hyzy R, Welsh R, Roth G, Bander J, tion in patients with malignancies: A prospective study of Kepros J, Goeschel C: An intervention to decrease catheter- short-course vancomycin prophylaxis. Bone Marrow Trans- related bloodstream infections in the ICU. N Engl J Med plant 1995; 15:77–80 2006; 355:2725–32 12. Raad II, Hohn DC, Gilbreath BJ, Suleiman N, Hill LA, Bruso 6. Schulman J, Stricof R, Stevens TP, Horgan M, Gase K, PA, Marts K, Mansfield PF, Bodey GP: Prevention of central Holzman IR, Koppel RI, Nafday S, Gibbs K, Angert R, Sim- venous catheter-related infections by using maximal sterile monds A, Furdon SA, Saiman L, New York State Regional barrier precautions during insertion. Infect Control Hosp Perinatal Care Centers: Statewide NICU central-line-associ- Epidemiol 1994; 15:231–8 13. Maki DG, Ringer M, Alvarado CJ: Prospective randomised ‡‡ A complete bibliography used to develop these Guidelines, trial of povidone-iodine, alcohol, and chlorhexidine for arranged alphabetically by author, is available as Supplemental prevention of infection associated with central venous and Digital Content 1, http://links.lww.com/ALN/A783. arterial catheters. Lancet 1991; 338:339–43

Anesthesiology 2012; 116:539–73 569 Practice Guidelines Practice Guidelines

14. Parienti JJ, du Cheyron D, Ramakers M, Malbruny B, Leclercq 29. George SJ, Vuddamalay P, Boscoe MJ: Antiseptic-impreg- R, Le Coutour X, Charbonneau P, Members of the NACRE nated central venous catheters reduce the incidence of Study Group: Alcoholic providone-iodine to prevent central bacterial colonization and associated infection in immuno- venous catheter colonization: A randomized unit crossover compromised transplant patients. Eur J Anaesthesiol 1997; study. Crit Care Med 2004; 32:708–13 14:428–31 15. Bach A, Darby D, Bo¨ttiger B, Bo¨hrer H, Motsch J, Martin E: 30. Hannan M, Juste RN, Umasanker S, Glendenning A, Night- Retention of the antibiotic teicoplanin on a hydromer- ingale C, Azadian B, Soni N: Antiseptic-bonded central ve- coated central venous catheter to prevent bacterial coloni- nous catheters and bacterial colonisation. Anaesthesia zation in postoperative surgical patients. Intensive Care 1999; 54:868–72 Med 1996; 22:1066–9 31. Heard SO, Wagle M, Vijayakumar E, McLean S, Bruegge- 16. Kamal GD, Pfaller MA, Rempe LE, Jebson PJ: Reduced mann A, Napolitano LM, Edwards LP, O’Connell FM, Puyana intravascular catheter infection by antibiotic bonding: A JC, Doern GV: Influence of triple-lumen central venous prospective, randomized, controlled trial. JAMA 1991; 265: catheters coated with chlorhexidine and silver sulfadiazine 2364–8 on the incidence of catheter-related bacteremia. Arch In- 17. Len C, Ruiz-Santana S, Rello J, de la Torre MV, Valls J, Alvarez- tern Med 1998; 158:81–7 Lerma F, Sierra R, Saavedra P, Alvarez-Salgado F: Benefits of 32. Maki DG, Stolz SM, Wheeler S, Mermel LA: Prevention of minocycline and rifampin-impregnated central venous catheters: central venous catheter-related bloodstream infection by use A prospective, randomized, double-blind, controlled, multicenter of an antiseptic-impregnated catheter. A randomized, con- trial. Intensive Care Med 2004; 30:1891–9 trolled trial. Ann Intern Med 1997; 127:257–66 18. Thornton J, Todd NJ, Webster NR: Central venous line 33. Ostendorf T, Meinhold A, Harter C, Salwender H, Egerer G, sepsis in the intensive care unit: A study comparing antibi- Geiss HK, Ho AD, Goldschmidt H: Chlorhexidine and silver- otic coated catheters with plain catheters. Anaesthesia sulfadiazine coated central venous catheters in haematological 1996; 51:1018–20 patients–a double-blind, randomised, prospective, controlled 19. Raad I, Darouiche R, Dupuis J, Abi-Said D, Gabrielli A, trial. Support Care Cancer 2005; 13:993–1000 Hachem R, Wall M, Harris R, Jones J, Buzaid A, Robertson C, 34. Rupp ME, Lisco SJ, Lipsett PA, Perl TM, Keating K, Civetta Shenaq S, Curling P, Burke T, Ericsson C: Central venous JM, Mermel LA, Lee D, Dellinger EP, Donahoe M, Giles D, catheters coated with minocycline and rifampin for the Pfaller MA, Maki DG, Sherertz R: Effect of a second-gener- prevention of catheter-related colonization and blood- ation venous catheter impregnated with chlorhexidine and stream infections: A randomized, double-blind trial. The silver sulfadiazine on central catheter-related infections: A Texas Medical Center Catheter Study Group. Ann Intern randomized, controlled trial. Ann Intern Med 2005; 143: Med 1997;127:267–74 570–80 20. Bong JJ, Kite P, Wilco MH, McMahon MJ: Prevention of 35. Tennenberg S, Lieser M, McCurdy B, Boomer G, Howington catheter related bloodstream infection by silver iontopho- E, Newman C, Wolf I: A prospective randomized trial of an retic central venous catheters: A randomised controlled antibiotic- and antiseptic-coated central venous catheter in trial. J Clin Pathol 2003; 56:731–5 the prevention of catheter-related infections. Arch Surg 21. Boswald M, Lugauer S, Regenfus A, Braun GG, Martus P, 1997; 132:1348–51 Geis C, Scharf J, Bechert T, Greil J, Guggenbichler J-P: 36. van Heerden PV, Webb SAR, Fong S, Golledge CL, Roberts Reduced rates of catheter-associated infection by use of a BL: Central venous catheters revisited: Infection rates and new silver-impregnated central venous catheter. Infection an assessment of the new fibrin analysing system brush. 1999;27:56–60 Anaesth Intens Care 1996; 24:330–3 22. Hagau N, Studnicska D, Gavrus RL, Csipak G, Hagau R, Slav- 37. Logghe C, Van Ossel C, D’Hoore W, Ezzedine H, Wauters G, covici AV: Central venous catheter colonization and catheter- Haxhe JJ: Evaluation of chlorhexidine and silver-sulfadiazine related bloodstream infections in critically ill patients: A com- impregnated central venous catheters for the prevention of parison between standard and silver-integrated catheters. Eur bloodstream infection in leukaemic patients: A randomized J Anaesthesiol 2009; 26:752–8 controlled trial. J Hosp Infect 1997; 37:145–56 23. Harter C, Salwender HJ, Bach A, Egerer G, Goldschmidt H, Ho 38. Pemberton LB, Ross V, Cuddy P, Kremer H, Fessler T, AD: Catheter-related infection and thrombosis of the internal McGurk E: No difference in catheter sepsis between stan- jugular vein in hematologic-oncologic patients undergoing dard and antiseptic central venous catheters. A prospective chemotherapy: A prospective comparison of silver-coated and randomized trial. Arch Surg 1996; 131:986–9 uncoated catheters. Cancer 2002; 94:245–51 39. Oda T, Hamasaki J, Kanda N, Mikami K: Anaphylactic shock 24. Kalfon P, de Vaumas C, Samba D, Boulet E, Lefrant JY, induced by an antiseptic-coated central venous catheter. Eyraud D, Lherm T, Santoli F, Naija W, Riou B: Comparison ANESTHESIOLOGY 1997; 87:1242–4 of silver-impregnated with standard multi-lumen central 40. Stephens R, Mythen M, Kallis P, Davies DW, Egner W, Rickards venous catheters in critically ill patients. Crit Care Med A: Two episodes of life-threatening anaphylaxis in the same 2007; 35:1032–9 patient to a chlorhexidine-sulphadiazine-coated central ve- 25. Bach A, Schmidt H, Bttiger B, Schreiber B, Bhrer H, Motsch J, nous catheter. Br J Anaesth 2001; 87:306–8 Martin E, Sonntag HG: Retention of antibacterial activity and 41. Terazawa E, Shimonaka H, Nagase K, Masue T, Dohi S: bacterial colonization of antiseptic-bonded central venous Severe anaphylactic reaction due to a chlorhexidine-im- catheters. J Antimicrob Chemother 1996; 37:315–22 pregnated central venous catheter. ANESTHESIOLOGY 1998; 26. Brun-Buisson C, Doyon F, Sollet JP, Cochard JF, Cohen Y, 89:1296–8 Nitenberg G: Prevention of intravascular catheter-related 42. Merrer J, De Jonghe B, Golliot F, Lefrant JY, Raffy B, Barre infection with newer chlorhexidine-silver sulfadiazine- E, Rigaud JP, Casciani D, Misset B, Bosquet C, Outin H, coated catheters: A randomized controlled trial. Intensive Brun-Buisson C, Nitenberg G, French Catheter Study Group Care Med 2004; 30:837–43 in Intensive Care: Complications of femoral and subclavian 27. Ciresi DL, Albrecht RM, Volkers PA, Scholten DJ: Failure of venous catheterization in critically ill patients: A random- antiseptic bonding to prevent central venous catheter-related ized controlled trial. JAMA 2001; 286:700–7 infection and sepsis. Am Surg 1996; 62:641–6 43. Parienti JJ, Thirion M, Me´garbane B, Souweine B, Ouchikhe 28. Collin GR: Decreasing catheter colonization through the use A, Polito A, Forel JM, Marque´ S, Misset B, Airapetian N, of an antiseptic-impregnated catheter: A continuous quality Daurel C, Mira JP, Ramakers M, du Cheyron D, Le Coutour improvement project. Chest 1999; 115:1632–40 X, Daubin C, Charbonneau P, Members of the Cathedia

Anesthesiology 2012; 116:539–73 570 Practice Guidelines SPECIAL ARTICLES

Study Group: Femoral vs jugular venous catheterization and DA, Pruett TL, Schwenzer KJ, Farr BM: A controlled trial of risk of nosocomial events in adults requiring acute renal scheduled replacement of central venous and pulmonary-ar- replacement therapy: A randomized controlled trial. JAMA tery catheters. N Engl J Med 1992; 327:1062–8 2008;299:2413–22 59. Eyer S, Brummitt C, Crossley K, Siegel R, Cerra F: Catheter- 44. Collignon P, Soni N, Pearson I, Sorrell T, Woods P: Sepsis related sepsis: Prospective, randomized study of three associated with central vein catheters in critically ill pa- methods of long-term catheter maintenance. Crit Care Med tients. Intensive Care Med 1988; 14:227–31 1990; 18:1073–9 45. Gil RT, Kruse JA, Thill-Baharozian MC, Carlson RW: Triple- 60. Kealey GP, Chang P, Heinle J, Rosenquist MD, Lewis RW: vs single-lumen central venous catheters: A prospective Prospective comparison of two management strategies of study in a critically ill population. Arch Intern Med 1989; central venous catheters in burn patients. J Trauma 1995; 149:1139–43 38:344–9 46. Gowardman JR, Robertson IK, Parkes S, Rickard CM: Influ- 61. Michel LA, Bradpiece HA, Randour P, Pouthier F: Safety of ence of insertion site on central venous catheter coloniza- central venous catheter change over guidewire for sus- tion and bloodstream infection rates. Intensive Care Med pected catheter-related sepsis. A prospective randomized 2008; 34:1038–45 trial. Int Surg 1988; 73:180–6 47. Lorente L, Henry C, Martín MM, Jime´nez A, Mora ML: 62. Snyder RH, Archer FJ, Endy T, Allen TW, Condon B, Kaiser Central venous catheter-related infection in a prospective J, Whatmore D, Harrington G, McDermott CJ: Catheter and observational study of 2,595 catheters. Crit Care 2005; infection: A comparison of two catheter maintenance tech- 9:R631–5 niques. Ann Surg 1988; 208:651–3 48. McKinley S, Mackenzie A, Finfer S, Ward R, Penfold J: 63. Casey AL, Burnell S, Whinn H, Worthington T, Faroqui MH, Incidence and predictors of central venous catheter related Elliott TS: A prospective clinical trial to evaluate the micro- infection in intensive care patients. Anaesth Intensive Care bial barrier of a needleless connector. J Hosp Infect 2007; 1999; 27:164–9 65:212–8 49. Ramos GE, Bolgiani AN, Patio O, Prezzavento GE, 64. Casey AL, Worthington T, Lambert PA, Quinn D, Faroqui MH, Guastavino P, Durlach R, Fernandez Canigia LB, Benaim F: Elliott TS: A randomized, prospective clinical trial to assess the Catheter infection risk related to the distance between potential infection risk associated with the PosiFlow needle- insertion site and burned area. J Burn Care Rehabil 2002; less connector. J Hosp Infect 2003; 54:288–93 23:266–71 65. Lucet JC, Hayon J, Bruneel F, Dumoulin JL, Joly-Guillou ML: 50. Levy I, Katz J, Solter E, Samra Z, Vidne B, Birk E, Ashkenazi Microbiological evaluation of central venous catheter ad- S, Dagan O: Chlorhexidine-impregnated dressing for pre- ministration hubs. Infect Control Hosp Epidemiol 2000; vention of colonization of central venous catheters in in- 21:40–2 fants and children: A randomized controlled study. Pediatr 66. Ybenes JC, Vidaur L, Serra-Prat M, Sirvent JM, Batlle J, Motje Infect Dis J 2005; 24:676–9 M, Bonet A, Palomar M: Prevention of catheter-related 51. Roberts B, Cheung D: Biopatch: A new concept in antimi- bloodstream infection in critically ill patients using a disin- crobial dressings for invasive devices. Aust Crit Care 1998; fectable, needle-free connector: A randomized controlled 11:16–9 trial. Am J Infect Control 2004; 32:291–5 52. Timsit JF, Schwebel C, Bouadma L, Geffroy A, Garrouste- 67. Kaiser CW, Koornick AR, Smith N, Soroff HS: Choice of Orgeas M, Pease S, Herault MC, Haouache H, Calvino- route for central venous cannulation: Subclavian or internal Gunther S, Gestin B, Armand-Lefevre L, Leflon V, Chaplain jugular vein? A prospective randomized study. J Surg Oncol C, Benali A, Francais A, Adrie C, Zahar JR, Thuong M, 1981; 17:345–54 Arrault X, Croize J, Lucet JC, Dressing Study Group: Chlo- 68. Eisenhauer ED, Derveloy RJ, Hastings PR: Prospective eval- rhexidine-impregnated sponges and less frequent dressing uation of central venous pressure (CVP) catheters in a large changes for prevention of catheter-related infections in city-county hospital. Ann Surg 1982; 196:560 critically ill adults: A randomized controlled trial. JAMA 2009; 301:1231–41 69. Molgaard O, Nielsen MS, Handberg BB, Jensen JM, Kjaer- gaard J, Juul N: Routine X-ray control of upper central 53. Madeo M, Martin CR, Turner C, Kirkby V, Thompson DR: A venous lines: Is it necessary? Acta Anaesthesiol Scand 2004; randomized trial comparing Arglaes (a transparent dressing 48:685–9 containing silver ions) to Tegaderm (a transparent polyurethane dressing) for dressing peripheral arterial catheters 70. Sznajder JI, Zveibil FR, Bitterman H, Weiner P, Bursztein S: and central vascular catheters. Intensive Crit Care Nurs Central vein catheterization: Failure and complication rates 1998; 14:187–91 by three percutaneous approaches. Arch Intern Med 1986; 146:259–61 54. Garland JS, Alex CP, Mueller CD, Otten D, Shivpuri C, Harris MC, Naples M, Pellegrini J, Buck RK, McAuliffe TL, Goldmann 71. Armstrong PJ, Sutherland R, Scott DH: The effect of posi- DA, Maki DG: A randomized trial comparing povidone-iodine tion and different manoeuvres on the internal jugular vein to a chlorhexidine gluconate-impregnated dressing for preven- diameter size. Acta Anaesth Scand 1994; 38:229–31 tion of central venous catheter infections in neonates. Pediat- 72. Bellazzini MA, Rankin PM, Gangnon RE, Bjoernsen LP: Ul- rics 2001; 107:1431–6 trasound validation of maneuvers to increase internal jugu- 55. Moro ML, Vigano` EF, Cozzi Lepri A: Risk factors for central lar vein cross-sectional area and decrease compressibility. venous catheter-related infections in surgical and intensive Am J Emerg Med 2009; 27:454–9 care units: The Central Venous Catheter-Related Infections 73. Modeliar SS, Sevestre MA, de Cagny B, Slama M: Ultrasound Study Group. Infect Control Hosp Epidemiol 1994; 15: evaluation of central veins in the intensive care unit: Effects of 253–64 dynamic manoeuvres. Intensive Care Med 2008; 34:333–8 56. Bonawitz SC, Hammell EJ, Kirkpatrick JR: Prevention of 74. Parry G: Trendelenburg position, head elevation and a mid- central venous catheter sepsis: A prospective randomized line position optimize right internal jugular vein diameter. trial. Am Surg 1991; 57:618–23 Can J Anaesth 2004; 51:379–81 57. Kowalewska-Grochowska K, Richards R, Moysa GL, Lam K, 75. Suarez T, Baerwald JP, Kraus C: Central venous access: The Costerton JW, King EG: Guidewire catheter change in cen- effects of approach, position, and head rotation on internal tral venous catheter biofilm formation in a burn population. jugular vein cross-sectional area. Anesth Analg 2002; 95: Chest 1991; 100:1090–5 1519–24 58. Cobb DK, High KP, Sawyer RG, Sable CA, Adams RB, Lindley 76. Tugrul M, Camci E, Pembeci K, Al-Darsani A, Telci L: Relation-

Anesthesiology 2012; 116:539–73 571 Practice Guidelines Practice Guidelines

ship between peripheral and central venous pressures in dif- insertion as a temporary hemodialysis access. Ren Fail 2005; ferent patient positions, catheter sizes, and insertion sites. 27:561–4 J Cardiothorac Vasc Anesthesia 2004; 18:446–50 95. Cajozzo M, Quintini G, Cocchiera G, Greco G, Vaglica R, 77. Sayin MM, Mercan A, Koner O, Ture H, Celebi S, Sozubir S, Pezzano G, Barbera V, Modica G: Comparison of central Aykac B: Internal jugular vein diameter in pediatric pa- venous catheterization with and without ultrasound guide. tients: Are the J-shaped guidewire diameters bigger than Transfus Apher Sci 2004; 31:199–202 internal jugular vein? An evaluation with ultrasound Paedi- 96. Grebenik CR, Boyce A, Sinclair ME, Evans RD, Mason DG, atr Anaesth 2008; 18:745–51 Martin B: NICE guidelines for central venous catheterization 78. Brown CQ: Inadvertent prolonged cannulation of the ca- in children. Is the evidence base sufficient? Br J Anaesth rotid artery. Anesth Analg 1982; 61:150–2 2004; 92:827–30 79. Digby S: Fatal respiratory obstruction following insertion of 97. Karakitsos D, Labropoulos N, De Groot E, Patrianakos AP, a central venous line. Anaesthesia 1994; 49:1013–4 Kouraklis G, Poularas J, Samonis G, Tsoutsos DA, Konsta- 80. Guilbert MC, Elkouri S, Bracco D, Corriveau MM, Beaudoin doulakis MM, Karabinis A: Real-time ultrasound-guided cath- N, Dubois MJ, Bruneau L, Blair JF: Arterial trauma during eterisation of the internal jugular vein: A prospective com- central venous catheter insertion: Case series, review and parison with the landmark technique in critical care proposed algorithm. J Vasc Surg 2008; 48:918–25, discus- patients. Crit Care 2006; 10:R162 sion 925 98. Koroglu M, Demir M, Koroglu BK, Sezer MT, Akhan O, 81. Farhat K, Nakhjavan K, Cope C, Yazdanfar S, Fernandez J, Yildiz H, Yavuz L, Baykal B, Oyar O: Percutaneous place- Gooch A, Goldberg H: Iatrogenic arteriovenous fistula: A ment of central venous catheters: Comparing the anatomi- complicaton of percutaneous subclavian vein puncture. cal landmark method with the radiologically guided tech- Chest 1975; 67:480–2 nique for central venous catheterization through the internal jugular vein in emergency hemodialysis patients. 82. Kulvatunyou N, Heard SO, Bankey PE: A subclavian artery Acta Radiol 2006; 47:43–7 injury, secondary to internal jugular vein cannulation, is a predictable right-sided phenomenon. Anesth Analg 2002; 99. Mallory DL, McGee WT, Shawker TH, Brenner M, Bailey KR, 95:564–6 Evans RG, Parker MM, Farmer JC, Parillo JE: Ultrasound guidance improves the success rate of internal jugular vein 83. Maschke SP, Rogove HJ: Cardiac tamponade associated with cannulation: A prospective, randomized trial. Chest 1990; a multilumen central venous catheter. Crit Care Med 1984; 98:157–60 12:611–3 100. Slama M, Novara A, Safavian A, Ossart M, Safar M, Fagon JY: 84. Nicholson T, Ettles D, Robinson G: Managing inadvertent Improvement of internal jugular vein cannulation using an arterial catheterization during central venous access proce- ultrasound-guided technique. Intensive Care Med 1997; 23: dures. Cardiovasc Intervent Radiol 2004; 27:21–5 916–9 85. Powell H, Beechey AP: Internal jugular catheterisation: 101. Teichgrber UKM, Benter T, Gebel M, Manns MP: A sono- Case report of a potentially fatal hazard. Anaesthesia 1990; graphically guided technique for central venous access. AJR 45:458–9 Am J Roentgenol 1997; 169:731–3 86. Shah PM, Babu SC, Goyal A, Mateo RB, Madden RE: Arterial 102. Troianos CA, Jobes DR, Ellison N: Ultrasound-guided can- misplacement of large-caliber cannulas during jugular vein nulation of the internal jugular vein: A prospective, ran- catheterization: Case for surgical management. Am Coll domized study. Anesth Analg 1991; 72:823–6 Surg 2004; 198:939–44 103. Verghese ST, McGill WA, Patel RI, Sell JE, Midgley FM, 87. Sloan MA, Mueller JD, Adelman LS, Caplan LR: Fatal brain- Ruttimann UE: Comparison of three techniques for internal stem stroke following internal jugular vein catheterization. jugular vein cannulation in infants. Paediatr Anaesth 2000; Neurology 1991; 41:1092–5 10:505–11 88. Zaidi NA, Khan M, Naqvi HI, Kamal RS: Cerebral infarct 104. Verghese ST, McGill WA, Patel RI, Sell JE, Midgley FM, following central venous cannulation. Anaesthesia 1998; Ruttimann UE: Ultrasound-guided internal jugular venous 53:186–91 cannulation in infants: A prospective comparison with the 89. Reeves ST, Roy RC, Dorman BH, Fishman RL, Pinosky ML: The traditional palpation method. ANESTHESIOLOGY 1999; 91:71–7 incidence of complications after the double-catheter technique 105. Gualtieri E, Deppe SA, Sipperly ME, Thompson DR: Subcla- for cannulation of the right internal jugular vein in a university vian venous catheterization: Greater success rate for less teaching hospital. Anesth Analg 1995; 81:1073–6 experienced operators using ultrasound guidance. Crit Care 90. Milling TJ Jr., Rose J, Briggs WM, Birkhahn R, Gaeta TJ, Bove Med 1995; 23:692–7 JJ, Melniker LA: Randomized, controlled clinical trial of 106. Fragou M, Gravvanis A, Dimitriou V, Papalois A, Kouraklis G, point-of-care limited ultrasonography assistance of central Karabinis A, Saranteas T, Poularas J, Papanikolaou J, Davlouros venous cannulation: The Third Sonography Outcomes As- P, Labropoulos N, Karakitsos D: Real-time ultrasound-guided sessment Program (SOAP-3) Trial. Crit Care Med 2005; subclavian vein cannulation versus the landmark method in 33:1764–9 critical care patients: A propsective randomized study. Crit 91. Alderson PJ, Burrows FA, Stemp LI, Holtby HM: Use of Care Med 2011; 39:1607–12 ultrasound to evaluate internal jugular vein anatomy and to 107. Aouad MT, Kanazi GE, Abdallah FW, Moukaddem FH, Tur- facilitate central venous cannulation in paediatric patients. bay MJ, Obeid MY, Siddik-Sayyid SM: Femoral vein cannu- Br J Anaesth 1993; 70:145–8 lation performed by residents: A comparison between ul- 92. Hayashi H, Amano M: Does ultrasound imaging before trasound-guided and landmark technique in infants and puncture facilitate internal jugular vein cannulation? Pro- children undergoing cardiac surgery. Anesth Analg 2010; spective randomized comparison with landmark-guided 111:724–8 puncture in ventilated patients. J Cardiothorac Vasc Anesth 108. Ezaru CS, Mangione MP, Oravitz TM, Ibinson JW, Bjerke RJ: 2002; 16:572–5 Eliminating arterial injury during central venous catheteriza- 93. Mansfield PF, Hohn DC, Fornage BD, Gregurich MA, Ota tion using manometry. Anesth Analg 2009; 109:130–4 DM: Complications and failures of subclavian-vein catheter- 109. Gillman LM, Blaivas M, Lord J, Al-Kadi A, Kirkpatrick W: ization. N Engl J Med 1994; 331:1735–8 Ultrasound confirmation of guidewire position may elimi- 94. Bansal R, Agarwal SK, Tiwari SC, Dash SC: A prospective nate accidental arterial dilatation during central venous randomized study to compare ultrasound-guided with non- cannulation. Scand J Trauma 2010; 18:39 ultrasound-guided double lumen internal jugular catheter 110. Gu X, Paulsen W, Tisnado J, He Y, Li Z, Nixon JV: Malpo-

Anesthesiology 2012; 116:539–73 572 Practice Guidelines SPECIAL ARTICLES

sition of a central venous catheter in the right main pulmo- after central venous catheter insertion? Am J Respir Crit nary artery detected by transesophageal echocardiography. Care Med 2001; 164:403–5 J Am Soc Echocardiogr 2009; 22:1420 123. Riblet JL, Shillinglaw W, Goldberg AJ, Mitchell K, Sedani 111. Mahmood F, Sundar S, Khabbaz K: Misplacement of a guide- KH, Davis FE, Reynolds HN: Utility of the routine chest wire diagnosed by transesophageal echocardiography. X-ray after “over-wire” venous catheter changes. Am Surg J Cardiothorac Vasc Anesth 2007; 21:420–1 1996; 62:1064–5 112. Sawchuk C, Fayad A: Confirmation of internal jugular guide 124. Weil BR, Ladd AP, Yoder K: Pericardial effusion and cardiac wire position utilizing transesophageal echocardiography. tamponade associated with central venous catheters in chil- Can J Anaesth 2001; 48:688–90 dren: An uncommon but serious and treatable condition. J Pediatr Surg 2010; 45:1687–92 113. Janik JE, Cothren CC, Janik JS, Hendrickson RJ, Bensard DD, Partrick DA, Karrer FM: Is a routine chest x-ray necessary 125. Wirsing M, Schummer C, Neumann R, Steenbeck J, Schmidt for children after fluoroscopically assisted central venous P, Schummer W: Is traditional reading of the bedside chest radiograph appropriate to detect intraatrial central ve- access? J Pediatr Surg 2003; 38:1199–202 nous catheter position? Chest 2008; 134:527–33 114. Lucey B, Varghese JC, Haslam P, Lee MJ: Routine chest 126. Francis KR, Picard DL, Fajardo MA, Pizzi WF: Avoiding radiographs after central line insertion: Mandatory postpro- complications and decreasing costs of central venous cath- cedural evaluation or unnecessary waste of resources? Car- eter placement utilizing electrocardiographic guidance. diovasc Intervent Radiol 1999; 22:381–4 Surg Gynecol Obstet 1992; 175:208–11 115. Gebhard RE, Szmuk P, Pivalizza EG, Melnikov V, Vogt C, 127. McGee WT, Ackerman BL, Rouben LR, Prasad VM, Bandi V, Warters RD: The accuracy of electrocardiogram-controlled Mallory DL: Accurate placement of central venous cathe- central line placement. Anesth Analg 2007; 104:65–70 ters: A prospective, randomized, multicenter trial. Crit Care 116. Abood GJ, Davis KA, Esposito TJ, Luchette FA, Gamelli RL: Med 1993; 21:1118–23 Comparison of routine chest radiograph versus clinician 128. O’Grady NP, Alexander M, Burns LA, Dellinger EP, Garland J, judgment to determine adequate central line placement in Heard SO, Lipsett PA, Masur H, Mermel LA, Pearson ML, Raad critically ill patients. J Trauma 2007; 63:50–6 II, Randolph AG, Rupp ME, Saint S, Healthcare Infection Con- 117. Amshel CE, Palesty JA, Dudrick SJ: Are chest X-rays mandatory trol Practices Advisory Committee. Guidelines for the preven- following central venous recatheterization over a wire? Am tion of intravascular catheter related infections. Am J Infect Surg 1998; 64:499–501; discussion 501–2 Control 2011; 39(4 Suppl 1):S1–34 118. Bailey SH, Shapiro SB, Mone MC, Saffle JR, Morris SE, Barton 129. Marschall J, Mermel LA, Classen D, Arias KM, Podgorny K, An- RG: Is immediate chest radiograph necessary after central derson DJ, Burstin H, Calfee DP, Coffin SE, Dubberke ER, Fraser venous catheter placement in a surgical intensive care unit? V, Gerding DN, Griffin FA, Gross P, Kaye KS, Klompas M, Lo E, Am J Surg 2000; 180:517–21; discussion 521–2 Nicolle L, Pegues DA, Perl TM, Saint S, Salgado CD, Weinstein RA, Wise R, Yokoe DS: Strategies to prevent central line-associated 119. Cullinane DC, Parkus DE, Reddy VS, Nunn CR, Rutherford bloodstream infections in acute care hospitals. Infect Control EJ: The futility of chest roentgenograms following routine Hosp Epidemiol 2008; 29(Suppl 1):S22–30 central venous line changes. Am J Surg 1998; 176:283–5 130. Institute for Healthcare Improvement: Prevent Central Line 120. Frassinelli P, Pasquale MD, Cipolle MD, Rhodes M: Utility Infections How-to Guide. Cambridge, MA; 2008 of chest radiographs after guidewire exchanges of 131. The Joint Commission Accreditation Program: Hospital: Na- central venous catheters. Crit Care Med 1998; 26:611–5 tional patient safety goals. http://www.jointcommission. 121. Lessnau KD: Is chest radiography necessary after uncom- org/assets/1/6/2011_NPSGs_HAP.pdf. Accessed January 1, plicated insertion of a triple-lumen catheter in the right 2011 internal jugular vein, using the anterior approach? Chest 132. National Institute for Clinical Excellence: Guidance on the use of 2005; 127:220–3 ultrasound locating devices for placing central venous catheters. 122. Maury E, Guglielminotti J, Alzieu M, Guidet B, Offenstadt G: http://www.nice.org.uk/nicemedia/pdf/49_English_patient. Ultrasonic examination: An alternative to chest radiography pdf. Accessed September 2012

Anesthesiology 2012; 116:539–73 573 Practice Guidelines European Journal of Radiology 76 (2010) 20–23

Contents lists available at ScienceDirect

European Journal of Radiology

journal homepage: www.elsevier.com/locate/ejrad

Radiation protection of medical staff

John Le Heron a,∗, Renato Padovani b, Ian Smith c, Renate Czarwinski a a Radiation Safety & Monitoring Section, Division of Radiation, Transport and Waste Safety, International Atomic Energy Agency, Wagramer Strasse 5, P.O. Box 100, 1400 Vienna, Austria b Medical Physics Department, University Hospital, Udine, Italy c St Andrew’s Medical Institute, St Andrew’s War Memorial Hospital, Brisbane, Australia article info abstract

Article history: The continuing increase in the worldwide use of X-ray imaging has implications for radiation protection of Received 14 June 2010 medical staff. Much of the increased usage could be viewed as simply a workload issue with no particular Accepted 15 June 2010 new challenges. However, advances in technology and developments in techniques have seen an increase in the number of X-ray procedures in which medical personnel need to maintain close physical contact Keywords: with the patient during radiation exposures. The complexity of many procedures means the potential Radiation protection for signiﬁcant occupational exposure is high, and appropriate steps must be taken to ensure that actual Occupational exposure occupational exposures are as low as reasonably achievable. Further attention to eye protection may be Image-guided interventional procedures Protective tools necessitated if a lowering of the dose limit for the lens of the eye is implemented in the near future. Protective clothing Education and training in radiation protection as it applies to speciﬁc situations, established working Monitoring procedures, availability and use of appropriate protective tools, and an effective monitoring programme are all essential elements in ensuring that medical personnel in X-ray imaging are adequately and acceptably protected. © 2010 Elsevier Ireland Ltd. All rights reserved.

1. Introduction dose limitation that is arguably the most important. However, it should be recognized that the lack of rigorous application of the The use of radiation in medical applications continues to justification principle is resulting in the performance of significant increase worldwide. Latest UNSCEAR estimates suggest that there numbers of unnecessary X-ray imaging examinations, and these are about 4 billion X-ray examinations per year, worldwide [1]. All add to occupational exposure. these procedures need to be performed by medical personnel with This paper cannot cover all the aspects of radiation protection the accompanying potential for occupational radiation exposure. of medical staff, but instead focuses more on the challenges in the Does the increasing demand for X-ray imaging have implications never ending story of occupational radiation protection, including for radiation protection of medical staff? The increasing usage could a brief discussion of factors that determine the level of occupational be viewed as simply a workload issue with no particular new chal- exposure in X-ray imaging, the implications of recent trends in X- lenges. However, there is also a change in the types of X-ray imaging ray imaging for occupational radiation protection, and the basis for procedures being performed and by whom – procedures requiring ensuring effective radiation protection of medical staff in the light personnel to be close to the patient, and these procedures present of these new developments. a challenge to ensure appropriate radiation protection of medical staff. The term X-ray imaging is used in this paper to cover both 2. What determines the level of occupational exposure in diagnostic radiology and image-guided interventional procedures. X-ray imaging? Occupational radiation protection is achieved by application of the three ICRP principles of justification, optimization and dose lim- The level of occupational exposure associated with X-ray imag- itation [2]. In practice, to afford radiation protection for medical ing procedures is highly variable and ranges from potentially staff in X-ray imaging, it is the application of optimization and negligible in the case of simple chest X-rays, to significant for complex interventional procedures. What are the reasons for this? From the occupational perspective, there are two “sources” of radiation ∗ exposure. Clearly the X-ray tube is the true source of radiation, Corresponding author. Tel.: +43 1 2600 21416. but in practice there should be very few situations in which per- E-mail addresses: [email protected] (J. Le Heron), [email protected] (R. Padovani), [email protected] sonnel have the potential to be directly exposed to the primary (I. Smith), [email protected] (R. Czarwinski). beam. This leaves the other “source”, which is the patient. Interac-

0720-048X/$ – see front matter © 2010 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejrad.2010.06.034 J. Le Heron et al. / European Journal of Radiology 76 (2010) 20–23 21 tion of the primary X-ray beam with the part of the patient’s body protective value. Gloves also may slow down the procedure and being imaged produces scattered radiation, which emanates from engender a false sense of safety: it is much better to train person- the patient in all directions. So, in most cases, the main determinant nel to keep their hands out of the primary beam. Ensuring the X-ray for occupational exposure is proximity of personnel to the patient tube is under the table provides the best protection when the hands when exposures are being made. Furthermore, the level of scat- must be near the X-ray field, as the primary beam is then attenuated ter is determined largely by the dose to the patient, resulting in an by the patient’s body. important corollary that reducing the patient dose to the minimum Ceiling-suspended protective screens can provide significant necessary to achieve the required medical outcome also lowers protection, but their effectiveness depends on them being posi- potential occupational exposures. A common and useful maxim is tioned correctly. Screens provide protection to only part of the body that when personnel take care of the patient, they will also take (typically the upper body, head and eyes) and their use is additional care of their occupational exposure. to wearing protective clothing. Sometimes a protective screen cannot be deployed for clinical reasons. Table-mounted protective 2.1. Working at some distance from the patient curtains also provide additional shielding, typically to the lower body and legs. A further option for complex image-guided inter- For many situations, such as radiography, mammography and ventional procedures is the use of disposable protective patient general CT, there is usually no need for personnel to be physi- drapes. These are attenuating drapes placed on the patient after cally close to the patient. This provides two opportunities to afford the operative site has been prepared. good occupational radiation protection. First, being distant from the Because radiological workloads can be very different for the dif- patient reduces the levels of scatter reaching personnel. Divergence ferent specialties, the necessary protective tools should be specified of the scattered X-rays from the patient results in radiation inten- by a medical physicist or a RPE. For example, a person with a high sities falling off rapidly with distance. This is the so-called “inverse workload in a cardiac laboratory should use all the described pro- square law” where, for example, a doubling of the distance results tective tools, while a person in an orthopedic suite may need only in a four-fold reduction in radiation intensity. a simple front-only protective apron. The second factor is that structural shielding can be placed Many factors can influence how fluoroscopic examinations and between the patient and personnel. For example, shielding may image-guided interventional procedures are performed, and hence be incorporated into the wall and window of the control room the patient dose and occupational exposure. It is therefore essen- barrier, resulting in negligible levels of scatter reaching personnel. tial that personnel performing such procedures have had effective Appropriate room design with shielding specification by a medical training in radiation protection so they understand the implications physicist or radiation protection expert (RPE) also should ensure of these factors. Moreover, because of the wide variability in poten- that occupational exposure will be essentially zero for these X-ray tial occupational exposures from these procedures, it is also crucial imaging situations. that personal monitoring is performed continuously and correctly. Both these aspects are discussed in this chapter. 2.2. Working close to the patient 3. Implications of recent trends in X-ray imaging for There are some situations, typically in fluoroscopic examina- medical staff tions and in image-guided interventional procedures, where it is necessary to maintain close physical contact with the patient when In the area of radiography, including mammography, there has radiation is being used. Distance and structural shielding are not been a shift from film-based to digital systems. Patient doses options, so what can be done for these personnel? remain broadly similar, but patient throughput for a given imaging Scattered radiation can be attenuated by protective cloth- room may have increased. Overall, the implications for occupa- ing worn by personnel, such as aprons, spectacles, and thyroid tional radiation protection have been minimal. shields, and by protective tools, such as ceiling-suspended pro- CT imaging has seen the advent of multi-detector CT scanners tective screens or table-mounted protective curtains or wheeled (MDCT). Again, depending on how well imaging protocols have screens, which are able to be placed between the patient and per- been optimized, patient doses are broadly similar to their single sonnel. Depending on its lead equivalence and the energy of the slice antecedents and future trend are likely to lower patient doses X-rays, an apron will attenuate 90% or more of the incident scat- as manufacturers continue to introduce more dose-saving features. tered radiation. Protective aprons come in different thicknesses and Image acquisition is considerably faster and usage of CT is increas- shapes, ranging from the simple front-only apron to a full coat: the ing, resulting in higher workloads in the CT room. This has often former being effective only if the wearer is always facing the source resulted in a need for increased structural shielding to maintain of the scattered radiation. acceptable standards of occupational radiation protection. The lens of the eye is highly radiation sensitive. For per- New technologies and techniques have also allowed the per- sons working close to the patient, doses to the eyes can become formance of complex diagnostic and interventional procedures unacceptably high. Wearing protective eyewear, particularly those with the advantage of avoiding more risky surgical interventions incorporating side protection, can reduce the dose to the eyes from in several cases. These procedures present several challenges for scatter by 90%, but to achieve maximum effectiveness, careful con- radiation protection of medical staff. sideration needs to be given to issues such as the placement of the viewing monitor to ensure the eyewear intercepts the scatter from the patient. It is likely that measures to protect the eyes will receive 3.1. CT-fluoroscopy further attention if the dose limit for the lens of the eye is lowered as a result of a review of current scientific evidence [2]. CT-fluoroscopy facilitates the performance of biopsies by gen- There are some situations, usually associated with image- erating live tomographic images to better guide the intervention. guided interventional procedures, when the hands of the operator However, as the interventionalist is very near to the patient, there may inadvertently be placed in the primary X-ray beam. Protective is the potential for the hands to be exposed to the primary beam gloves may appear to be indicated, but such gloves can prove to and hand doses of up to 0.6 mSv for a procedure lasting only 20 s be counter-productive as their presence in the primary beam leads have been reported [3]. Without protective measures, dose limits to an automatic increase in the radiation dose rate, offsetting any (see Table 1) for the hands could easily be exceeded. 22 J. Le Heron et al. / European Journal of Radiology 76 (2010) 20–23

Table 1 requirements for such training. In X-ray imaging, personnel need Recommended dose limits for occupational exposure in X-ray imaging (adapted training not only in occupational radiation protection, but also in from ICRP[2]). patient radiation protection as the latter can inﬂuence occupational Dose quantity Occupational dose limita exposure. Guidance has been published on what the training for Effective dose 20 mSv per year averaged over 5 consecutive medical staff should cover, such as by the European Commission [7]. years (100 mSv in 5 years), and 50 mSv in any Several resources for training are freely available, including IAEA single year material for diagnostic and interventional radiology and cardiology Equivalent dose in: [8], and the MARTIR project (multimedia tool for training in inter- Lens of the eyeb 150 mSv in a year Skinc 500 mSv in a year ventional radiology) [9], which is particularly useful for distance Hands and feet 500 mSv in a year learning purposes.

a Dose limits can be different in national regulations. For the reasons noted previously, it is crucial that programmes b This limit is currently being reviewed by an ICRP Task Group. for radiation protection training in hospital settings include those c Averaged over 1 cm2 of the most highly irradiated area of the skin. medical personnel who are outside the radiology department but who are involved in X-ray imaging procedures. 3.2. Interventional radiology and cardiology 4.2. Monitoring of occupational exposure in X-ray imaging In interventional radiology or cardiology the main operators Occupational exposure is subject to dose limits as given in stay very close to the patient table where the intensity of scattered Table 1, established to ensure that the risks arising from occupa- radiation can be very high. For example, in a cardiology procedure tional exposure are not unacceptable [2]. The dose limit expressed when the X-ray tube is under the table, the dose rate of incident as effective dose addresses cancers and hereditary effects, while the scattered radiation at the position of the operator can typically other dose limits (in terms of equivalent dose) are to prevent radi- range from around 0.5 to 10 mSv/h, from head height down towards ation effects in particular tissues or areas of the body. In addition, the legs. It is easy to see how cumulative doses can exceed dose the application of the principle of optimization of protection should limits and reach hundreds of mSv per year if protective measures ensure that occupational doses to medical staff in X-ray imaging are were not taken, as described briefly in Section 2.2. This is described as low as reasonably achievable (ALARA principle), and typically more comprehensively, in the case of interventional radiology, in well below the dose limits. the recent guidelines on occupational radiation protection pub- The purpose of individual radiation monitoring is to ensure the lished by the Cardiovascular and Interventional Radiology Society dose limits are not exceeded. Further, through regular review, the of Europe and the Society of Interventional Radiology [4]. results of individual monitoring are used to assess the effectiveness of strategies for optimization. It is always hoped that the monitor- 3.3. Image-guided procedures outside radiology departments ing results confirm that good practice is taking place. Estimates of effective dose and equivalent dose can be made Many fluoroscopy-guided procedures today are performed out- using appropriately designed and calibrated dosimeters, such as side radiology departments, using mobile C-arm or O-arm units those provided by an accredited individual monitoring service, in, for example, orthopaedic, gastroenterology, urology, and vas- worn by medical personnel when performing X-ray imaging. It is cular operating theatres. The use of radiation in these procedures important that personal dosimeters are worn correctly. In X-ray can range from very short fluoroscopy times (<0.5 min for a typi- imaging this can become complicated, depending on the roles of cal orthopaedic procedure) to very long fluoroscopy times, such as the person being monitored and whether they are sometimes or 60–90 min for an aortic aneurysm treatment. The main issue is that even always wearing protective clothing. If a person is normally dis- the non-radiologist specialists and nurses involved in these proce- tant from the patient and is being protected by structural shielding, dures are often outside the umbrella of the radiology department, then wearing a single dosimeter on the front of the torso, between and hence may not receive adequate training (including training in the shoulders and the waist, would normally suffice. However, if radiation protection) in the use of modern equipment, which allows a person is working close to the patient during imaging and is not only simple fluoroscopy, but also pulsed fluoroscopy, digital wearing a protective apron, then a single dosimeter worn under acquisition, angiography and digital subtraction angiography. the apron will provide a good estimate of exposure to the shielded parts of the body, but will underestimate exposure to the organs 4. Ensuring effective radiation protection of medical staff and tissue outside the apron. In these cases, as dose to the eyes is often of most concern, it is recommended that two dosimeters are A radiation protection programme (RPP) is one means of worn, one under the apron and the other outside the apron, often implementing occupational radiation protection by the adoption at shoulder or collar level. The results of the two dosimeters can of appropriate management structures, policies, procedures and be combined to give a better estimate of effective dose, as well as organizational arrangements. For medical staff in X-ray imaging, providing an estimate for the dose to the eyes. topics would include the need for local rules and procedures for In some cases in X-ray imaging, monitoring of the fingers and personnel to follow, arrangements for the provision of personal hands may be indicated. This requires special dosimeters, such as protective equipment, a programme for education and training in ring or bracelet dosimeters. It is important to position the dosime- radiation protection, arrangements for individual monitoring, and ter in the position most likely to receive the highest exposure. This methods for periodically reviewing and auditing the performance may not be the dominant hand. In all cases of individual monitoring, of the RPP. The following sections describe aspects of two of these specific advice from a medical physicist or RPE should be obtained. topics. Further details on RPPs can be found in the BSS [5], and One of the practical difficulties in individual monitoring is ensuring IAEA SRS No 39 [6]. that monitored personnel actually always wear their dosimeters when working with radiation. Incomplete compliance means that 4.1. Radiation protection training a person’s occupational exposure is underestimated and, if this lack of compliance with monitoring is widespread, then doses to that Education and training underpin the implementation of radi- occupational group as a whole would also be underestimated. The ation protection in practice, and most countries have regulatory reality is that many medical personnel do not always wear their J. Le Heron et al. / European Journal of Radiology 76 (2010) 20–23 23 dosimeters, for a variety of reasons that range from simple neg- 5. Conclusions ligence to deliberate avoidance to ensure that monitored results remain below the threshold for administrative or regulatory inves- The continuing increase in the worldwide use of X-ray imaging tigation. It is important that monitoring is not viewed negatively, is creating new challenges for occupational radiation protection of but rather as a means for adding value to occupational radiation medical staff. Advances in technology and developments in tech- protection by confirming good practice when this is the case, and niques have seen an increase in the number of X-ray procedures for improvement through corrective actions when problems are in which medical personnel must maintain close physical contact identified. with the patient during radiation exposures. The complexity of Developments in individual dosimetry may help solve the many procedures means the potential for significant occupational monitoring compliance issue. For example, technology is being exposure is high, and appropriate steps must be taken to ensure developed that utilises compact personal electronic monitoring actual occupational exposures are as low as reasonably achievable. devices that wirelessly connect to a base station. Multiple devices Further attention to eye protection may be necessitated if a low- can be monitored simultaneously, so all staff involved in a com- ering of the dose limit for the lens of the eye is implemented in plex imaging procedure can monitor their doses and dose rates in the near future. Education and training in radiation protection as it real-time and use this information to modify their practice if indi- applies to specific situations, plus established working procedures, cated. Future technology may even negate the need for dosimeters availability and use of appropriate protective tools, and an effec- or devices to be worn by personnel. It is feasible that a person’s loca- tive monitoring programme are all essential elements in ensuring tion in a room can be monitored in real-time, and their real-time that medical staff in X-ray imaging are adequately and acceptably occupational exposure calculated from the real-time knowledge of protected. the X-ray equipment and how it is being used. References 4.3. Current occupational doses to medical staff [1] UNSCEAR 2008 Report: Sources of ionizing radiation, vols. I & II; in press. [2] International Commission on Radiological Protection. The 2007 recommen- There is a large number of publications giving occupational dations of the International Commission on Radiological Protection. ICRP doses per given procedure in X-ray imaging in a given facility, Publication 103. Ann ICRP 2007;37:1–332. and others that estimate likely annual doses. UNSCEAR provides a [3] Stoeckelhuber Beate M, Leibecke T, Schulz E, et al. Radiation dose to the radiol- comprehensive source of information on occupational doses world- ogists hand during continuous CT fluoroscopy-guided interventions. Cardiovasc Intervent Radiol 2005;28:589–94. wide, and in its 2010 report (covering the period 2002–2006) it [4] Miller DL, Vano E, Bartal G, et al. Occupational radiation protection in inter- concludes that over 80% of general and CT radiographers did not ventional radiology: a joint guideline of the Cardiovascular and Interventional receive measurable doses [1], confirming that for personnel who Radiology Society of Europe and the Society of Interventional Radiology. Cardio- vasc Intervent Radiol 2010;33:230–9. are protected by structural shielding, doses are very low. The report [5] Food and Agricultural Organization of the United Nations, International Atomic notes that only a few countries were able to provide data that distin- Energy Agency, International Labour Organization, OECD Nuclear Energy guished between conventional techniques in diagnostic radiology Agency, Pan American Health Organization, World Health Organization. Inter- national basic safety standards for protection against ionizing radiation and for and interventional procedures. From these limited data, for con- the safety of radiation sources. In: Safety Series No. 115. Vienna: IAEA; 1996. ventional diagnostic radiology the reported mean annual effective [6] International Atomic Energy Agency. Applying radiation safety standards in dose was about 0.5 mSv for monitored personnel, while for inter- diagnostic radiology and interventional procedures using X rays. In: Safety Report Series No. 39. Vienna: IAEA; 2006. ventional procedures it was about 1.6 mSv. A recent study with data [7] European Commission. Radiation protection 116. Guidelines on education from 23 countries gave an average median effective dose for inter- and training in radiation protection for medical exposures. Luxembourg: ventional cardiologists of 0.7 mSv, averaged per country (Padovani Directorate-General for the Environment, European Commission; 2000. [8] Accessible at: http://rpop.iaea.org/RPOP/RPoP/Content/AdditionalResources/ et al., unpublished results). These figures would seem to be under- Training/1 TrainingMaterial/index.htm. estimates when compared with published facility-specific data, [9] MARTIR. Multimedia and audio-visual radiation protection training in inter- again raising the issues of monitoring compliance. ventional radiology, radiation protection 119. Luxembourg: Office for Official Publications of the European Communities; 2002. Anesthesiology 2001; 94:239–44 © 2001 American Society of Anesthesiologists, Inc. Lippincott Williams & Wilkins, Inc. Chlorhexidine versus Povidone Iodine in Preventing Colonization of Continuous Epidural Catheters in Children A Randomized, Controlled Trial Brian Kinirons, M.D.,* Olivier Mimoz, M.D., Ph.D.,† Leila Lafendi, Pharm.D.,‡ Thierry Naas, Ph.D.,§ Jean-François Meunier, M.D.,* Patrice Nordmann, M.D., Ph.D.ʈ

Background: Chlorhexidine is better than povidone iodine scess is a recognized complication of the use of such for skin preparation before intravascular device insertion or catheters. Although rare, epidural abscess can progress blood culture collection, but it is not known whether chlorhexi- rapidly to meningitis, paralysis, or death.1–3 dine is superior in reducing colonization of continuous epidural catheters. Catheters kept in place for a short time are generally Methods: Children requiring an epidural catheter for postop- colonized by skin flora present on the insertion sites, erative analgesia longer than 24 h were randomly assigned to mostly during catheter placement.4 Colonization risk receive skin preparation with an alcoholic solution of 0.5% increases after a threshold level of skin colonization chlorhexidine or an aqueous solution of 10% povidone iodine occurs.5,6 Thus, effective cutaneous antisepsis before before catheter insertion. Using surgical aseptic techniques, catheters were inserted into either the lumbar or the thoracic epidural insertion may reduce catheter colonization, epi- epidural space based on the preferences of the anesthesia team, dural catheter sepsis, and, ultimately, deep space on clinical indication, or both. Immediately before epidural infection. catheter removal, their insertion site and hub were qualitatively The French Society for Anesthesiology recommends cultures. After their removal, the catheter tips were quantita- the use of either povidone iodine or chlorhexidine for tively cultured. Catheters were classified as colonized when their tips yielded 1,000 or more colony-forming units/ml in skin preparation before epidural catheter placement in 7 cultures. children. Iodine tincture is not used in this setting Results: Of 100 randomly assigned patients, 96 were evalu- because of its possible toxic effects, particularly on the able. The clinical characteristics of the patients and the risk thyroid gland. Although chlorhexidine has been found to factors for infection were similar in the two groups. Catheters be superior to povidone iodine for skin preparation were kept in place for a median (range) duration of 50 (range, 8–10 21–100) h. Catheters inserted after skin preparation with chlo- before intravascular catheter insertion or blood cul- 11 rhexidine were one sixth as likely and less quickly to be colo- ture collection, its value in preventing epidural cathe- nized as catheters inserted after skin preparation with povi- ter colonization remains unknown. done iodine (1 of 52 catheters [0.9 per 100 catheter days] vs. 5of The goal of this prospective trial therefore was to 44 catheters [5.6 per 100 catheter days]; relative risk, 0.2 [95% determine if an alcoholic solution of 0.5% chlorhexidine Coagulase-negative .(0.02 ؍ confidence interval, 0.1–1.0]; P staphylococci were the only colonizing microorganisms recov- is more effective than an aqueous solution of 10% povi- ered, and the skin surrounding the catheter insertion site was done iodine in reducing catheter colonization associated the origin of all the colonizing microorganisms. with short-term epidural catheter placement in children. Conclusions: Compared with aqueous povidone iodine, the use of alcoholic chlorhexidine for cutaneous antisepsis before epidural catheter insertion reduces the risk of catheter coloni- Methods zation in children. Patients THE use of epidural catheters to infuse analgesic agents The study was conducted between December, 4, 1998 continuously for pain management during the perioper- and December 20, 1999 in a 1,000-bed, university-affili- ative period has becoming an increasingly popular treat- ated hospital in France (Hoˆpital de Biceˆtre, Le Kremlin ment method in the pediatric population. Epidural ab- Biceˆtre). After ethics committee approval (Comite´ Con- sultatif de Protection des Personnes dans la Rechere Biome´dicale de l’hoˆpital Ambroise Pareá` Boulogne Bil- * Fellow of Anesthesiology, † Assistant Professor of Anesthesiology, ‡ Staff lancourt) and informed parental written consent, pa- Microbiologist, § Assistant Professor of Microbiology, ʈ Chairman, Department of Microbiology, University Paris XI. tients younger than 15 yr of age who were to have an Received from the Department of Anesthesia and Intensive Care, Hôpital de epidural catheter placed during surgery for abdominal, Bicêtre, Assistance Publique-Hôpitaux de Paris, Le Kremlin Bicêtre, France. Sub- lower extremity, or urologic surgical procedures were mitted for publication June 20, 2000. Accepted for publication September 18, 2000. Supported by a grant from Laboratoires Zeneca Pharma, Cergy, France. The eligible to be included in the study. funding agency was not involved in the design, conduct, or reporting of this Exclusion criteria included patients with history of study or in the decision to submit the manuscript for publication. Address correspondence to Dr. Mimoz: Département d’ Anesthésie-Réanima- allergy to one of the antiseptic solutions used, the pres- tion, CHU la Milétrie, BP 577, 86021 Poitiers Cedex, France. Address electronic ence of a clotting defect, neutropenia, neurologic dis- mail to: [email protected]. Reprints will not be available from the authors. Individual article reprints may be purchased through the Journal Web site, ease, local or generalized infection, and those receiving www.anesthesiology.org. immunosuppressive therapy. No antimicrobial prophy-

Anesthesiology, V 94, No 2, Feb 2001 239 240 KINIRONS ET AL. laxis was administered specifically for the epidural cath- were blind to the antiseptic solution used, to search for eter insertion. However, most of the patients received signs of infection (pus), inflammation (erythema, heat, antibiotics during the operative period that were tenderness), or cutaneous allergic events to the disinfec- stopped at the end of surgery (aminopenicillins or ceph- tant (edema, erythema). The decision to remove the alosporins given mostly alone, but sometimes in combi- catheter was made solely by the patient’s physician, who nation with aminoglycosides), according to the French kept the catheter in place until it was no longer required guidelines on antibiotic prophylaxis during surgery.12 or until an adverse event, such catheter-related infection or catheter migration out of the epidural space, necessi- Insertion and Maintenance of Catheters tated its removal. Catheter-related infection was sus- Patients were randomized to receive either 0.5% pected in a patient who became febrile without any chlorhexidine gluconate (Hibitane Champ; Zeneca other cause. Pharma, Cergy, France) or 10% povidone iodine (Béta- dine; Asta Médica, Marignane, France) for cutaneous Cultures antisepsis before epidural catheter insertion using After removal of the occlusive dressing, dry swabs computerized randomization lists. Catheters were in- were taken at the site surrounding catheter insertion and serted using the maximal sterile-barrier precautions at the catheter hub to help identify the sources of mi- into either the lumbar or the thoracic epidural space, croorganisms that may colonize catheters. The skin was based on the preferences of the anesthesia team, on then cleaned with 70% isopropyl alcohol. After the alco- clinical indication, or both. After surgical hand wash- hol was allowed to dry, the catheter was removed asep- ing, fellow or resident anesthesiologists were gowned, tically, and its tip (3–4 cm distal segment) was cut and gloved, and masked. The skin was cleaned twice (be- cultured quantitatively by a method previously described fore and after placement of sterile disposable drapes) for vascular catheters.14 The laboratory technicians were by vigorously applying the designed antiseptic solu- unaware of the antiseptic solution used for skin prepa- tion on an area more than 300 cm2 for at least 10 s and ration. Standard microbiologic methods and criteria allowing the area to dry between each antiseptic ap- identified recovered microorganisms. If a same bacterial plication. Epidural catheters were located using the species was isolated from different sites in a given pa- loss-of-resistance technique with either saline solution tient, their Sma-I-restricted pulse-field gel electrophore- or air (Tuohy 20-gauge needle). A 22-gauge polyamide sis patterns were visually analyzed for similarity.15 Bac- epidural catheter (Portex, Berck, France) was ad- terial isolates were considered similar when pulse-field vanced 3–4 cm into the epidural space to ensure gel electrophoresis patterns differed by no more than secure placement without subcutaneous tunneling. one band, according to Tenover criteria.15 All catheters were fixed in place with a sterile occlusive dressing (Opsite; Smith and Nephew Medical Ltd., Definitions Hull, UK). The proximal portion of the catheter was We adapted the definition of catheter colonization then directed cephalad and fixed on the back using proposed by the Centers for Disease Control and Preven- tape (Albupore; Smith and Nephew Medical Ltd.). tion for intravascular devices.16 Epidural catheter colo- Continuous infusions of 0.1% bupivacaine with or nization was defined as the growth of 1,000 or more without preservative-free morphine chlorhydrate colony-forming units/ml in cultures from the catheter (30–50 ␮g/kg) or clonidine (1 ␮g/kg) were adminis- tip. Colonization of the catheter was related to the skin tered through the catheters using an antimicrobial when the same bacterial isolate yielded from the cathe- filter. The anesthetic solutions were changed at least ter tip and the skin surrounding the insertion site, and every 24 h, and the connections were manipulated was related to the catheter hub when the same bacterial with gauze soaked with 70% isopropyl alcohol. Topi- isolate yielded from the catheter tip and the catheter hub. cal antibiotic or antiseptic ointments were not used on any catheter. The dressings were not changed until Statistical Analysis the catheters were removed to avoid catheter migra- Before undertaking this study, we estimated the num- tion out of the epidural space. The following informa- ber of catheters that would be required for an adequate tion was recorded when the epidural catheters were examination of the hypothesis that epidural catheters inserted: age, sex, weight, American Society of Anes- inserted after skin preparation with chlorhexidine are thesiologists classification score, Altemeir class13 and two thirds less likely to be colonized than catheters duration of surgery, status of doctor performing the inserted after skin preparation with povidone iodine. epidural (resident or fellow), epidural solution used, This was based on our experience with intravascular perioperative antimicrobial prophylaxis administra- devices.9 Supposing that the use of povidone iodine will tion, and duration of epidural infusion. be associated with a 20% incidence of catheter coloni- The insertion site and the dressing were inspected zation, randomly assigning approximately a total of 96 daily by nursing staff and the patient’s physician, who catheters would have allowed us to detect, with 80%

Anesthesiology, V 94, No 2, Feb 2001 CONTINUOUS EPIDURAL CATHETER COLONIZATION IN CHILDREN 241 power, a significant difference in the rates of coloniza- Table 1. Characteristics of the Patients and Epidural Catheters* tion between the two types of antiseptic solutions at a Povidone one-tailed significance level of 5%. iodine Chlorhexidine Data are expressed as medians with range for contin- (n ϭ 44) (n ϭ 52) P Value uous variables or percentages for categorical variables. No. of patients 44 52 The significance of the differences between the two Male sex, n (%) 17 (39) 29 (56) 0.11 study groups was determined with the Mann–Whitney U Age (months) 22 (1–180) 24 (1–180) 0.87 Weight (kg) 12 (3–56) 12 (2–42) 0.81 test for continuous variables and the Fisher exact test or ASA score, n (%) 0.79 the chi-square test for categorical variables. The propor- I 15 (34) 20 (38) tions of catheters that were free of colonization as a II 27 (61) 29 (56) function of time they had been in place were compared III 2 (5) 3 (6) Altemeier class, n (%) 0.32 between the groups with use of a log-rank test on Clean 33 (75) 33 (63) Kaplan-Meier estimates. All P values were based on two- Clean–contaminated 9 (20) 18 (35) tailed tests of significance with P Ͻ 0.05 used to deter- Contaminated 2 (5) 1 (2) mine significance. All computations were performed Antibioprophylaxis, n (%) Yes 39 (89) 50 (96) 0.26 with SPSS 10.0 for Windows software (SPSS Inc. Head- Amino-penicillin 23 (52) 30 (58) 0.28 17 quarters, Chicago, IL). Cephalosporin 9 (36) 20 (38) 0.56 Aminoglycoside 2 (5) 1 (2) 0.59 Type of clinician, n (%) 0.96 Resident 25 (57) 29 (56) Results Fellow 19 (43) 23 (44) Insertion site levels, n (%) 0.22 A total of 100 patients requiring an epidural catheter Thoracic 2 (5) 1 (2) L2–L3 11 (25) 11 (22) were enrolled in the trial and randomly assigned (48 in L3–L4 21 (47) 19 (36) the povidone iodine group and 52 in the chlorhexidine L4–L5 8 (18) 19 (36) group). Completed data could be evaluated for 96 cath- L5–S1 2 (5) 2 (4) eters (44 in the povidone iodine group and 52 in the Duration of surgery (min) 180 (80–550) 240 (90–840) 0.12 Duration of placement (h) 50 (21–74) 50 (24–100) 0.59 chlorhexidine group). The remaining four catheters Reason for removal, n (%) 0.53 sited after povidone iodine skin preparation were not Catheter no longer needed 41 (93) 49 (94) cultured (three were not inserted as a result of failure to Suspected catheter infection 2 (5) 1 (2) catheterize the epidural space and one migrated outside Catheter displacement 1 (2) 2 (4) the skin and was grossly contaminated before catheter * Continuous variables are expressed as medians with ranges and compared culture). The two groups of catheters were similar with with the Mann–Whitney U test. Categorical variables are expressed as num- respect to characteristics of patients and catheters, al- bers with percentages and compared with the Fisher exact test or the chi- though the duration of surgery was not significantly square test. ASA ϭ American Society of Anesthesiologists. higher in the chlorhexidine group (table 1). Neither local nor systemic hypersensitivity reactions were observed with the use of either antiseptic solution. 0.02, log-rank test). None of the catheters colonized Twenty cultures of catheter tips yielded microorganisms wassuspected of being infected before catheter removal. (table 2). The bacterial species isolated were coagulase- The presence of signs of inflammation (mainly ery- negative staphylococci (mainly methicillin resistant) in 19 thema) at the insertion site was not more frequent when cases and Enterococcus faecalis in one case. Cultures of the catheter was colonized (one catheter colonized catheter tips in the chlorhexidine group were significantly among the 11 patients [9%] with signs of inflammation less likely to yield microorganisms on removal (five events vs. five catheters colonized among the 85 patients [6%] [4.3 per 100 catheter days]) than were catheter tips in the without signs of inflammation; relative risk, 1.5 [95% povidone iodine group (15 events [16.7 per 100 catheter confidence interval, 0.2–11.8]; P ϭ 0.54, Fisher exact days]; relative risk, 0.2 [95% confidence interval, 0.1–0.7]; test). Methicillin-resistant coagulase-negative staphylo- P Ͻ 0.001, log-rank test). cocci were the only colonizing microorganisms recov- Six of the 20 catheters tips yielding microorganisms in ered. In no patient did an epidural abscess, meningitis, culture yielded more than 1,000 colony-forming units/ml, or any serious local or systemic infection develop. and the corresponding catheters were considered as Cultures of the sites surrounding catheter insertion in the colonized (table 2). As shown in figure 1, catheter colo- chlorhexidine group were less likely to yield microorgan- nization occurred less frequently and less quickly when isms at catheter removal than were insertion sites in the chlorhexidine was used for skin preparation (one event povidone iodine group (table 2), whereas the number of [0.9 per 100 catheter days]) than when povidone iodine catheter hubs contaminated were similar between the two was used (five events [5.6 per 100 catheter days]; rela- study groups. The skin surrounding the catheter insertion tive risk, 0.2 [95% confidence interval, 0.1–1.0]; P ϭ site was the origin of all the colonizing microorganisms.

Anesthesiology, V 94, No 2, Feb 2001 242 KINIRONS ET AL.

Table 2. Positive Bacteriologic Cultures pediatric population, with incidences varying from 12 to 1–3 Povidone Iodine Chlorhexidine 35%. Our results suggest that despite high coloniza- (n ϭ 44) (n ϭ 52) P Value tion, the risk of epidural space infection in this population is low. This finding is consistent with previous Catheter tip, n (%) 18 Ͻ1,000 cfu/ml 10 (23) 4 (8) 0.04 reports. Strafford et al., in a study of 1,620 children Ն1,000 cfu/ml 5 (11) 1 (2) 0.07 who received epidural analgesia, reported only one tho- Catheter hub, n (%) 1 (2) 4 (8) 0.4 racic epidural infection occurring with very prolonged Insertion site, n (%) 20 (45) 12 (23) 0.03 use in palliative care of a patient with immunosuppres- 2 cfu ϭ colony-forming units. sion. Likewise, Kost-Byerly et al., in a study of 210 children who received caudal or lumbar epidural anal- Discussion gesia, demonstrated that despite a 35% colonization rate, no serious systemic infection occurred. McNeely et al.1 Our study demonstrates that in the pediatric popula- reported an overall colonization rate of 20% in caudal tion studied, an alcohol solution of 0.5% chlorhexidine epidurals and similarly reported no deep tissue infec- used for skin disinfection before short-term epidural tions. Although rare, catheter infection can rapidly catheter insertion was more effective than an aqueous progress to epidural abscess, meningitis, paralysis, or solution of 10% povidone iodine in preventing catheter death.1–3 Such complications are unacceptable in this colonization. No patient had an epidural space infection. setting. Although there are several studies assessing coloniza- Although the precise mechanism of epidural infection tion of epidural catheters in the pediatric population, associated with epidural block has yet to be defined, interstudy comparisons are difficult to make because of several possible mechanisms have been proposed. Con- differences in definition of catheter colonization, route tamination of drug or material may be a factor. Raedler et of epidural used (caudal versus lumbar) and differences al.19 assessed bacterial contamination of 114 spinal and in the choice of antiseptic solution used before epidural 20 epidural needles after subarachnoid or epidural block catheter insertion. The presence or absence of prophy- performed under strict aseptic guidelines. Bacterial con- lactic antibiotics and the varying duration of epidural tamination occurred in 18% of the needles, suggesting catheterization also make interstudy comparisons diffi- than even when following strict aseptic guidelines, nee- cult. Colonization of epidural catheters is common in the dle contamination by skin pathogens is common. Simi- larly, skin flora introduced either at the time of puncture or as a result of bacterial migration along a catheter or needle tract has been implicated as a potential source of epidural abscess. Sato et al.4 assessed 69 paired skin specimens that had been excised from the incisional site (laminectomies) after disinfection with 10% povidone iodine or an alcoholic solution of 0.5% chlorhexidine. They found viable microorganisms in 13 biopsies, mainly in the povidone iodine group (32% versus 6%; P Ͻ 0.01). They explained the superior performance of chlorhexidine by its more potent bactericidal activity and its high permeability into the hair follicles. Hematogenous spread from another site of infection has been postu- lated as another potential source of epidural infection. Darchy et al.20 reviewed infectious complications in 75 intensive care patients who received epidural analgesia. Twenty-four other site infections occurred in 21 patients. The microorganisms isolated were different from those isolated from swab or catheter cultures. They stated that the presence of other infection sites did not increase the incidence of epidural infection. Finally, catheter colonization can arises from clinicians’ and Fig. 1. Kaplan-Meier curves for freedom from catheter colonization after skin preparation with either chlorhexidine or po- nurses’ handling of syringes and solutions, via the cath- vidone iodine. The number of catheters in each group that were eter hub. To determine if catheter colonization was the at risk to be colonized at various times are shown below the result of unsatisfactory skin antisepsis or septic manipu- figure. The risk of catheter colonization was significantly lower when skin preparation was performed with chlorhexidine so- lation of the catheter line, we cultured the insertion site .log-rank test). and the catheter hub at catheter removal ,0.02 ؍ lution than with povidone iodine (P

Anesthesiology, V 94, No 2, Feb 2001 CONTINUOUS EPIDURAL CATHETER COLONIZATION IN CHILDREN 243

There are few studies comparing the efficacy of chlo- the patient’s physician who decided to remove the cath- rhexidine versus povidone iodine used as skin disinfec- eters were not aware of the antiseptic solutions used. tant in reducing epidural catheter colonization. Adam et The duration of epidural use was short, with a median of al.,21 in a study of 294 obstetric epidurals, were unable 50 h in both groups. Although a short duration may to demonstrate any benefit of chlorhexidine over povi- mitigate against colonization, this duration represents done iodine. Methodologic deficiencies of this study current practice in our unit. We did not evaluate include lack of randomization and the fact that the skin whether the use of chlorhexidine reduces colonization was not cleaned before catheter removal. Thus, bias and of caudal catheters, despite the fact that caudal catheters unavoidable contamination may have altered the results. 22 are frequently placed in the pediatric population and In contrast, Shapiro et al. demonstrated that use of a that the incidence of catheter colonization is higher after chlorhexidine dressing reduced microbial colonization caudal than lumbar catheter placement, because of the of epidural catheters. The chlorhexidine group was one few number of caudal catheters inserted in our unit. The seventh as likely to be colonized as those in the control widespread use of antibiotics in this study may have group. influenced culture results. However, prophylactic anti- Our study supports these findings. Catheters inserted after skin preparation with chlorhexidine were one sixth biotic use was guided by national guidelines, and their as likely and less quickly colonized as catheters inserted use was similar in both groups. Moreover, no patient after skin preparation with povidone iodine. Our results received antibiotic potent against methicillin-resistant are in accordance with those from the previously quoted coagulase-negative staphylococci, the causative agent for study of Sato et al.,4 indicating that fewer viable micro- catheter colonization. Finally, we could not evaluate if organisms were cultured from skin biopsies after cuta- the use of chlorhexidine reduces the risk of lumbar neous disinfection with chlorhexidine. Chlorhexidine catheter infection because of the very low incidence of gluconate is a potent broad-spectrum germicide, which subcutaneous or epidural space infection with short- is effective against nearly all nosocomial bacteria and term catheterization. However, the use of catheter col- yeasts.8,9,11 In addition, chlorhexidine has a low skin onization as a surrogate marker of true risk of infection irritancy and sensitization potential. It has a strong affin- is reasonable, because colonization generally precedes ity for the skin and demonstrates prolonged duration of infection. antimicrobial effect. Unlike povidone iodine, the anti- In conclusion, the use of chlorhexidine solution rather bacterial activity of chlorhexidine persists for hours after than povidone iodine may be a better choice for cuta- topical application. In contrast to iodine-containing com- neous antisepsis before short-term epidural catheter pounds, chlorhexidine is not neutralized by contact with placement in children. Whether this antiseptic agent proteinous solutions. Iodine-containing compounds lack reduces colonization for longer lumbar catheterization persistence and may induce allergic reactions in sensi- or of caudal catheters requires further investigation. tive individuals. Finally, bacterial resistance to chlorhexidine is rare. Coagulase-negative staphylococci were the only colonizing microorganisms recovered in our study. This con- curs with previous studies indicating that coagulase- References negative staphylococci, the predominant species on the 1. McNeely JK, Trentadue NC, Rusy LM, Farber NE: Culture of bacteria from human skin, are the most common agents of cannula- lumbar and caudal epidural catheters used for postoperative analgesia in chil- 1,9,19 dren. Reg Anesth 1997; 22:428–31 related infections. The superiority of chlorhexidine 2. Kost-Byerly S, Tobin JR, Greenberg RS, Billett C, Zahurak M, Yaster M: over povidone iodine in preventing catheter coloniza- Bacterial colonization and infection rate of continuous epidural catheters in children. Anesth Analg 1998; 86:712–6 tion and catheter-related sepsis as a result of Gram- 3. Abouleish E, Orig T, Amortegui AJ: Bacterial comparisons between epidural positive bacteria has already been noted in various stud- and caudal techniques. ANESTHESIOLOGY 1980; 53:511–4 ies,9,23 and was reported to be the result of a more 4. Sato S, Sakuragi T, Dan K: Human skin flora as a potential source of epidural abscess. ANESTHESIOLOGY 1996; 85:1276–82 prolonged activity of chlorhexidine against staphylococ- 5. Snydman DR, Gorbea HF, Pober BR, Majka JA, Murray SA, Perry LK: Pre- ci.23 This superiority may explain, at least in part, the dictive value of surveillance skin cultures in total-parenteral-nutrition-related infection. Lancet 1982; 2:1385–8 lower epidural catheter colonization rate that we ob- 6. Bjornson HS, Colley R, Bower RH, Duty VP, Schwartz-Fulton JT, Fischer JE: served after chlorhexidine disinfection. Association between microorganism growth at the catheter site and colonization of the catheter in patients receiving total parenteral nutrition. Surgery 1982; Our trial had several limitations. Because color differ- 92:720–7 ences between the two solutions used, the physician 7. Auboyer C: Risque infectieux et anesthésie locorégionale. Ann Fr Anesth was not blinded, and this may have introduced bias into Réanim 1998; 17:1257–60 8. Maki D, Ringer M, Alvarado CJ: Prospective randomised trial of povidone the study. However, both solutions are in common use, iodine, alcohol, and chlorhexidine for prevention of infection associated with and we do not believe that physicians knew that one central venous and arterial catheters. Lancet 1991; 338:339–43 9. Mimoz O, Pieroni L, Lawrence C, Edouard A, Costa Y, Samii K, Brun-Buisson solution was better than the other. Furthermore, the C: Prospective randomized trial of two antiseptic solutions in prevention of microbiologists who performed the cultures as well as central venous or arterial catheter infection. Crit Care Med 1996; 24:1818–23

Anesthesiology, V 94, No 2, Feb 2001 244 KINIRONS ET AL.

10. Garland JS, Buck RK, Maloney P, Durkin DM, Toth-Lloyd S, Duffy M, Szocik Guidelines for prevention of intravascular device-related infections. I. Intravas- P, Mcauliffe TL, Goldman D: Comparison of 10% povidone iodine and 0.5% cular device-infection overview. Am J Infect Control 1996; 24:277–93 chlorhexidine gluconate for the prevention of peripheral intravenous catheter 17. SPSS Inc.: SPSS Users Guide: Statistics, Version 10.0. Chicago, SPSS Inc., 1999 colonization in neonates: a prospective trial. Pediatr Infect Dis J 1995; 14:510–6 18. Strafford MA, Wilder RT, Berde CB: The risk of infection from epidural 11. Mimoz O, Karim A, Mercat A, Cosseron M, Falissard B, Parker F, Richard C, analgesia in children: A review of 1620 cases. Anesth Analg 1995; 80:234–8 Samii K, Nordmann P: Chlorhexidine compared with povidone iodine as skin 19. Raedler C, Lass-Flörl C, Pühringer F, Kolbitsch Ch, Lingnau W, Benzer A: preparation before blood culture. Ann Intern Med 1999; 131:834–7 Bacterial contamination of needles used for spinal and epidural anaesthesia. Br J 12. Société Française d’Anesthésie et de Réanimation, Consensus Conference: Anaesth 1999; 83:657–8 ‘Antibioprophylaxis in a surgical setting.’ Réactualisation 1999. Ann Fr Anesth 20. Darchy B, Forceville X, Bavoux E, Soriot F, Domart Y: Clinical and bacte- Réanim 1999; 18: fi 75–85 riologic survey of epidural analgesia in patients in the intensive care unit. 13. Altemeir WA, Culbertson WR: Surgical infections, Surgery Principles and ANESTHESIOLOGY 1996; 85:988–98 Practice. Edited by Moyer CR, Disterhoft JF. Philadelphia, JB Lippincott, 1965, pp 185–218 21. Adam MN, Dinulescu T, Mathieu P, Giacomini T, Le Pennec MP: Com- 14. Brun-Buisson C, Abrouk F, Legrand P, Huet Y, Larabi S, Rapin M: Diagnosis paraison de l’efficacité de deux antiseptiques dans la prévention de l’infection of central venous catheter-related sepsis. Critical level of quantitative tip cultures. liée aux cathéters périduraux. Cah Anesth 1996; 44:465–7 Arch Intern Med 1987; 147:873–7 22. Shapiro JM, Bond EL, Garman JK: Use of a chlorhexidine dressing to 15. Tenover FC, Arbeit RD, Goering RV, Mickelsen PA, Murray BE, Persing DH, reduce microbial colonization of epidural catheters. ANESTHESIOLOGY 1990; 73: Swaminathan B: Interpreting chromosomal DNA restriction patterns produced 625–31 by pulse-field gel electrophoresis criteria for bacterial strain typing. J Clin Micro- 23. Shelton DM: A comparison of the effects of two antiseptic agents on biol 1995; 33:2233–9 Staphylococcus epidermidis colony forming units at the peritoneal dialysis cath- 16. Pearson ML: Hospital Infection Control Practices Advisory Committee. eter exit site. Adv Perit Dial 1991;7:120–4

Anesthesiology, V 94, No 2, Feb 2001 Art: ALN200979 Input-mn 9:06 10/5/1 11؍balt6/z7i-anesth/z7i-anesth/z7i00210/z7i7551-10a dasn S

Review Article Review Article David S. Warner, M.D., and Editor

David S. Warner, M.D., Editor Ultrasound Imaging for Regional Anesthesia in Infants, ManagingChildren, and New Adolescents Oral Anticoagulants in the Perioperative and Intensive Care Unit Setting A Review of Current Literature and Its Application in the Practice of Extremity and Trunk Blocks Jerrold H. Levy, M.D., F.A.H.A., F.C.C.M.,* David Faraoni, M.D.,† Jenna L. Spring, M.S.,‡ JamesBan C. D. H. Douketis, Tsui, M.D., M.D.,§ F.R.C.P.C.,* Charles SanthanamM. Samama, Suresh, M.D., Ph.D., M.D., F.A.A.P.†F.C.C.P.||

This article has been selected for the ANESTHESIOLOGY CME Program. Learning objectives and disclosure and ordering information can be found in the CME section at the front of this issue.

ABSTRACTABSTRACT warfarin)NE should of the mostbe considered. exciting recent This advances review insummarizes technology the in The use of ultrasound guidance has provided an opportunity to pediatric regional anesthesia has been the introduction of perform many peripheral nerve blocks that would have been difficult availableO data regarding the management of bleeding with to perform in children based on pure landmark techniques due to the novelanatomically oral anticoagulation based ultrasound agents. imaging Hemodialysis for facilitating is a therapeu nerve lo-- Managingpotential for patients injection in into the contiguous perioperative sensitive setting vascular receiving areas. novel This ticcalization. option for This dabigatran-related is because regional bleeding, anesthesia while techniques in vitro instudies chil- oralreview anticoagulation article provides agents the readers for thromboprophylaxis with techniques on ultrasound- or stroke showeddren have that been prothrombin considered challengingcomplex concentrates due to (1) targetare reported neural 10.1097/ALN.0b013e318289bcba preventionguided peripheral with atrial nerve fibrillation blocks of is the an extremities important andconsideration trunk with tostructures be useful that for often rivaroxaban-related course very close bleeding. to critical Additional structures clini (e.g.-, forcurrently clinicians. available The literature novel to oral substantiate anticoagulation the available agents evidence include for nerves of the brachial plexus run close to the pleura as they the use of these techniques. Ultrasound images of the blocks with cal studies are needed to determine the best method for reversal Deepa koshi directcorresponding Factor Xa line inhibitors diagrams rivaroxaban to demonstrate and the apixaban, placement and of the the oftraverse the novel the supraclavicularoral anticoagulation region), agents and particularlywhen bleeding during occurs. cen- directultrasound thrombin probe haveinhibitor been provideddabigatran. for all In the elective relevant surgery, nerve blocks dis- tral neuraxial blocks where the safety margin is narrow for needle continuingin children. Thetheir authors use is important, hope that this but review renal function will stimulate must further also placementNTICOAGULATION particularly close to is the routinely spinal cord, used (2) in thediverse prerequi- clini- Perioperative Management of New Oral Anticoagulants beresearch considered into ultrasound-guided because elimination regional is highly anesthesia dependent in infants, on renal chil- Asite forcal sedation settings, or including general anesthesia perioperative masking venous potential thromboem warning- dren, and adolescents and stimulate more randomized controlled signs (paresthesia), and (3) the need for limiting the volume of elimination.trials to provide If ableeding greater understandingoccurs in patients of the who anatomy have and received physi- bolism (VTE) prophylaxis and stroke prevention in patients Levy et al. theseology agents, of regional common anesthesia principles in pediatrics. of bleeding management as withlocal anestheticatrial fibrillation solution below(AF). toxic Unfractionated levels. With theheparin, possibility low- with any anticoagulant (including the known principles for molecular-weightof visualizing the targetheparin, structures, and vitamin ultrasound K antagonists technology such may asencourage warfarin manyare widely anesthesiologists used but have who disadvantages. had previously Although aban- June ** Professor,Professor, DepartmentDepartment ofof AnesthesiologyAnesthesiology/Critical and Pain Care, Medicine, Duke UniversityUniversity School of Alberta. of Medicine, † Professor, Durham, Department North ofCarolina. Pediatric † Assistant Anesthe- theredoned is regionalno “ideal” techniques anticoagulant, to resume an optimal or increase profile their includes, use of Professor,siology and Queen Pediatrics, Fabiola Children’s Children’s Memorial University Hospital, Hospital, Northwestern Brussels, inregional addition anesthesia to established in children. efficacy and safety, oral adminis- 2013 Belgium.University, ‡ Medical Feinberg Student, School ofEmory Medicine. University School of Medicine, tration,Although no routine literature monitoring evaluating requirement, the evidence a forpredictable success Atlanta,Received Georgia. from Department§ Professor of of Medicine, Anesthesiology Division and of Pain Hematology Medicine, and Thromboembolism, McMaster University, Hamilton, Ontario, anticoagulantand safety of ultrasoundeffect, a rapid in regional onset and anesthesia offset hasof action, begun toand University of Alberta, Edmonton, Alberta, Canada, and the Depart- 1 Canada.ment of || Pediatric Professor, Anesthesiology, Department of Northwestern Anesthesiology University, and Intensive Fein- reversibility.emerge, a comprehensive Only a few anticoagulants narrative review are acutely of the reversible, literature 2013 Care,berg Hotel-Dieu School of Medicine, University Chicago, Hospital, Illinois. Paris, France. Submitted for publica- includingpertaining unfractionated to techniques describedheparin with and outcomesprotamineevaluating and vita- tionReceived April 17, from 2009. the Accepted Department for publication of Anesthesiology, September Emory 30, 2009.Uni- min K antagonists with four-component prothrombin com- versitySupported School by aof Career Medicine, Scientist Atlanta, Award Georgia. in Anesthesia Submitted (to for Dr. publi Tsui),- ultrasound guidance in pediatric regional anesthesia was not Anesthesiology cationCanadian July Anesthesiologists’30, 2012. Accepted Society-Abbott for publication Laboratories December 20, Ltd., 2012. To- plexavailable concentrates. at the time The of writingnewer thisoral article.anticoagulants This review (NOACs) aims Supportronto, Ontario, was provided Canada, solely a Clinical from Scholarinstitutional Award and/or (to Dr. departmen Tsui) from- includeto provide the the direct pediatric thrombin anesthesiologist inhibitor withdabigatran an overall etexilate sum- talthe sources. Alberta HeritageJHL serves Foundation on Steering for MedicalCommittees Research, for Boehringer- Edmonton, (Pradaxamary of®,the Boehringer-Ingelheim techniques used and Pharma of the GmbH, outcomes Ingelheim found 118 Ingelheim,Alberta, Canada, Ingelheim and am Foundation Rhein, Germany; for Anesthesia CSL Behring, Education King and of Prussia,Research Pennsylvania; (FAER) Research and Johnson in Education and Johnson, grant (to Dr.New Suresh). Brunswick, am(based Rhein, on controlledGermany) orand comparative the direct factor studies) Xa asinhibitors described riva in- NewMark Jersey. A. Warner, Figures M.D.,1 and served 2 were as created Handling by EditorAnnemarie for this B. article.John- roxabanthe literature (Xarelto, on ultrasoundJohnson and guidance Johnson/Bayer of peripheral HealthCare nerve son, C.M.I., Medical Illustrator, Wake Forest University School of 6 MedicineAddress Creative correspondence Communications, to Dr. Suresh: Wake DepartmentForest University of Pediatric Medi- AG,blocks Leverkusen, of the extremities Germany) and trunkand apixaban in pediatrics. (Eliquis, A compan- Bristol calAnesthesiology, Center, Winston-Salem, Children’s MemorialNorth Carolina. Hospital, 2300 Children’s Plaza, Myersion article Squibb/Pfizer, with similar Bristol-Myers objectives related Squibb to neuraxialHouse, Uxbridge, blocks P. O. Box 19, Chicago, Illinois 60614. [email protected]. United Kingdom).2 Advantages of these new agents include1 1466 ThisAddress article maycorrespondence be accessed for to personal Dr. Levy: use Duke at no chargeUniversity through Medi the- will be published in the next issue of ANESTHESIOLOGY. In calJournal Center, Web 2301 site, www.anesthesiology.org.Erwin Rd., 5691H, HAFS, Box 3094, Durham, theiraddition relatively to case rapid series onset and clinicaland offset studies, of action descriptions and predict from- North Carolina 27710. [email protected]. This article may be accessed for personal use at no charge through the Journal Web site, able anticoagulant effect so that routine coagulation moni- 74 www.anesthesiology.org. toring is not required. However, laboratory monitoring may Anesthesiology, V 112 • No 2 473 February 2010 Copyright © 2013, the American Society of Anesthesiologists, Inc. Lippincott be relevant in certain clinical situations, where an assessment Williams & Wilkins. Anesthesiology 2013; 118:1466-74 of the anticoagulation status is needed. As the NOACs are

Anesthesiology, V 118 • No 6 1466 June 2013 EDUCATION

increasingly replacing older parenteral agents and vitamin K other NOACs that are highly protein bound, the relatively antagonists in clinical practice, it is important to consider that low protein binding of dabigatran allows it to be eliminated patients treated with these agents will be exposed to differ- to a large extent by hemodialysis.6 In cases of moderate ent clinical situations (spontaneous or postoperative bleeding, hepatic impairment, dabigatran can be administered safely overdose, trauma, and elective or emergent surgical proce- and no dose adjustment is necessary.7 dures) that require an intervention. There are also increasing Dabigatran is approved in the United States, Canada, concerns about managing patients on these therapeutic agents Europe, and Japan for stroke prevention in patients with following trauma or in a perioperative setting. The purpose non-valvular AF based on the results of the Randomized of this review is (1) to examine the NOACs, focusing on key Evaluation of Long-term anticoagulant therapY (RE-LY) pharmacologic properties, and (2) to provide management trial in which 150 mg of dabigatran twice-daily was superior approaches for users of NOACs in the perioperative and criti- to dose-adjusted warfarin with a similar rate of major bleed- cal care settings based on the available literature. ing.8 Dabigatran, 75 mg twice-daily, is approved for use in the United States for patients with severe renal insufficiency Oral Direct Thrombin Inhibitors (CrCl 15–30 ml/min), based on indirect pharmacokinetic modeling and the assumed anticoagulant effect with this Thrombin has a pivotal role in hemostasis, making it an level of renal dysfunction. In Europe and Canada, the 75-mg appealing target for anticoagulant drugs. When thrombin is dose is not approved for clinical use and dabigatran is con- activated from prothrombin, it converts soluble fibrinogen traindicated in patients with a CrCl < 30 ml/min. Dabiga- to insoluble fibrin; activates coagulation factors V, VIII, and tran is also is approved for VTE prophylaxis following total XI (which generate more thrombin); and activates platelets hip or knee replacement surgery in Europe and Canada, but (fig. 1).3 Dabigatran is a reversible direct thrombin inhibitor not the United States. A recent indirect network meta-analy- that directly inhibits free and fibrin-bound thrombin without sis suggests that treatment with dabigatran offers benefit for the need for antithrombin. Dabigatran etexilate is a prodrug the prevention of stroke, systemic embolism, and mortality that has a rapid onset of action, no reported food interactions, over antiplatelets and placebo without increased intracranial few drug interaction, and does not require routine or extracranial hemorrhage compared to antiplatelet agents.9 coagulation monitoring. The peak plasma concentration Further investigations are needed to confirm these results. is reached 1.25–3 h after administration, and it has a half- 4 life of 12–14 h in healthy volunteers. Dabigatran is 35% Oral Direct Factor Xa Inhibitors bound to plasma proteins and undergoes renal excretion, with 80% of the drug entering the urine unchanged. The Factor Xa is another important target for anticoagulant anticoagulant effect of dabigatran accumulates in the setting drugs due to its role as the rate-limiting factor in throm- of renal insufficiency, and such bioaccumulation correlates bin generation and amplification, generating the Xa com- 2 well with the degree of renal dysfunction.5 In contrast to plex that converts prothrombin to thrombin (fig. 1). The direct factor Xa inhibitors inhibit free Factor Xa, Factor Xa in the prothrombinase complex, and Factor Xa found in clots, independent of an antithrombin cofactor.2,10 This is in contrast to low-molecular-weight heparin, unfractionated heparin, and fondaparinux, which all are dependent on antithrombin to inhibit Factor Xa.

Rivaroxaban Rivaroxaban is an oral, direct Factor Xa inhibitor that has good bioavailability (80%), is highly protein-bound, and has few drug interactions. Peak plasma concentrations occur within 2–4 h of administration, and rivaroxaban has a half- life of 5–9 h in healthy subjects and 11–13 h in the elderly.10 It is selective for Factor Xa in relation to other serine prote- Fig. 1. Effect sites of anticoagulation agents. The new oral an- ases.2 Clearance of rivaroxaban may be decreased to some ticoagulation agents directly inhibit one of two major targets in extent in patients with renal impairment,11 but its primary the coagulation cascade. Rivaroxaban and apixaban directly mode of clearance is by non-renal mechanisms. It should be inhibit factor Xa, and dabigatran directly inhibits thrombin. noted that although some reports may indicate that approxi- The parenteral anticoagulants that inhibit factor Xa include low-molecular-weight heparin (LMWH) and fondaparinux mately 67% of rivaroxaban is eliminated by the kidney, such by antithrombin (AT)-dependent binding. Parenteral direct total renal clearance reflects 33% clearance of active drug thrombin inhibitors include argatroban, bivalirudin, and de- and 33% clearance of inactive rivaroxaban, which is not clin- sirudin that also directly inhibit thrombin independent of AT. ically important. Thus, two-thirds of the active rivaroxaban

Anesthesiology 2013; 118:1466-74 1467 Levy et al. Perioperative Management of New Oral Anticoagulants

are cleared by nonrenal mechanisms. Based on pharmaco- recommended in patients with mild (CrCl, 50–80 ml/min) kinetic study, 10-mg rivaroxaban administered once daily or moderate (CrCl, 30–50 ml/min) renal impairment. The offers the best efficacy profile while avoiding excessive bleed- half-life in healthy subjects is 8–15 h.10 Apixaban has been ing complications.12 In patients with AF, the recommended approved in Canada and Europe for VTE prophylaxis after rivaroxaban dose is 20 mg daily, although a reduced dose total hip and knee replacement surgery based on the results (15 mg daily) is recommended in patients with a CrCl 15– of the ADVANCE trials.22,23 Apixaban is currently under 30 ml/min.13 Due to high plasma protein binding (>90%), Food and Drug Administration review in the United States rivaroxaban cannot be eliminated during hemodialysis. for this indication. In the ARISTOLE trial, apixaban was Rivaroxaban is approved in the United States, Canada, superior to dose-adjusted warfarin in preventing stroke and and Europe for VTE prophylaxis after hip or knee systemic embolism, with a decrease in bleeding complica- replacement surgery and for stroke prevention in patients tions, and a lower mortality.24 The ApixabanVersus Acetyl- with non-valvular AF. Rivaroxaban was recently approved for salicylic Acid to Prevent Stroke in Atrial Fibrillation Patients treatment of deep vein thrombosis, pulmonary embolism, Who Have Failed or Are Unsuitable for Vitamin K Antago- and reduction in the risk of recurrence. In the Regulation nist Treatment (AVERROES) study is noteworthy as it is the of Coagulation in Orthopedic surgery to pRevent Deep only trial comparing treatment with a NOAC, in this case venous thrombosis and pulmonary embolism (RECORD) apixaban 5 mg twice-daily, against aspirin (81–325 mg) for trials, rivaroxaban, 10 mg daily, was superior to enoxaparin stroke prevention in AF.25 Although it would be expected 30 mg twice-daily, and 40 mg once-daily for the prevention that apixaban conferred a lower risk for stroke, what is sur- of VTE after knee and hip replacement, respectively, without prising, perhaps, is that this treatment was not associated a significant increase in the rate of major bleeding.14–17 In with more bleeding compared with aspirin.25 terms of stroke prevention, the Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared with Vitamin K Monitoring Anticoagulation with the NOACs Antagonism for Prevention of Stroke and Embolism Trial in Although routine laboratory testing is not required in Atrial Fibrillation (ROCKET AF) trial randomly allocated NOAC-treated patients, except for periodic monitoring of (in a double-blind manner) 14,264 patients with AF to renal function (especially in patients with preexisting renal rivaroxaban 20 mg daily (15 mg daily if CrCl 15–50 ml/ impairment), coagulation function tests should be ordered min) compared to dose-adjusted warfarin and found that for any anticoagulated patient presenting with an acute rivaroxaban was not inferior to warfarin in efficacy, with no 26 18 bleed, suspected overdose, or requiring emergency surgery. significant difference in major bleeding events. In patients Due to the effect of dabigatran on thrombin-mediated con- with an acute coronary syndrome, the Anti-Xa Therapy to version of fibrinogen to fibrin, most of the routine coagula- Lower Cardiovascular Events in Addition to Standard Therapy tion assays will be prolonged except the prothrombin time.27 in Subjects with Acute Coronary Syndrome–Thrombolysis The maximum effect of dabigatran on coagulation param- in Myocardial Infarction 51 (ATLAS ACS2–TIMI 51) eters occurs at the same time as peak plasma concentration. compared rivaroxaban, 2.5 mg or 5 mg daily, to placebo in For this reason, the delay between the last dabigatran dose patients who were receiving aspirin and a thienopyridine and the time of blood sampling is needed to interpret the (usually clopidogrel). Although the 2.5-mg dose regimen tests. The effects of dabigatran can be best measured using conferred a significant reduction in cardiovascular and all- the thrombin time or a dilute thrombin time, available as cause mortality (and also led to more bleeding), the Food the Hemoclot assay.28 Other assays that have been studied and Drug Administration issued a “complete response letter” 19 include the ecarin clotting time, although this test is not and requested additional data. widely available.2,29 The thrombin time provides a direct assessment of thrombin activity and increases linearly with Apixaban increasing dabigatran concentration; however, experience Apixaban is another oral, direct Factor Xa inhibitor with with this assay indicates that it is overly sensitive to dabiga- good oral bioavailability (80%), is highly protein bound, tran levels and, consequently, the thrombin time may be pro- reaches peak plasma concentration within 2–3 h after intake, longed in the setting of a clinically insignificant dabigatran and has limited potential for drug interactions.10 Apixaban effect.28 The Hemoclot assay has better linear correlation to 2.5 mg twice-daily is the recommended dose for VTE pro- plasma levels of dabigatran and is probably the most reliable phylaxis based on pharmacokinetic study.20 In patients who method to measure the anticoagulant effect of dabigatran.28 received apixaban 2.5 mg twice-daily for VTE prophylaxis, In the ecarin clotting time assay, prothrombin is converted the risk of major bleeding was not influenced by renal func- to meizothrombin, a prothrombin–thrombin intermediate, tion.21 Moreover, in the Apixaban for Reduction in Stroke by the snake venom ecarin, and dabigatran directly inhibits and Other Thromboembolic Events in Atrial Fibrillation this meizothrombin intermediate in a dose-dependent fash- (ARISTOTLE) trial, assessing apixaban 5 mg twice-daily for ion but this assay is not frequently used.30 A more universally stroke prevention, patients were excluded only if they had a available test, the activated partial thromboplastin time, can CrCl < 25 ml/min. For these reasons, no dose adjustment is also be used; however, the relationship between dabigatran

Anesthesiology 2013; 118:1466-74 1468 Levy et al. EDUCATION

Table 1. Preoperative Discontinuation of Dabigatran point-of-care monitoring or rapid laboratory assays might be Based on Renal Function required. Timing of Last Dose before Surgery Temporary Discontinuation before Surgery and Neuraxial Anesthesia Normal High Renal Function Bleeding Bleeding Before discontinuing any anticoagulant medication, the risk (CrCl, ml/min) Half-life, h Risk Risk of bleeding must be carefully weighed against the risk of thrombosis. For dabigatran, which is eliminated primarily >80 13 (11–12) 1 d 2–4 d 50–80 15 (12–34) 1 d 2–4 d by renal mechanisms, the timing of discontinuation should 30–50 18 (13–23) >2 d >4 d be based on patients’ CrCl and the bleeding risk associated 34,35 <30 27 (22–35) 2–5 d >5 d with the procedure (table 1). Renal impairment may be less important in patients taking rivaroxaban, in which CrCl = creatinine clearance (ml/min); d = days. a decreased creatinine clearance appears to have a limited effect on the half-life of the drug. In a study of patients with concentration and partial thromboplastin time is curvilinear, renal impairment who are receiving a single 10-mg dose of resulting in decreased precision of the assay as plasma dabi- rivaroxaban, the mean half-life only increased very slightly gatran concentrations increase,29 and there may be interassay from 8.3 h in healthy controls to 9.5 h in patients with severe variability in measurements. Nonetheless, the partial throm- renal impairment (CrCl < 30 ml/min).11 boplastin time provides a reasonable alternative if other tests The Working Group on perioperative hemostasis and are not available and a normal partial thromboplastin time the French Study Group on thrombosis and hemostasis will likely indicate the absence of a clinically important anti- published recommendations about the perioperative man- coagulant effect. Additional laboratory studies are urgently agement of NOACs.36,37 For scheduled surgery or invasive needed to correlate coagulation assay results with varying procedures with low risk of bleeding, they recommend plasma levels of dabigatran. interruption 24 h (≈2 half-lives) before the procedure and Rivaroxaban and apixaban inhibit factor Xa directly, which to restart 24 h after. In case of scheduled surgery or invasive 2 is in complex with FVa and independ of antithrombin. procedures at moderate or high risk of bleeding, a 5 days Rivaroxaban causes a prolongation of the prothrombin time, interruption before surgery is recommended, while treat- although there may be considerable inter-assay variability in ment should be restarted according to the bleeding risk. For such measurements, and has less of an effect on the partial patients at higher thrombosis risk, unfractionated heparin thromboplastin time. However, these tests are not useful for or low-molecular-weight heparin at curative dose should measuring the pharmacodynamic effects of oral factors Xa be initiated 12 h after the last dose of oral anticoagulants. 31 inhibitors. More recently, a specific assay has been developed Although these recommendations are easy to use and appear for the direct Xa inhibitors that is different from an antiXa sensible, there is a need for prospective studies assessing the assay used to monitor low-molecular-weight heparin, and efficacy and safety of these (and other) perioperative man- may provide the optimal method for determining the effect agement protocols for NOAC-treated patients who require of rivaroxaban, although further studies are needed.2,32,33 an elective surgery/procedure. Rivaroxaban produces a concentration-dependent The safety of neuraxial anesthesia for patients treated with prolongation of clotting parameters on thromboelastometry, NOACs will be based on the pharmacokinetic properties of including R and K time without significant modification the anticoagulant (table 2).36 Catheter placement and, to a of maximal amplitude, making this assay not useful for lesser extent, removal should be considered when anticoagu- routine monitoring.31 Until now, the lack of readily available lant concentrations are at their lowest, and patients should means for assessing the degree of anticoagulation remains a be monitored closely for signs of hematoma in the initial notable concern, especially in a life-threatening bleed where days after catheter removal. Specific recommendations for

Table 2. Pharmacokinetics of the New Oral Anticoagulation Agents

Dabigatran Rivaroxaban Apixaban

Route of administration Oral twice daily Oral once daily Oral twice daily Bioavailability 6.5% 80% 66% Time to maximal concentration (Tmax) 1.25–3 h 2–4 h 1–3 h Half-life 12–14 h 5–13 h 8–15 h Renal excretion 80% 66% 25% Plasma protein binding 35% >90% 87%

Anesthesiology 2013; 118:1466-74 1469 Levy et al. Perioperative Management of New Oral Anticoagulants

Table 3. Recommendations for Novel Anticoagulants for Venous Thromboembolic Prophylaxis in the Setting of Peridural/Regional Anesthesia36

Dabigatran Rivaroxaban Apixaban

Time between epidural anesthetic technique and next anticoagulant dose 2–4 h 4–6 h 6 h Time before last anticoagulant dose and epidural catheter removal NR* 22–26 h 26–30 h Time between removal of epidural catheter and next anticoagulant dose 6 h 4–6 h 4–6 h

* Dabigatran in not recommended in patients undergoing anesthesia with postoperative indwelling catheters. NR = not recommended. managing these agents for venous thromboembolic prophy- following: pacemaker/defibrillator insertion (10.3%), dental laxis in the setting of peridural/regional anesthesia are listed procedures (10.0%), diagnostic procedures (10.0%), cataract in table 3. Rosencher38 suggests allowing at least two half- removal (9.3%), colonoscopy (8.6%), and joint replacement lives to pass before catheter removal, at which point only (6.2%). The last dose of dabigatran was given a mean of 49 25% of the drug remains active. Allowing a longer interval (range: 35–85) h before the procedure, compared to a mean would only slightly reduce the drug concentration, because of 114 (range: 87–144) h for the last preprocedure dose of elimination slows after this point.38 The risk of the residual warfarin (P < 0.001). There was no significant difference in anticoagulant activity and neuraxial hematoma needs to be the rates of periprocedural major bleeding between patients weighed against the risk of VTE. However, the authors sug- receiving dabigatran, 110 mg (3.8%), or dabigatran, 150 mg gest that anticoagulation should be restarted after 8 h minus (5.1%), or warfarin (4.6%). The relative risk for major the time to reach maximum activity (Tmax), based on the bleeding with dabigatran-110 mg versus warfarin was 0.83 their suggestion that it takes 8 h to establish a stable clot, and (95% CI, 0.59–1.17; P = 0.28), and with dabigatran-150 mg allowing time for the peak of anticoagulation to be reached.38 versus warfarin, it was 1.09 (95% CI, 0.80–1.49; P = 0.58). However, there may be considerable variability in the time Among patients having urgent surgery, major bleeding needed to attain a dry vascular bed, especially after major occurred in 17.8% with dabigatran-110, 17.7% with orthopedic or oncologic surgery, and longer times may be dabigatran-150, and 21.6% with warfarin: dabigatran-110. needed because of the risk of bleeding compared to the risk There are no published data on perioperative outcomes of VTE.38 Recommendations for the use of the new anti- in patients receiving rivaroxaban or apixaban who require coagulants in the setting of neuraxial anesthesia have been elective or urgent surgery/procedures, although the same proposed by Llau et al.39 based on existing guidelines and the management principles should apply that incorporate pharmacokinetics of each drug (table 3). procedure bleeding risk and drug elimination as with A recent report evaluated bleeding rates from 7 days dabigatran-treated patients. prior until 30 days following invasive procedures for patients receiving dabigatran.40 Based on 4,591 patients Reversal of the Novel Anticoagulants and who had a first treatment interruption for an elective surgery Management of Acute Bleeding or invasive procedure, 24.7% of patients were receiving Immediate reversal of anticoagulation is often needed in the dabigatran—110 mg, 25.4% were on dabigatran—150 mg, bleeding patient or patient requiring emergency surgery. and 25.9% were on warfarin. The procedures included the Current dosing and indications for these agents are listed in

Table 4. Current Dosing Guidelines for New Oral Anticoagulation Agents

Dabigatran Rivaroxaban Apixiban*

Dosing for atrial fibrillation: 150 mg BID 20 mg QD 5 mg BID normal renal function Dosing for atrial fibrillation: 110 mg BID 15 mg QD with CrCl 15–50 ml/min 2.5 mg BID renal dysfunction United States: 75 mg BID with CrCl 15–30 ml/min DVT prophylaxis 220 mg QD 10 mg QD† 2.5 mg BID Renal dysfunction 150 mg QD Avoid with CrCl < 30 ml/min Not approved in the United States for this indication

* For apixaban, no data available for use with CrCl < 15 ml/min or on dialysis. † For hip and knee surgery. BID = twice a day; CrCl = creatinine clearance; QD = daily.

Anesthesiology 2013; 118:1466-74 1470 Levy et al. EDUCATION

table 4. Although managing any anticoagulation agent in a decrease mortality. There is insufficient evidence to recom- bleeding patient is a challenge, it is important to note that mend its use.43,44 warfarin and other vitamin K antagonist agents are not easily Recombinant factor VIIa (rFVIIa) is increasingly used in reversible with therapies available in the United States, such an “off-label” manner as a universal hemostatic and rever- as vitamin K and/or fresh frozen plasma. Four-component sal agent. However, it has not been studied in humans for prothrombin complex concentrates (PCCs) are currently reversal of NOACs, and the results of studies in animal mod- preferred in Canada and most European countries and rec- els are inconclusive. The benefit-to-risk balance for rFVIIa ommended in recent guidelines.37,41,42 Stopping a NOAC must be carefully weighed, as rFVIIa has been associated and providing supportive care are the most important con- with an increased risk of arterial thrombosis among elderly sideration and are often sufficient if the bleeding is not severe patients.45,46 In several animal models, rFVIIa reversed bleed- or if surgery can be delayed. However, when patients present ing time prolongation associated with dabigatran and rivar- with a major bleeding episode related to these agents and/or oxaban, but it did not correct the underlying coagulopathy 44 require emergency surgery, other measures must be taken. as suggested by other laboratory markers. In one study, rats For any significant bleeding event, initial measures should received high-dose dabigatran before a standard tail incision, include volume resuscitation with fluids and/or packed red prolonging the bleeding time from 125 s in controls to 1455 blood cells, identification of the bleeding source, and attempts s. A 0.5 mg/kg dose of rFVIIa decreased this bleeding time to at local hemostatic control. If an anticoagulant overdose is 135 s. The partial thromboplastin time was 58 s after dabi- the suspected cause, activated charcoal may be effective in gatran exposure, versus 7 s in controls, and administration of preventing additional drug absorption when administered rFVIIa reduced this to 27 s. In a similar study that exposed within 1–2 h of ingestion. Activated charcoal has not been rats to supratherapeutic dabigatran levels, administration of rFVIIa rapidly corrected the bleeding time and preserved used in the clinical setting and is limited by its narrow win- 47 dow of use and inability to use in a perioperative setting. this effect for the entire 2-h study period. However, of the coagulation markers examined, only the prothrombin time Hemodialysis or hemoperfusion is another potential was completely corrected. The partial thromboplastin time, option for the emergent removal of anticoagulants. the ecarin clotting time, and the thrombin time all failed to Rivaroxaban and apixaban are too highly protein bound to normalize.47 be effectively removed by these methods, but dabigatran is PCCs are available as three-factor (II, IX, X) and four- an appropriate candidate for these therapies.10 In a study of factor (II, VII, IX, X) varieties that are procoagulant and six volunteers with end-stage renal disease who were given a enhance thrombin generation.42 The four-factor PCCs have 50-mg dose of dabigatran before routine hemodialysis, an activated and nonactivated forms, and only three-factor average of 62% of the active dabigatran was removed after PCCs are available in the United States.44 Small quantities of 2 h and 68% after 4 h.5 Unfortunately, attempting to perform heparin, antithrombin, protein C, and protein S are added either of these procedures in a bleeding patient in shock may 29 to the concentrates to reduce coagulation activation through not be possible. Therefore, the use of procoagulant agents endogenous pathways,48 but caution is required for off-label should be considered for a life-threatening bleed. However, use as thrombotic events in 1–3% of treated patients have unlike when fresh frozen plasma or PCCs are used to replace been reported with both formulations.49,50 depleted factors II, VII, IX, and X in warfarin-treated Unlike the other procoagulant agents, four-factor PCCs patients, the effectiveness of such clotting factor replacement have been studied as potential reversal agents in humans. In a therapies may be limited in NOAC-treated patients who small, randomized, double-blinded, placebo-controlled trial, do not harbor deficiencies of clotting factors but in whom 12 healthy men were given dabigatran 150 mg twice-daily there is an ongoing clotting factor inhibitory effect. It may or (supratherapeutic) rivaroxaban 20 mg twice-daily for 2.5 be argued, therefore, that providing supraphysiologic levels days, and subsequently received either a 50 IU/kg bolus of a of clotting factors may be ineffective in the setting of an four-factor PCC or saline.51 They were switched to the other ongoing NOAC-related inhibitory effect. On the contrary, anticoagulant following a wash-out period and the procedure such clotting factors may overwhelm such an ongoing was repeated. The addition of the PCC completely reversed inhibitory effect and, in cases of severe bleeding, may replace both the prothrombin time prolongation and inhibition of clotting factors that are depleted due to consumption. endogenous thrombin potential associated with rivaroxaban, but Although fresh frozen plasma is commonly administered only laboratory parameters were evaluated in this study, and no for initial control of bleeding in anticoagulated patients, its bleeding outcomes were measured in the volunteers. However, use as a reversal agent for the NOACs has not been studied dabigatran-associated prolongations in the clotting assays and in humans.43 In a study of mice pretreated with dabigatran endogenous thrombin potential lag time were not corrected by (4.5 mg/kg or 9.0 mg/kg) before induction of intracranial the administration of the PCC. Recently, in an animal model, hemorrhage, fresh frozen plasma administration successfully although both rFVIIa and PCC partially corrected laboratory limited hematoma expansion in the low-dose group but had assays (thromboelastography and thrombin generation assay), no effect in the high-dose group and did not significantly none of them reduced bleeding induced by rivaroxaban.52

Anesthesiology 2013; 118:1466-74 1471 Levy et al. Perioperative Management of New Oral Anticoagulants

Fig. 2. Management strategies for patients bleeding who have received the novel oral anticoagulation agents. In cases of mild bleeding, stopping or delaying the next dose should be considered. The new agents including dabigatran, rivaroxaban, and recently approved apixiban have relatively short half-lives, so stopping the drug in patients with normal renal function, the anticoagulant effect rapidly decreases compared to warfarin. In patients with moderate to severe bleeding, standard therapeutic approaches should be considered, including supportive care that includes volume resuscitation, hemodynamic support with vasoactive therapy, blood product transfusions as determined by testing, and identification of bleeding source that may require surgical or another intervention. If the agents were taken within ≈2 h of admission, administration of oral activated charcoal should be considered. For dabigatran,† hemodialysis can remove ≈60% of the drug after several hours of dialysis and should be considered in patients with impaired renal function who are bleeding and will have altered clearance. Apixaban and rivaroxaban are highly protein bound and will not be cleared by hemodialysis. However, emergency access for hemodialysis requires vascular access with large bore catheters that may pose additional risk in the anticoagulated patient. For patients with life-threatening bleeding, hemodynamic and hemostatic resuscitation should be considered, with therapy similar to that for a trauma patient including the use of a massive transfusion protocol. Based on current data as discussed in the manuscript, the use of either three-factor or four-factor prothrombin complex concentrates (PCCs) depending on their availability should be considered as they have been shown to reverse or partially reverse the anticoagulation effect of the newer agents. ‡ In patients receiving dabigatran, the use of an activated PCC may be more effective. † However, there are no studies reporting the use of PCCs on actual bleeding in patients, and further studies including the development of specific reversal agents are underway currently. In hypotensive patients, hemodialysis is unlikely to be tolerated, and alternate methods for hemofiltration should be considered if needed. † The use of recombinant activated factor VIIa (rVIIa) decreases bleeding times in animal models, but there are no human studies to determine if this is effective. † = for dabigatran. ‡ = for rivaroxaban and apixiban.

Marlu et al.53 evaluated dabigatran and rivaroxaban rever- Ongoing clinical studies are needed to determine the sal using thrombin generation tests evaluating 10 healthy best method for reversal of the NOACs when bleeding volunteers randomized to receive rivaroxaban (20 mg) or occurs. Based on the available evidence, supportive care dabigatran (150 mg) orally in a cross-over study, and blood and interventions as discussed including dialysis for dabi- was collected 2 h post-ingestion. Anticoagulation reversal gatran should be considered in a bleeding patient and was tested in vitro using PCC, rFVIIa, or factor eight inhibi- potential therapeutic approaches as listed in figure 2. For tory bypass activity at different concentrations.53 In rivar- rivaroxaban-treated patients, studies in volunteers sug- oxaban-treated patients, PCC and factor eight inhibitory gest the PCCs may be effective, but additional studies are bypass activity corrected thrombin generation, but rFVIIa needed. When possible, these drugs should be stopped only modified the kinetic parameters. In dabigatran-treated preoperatively at times based on renal function and pro- patients, PCC increased thrombin generation as determined cedure.54 Additional drug-specific antidotes are also under by area under the curve, but only rFVIIa and factor eight investigation. inhibitory bypass activity corrected the altered lag time.53 ‍References Summary 1. Eikelboom JW, Weitz JI: New anticoagulants. Circulation 2010; 121:1523–32 Common principles of bleeding management as with any 2. Levy JH, Key NS, Azran MS: Novel oral anticoagulants: anticoagulant (including the known principles for warfa- Implications in the perioperative setting. Anesthesiology rin) should be followed in patients receiving dabigatran. 2010; 113:726–45 3. Weitz JI: Factor Xa or thrombin: Is thrombin a better target? Hemodialysis is an additional, unique therapeutic option J Thromb Haemost 2007; 5:65–7 for urgently reducing exposure to dabigatran that has not 4. Eriksson BI, Quinlan DJ, Weitz JI: Comparative pharmaco- been shown to be useful for other new oral anticoagulants. dynamics and pharmacokinetics of oral direct thrombin and

Anesthesiology 2013; 118:1466-74 1472 Levy et al. EDUCATION

factor xa inhibitors in development. Clin Pharmacokinet 18. Patel MR, Mahaffey KW, Garg J, Pan G, Singer DE, Hacke 2009; 48:1–22 W, Breithardt G, Halperin JL, Hankey GJ, Piccini JP, Becker 5. Stangier J, Rathgen K, Stähle H, Mazur D: Influence of renal RC, Nessel CC, Paolini JF, Berkowitz SD, Fox KA, Califf RM; impairment on the pharmacokinetics and pharmacodynam- ROCKET AF Investigators: Rivaroxaban versus warfarin in ics of oral dabigatran etexilate: An open-label, parallel-group, nonvalvular atrial fibrillation. N Engl J Med 2011; 365:883–91 single-centre study. Clin Pharmacokinet 2010; 49:259–68 19. Mega JL, Braunwald E, Wiviott SD, Bassand JP, Bhatt DL, 6. Khadzhynov D, Wagner F, Formella S, Wiegert E, Moschetti Bode C, Burton P, Cohen M, Cook-Bruns N, Fox KA, Goto S, V, Slowinski T, Neumayer HH, Liesenfeld KH, Lehr T, Hartter Murphy SA, Plotnikov AN, Schneider D, Sun X, Verheugt FW, S, Friedman J, Peters H, Clemens A: Effective elimination of Gibson CM; ATLAS ACS 2–TIMI 51 Investigators: Rivaroxaban dabigatran by haemodialysis: A phase I single-centre study in in patients with a recent acute coronary syndrome. N Engl J patients with end-stage renal disease. Thromb Haemost 2013 Med 2012; 366:9–19 Feb 7; 109 [Epub ahead of print] 20. Leil TA, Feng Y, Zhang L, Paccaly A, Mohan P, Pfister M: 7. Stangier J, Stähle H, Rathgen K, Roth W, Shakeri-Nejad K: Quantification of apixaban’s therapeutic utility in preven- Pharmacokinetics and pharmacodynamics of dabigatran tion of venous thromboembolism: Selection of phase III trial etexilate, an oral direct thrombin inhibitor, are not affected dose. Clin Pharmacol Ther 2010; 88:375–82 by moderate hepatic impairment. J Clin Pharmacol 2008; 21. DeLoughery TG: Practical aspects of the oral new anticoagu- 48:1411–9 lants. Am J Hematol 2011; 86:586–90 8. Connolly SJ, Ezekowitz MD, Yusuf S, Eikelboom J, Oldgren J, 22. Lassen MR, Raskob GE, Gallus A, Pineo G, Chen D, Hornick Parekh A, Pogue J, Reilly PA, Themeles E, Varrone J, Wang S, P; ADVANCE-2 investigators: Apixaban versus enoxaparin for Alings M, Xavier D, Zhu J, Diaz R, Lewis BS, Darius H, Diener thromboprophylaxis after knee replacement (ADVANCE-2): HC, Joyner CD, Wallentin L; RE-LY Steering Committee and A randomised double-blind trial. Lancet 2010; 375:807–15 Investigators: Dabigatran versus warfarin in patients with 23. Lassen MR, Gallus A, Raskob GE, Pineo G, Chen D, Ramirez atrial fibrillation. N Engl J Med 2009; 361:1139–51 LM; ADVANCE-3 Investigators: Apixaban versus enoxaparin 9. Roskell NS, Lip GY, Noack H, Clemens A, Plumb JM: for thromboprophylaxis after hip replacement. N Engl J Med Treatments for stroke prevention in atrial fibrillation: A net- 2010; 363:2487–98 work meta-analysis and indirect comparisons versus dabiga- 24. Granger CB, Alexander JH, McMurray JJ, Lopes RD, Hylek tran etexilate. Thromb Haemost 2010; 104:1106–15 EM, Hanna M, Al-Khalidi HR, Ansell J, Atar D, Avezum A, 10. Eriksson BI, Quinlan DJ, Weitz JI: Comparative pharmaco- Bahit MC, Diaz R, Easton JD, Ezekowitz JA, Flaker G, Garcia D, Geraldes M, Gersh BJ, Golitsyn S, Goto S, Hermosillo AG, dynamics and pharmacokinetics of oral direct thrombin and Hohnloser SH, Horowitz J, Mohan P, Jansky P, Lewis BS, factor xa inhibitors in development. Clin Pharmacokinet Lopez-Sendon JL, Pais P, Parkhomenko A, Verheugt FW, Zhu 2009; 48:1–22 J, Wallentin L; ARISTOTLE Committees and Investigators: 11. Kubitza D, Becka M, Mueck W, Halabi A, Maatouk H, Klause Apixaban versus warfarin in patients with atrial fibrillation. N, Lufft V, Wand DD, Philipp T, Bruck H: Effects of renal N Engl J Med 2011; 365:981–92 impairment on the pharmacokinetics, pharmacodynamics 25. Connolly SJ, Eikelboom J, Joyner C, Diener HC, Hart R, and safety of rivaroxaban, an oral, direct Factor Xa inhibitor. Golitsyn S, Flaker G, Avezum A, Hohnloser SH, Diaz R, Br J Clin Pharmacol 2010; 70:703–12 Talajic M, Zhu J, Pais P, Budaj A, Parkhomenko A, Jansky 12. Eriksson BI, Borris LC, Dahl OE, Haas S, Huisman MV, P, Commerford P, Tan RS, Sim KH, Lewis BS, Van Mieghem Kakkar AK, Muehlhofer E, Dierig C, Misselwitz F, Kälebo W, Lip GY, Kim JH, Lanas-Zanetti F, Gonzalez-Hermosillo P; ODIXa-HIP Study Investigators: A once-daily, oral, direct A, Dans AL, Munawar M, O’Donnell M, Lawrence J, Lewis Factor Xa inhibitor, rivaroxaban (BAY 59-7939), for throm- G, Afzal R, Yusuf S; AVERROES Steering Committee and boprophylaxis after total hip replacement. Circulation 2006; Investigators: Apixaban in patients with atrial fibrillation. N 114:2374–81 Engl J Med 2011; 364:806–17 13. Fox KA, Piccini JP, Wojdyla D, Becker RC, Halperin JL, Nessel 26. Garcia D, Libby E, Crowther MA: The new oral anticoagu- CC, Paolini JF, Hankey GJ, Mahaffey KW, Patel MR, Singer lants. Blood 2010; 115:15–20 DE, Califf RM: Prevention of stroke and systemic embolism 27. van Ryn J, Baruch L, Clemens A: Interpretation of point-of- with rivaroxaban compared with warfarin in patients with care INR results in patients treated with dabigatran. Am J non-valvular atrial fibrillation and moderate renal impair- Med 2012; 125:417–20 ment. Eur Heart J 2011; 32:2387–94 28. Stangier J, Feuring M: Using the HEMOCLOT direct thrombin 14. Eriksson BI, Borris LC, Friedman RJ, Haas S, Huisman inhibitor assay to determine plasma concentrations of dabi- MV, Kakkar AK, Bandel TJ, Beckmann H, Muehlhofer E, gatran. Blood Coagul Fibrinolysis 2012; 23:138–43 Misselwitz F, Geerts W; RECORD1 Study Group: Rivaroxaban versus enoxaparin for thromboprophylaxis after hip arthro- 29. van Ryn J, Stangier J, Haertter S, Liesenfeld KH, Wienen W, plasty. N Engl J Med 2008; 358:2765–75 Feuring M, Clemens A: Dabigatran etexilate–a novel, reversible, oral direct thrombin inhibitor: Interpretation of coagu- 15. Kakkar AK, Brenner B, Dahl OE, Eriksson BI, Mouret P, lation assays and reversal of anticoagulant activity. Thromb Muntz J, Soglian AG, Pap AF, Misselwitz F, Haas S; RECORD2 Haemost 2010; 103:1116–27 Investigators: Extended duration rivaroxaban versus short- term enoxaparin for the prevention of venous thromboembo- 30. Stangier J, Rathgen K, Stähle H, Gansser D, Roth W: The lism after total hip arthroplasty: A double-blind, randomised pharmacokinetics, pharmacodynamics and tolerability of controlled trial. Lancet 2008; 372:31–9 dabigatran etexilate, a new oral direct thrombin inhibitor, in healthy male subjects. Br J Clin Pharmacol 2007; 64:292–303 16. Lassen MR, Ageno W, Borris LC, Lieberman JR, Rosencher N, Bandel TJ, Misselwitz F, Turpie AG; RECORD3 Investigators: 31. Samama MM, Martinoli JL, LeFlem L, Guinet C, Plu-Bureau Rivaroxaban versus enoxaparin for thromboprophylaxis after G, Depasse F, Perzborn E: Assessment of laboratory assays total knee arthroplasty. N Engl J Med 2008; 358:2776–86 to measure rivaroxaban–an oral, direct factor Xa inhibitor. 17. Turpie AG, Lassen MR, Davidson BL, Bauer KA, Gent M, Thromb Haemost 2010; 103:815–25 Kwong LM, Cushner FD, Lotke PA, Berkowitz SD, Bandel TJ, 32. Favaloro EJ, Lippi G, Koutts J: Laboratory testing of anti- Benson A, Misselwitz F, Fisher WD; RECORD4 Investigators: coagulants: The present and the future. Pathology 2011; Rivaroxaban versus enoxaparin for thromboprophylaxis 43:682–92 after total knee arthroplasty (RECORD4): A randomised trial. 33. Samama MM, Amiral J, Guinet C, Perzborn E, Depasse F: An Lancet 2009; 373:1673–80 optimised, rapid chromogenic assay, specific for measuring

Anesthesiology 2013; 118:1466-74 1473 Levy et al. Perioperative Management of New Oral Anticoagulants

direct factor Xa inhibitors (rivaroxaban) in plasma. Thromb 44. Kaatz S, Kouides PA, Garcia DA, Spyropolous AC, Crowther Haemost 2010; 104:1078–9 M, Douketis JD, Chan AK, James A, Moll S, Ortel TL, Van 34. Stangier J, Rathgen K, Stähle H, Mazur D: Influence of renal Cott EM, Ansell J: Guidance on the emergent reversal of impairment on the pharmacokinetics and pharmacodynam- oral thrombin and factor Xa inhibitors. Am J Hematol 2012; ics of oral dabigatran etexilate: An open-label, parallel-group, 87:S141–5 single-centre study. Clin Pharmacokinet 2010; 49:259–68 45. Simpson E, Lin Y, Stanworth S, Birchall J, Doree C, Hyde 35. Huisman MV, Lip GY, Diener HC, Brueckmann M, van Ryn C: Recombinant factor VIIa for the prevention and treat- J, Clemens A: Dabigatran etexilate for stroke prevention in ment of bleeding in patients without haemophilia. Cochrane patients with atrial fibrillation: Resolving uncertainties in Database Syst Rev 2012; 3:CD005011 routine practice. Thromb Haemost 2012; 107:838–47 46. Levi M, Levy JH, Andersen HF, Truloff D: Safety of recombi- 36. Gogarten W, Vandermeulen E, Van Aken H, Kozek S, Llau JV, nant activated factor VII in randomized clinical trials. N Engl Samama CM; European Scoeity of Anaesthesiology: Regional J Med 2010; 363:1791–800 anaesthesia and antithrombotic agents: Recommendations of 47. van Ryn J, Schurer J, Knk-Elband M, Clemens A: The success- the European Society of Anaesthesiology. Eur J Anaesthesiol ful reversal of dabigatran induced bleeding by coagulation 2010; 27:999–1015 factor concentrates in a rat tail bleeding model do not cor- 37. Sié P, Samama CM, Godier A, Rosencher N, Steib A, Llau JV, relate with ex vivo markers of anticoagulation. Blood 2011; Van der Linden P, Pernod G, Lecompte T, Gouin-Thibault I, 118: Abst 2318 Albaladejo P; Working Group on Perioperative Haemostasis; 48. Patanwala AE, Acquisto NM, Erstad BL: Prothrombin com- French Study Group on Thrombosis and Haemostasis: plex concentrate for critical bleeding. Ann Pharmacother Surgery and invasive procedures in patients on long-term 2011; 45:990–9 treatment with direct oral anticoagulants: Thrombin or fac- 49. Arnold DM, Dentali F, Crowther MA, Meyer RM, Cook RJ, tor-Xa inhibitors. Recommendations of the Working Group Sigouin C, Fraser GA, Lim W, Kelton JG: Systematic review: on Perioperative Haemostasis and the French Study Group Efficacy and safety of rituximab for adults with idiopathic on Thrombosis and Haemostasis. Arch Cardiovasc Dis 2011; thrombocytopenic purpura. Ann Intern Med 2007; 146:25–33 104:669–76 50. Dentali F, Marchesi C, Pierfranceschi MG, Crowther M, 38. Rosencher N, Bonnet MP, Sessler DI: Selected new anti- Garcia D, Hylek E, Witt DM, Clark NP, Squizzato A, Imberti thrombotic agents and neuraxial anaesthesia for major D, Ageno W: Safety of prothrombin complex concentrates orthopaedic surgery: Management strategies. Anaesthesia for rapid anticoagulation reversal of vitamin K antagonists. A 2007; 62:1154–60 meta-analysis. Thromb Haemost 2011; 106:429–38 39. Llau JV, Ferrandis R: New anticoagulants and regional anes- 51. Eerenberg ES, Kamphuisen PW, Sijpkens MK, Meijers JC, thesia. Curr Opin Anaesthesiol 2009; 22: 661–6 Buller HR, Levi M: Reversal of rivaroxaban and dabigatran by 40. Healey JS, Eikelboom J, Douketis J, Wallentin L, Oldgren J, prothrombin complex concentrate: A randomized, placebo- Yang S, Themeles E, Heidbuchel H, Heidbuchle H, Avezum controlled, crossover study in healthy subjects. Circulation A, Reilly P, Connolly SJ, Yusuf S, Ezekowitz M; RE-LY 2011; 124:1573–9 Investigators: Periprocedural bleeding and thromboembolic 52. Godier A, Miclot A, Le Bonniec B, Durand M, Fischer AM, events with dabigatran compared with warfarin: Results from Emmerich J, Marchand-Leroux C, Lecompte T, Samama CM: the Randomized Evaluation of Long-Term Anticoagulation Evaluation of prothrombin complex concentrate and recom- Therapy (RE-LY) randomized trial. Circulation 2012; binant activated factor VII to reverse rivaroxaban in a rabbit 126:343–8 model. Anesthesiology 2012; 116:94–102 41. Garcia D: Rethinking warfarin reversal. Blood 2010; 53. Marlu R, Hodaj E, Paris A, Albaladejo P, Crackowski JL, 116:675–6 Pernod G: Effect of non-specific reversal agents on antico- 42. Levy JH, Tanaka KA, Dietrich W: Perioperative hemostatic agulant activity of dabigatran and rivaroxaban: A randomised management of patients treated with vitamin K antagonists. crossover ex vivo study in healthy volunteers. Thromb Anesthesiology 2008; 109:918–26 Haemost 2012; 108:217–24 43. van Ryn J, Stangier J, Haertter S, Liesenfeld KH, Wienen W, 54. Kaatz S, Kouides PA, Garcia DA, Spyropolous AC, Crowther Feuring M, Clemens A: Dabigatran etexilate–a novel, revers- M, Douketis JD, Chan AK, James A, Moll S, Ortel TL, Van ible, oral direct thrombin inhibitor: Interpretation of coagu- Cott EM, Ansell J: Guidance on the emergent reversal of lation assays and reversal of anticoagulant activity. Thromb oral thrombin and factor Xa inhibitors. Am J Hematol 2012; Haemost 2010; 103:1116–27 87:S141–5

Anesthesiology 2013; 118:1466-74 1474 Levy et al. T h e new england journal o f medicine

original article

A Trial of Intraoperative Low-Tidal-Volume Ventilation in Abdominal Surgery Emmanuel Futier, M.D., Jean-Michel Constantin, M.D., Ph.D., Catherine Paugam-Burtz, M.D., Ph.D., Julien Pascal, M.D., Mathilde Eurin, M.D., Arthur Neuschwander, M.D., Emmanuel Marret, M.D., Marc Beaussier, M.D., Ph.D., Christophe Gutton, M.D., Jean-Yves Lefrant, M.D., Ph.D., Bernard Allaouchiche, M.D., Ph.D., Daniel Verzilli, M.D., Marc Leone, M.D., Ph.D., Audrey De Jong, M.D., Jean-Etienne Bazin, M.D., Ph.D., Bruno Pereira, Ph.D., and Samir Jaber, M.D., Ph.D., for the IMPROVE Study Group*

ABSTRACT

From the Département d’Anesthésie et Ré BACKGROUND animation, Hôpital Estaing (E.F., J.-M.C., Lung-protective ventilation with the use of low tidal volumes and positive end- J.P., J.-E.B.), Université de Clermont-Fer- rand, Retinoids, Reproduction, and Devel- expiratory pressure is considered best practice in the care of many critically ill opmental Diseases Unit, Équipe Accueil patients. However, its role in anesthetized patients undergoing major surgery is 7281 (E.F., J.-M.C.), and the Biostatistics not known. Unit, Direction de la Recherche Clinique (B.P.), Centre Hospitalier Universitaire (CHU) de Clermont-Ferrand, Clermont- METHODS Ferrand; Assistance Publique–Hôpitaux de In this multicenter, double-blind, parallel-group trial, we randomly assigned 400 Paris (AP-HP), Département d’Anesthé adults at intermediate to high risk of pulmonary complications after major ab- sie et Réanimation, Hôpital Beaujon, Hôpi- taux Universitaires Paris Nord Val de Seine dominal surgery to either nonprotective mechanical ventilation or a strategy of and Université Paris Diderot, Sorbonne lung-protective ventilation. The primary outcome was a composite of major pul- Paris Cité (C.P.-B., M.E., A.N.), Départe ment d’Anesthésie et Réanimation, Hôpi- monary and extrapulmonary complications occurring within the first 7 days after tal Tenon (E.M.), and AP-HP, Départe surgery. ment d’Anesthésie et Réanimation, Hôpital Saint-Antoine (M.B., C.G.), Paris; RESULTS CHU de Nîmes, Section d’Anesthésie and Département d’Anesthésie et Réani- The two intervention groups had similar characteristics at baseline. In the inten- mation, Nîmes (J.-Y.L.); CHU de Lyon, tion-to-treat analysis, the primary outcome occurred in 21 of 200 patients (10.5%) Département d’Anesthésie et Réanima- assigned to lung-protective ventilation, as compared with 55 of 200 (27.5%) assigned tion, Hôpital Edouard Herriot, Lyon (B.A.); CHU de Montpellier, Département d’Anes to nonprotective ventilation (relative risk, 0.40; 95% confidence interval [CI], 0.24 thésie et Réanimation B, Hôpital Saint- to 0.68; P = 0.001). Over the 7-day postoperative period, 10 patients (5.0%) assigned Eloi, and INSERM Unité 1046 and Univer- to lung-protective ventilation required noninvasive ventilation or intubation for sité Montpellier 1, Montpellier (D.V., A.D.J., S.J.); and Assistance Publique–Hôpital acute respiratory failure, as compared with 34 (17.0%) assigned to nonprotective de Marseille, Département d’Anesthésie ventilation (relative risk, 0.29; 95% CI, 0.14 to 0.61; P = 0.001). The length of the hospi- et Réanimation, Hôpital Nord, Marseille tal stay was shorter among patients receiving lung-protective ventilation than among (M.L.) — all in France. Address reprint requests to Dr. Jaber at the Département those receiving nonprotective ventilation (mean difference, −2.45 days; 95% CI, d’Anesthésie et Réanimation B (DAR B), −4.17 to −0.72; P = 0.006). 80 Ave. Augustin Fliche, 34295 Montpellier, France, or at [email protected]. CONCLUSIONS * Additional investigators in the Intraop- As compared with a practice of nonprotective mechanical ventilation, the use of erative Protective Ventilation (IMPROVE) Study Group are listed in the Supplemen- a lung-protective ventilation strategy in intermediate-risk and high-risk patients tary Appendix, available at NEJM.org. undergoing major abdominal surgery was associated with improved clinical out- N Engl J Med 2013;369:428-37. comes and reduced health care utilization. (IMPROVE ClinicalTrials.gov number, DOI: 10.1056/NEJMoa1301082 NCT01282996.) Copyright © 2013 Massachusetts Medical Society.

428 n engl j med 369;5 nejm.org august 1, 2013 The New England Journal of Medicine Downloaded from nejm.org at MERCY HOSP ST LOUIS on August 12, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Intraoperative Low-Tidal-Volume Ventilation

orldwide, more than 230 million METHODS patients undergoing major surgery each year require general anesthesia and TRIAL DESIGN AND OVERSIGHT W 1 mechanical ventilation. Postoperative pulmo- The IMPROVE trial was an investigator-initiated, nary complications adversely affect clinical out- multicenter, double-blind, stratified, parallel-group, comes and health care utilization,2 so preven- clinical trial. Randomization was performed with tion of these complications has become a the use of a computer-generated assignment se- measure of the quality of hospital care.3 Previ- quence and a centralized telephone system. The ous, large cohort studies have shown that 20 to study protocol and statistical analysis plan were 30% of patients undergoing surgery with gen- approved for all centers by a central ethics commit- eral anesthesia are at intermediate to high risk tee (Comité de Protection des Personnes Sud-Est I, for postoperative pulmonary complications.4,5 Saint-Etienne, France) according to French law. Mechanical ventilation with the use of high The protocol, including the statistical analysis tidal volumes (10 to 15 ml per kilogram of pre- plan, is available with the full text of this article at dicted body weight) has traditionally been rec- NEJM.org. An independent data and safety moni- ommended to prevent hypoxemia and atelecta- toring committee oversaw the study conduct and sis in anesthetized patients.6 There is, however, reviewed blinded safety data. The members of considerable evidence from experimental and the steering committee (see the Supplementary observational studies that mechanical ventila- Appendix, available at NEJM.org) vouch for the tion — in particular, high tidal volumes that accuracy and completeness of the data and analy- cause alveolar overstretching — can initiate ses and the fidelity of the study to the protocol. ventilator-associated lung injury7 and contrib- There was no industry support or involvement in ute to extrapulmonary organ dysfunction the trial. through systemic release of inflammatory me- Patients were screened and underwent random- diators.8,9 ization between January 31, 2011, and August 10, Lung-protective ventilation, which refers to 2012, at seven French university teaching hospitals. the use of low tidal volumes and positive end- Written informed consent was obtained before expiratory pressure (PEEP), and which may also randomization from each patient, on the day be- include the use of recruitment maneuvers (peri- fore surgery. Randomization was stratified accord- odic hyperinflation of the lungs),10 has been ing to study site and the planned use or nonuse of shown to reduce mortality among patients with postoperative epidural analgesia, which is a factor the acute respiratory distress syndrome11 and is that may influence outcomes.21 Treatment assign- now considered best practice in the care of ments were concealed from patients, research staff, many critically ill patients.12 Although this ap- the statistician, and the data and safety monitor- proach may be beneficial in a broader popula- ing committee. Although the staff members who tion,13,14 some physicians have questioned the collected data during surgery were aware of the benefits of using lung-protective ventilation in group assignments, outcome assessors were un- the surgical setting,15-18 especially since the use aware of these assignments throughout the study. of high tidal volumes and no PEEP is still com- monplace and less than 20% of patients receive PATIENTS protective ventilation in routine anesthetic Patients were eligible for participation in the practice.19,20 study if they were older than 40 years of age, We conducted the Intraoperative Protective were scheduled to undergo laparoscopic or non- Ventilation (IMPROVE) trial to determine wheth- laparoscopic elective major abdominal surgery1 er a multifaceted strategy of prophylactic lung- with an expected duration of at least 2 hours, and protective ventilation that combined low tidal had a preoperative risk index for pulmonary volumes, PEEP, and recruitment maneuvers complications5 of more than 2. The risk index could improve outcomes after abdominal sur- uses risk classes that range from 1 to 5, with gery, as compared with the standard practice of higher risk classes indicating a higher risk of nonprotective mechanical ventilation. postoperative pulmonary complications (see the

n engl j med 369;5 nejm.org august 1, 2013 429 The New England Journal of Medicine Downloaded from nejm.org at MERCY HOSP ST LOUIS on August 12, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. T h e new england journal o f medicine

Supplementary Appendix). Patients were ineligible (defined according to standard criteria; see the if they had received mechanical ventilation within Supplementary Appendix) or the need for invasive the 2 weeks preceding surgery, had a body-mass or noninvasive ventilation for acute respiratory fail- index (the weight in kilograms divided by the ure. Major extrapulmonary complications were de- square of the height in meters) of 35 or higher, fined as sepsis, severe sepsis and septic shock (de- had a history of respiratory failure or sepsis with- fined according to consensus criteria),22 or death. in the 2 weeks preceding surgery, had a require- Secondary outcomes within the 30-day fol- ment for intrathoracic or emergency surgery, or low-up period were the incidence of pulmonary had a progressive neuromuscular illness. complications due to any cause, graded on a scale from 0 (no pulmonary complications) to 4 (the INTERVENTIONS most severe complications)23 (see the Supplemen- Patients were assigned to receive volume-con- tary Appendix); ventilation-related adverse events trolled mechanical ventilation according to one during surgery; postoperative gas exchange; un- of two strategies: nonprotective ventilation with expected need for admission to the intensive a tidal volume of 10 to 12 ml per kilogram of care unit (ICU); extrapulmonary complications; predicted body weight, with no PEEP and no re- durations of ICU and hospital stays; and the rate cruitment maneuvers, as previously described20 of death from any cause 30 days after surgery. (the nonprotective-ventilation group), or lung- Pulmonary complications were analyzed sepa- protective ventilation with a tidal volume of 6 to rately; in particular, the need for invasive or 8 ml per kilogram of predicted body weight, a noninvasive ventilation because of acute respira- PEEP of 6 to 8 cm of water, and recruitment matory failure, the development of postoperative neuvers repeated every 30 minutes after tracheal atelectasis, pneumonia, acute lung injury, and the intubation (the protective-ventilation group). Each acute respiratory distress syndrome, defined ac- recruitment maneuver consisted of applying a cording to standard criteria (see the Supplementary continuous positive airway pressure of 30 cm of Appendix). Extrapulmonary complications included water for 30 seconds. During anesthesia, a plateau the systemic inflammatory response syndrome pressure of no more than 30 cm of water was (SIRS); sepsis; severe sepsis and septic shock; and targeted in each group. All other ventilation pro- surgical complications, including intraabdomi- cedures were identical in the two study groups nal abscess, anastomotic leakage, and unplanned (see the Supplementary Appendix). reoperation (all defined according to consensus The predicted body weight was calculated for criteria22,24). each patient with the use of previously defined formulas.11 For episodes of arterial desaturation STATISTICAL ANALYSIS (defined as a peripheral oxygen saturation of We calculated that a sample of 400 patients ≤92%), a transient increase in the fraction of in- would provide 80% power to detect a relative dif- io spired oxygen (F 2) to 100% was permitted, and ference of 50% in the primary outcome, at a two- in patients assigned to nonprotective ventilation, sided alpha level of 0.05, assuming a 20% rate of the use of PEEP, recruitment maneuvers, or both postoperative complications in the nonprotective- was allowed, if required. Decisions about all ventilation group.25 For safety reasons, an interim other aspects of patient care during the intraop- analysis was conducted after the enrollment of erative and postoperative periods, including gen- the first 200 patients, according to the a priori eral anesthesia, administration of fluids, use of statistical analysis plan. The data and safety mon- prophylactic antibiotic agents, and postoperative itoring committee did not recommend discontin- pain management, were made by the attending uation of the trial on the basis of that analysis, physician according to the expertise of the staff and 400 patients were therefore included. A total at each center and routine clinical practice. of 3 patients were excluded after randomization; surgery was stopped prematurely in 2 of the 3 pa- OUTCOMES tients because of extensive illness (duration of sur- The primary outcome was a composite of major gery, <2 hours), and 1 had undergone random- pulmonary and extrapulmonary complications ization in error (violation of exclusion criteria). occurring by day 7 after surgery. Major pulmo- An additional 3 patients were thus randomly as- nary complications were defined as pneumonia signed to a study group to obtain the full sample.

430 n engl j med 369;5 nejm.org august 1, 2013 The New England Journal of Medicine Downloaded from nejm.org at MERCY HOSP ST LOUIS on August 12, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Intraoperative Low-Tidal-Volume Ventilation

All analyses were conducted on data from the use of robust Poisson generalized-linear-model re modified intention-to-treat population, which in- gression26 and are presented as relative risks with cluded all patients who underwent randomization 95% confidence intervals. A chi-square test (or except the three who were excluded (Fig. 1). An un Fisher’s exact test, as appropriate) was used for adjusted chi-square test was used for the primary secondary binary outcomes. The Hochberg pro- outcome analysis. Multiple logistic-regression anal- cedure was used to adjust for multiple testing of ysis was used to identify relevant baseline covariates components of the composite primary outcome.27 associated with the primary outcome, in addition Adjusted analyses were performed with the use to the stratification variables (use or nonuse of of the same adjustment variables that were used epidural analgesia and study center). Variables in the robust Poisson regression analysis. Con- tested in the model were selected if the P value tinuous variables were compared with the use of was less than 0.10 and if they were clinically rel- an unpaired t-test or the Mann–Whitney U test. evant. Adjusted analyses were performed with the Adjusted analyses were performed with the use

1803 Patients were scheduled to undergo elective abdominal surgery

1202 Were ineligible 47 Were ≤40 yr of age 102 Had expected duration of surgery of <2 hr 1053 Had preoperative risk index for pulmonary complications of <2

601 Underwent screening

198 Were excluded 49 Declined to participate 149 Met exclusion criteria

400 Underwent randomization

3 Were excluded after randomization 2 Had extensive illness (surgery <2 hr) 1 Had violation of exclusion criteria (age ≤40 yr)

An additional 3 underwent randomization

200 Were assigned to receive nonprotective 200 Were assigned to receive lung-protective mechanical ventilation mechanical ventilation

200 (100%) Were included in 30-day analysis 200 (100%) Were included in 30-day analysis

Figure 1. Assessment, Randomization, and Follow-up of Patients. A total of 1803 patients awaiting abdominal surgery were assessed preoperatively for trial eligibility by the research staff. A total of 400 patients were included in the modified intention-to-treat analysis and were followed for 30 days after surgery. After randomization, 3 patients were excluded; 2 patients were excluded because surgery was stopped prematurely (duration, <2 hours), owing to extensive illness, and 1 had undergone randomization in error. An additional 3 patients were then enrolled in the study.

n engl j med 369;5 nejm.org august 1, 2013 431 The New England Journal of Medicine Downloaded from nejm.org at MERCY HOSP ST LOUIS on August 12, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. T h e new england journal o f medicine

of the same adjustment variables that were used Table 1. Baseline Characteristics of the Patients.* in the linear-regression model. The time-to-event Nonprotective Lung-Protective curves were calculated with the Kaplan–Meier Ventilation Ventilation method. Details regarding the handling of missing Characteristic (N = 200) (N = 200) data are provided in the Supplementary Appendix. Age — yr 63.4±10.0 61.6±11.0 All analyses were conducted with the use of Male sex — no. (%) 121 (60.5) 116 (58.0) Stata software, version 12 (StataCorp). A two-sided Height — cm 169.5±9.0 169.1±8.8 P value of less than 0.05 was considered to indicate Body weight — kg statistical significance. Actual 71.3±13.9 71.4±14.2 Predicted† 63.8±9.9 63.3±9.7 RESULTS Body-mass index‡ STUDY POPULATION Mean 24.7±3.8 24.8±3.8 From January 2011 through August 2012, a total 25–35 — no. (%) 88 (44.0) 99 (49.5) of 1803 patients awaiting abdominal surgery were Preoperative risk index — no. (%)§ assessed for trial eligibility. A total of 400 pa- Risk class 2 100 (50.0) 101 (50.5) tients were included in the modified intention- Risk class 3 94 (47.0) 93 (46.5) to-treat analysis and were followed for 30 days Risk class 4 or 5 6 (3.0) 6 (3.0) after surgery (Fig. 1). One patient in the nonpro- Coexisting condition — no. (%)¶ tective-ventilation group received lung-protective Current smoking 50 (25.0) 51 (25.5) ventilation but was included in the analysis for Any alcohol intake 10 (5.0) 21 (10.5) the group to which he was assigned. Data on the primary outcome were available for all patients. Not fully independent in activities 8 (4.0) 8 (4.0) of daily living Baseline characteristics were similar between the Chronic obstructive pulmonary 20 (10.0) 20 (10.0) two groups (Table 1). Open laparotomy, mainly disease for cancer resection, was performed in 156 patients Loss of >10% of body weight in 44 (22.0) 40 (20.0) (78.0%) in the nonprotective-ventilation group and previous 6 mo in 159 (79.5%) in the protective-ventilation group Long-term glucocorticoid use 4 (2.0) 7 (3.5) (P = 0.80). Laparoscopic surgery — no. (%) 44 (22.0) 41 (20.5) INTRAOPERATIVE PROCEDURES Type of surgery — no. (%) Table 2 shows the distribution of the main intra- Pancreaticoduodenectomy 80 (40.0) 84 (42.0) operative procedures. Mean (±SD) tidal volumes Liver resection 52 (26.0) 44 (22.0) were 11.1±1.1 ml per kilogram in the nonprotec- Gastrectomy 17 (8.5) 15 (7.5) tive-ventilation group, as compared with 6.4±0.8 Colorectal resection 40 (20.0) 47 (23.5) ml per kilogram in the protective-ventilation Other procedure 11 (5.5) 10 (5.0) group (P<0.001), and values remained within tar- Diagnosis — no. (%) get ranges throughout the intraoperative period. Cancer 164 (82.0) 155 (77.5) In the protective-ventilation group, the median Diagnosis other than cancer 36 (18.0) 45 (22.5) PEEP was 6 cm of water (interquartile range, 6 to 8), and the median number of recruitment maneu- * Plus–minus values are means ±SD. There were no significant differences between vers was 9 (interquartile range, 6 to 12); in the the two groups (P>0.05). nonprotective-ventilation group, the value for each † The predicted body weight was calculated as follows: for men, 50+0.91(height in centimeters − 152.4); and for women, 45.5 + 0.91(height in centimeters − 152.4).11 of these measures was 0 (interquartile range, 0 to ‡ The body-mass index is the weight in kilograms divided by the square of the 0) (Table 2). There were no significant between- height in meters. group differences in type and duration of sur- § The preoperative risk index for pulmonary complications5 uses risk classes that range from 1 to 5, with higher risk classes indicating a higher risk of gery, use or nonuse of epidural analgesia, blood postoperative complications. Patients with a risk class of 2 or more were eligi- loss, volume of fluids administered, and need for ble for participation in the study. vasopressor administration. Five patients in the ¶ All factors listed as coexisting conditions were included in the preoperative risk index as predictors of postoperative pulmonary complications. nonprotective-ventilation group required at least one intraoperative rescue therapy for arterial de-

432 n engl j med 369;5 nejm.org august 1, 2013 The New England Journal of Medicine Downloaded from nejm.org at MERCY HOSP ST LOUIS on August 12, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Intraoperative Low-Tidal-Volume Ventilation

Table 2. Intraoperative Procedures.*

Nonprotective Lung-Protective Ventilation Ventilation Variable (N = 200) (N = 200) P Value

Tidal volume — ml 719.0±127.8 406.7±75.6 <0.001 Tidal volume — ml/kg of predicted body weight 11.1±1.1 6.4±0.8 <0.001 PEEP — cm of water Baseline <0.001 Median 0 6 Interquartile range 0–0 6–8 End of surgery <0.001 Median 0 6 Interquartile range 0–0 6–8 No. of recruitment maneuvers <0.001 Median 0 9 Interquartile range 0–0 6–12 Peak pressure — cm of water Baseline 20.1±4.9 18.9±3.6 0.04 End of surgery 20.6±4.4 20.0±4.0 0.15 Plateau pressure — cm of water Baseline 16.1±4.3 15.2±3.0 0.02 End of surgery 16.6±3.5 15.2±2.6 <0.001 Respiratory system compliance — ml/cm of water Baseline 48.4±17.8 55.2±26.6 0.06 End of surgery 45.1±12.9 55.2±26.7 <0.001

Fio2 — % 47.2±7.6 46.4±7.3 0.27 Volume of fluids administered — liters Crystalloid 0.47 Median 2.0 1.5 Interquartile range 1.5–3.5 2.0–3.0 Colloid 0.97 Median 0.5 0.5 Interquartile range 0.25–1.0 0.50–1.0 Duration of surgery — no./total no. (%)† 0.95 2–4 hr 76/192 (39.6) 75/195 (38.5) >4–6 hr 75/192 (39.1) 76/195 (39.0) >6 hr 41/192 (21.4) 44/195 (22.6) Duration of mechanical ventilation — min 344±127.9 319±139.4 0.84 Epidural analgesia — no. (%) 77 (38.5) 83 (41.5) 0.61

* Plus–minus values are means ±SD. Detailed data on intraoperative procedures are given in Table S2 in the Supplementary Appendix. Fio2 denotes inspired oxygen fraction, and PEEP positive end-expiratory pressure. † The duration of surgery was calculated as the time between skin incision and closure of the incision.

n engl j med 369;5 nejm.org august 1, 2013 433 The New England Journal of Medicine Downloaded from nejm.org at MERCY HOSP ST LOUIS on August 12, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. T h e new england journal o f medicine

saturation (PEEP in one patient, recruitment ma- ventilation (P<0.001 by the log-rank test) (Fig. S1 neuvers in two, and both in two), as compared in the Supplementary Appendix). with no patients in the protective-ventilation There was no significant difference between group (P = 0.06). the protective-ventilation group and the nonprotective-ventilation group with respect to the pro- OUTCOMES portion of patients who were unexpectedly ad- Primary Outcome mitted to the ICU during the 30-day period after Major pulmonary and extrapulmonary complica- surgery (11.0 and 12.5%, respectively; adjusted rela- tions occurred within the first 7 days after sur- tive risk with protective ventilation, 0.88; 95% CI, gery in 21 patients (10.5%) in the protective-ven- 0.49 to 1.59; P = 0.67), nor was there a significant tilation group, as compared with 55 (27.5%) in the between-group difference in the rate of adverse nonprotective-ventilation group (adjusted relative events (Table S3 in the Supplementary Appendix). risk, 0.40; 95% confidence interval [CI], 0.24 to Mortality at 30 days in the protective-ventilation 0.68; P = 0.001) (Table 3). The results of associated group was similar to that in the nonprotective- univariate and multivariate analyses are provided ventilation group (3.0% and 3.5%, respectively; in Table S1 in the Supplementary Appendix. adjusted relative risk with protective ventilation, 1.13; 95% CI, 0.36 to 3.61; P = 0.83). However, the Secondary Outcomes median hospital stay was shorter in the protec- One or more pulmonary complications developed tive-ventilation group than in the nonprotective- within the first 7 days after surgery in 35 patients ventilation group (Table 3). (17.5%) in the protective-ventilation group, as compared with 72 (36.0%) in the nonprotective-ventila- DISCUSSION tion group (adjusted relative risk, 0.49; 95% CI, 0.32 to 0.74; P<0.001). More patients in the non- In this trial, intraoperative lung-protective me- protective-ventilation group than in the protec- chanical ventilation, as compared with non tive-ventilation group had major (grade ≥3) pul- protective ventilation, led to improved clinical monary complications (Table 3, and Tables S3 and outcomes and reduced health care utilization S4 in the Supplementary Appendix) and major after abdominal surgery. The observed rate of pulmonary and extrapulmonary complications dur- postoperative complications in our study was ing the 30 days after surgery (P<0.001 by the log- slightly higher than predicted.25 This was due, rank test) (Fig. 2). There were no relevant between- in part, to the exclusion of patients with a low group differences in gas exchange after extubation risk of complications, as well as the large pro- and on day 1 after surgery (Table S5 in the Supple- portion of patients who underwent major ab- mentary Appendix). dominal procedures, which are associated with The proportion of patients who required post- increased morbidity rates. Of the 400 patients operative ventilatory assistance (noninvasive ven- enrolled, 19 had postoperative pneumonia and tilation or intubation) for acute respiratory failure 47 had respiratory failure requiring intubation was lower in the protective-ventilation group than or noninvasive ventilation. These rates are con- in the nonprotective-ventilation group during the sistent with previously reported rates of pulmo- first 7 days after surgery (10 of 200 patients [5.0%], nary complications25,28 and mortality.29 Our lung- vs. 34 of 200 [17.0%]; adjusted relative risk, 0.29; protective ventilation strategy resulted in a 69% 95% CI, 0.14 to 0.61; P = 0.001), and the proportion reduction in the number of patients requiring was also lower with protective ventilation during ventilatory support within the first 7 days after the first 13 days after surgery (6.5% vs. 18.5%; surgery. adjusted relative risk, 0.36; 95% CI, 0.19 to 0.70; Several hypotheses could explain some of P = 0.003) (Table 3). In addition, the cumulative the differences between the results of the pres- 30-day probability of an event requiring intuba- ent study and findings in other trials of lung- tion or noninvasive ventilation for postoperative protective ventilation during high-risk surgery. acute respiratory failure was lower among pa- Previous trials have included small numbers of tients who received lung-protective ventilation patients, have focused on different (and not than among those who received nonprotective necessarily clinically relevant) outcomes,17 and

434 n engl j med 369;5 nejm.org august 1, 2013 The New England Journal of Medicine Downloaded from nejm.org at MERCY HOSP ST LOUIS on August 12, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Intraoperative Low-Tidal-Volume Ventilation

Table 3. Results of Unadjusted and Adjusted Outcome Analyses.*

Unadjusted Adjusted Relative Risk or Relative Risk or Nonprotective Lung-Protective Between-Group Between-Group Ventilation Ventilation Difference Difference Variable (N = 200) (N = 200) (95% CI) P Value† (95% CI)‡ P Value Primary composite outcome — no. (%) Within 7 days§ 55 (27.5) 21 (10.5) 0.38 (0.24–0.61) <0.001 0.40 (0.24–0.68) 0.001 Within 30 days 58 (29.0) 25 (12.5) 0.43 (0.28–0.66) <0.001 0.45 (0.28–0.73) <0.001 Secondary outcomes — no. (%) Pulmonary complication within 7 days¶ Grade 1 or 2 30 (15.0) 25 (12.5) 0.69 (0.42–1.13) 0.14 0.67 (0.39–1.16) 0.16 Grade ≥3 42 (21.0) 10 (5.0) 0.24 (0.12–0.46) <0.001 0.23 (0.11–0.49) <0.001 Atelectasis within 7 days‖ 34 (17.0) 13 (6.5) 0.38 (0.21–0.70) 0.001 0.37 (0.19–0.73) 0.004 Pneumonia within 7 days 16 (8.0) 3 (1.5) 0.19 (0.05–0.63) 0.01 0.19 (0.05–0.66) 0.009 Acute lung injury or ARDS within 7 days 6 (3.0) 1 (0.5) 0.17 (0.02–1.37) 0.12 0.21 (0.02–1.71) 0.14 Need for ventilation within 7 days Invasive 7 (3.5) 2 (1.0) 0.29 (0.06–1.36) 0.51 0.40 (0.08–1.97) 0.26 Noninvasive 29 (14.5) 9 (4.5) 0.31 (0.15–0.64) 0.006 0.29 (0.13–0.65) 0.002 Extrapulmonary complication within 7 days SIRS 100 (50.0) 86 (43.0) 0.86 (0.70–1.06) 0.16 0.87 (0.65–1.17) 0.37 Sepsis 29 (14.5) 13 (6.5) 0.45 (0.24–0.84) 0.04 0.48 (0.25–0.93) 0.03 Severe sepsis or septic shock 9 (4.5) 8 (4.0) 0.89 (0.35–2.26) 0.80 1.48 (0.51–4.32) 0.47 Death within 30 days 7 (3.5) 6 (3.0) 0.86 (0.29–2.51) 0.80 1.13 (0.36–3.61) 0.83 Duration of stay in hospital and ICU — days Hospital 0.02 0.006 Median 13 11 −2.25 (−4.04 to −0.47) −2.45 (−4.17 to −0.72) Interquartile range 8–20 8–15 ICU 0.58 0.69 Median 7 6 −1.48 (−6.87 to 3.91) −1.21 (−4.98 to 7.40) Interquartile range 4–9 4–8

* All postoperative complications were defined according to consensus criteria (see the Supplementary Appendix). For additional data on postoperative outcomes, see Tables S3 and S4 in the Supplementary Appendix. ARDS denotes acute respiratory distress syndrome, CI confidence interval, ICU intensive care unit, and SIRS systemic inflammatory response syndrome. Relative risks are shown for outcome variables, and differences between groups are shown for the duration of stays in the hospital and ICU. † Adjustment was performed for stratification variables (use or nonuse of epidural analgesia and study center), preoperative risk index for postoperative pulmonary complications, sex, duration of surgery, and need for blood transfusion (yes or no). ‡ The Hochberg procedure was used to adjust for multiple testing of components of the composite primary outcome.27 § The primary outcome was a composite of major pulmonary complications (defined as pneumonia or need for invasive or noninvasive ventilation for acute respiratory failure) and extrapulmonary complications (defined as sepsis, septic shock, or death) within the first 7 days after surgery. ¶ Postoperative pulmonary complications were scored with the use of a graded scale23 from 0 (no pulmonary complications) to 4 (the most severe complications) (see the Supplementary Appendix). ‖ Atelectasis was defined as opacification of the lung with shift of the mediastinum, hilum, or hemidiaphragm toward the affected area and compensatory overinflation in the adjacent, nonatelectatic lung. have used either very low levels of PEEP or no Mechanical ventilation itself can induce an in- PEEP.15,16,18 One strength of the present trial is our flammatory response30 and can synergize with use of a robust composite outcome that is highly the response induced by major surgery at both pertinent to this high-risk surgical population.5 local and systemic levels. This amplification of

n engl j med 369;5 nejm.org august 1, 2013 435 The New England Journal of Medicine Downloaded from nejm.org at MERCY HOSP ST LOUIS on August 12, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. T h e new england journal o f medicine

ment maneuvers, in which hemodynamic effects 0.50 are potentially influenced by the applied level of alveolar pressure,36 should be used with caution 0.40 in patients with hemodynamic instability. There are several limitations to our study. The Nonprotective ventilation 0.30 trial design did not include standardization of the administration of fluids. However, this limi- 0.20 tation is unlikely to have affected our results, Lung-protective ventilation

Probability of Event since the volume of fluids administered was 0.10 similar in the two groups. The definition of nonprotective ventilation was arbitrary but is 0.00 19,20 1 3 7 15 30 supported in the literature. The trial protocol Days since Randomization did not include standardization of requirements No. at Risk for noninvasive ventilation; however, it was rec- Nonprotective 200 182 163 145 142 142 ommended that the study centers follow clinical- ventilation practice guidelines,37,38 and postoperative care Lung-protective 200 192 184 179 176 175 ventilation was conducted by health care workers who were unaware of the study assignments. The utiliza- Figure 2. Kaplan–Meier Estimates of the Probability of the Composite tion of noninvasive ventilation in our trial is Primary Outcome. close to that reported in earlier studies.37 We Data for the Kaplan–Meier estimates of the probability of the composite primary outcome of major pulmonary or extrapulmonary complications therefore consider it unlikely that any imbalance were censored at 30 days after surgery. Major pulmonary complications in interventions affected our results. included pneumonia or the need for invasive or noninvasive ventilation for In conclusion, our study provides evidence that acute respiratory failure. Major extrapulmonary complications were sepsis, a multifaceted strategy of prophylactic lung-pro- severe sepsis, septic shock, and death. P<0.001 by the log-rank test for the tective ventilation during surgery, as compared between-group difference in the probability of the primary outcome. with a practice of nonprotective mechanical ventilation, results in fewer postoperative complications and reduced health care utilization. the inflammation cascade contributes to the 31 Dr. Futier reports receiving consulting fees from General subsequent development of lung injury and sys Electric Medical Systems, lecture fees from Fresenius Kabi, and temic organ failure.8,32 reimbursement of travel expenses from Fisher and Paykel The use of very low levels of PEEP in previous Healthcare. Dr Constantin reports receiving consulting fees from Baxter, Fresenius Kabi, Dräger, and General Electric Medi- trials may have promoted the repeated opening cal Systems, payment for expert testimony from Baxter, Dräger, and closing of small airways, leading to atelec- and Fresenius Kabi, lecture fees from General Electric Medical tasis, which can precipitate the development of Systems, Dräger, Fresenius Kabi, Baxter, Hospal, Merck Sharp & 6,33 Dohme, and LFB Biomedicaments, payment for the development pulmonary complications. We used a multi- of educational presentations from Dräger, General Electric faceted strategy of lung-protective ventilation Medical Systems, Baxter, and Fresenius Kabi, and reimburse- that combined low tidal volumes, recruitment ment of travel expenses from Bird, Astute Medical, Astellas, Fresenius Kabi, Baxter, and Hospal. Dr. Paugam-Burtz reports maneuvers to open collapsed alveoli, and moder- receiving consulting fees from Fresenius Kabi, lecture fees and ate levels of PEEP to prevent further collapse.34 reimbursement of travel expenses from Astellas, and payment Other strengths of the present trial include the for the development of educational presentations from LFB Bio- medicaments and Merck Sharp & Dohme. Dr. Allaouchiche re- methods used to minimize bias (blinded and ports receiving consulting fees from Fresenius Kabi and lecture centralized randomization, complete follow-up, fees from Novartis and Astellas. Dr. Leone reports receiving and intention-to-treat analyses); the pragmatic consulting fees from LFB Biomedicaments and lecture fees from Fresenius Kabi and Novartis. Dr. Jaber reports receiving consult- nature of the trial protocol, with routine practice ing fees from Dräger France and Maquet France, lecture fees being maintained; and the enrollment of pa- from Fisher and Paykel Healthcare, Abbott, and Philips, and re- tients with characteristics similar to those of imbursement of travel expenses from Pfizer. No other potential conflict of interest relevant to this article was reported. patients enrolled in other studies analyzing out- Disclosure forms provided by the authors are available with comes after major surgery.29 the full text of this article at NEJM.org. Our findings are consistent with the observa- We thank all the patients who participated in the study; the clinical and research staff at all the trial sites, without whose as- tion of transient arterial hypotension during resistance the study would never have been completed; and Mervyn cruitment maneuvers.35 Consequently, recruit- Singer for valuable advice during the preparation of the manuscript.

436 n engl j med 369;5 nejm.org august 1, 2013 The New England Journal of Medicine Downloaded from nejm.org at MERCY HOSP ST LOUIS on August 12, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Intraoperative Low-Tidal-Volume Ventilation

References

1. Weiser TG, Regenbogen SE, Thomp- tilation strategies and outcomes of the of 6336 patients and results of a survey. son KD, et al. An estimation of the global acute respiratory distress syndrome and Ann Surg 2004;240:205-13. volume of surgery: a modelling strategy acute lung injury. Ann Intern Med 2009; 25. Thompson JS, Baxter BT, Allison JG, based on available data. Lancet 2008;372: 151:566-76. [Erratum, Ann Intern Med Johnson FE, Lee KK, Park WY. Temporal 139-44. 2009;151:897.] patterns of postoperative complications. 2. Khuri SF, Henderson WG, DePalma 13. Schultz MJ, Haitsma JJ, Slutsky AS, Arch Surg 2003;138:596-602. RG, et al. Determinants of long-term sur- Gajic O. What tidal volumes should be 26. Zou G. A modified Poisson regression vival after major surgery and the adverse used in patients without acute lung injury? approach to prospective studies with bi- effect of postoperative complications. Ann Anesthesiology 2007;106:1226-31. nary data. Am J Epidemiol 2004;159:702-6. Surg 2005;242:326-41. 14. Severgnini P, Selmo G, Lanza C, et al. 27. Hochberg Y. A sharper Bonferroni 3. Shander A, Fleisher LA, Barie PS, Big- Protective mechanical ventilation during procedure for multiple tests of signifi- tello LM, Sladen RN, Watson CB. Clinical general anesthesia for open abdominal cance. Biometrika 1988;75:800-2. and economic burden of postoperative surgery improves postoperative pulmonary 28. Lawrence VA, Dhanda R, Hilsenbeck pulmonary complications: patient safety function. Anesthesiology 2013 March 29 SG, Page CP. Risk of pulmonary compli- summit on definition, risk-reducing inter- (Epub ahead of print). cations after elective abdominal surgery. ventions, and preventive strategies. Crit 15. Wrigge H, Uhlig U, Zinserling J, et al. Chest 1996;110:744-50. Care Med 2011;39:2163-72. The effects of different ventilatory set- 29. Pearse RM, Moreno RP, Bauer P, et al. 4. Arozullah AM, Daley J, Henderson tings on pulmonary and systemic inflam- Mortality after surgery in Europe: a 7 day WG, Khuri SF. Multifactorial risk index matory responses during major surgery. cohort study. Lancet 2012;380:1059-65. for predicting postoperative respiratory Anesth Analg 2004;98:775-81. 30. Ranieri VM, Suter PM, Tortorella C, failure in men after major noncardiac sur- 16. Wrigge H, Uhlig U, Baumgarten G, et al. Effect of mechanical ventilation on gery: the National Veterans Administra- et al. Mechanical ventilation strategies and inflammatory mediators in patients with tion Surgical Quality Improvement Pro- inflammatory responses to cardiac sur- acute respiratory distress syndrome: a gram. Ann Surg 2000;232:242-53. gery: a prospective randomized clinical randomized controlled trial. JAMA 1999; 5. Arozullah AM, Khuri SF, Henderson trial. Intensive Care Med 2005;31:1379-87. 282:54-61. WG, Daley J. Development and validation 17. Hong CM, Xu DZ, Lu Q, et al. Low 31. Mascia L, Pasero D, Slutsky AS, et al. of a multifactorial risk index for predict- tidal volume and high positive end-expira- Effect of a lung protective strategy for ing postoperative pneumonia after major tory pressure mechanical ventilation re- organ donors on eligibility and availabil- noncardiac surgery. Ann Intern Med sults in increased inflammation and ven- ity of lungs for transplantation: a ran- 2001;135:847-57. tilator-associated lung injury in normal domized controlled trial. JAMA 2010;304: 6. Bendixen HH, Hedley-Whyte J, Laver lungs. Anesth Analg 2010;110:1652-60. 2620-7. MB. Impaired oxygenation in surgical pa- 18. Treschan TA, Kaisers W, Schaefer MS, 32. Bouadma L, Dreyfuss D, Ricard JD, tients during general anesthesia with con et al. Ventilation with low tidal volumes Martet G, Saumon G. Mechanical ventila- trolled ventilation: a concept of atelecta- during upper abdominal surgery does not tion and hemorrhagic shock-resuscitation sis. N Engl J Med 1963;269:991-6. improve postoperative lung function. Br J interact to increase inflammatory cyto- 7. Serpa Neto A, Cardoso SO, Manetta Anaesth 2012;109:263-71. kine release in rats. Crit Care Med 2007; JA, et al. Association between use of lung- 19. Jaber S, Coisel Y, Chanques G, et al. A 35:2601-6. protective ventilation with lower tidal vol- multicentre observational study of intra- 33. Duggan M, Kavanagh BP. Pulmonary umes and clinical outcomes among pa- operative ventilatory management during atelectasis: a pathogenic perioperative en- tients without acute respiratory distress general anaesthesia: tidal volumes and re- tity. Anesthesiology 2005;102:838-54. syndrome: a meta-analysis. JAMA 2012; lation to body weight. Anaesthesia 2012; 34. Futier E, Constantin JM, Pelosi P, et al. 308:1651-9. 67:999-1008. Noninvasive ventilation and alveolar re- 8. Imai Y, Parodo J, Kajikawa O, et al. 20. Hess DR, Kondili D, Burns E, Bittner cruitment maneuver improve respiratory Injurious mechanical ventilation and end- EA, Schmidt UH. A 5-year observational function during and after intubation of organ epithelial cell apoptosis and organ study of lung-protective ventilation in the morbidly obese patients: a randomized dysfunction in an experimental model of operating room: a single-center experi- controlled study. Anesthesiology 2011;114: acute respiratory distress syndrome. ence. J Crit Care 2013 January 29 (Epub 1354-63. JAMA 2003;289:2104-12. ahead of print). 35. Fan E, Wilcox ME, Brower RG, et al. 9. Lellouche F, Dionne S, Simard S, Bus- 21. Pöpping DM, Elia N, Marret E, Remy Recruitment maneuvers for acute lung in- sieres J, Dagenais F. High tidal volumes in C, Tramèr MR. Protective effects of epidu- jury: a systematic review. Am J Respir Crit mechanically ventilated patients increase ral analgesia on pulmonary complications Care Med 2008;178:1156-63. organ dysfunction after cardiac surgery. after abdominal and thoracic surgery: a 36. Lim SC, Adams AB, Simonson DA, Anesthesiology 2012;116:1072-82. meta-analysis. Arch Surg 2008;143:990-9. et al. Transient hemodynamic effects of 10. Amato MB, Barbas CS, Medeiros DM, 22. Bone RC, Sibbald WJ, Sprung CL. The recruitment maneuvers in three experi- et al. Effect of a protective-ventilation ACCP-SCCM consensus conference on mental models of acute lung injury. Crit strategy on mortality in the acute respira- sepsis and organ failure. Chest 1992;101: Care Med 2004;32:2378-84. tory distress syndrome. N Engl J Med 1998; 1481-3. 37. Jaber S, Chanques G, Jung B. Post 338:347-54. 23. Hulzebos EH, Helders PJ, Favié NJ, operative noninvasive ventilation. Anes- 11. The Acute Respiratory Distress Syn- et al. Preoperative intensive inspiratory mus- thesiology 2010;112:453-61. drome Network. Ventilation with lower tid- cle training to prevent postoperative pul- 38. Keenan SP, Sinuff T, Burns KE, et al. al volumes as compared with traditional monary complications in high-risk patients Clinical practice guidelines for the use of tidal volumes for acute lung injury and the undergoing CABG surgery: a randomized noninvasive positive-pressure ventilation acute respiratory distress syndrome. N Engl clinical trial. JAMA 2006;296:1851-7. and noninvasive continuous positive air- J Med 2000;342:1301-8. 24. Dindo D, Demartines N, Clavien PA. way pressure in the acute care setting. 12. Putensen C, Theuerkauf N, Zinserling Classification of surgical complications: a CMAJ 2011;183(3):E195-E214. J, Wrigge H, Pelosi P. Meta-analysis: ven- new proposal with evaluation in a cohort Copyright © 2013 Massachusetts Medical Society.

n engl j med 369;5 nejm.org august 1, 2013 437 The New England Journal of Medicine Downloaded from nejm.org at MERCY HOSP ST LOUIS on August 12, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved.