Prosthetic Valve Endocarditis

Heart 2001;85:590–593

be classified as “polymer associated infections” VALVE DISEASE rather than PVE. PVE should be classified as either being Heart: first published as 10.1136/hrt.85.5.590 on 1 May 2001. Downloaded from acquired perioperatively, and thus nosocomial Prosthetic valve endocarditis (early PVE), or as community acquired (late PVE).8 Because of significant diVerences in C Piper, R Körfer*, D Horstkotte microbiology of PVE observed within the first 590 Department of Cardiology, *Department of Thoracic and year of operation and later on, the time cut oV point between early and late PVE should be Cardiovascular Surgery, Heart Center North Rhine-Westphalia, Ruhr 9 University, Bad Oeynhausen, Germany regarded as one year. The risk for early PVE is higher (approximately 5%) in patients with replacement fter 40 years of continuous improve- surgery during active infective endocarditis, ments in the design and materials used especially if the causal organism is unknown or Afor prosthetic heart valves, valve re- the antibiotic treatment is insuYcient. The placement surgery is now performed with low incidence of late PVE is lower for mechanical morbidity and mortality. These advantages prostheses than for bioprostheses. The have been hampered by a few but severe weighted mean incidence for infections of bio- adverse eVects; in particular, infections of the prostheses calculated from published series is prosthetic material continue to be an extremely 0.49% per patient year for mitral valves and serious complication occurring with a relatively 0.91% per patient year for aortic valves. For low but increasing frequency ranging from mechanical prostheses, the incidence is 0.18% 0.1–2.3% per patient year.1–3 The prosthesis per patient year for mitral, 0.27% per patient obviously predisposes to device related infec- year for aortic, and 0.29% per patient year for 10 tions, especially those caused by novobiocin multiple implants. susceptible, coagulase negative staphylococci, Comparing diVerent periods of implanta- which are able to adhere to a variety of tion, hazard functions reveal a significant surfaces4 and produce an antibiotic resistant decline in early PVE cases in recent years, con- biofilm.56 trasting with a slight increase in the hazard for late PVE (fig 1). Definition and frequency Pathogenesis Prosthetic valve endocarditis (PVE) is an endovascular, microbial infection occurring on Prostheses made from metal, pyrolyte or other http://heart.bmj.com/ parts of a valve prosthesis or on reconstructed materials do not allow adherence of micro- native heart valves.7 It is recommended to organisms as long as they are free from throm- determine whether (a) a mechanical prosthesis, botic material. Infections of mechanical pros- (b) a bioprosthetic xenograft, stented or theses generally originate from the sewing cuV Correspondence to: C Piper, MD, Heart unstented, (c) an allograft, (d) a homograft, or or from thrombi located near the sewing ring Center, North (e) a repaired native valve with or without downstream in recirculation areas. Inflamma- Rhine-Westphalia, implantation of an annular ring is involved.8 tory periprosthetic leaks, ring abscesses, and Department of Although clinical relevance and therapeutic invasion of the infective process into the Cardiology, Georgstr. considerations may be similar, infections of adjacent tissue are common findings. The on September 25, 2021 by guest. Protected copyright. 11, D-32545 Bad pathogenesis of bioprosthetic infections may be Oeynhausen, Germany devices or lines placed inside the heart but not [email protected] connected to the endocardial structures should similar to that of native valves. In these cases, the infection is restricted to the cusps, eventu- ally initiating secondary bioprosthetic failure but with only a low tendency to invade the sewing cuV or to result in periprosthetic abscesses.10 If the sewing cuV, however, is involved, the pathogenesis and clinical course are more or less the same as in PVE involving mechanical prostheses.

Microbiology

The microbiology of PVE is very diVerent from that of native valve endocarditis (NVE). Strep- tococci and enterococci occur less frequently, while staphylococci, bacteria of the HACEK group (Haemophilus, Actinobacillus, Cardiobac- terium, Eikinella, and Kingella), and fungi are found more frequently in cases of PVE. Novo- biocin susceptible, coagulase negative staphylococci have a particularly high aYnity for Figure 1. Hazard functions for prosthetic valve endocarditis in 4189 consecutive implanted or indwelling foreign surfaces, espe- patients during three successive follow up periods. cially polymers.4 They are the most frequent

www.heartjnl.com Education in Heart

pathogens causing PVE. The usual spectrum Deﬁnition, pathogenesis, and microbiology for Europe is given in table 1. of prosthetic valve endocarditis (PVE) Heart: first published as 10.1136/hrt.85.5.590 on 1 May 2001. Downloaded from

x Microbial infection of parts of a prosthetic Diagnostic approach valve or reconstructed native heart valve x Early PVE is usually acquired The diagnostic approach in PVE does not dif- 591 perioperatively (nosocomial) fer from that in NVE, as both are systemic infections maintaining a continuous bacterae- x Late PVE is mostly community acquired mia. Hence, the diagnosis is established if in addition to typical clinical signs and symptoms Time cut o point between early and late x V and positive blood cultures, the device can be PVE should be one year (notable shown to be a ected by echocardiography, diVerences in microbiology) V preferably using multiplane transoesophageal x The risk of early PVE is higher (TOE) probes. TOE should be performed (approximately 5%) in patients with without delay in all patients with suspicion of replacement surgery during active infective PVE.11 For the diagnosis of PVE, TOE is of endocarditis such immense importance that institutions without this facility are best advised to ask for x Mechanical prosthesis infections originate assistance from a specialised centre. With TOE from the sewing cuV or from nearby the size of vegetation can be defined more pre- located thrombi → periprosthetic leaks, cisely than with transthoracic echocardio- ring abscesses, invasion of adjacent tissue graphy (TTE), and periannular complications indicating a locally uncontrolled infection (for x Bioprosthesis infections mostly are example, abscesses, dehiscence, fistulas) may restricted to the cusps → secondary be detected earlier. Both size of vegetations and bioprosthetic failure infection morphology significantly influence x Staphylococci (especially novobiocin therapeutic decisions (namely duration of anti- susceptible, coagulase negative microbial treatment and the need for urgent staphylococci), bacteria of the HACEK surgical intervention). group, and fungi occur more frequently in In otherwise unproven cases, gallium-67 PVE scans or indium-111 leucocyte scintigraphy have been reported to be useful in detecting x Streptococci and enterococci are found myocardial abscesses or diVuse tissue infiltra- more frequently in native valve endocarditis tions. Their diagnostic impact has not been http://heart.bmj.com/ established so far.12

Treatment Table 1 Microbiology of early and late PVE. Authors’ own ﬁndings compared to a recent European literature review3 Antimicrobial treatment Early PVE (%) Late PVE (%) The basic principles of antimicrobial treatment

Own experience Europe Own experience Europe in PVE do not diVer from those for NVE. Some on September 25, 2021 by guest. Protected copyright. (n=34) (n=68) (n=132) (n=194) special aspects need to be considered, however. PVE is usually associated with vegetations Staphylococcus epidermidis 29 43 21 28 larger than those found in NVE. Consequently, Staphylococcus aureus 18 13 19 13 Streptococci 6 3 15* 20 antibiotics have to be used in dosages which Enterococci 6 2 18 7.5 result in maximum, non-toxic serum concen- HACEK 18 17 8 7 Fungi 9 6 5 4 trations in order to penetrate the total vegeta- Mixed infections 6 – 3 – tion. The duration of treatment usually has to Others 6 12 7 9 be longer than for the treatment of NVE and Culture negative 3 4 4 12 should consider vegetation size as determined *Viridans group n=13 (10%), â haemolytic streptococci n=3 (2%), and Streptococcus bovis n=4 by TOE as well as the minimal inhibitory con- (3%). centration (MIC) of the most eYcient combi- HACEK, Haemophilus, Actinobacillus, Cardiobacterium, Eikinella, Kingella. nation of antibiotics (table 2). Antibiotic sterilisation of large vegetations is unlikely with an MIC > 4 µg/ml. Table 2 Duration of antimicrobial treatment in prosthetic valve endocarditis with respect to vegetation size and minimal inhibitory concentration (MIC) In PVE caused by coagulase negative staphylococci, complex interactions between the Vegetation size* microorganism and the synthetic material—for example, irreversible adhesion and production <4mm 5–9mm >10mm of a bioﬁlm, which inhibit the host defence MIC > 4 µg/l Antibiotic cure unlikely mechanisms—protects against antimicrobial 4 µg/ml > MIC > 2 µg/ml > 6 weeks Antibiotic cure unlikely treatment and makes antibiotic sterilisation 2 µg/ml > MIC > 0.5 µg/ml 6 weeks > 6 weeks Antibiotic cure 56 unlikely extremely diYcult. 0.5 µg/ml > MIC > 0.1 µg/ml 6 weeks 6 weeks > 6 weeks The presence of (micro-) abscesses is likely MIC < 0.1 µg/ml 4 weeks 4 weeks > 6 weeks in PVE caused by coagulase negative staphylo- *Actual size of vegetation during treatment; MIC, minimal inhibitory concentration of the most cocci, and triple therapy including rifampicin eVective antibiotic combination. (900 mg/day divided into three doses) is

www.heartjnl.com Education in Heart

recommended.13 Rifampicin is actively taken up by granulocytes and becomes eVective

against intracellular staphylococci and staphy- Heart: first published as 10.1136/hrt.85.5.590 on 1 May 2001. Downloaded from lococci inside abscesses.4 When PVE is clinically apparent and blood cultures are not yet positive, empiric treatment should be initiated with vancomycin and 592 gentamicin.8 For PVE caused by penicillin sen-

sitive streptococci (MICPEN < 0.1 µg/ml), it is advisable to combine penicillin (20–24 million units/24 hours intravenously (iv) divided into 4–6 doses) with an aminoglycoside (preferably gentamicin, 3 mg/kg/24 hours iv, divided into 2–3 doses), treating for at least two weeks with this combination and at least a further two weeks with penicillin alone. In the case of penicillin allergy, vancomycin as a single drug treatment (30 mg/kg/24 hours iv divided into two doses) or ceftriaxone (2 g/24 hours iv as single Figure 2. Influence of the timing of surgery on the prognosis of patients with cerebral thromboembolism dose) in combination with gentamicin should complicating prosthetic valve endocarditis. be given. PVE caused by streptococci less sen-

sitive to penicillin (MICPEN > 0.5 µg/ml) or Antithrombotic treatment enterococci (MICPEN < 8 µg/ml) are best Antithrombotic management in patients with treated with a combination of penicillin and PVE has been discussed widely.14 There seems gentamicin for at least four weeks. Vancomycin to be a consensus that oral anticoagulation can replace penicillin, if the patient is allergic to treatment should be suspended and replaced

penicillin or if MICPEN > 8 µg/ml. Vancomycin by intravenous heparin. The dosage of heparin resistant strains susceptible to teicoplanin depends on the presence and nature of second-

(MICTEICOPL < 4 µg/ml) may be treated with ary complications (for example, thrombocyto- teicoplanin (10 mg/kg iv divided into two penia) and varies between 7–30 U/kg body doses) plus gentamicin. If the isolates are weight. Low molecular weight heparins may be highly resistant to gentamicin or multiresistant advantageous, as side eVects (especially throm- to the standard antimicrobial agents, alterna- bocytopenia) are less frequent. tive combinations of drugs (for example, quinolones) must be considered in consulta- http://heart.bmj.com/ tion with an expert in clinical microbiology. Surgical reintervention Enterococcal PVE is often complicated by periprosthetic dehiscence, annular abscesses or If PVE is complicated, it has to be decided fistulas. In these cases, if antibiotic treatment whether medical treatment should be contin- fails an early surgical intervention should be ued or urgent surgical intervention is required. considered. The indications for surgery in PVE are similar If the pathogen is an oxacillin susceptible to those in NVE: large (> 10 mm), mobile veg- staphylococcus (MIC < 0.1 µg/ml), genta- etations, thromboembolic events with vegeta- on September 25, 2021 by guest. Protected copyright. micin should be combined with dicloxacillin/ tions still demonstrable, sepsis persisting for flucloxacillin (12 g/24 hours iv, divided into six more than 48 hours despite eVective antibiotic doses) for two weeks; thereafter dicloxacillin/ treatment (guided by blood cultures and flucloxacillin should be given for an additional MICs), and acute renal failure. A cerebral four weeks. In oxacillin resistant strains, vanco- embolic event is not a contraindication for mycin (see above for doses) should replace open heart surgery provided that there is no oxacillin derivates. There is no valid evidence cerebral haemorrhage and the time between to prove that teicoplanin is superior to the embolic event and surgery is short (preferably established antistaphylococcal drugs. Early < 72 hours) so that the blood–brain barrier can surgery in most cases is indicated to prevent be expected not to be significantly disturbed secondary complications.813 (fig 2).15 Periprosthetic dehiscence with or

Table 3 Guidelines for endocarditis prophylaxis in patients with biological or prosthetic heart valves, categorised according to the patient population at high risk of acquiring endocarditis

Oropharynx, gastrointestinal tract, urogenital tract

No penicillin allergy In case of penicillin allergy Skin, heart catheterisation

1 hour before the procedure + Outpatient: 2 g (3 g > 70 kg + Outpatient: 600 mg clindamycin orally or 1 g + 600 mg clindamycin orally or bodyweight) amoxicillin orally vancomycin iv for 1 hour or 800 mg teicoplanin 1 g vancomycin iv for 1 hour + Hospitalised: 2 g amoxicillin iv + + Hospitalised: vancomycin 1 g iv for 1 hour + or 800 mg teicoplanin 1.5 mg/kg gentamicin 1.5 mg/kg gentamicin 6 hours after the procedure + Outpatient: 1 g amoxicillin orally + Outpatient: 300 mg clindamycin orally + 300 mg clindamycin + Hospitalised: 1 g amoxicillin + + Hospitalised: vancomycin 1 g iv for 1 hour + 1.5 mg/kg gentamicin 1.5 mg/kg gentamicin

iv, intravenous.

www.heartjnl.com Education in Heart

the published literature and on the authors’ personal Treatment and prophylaxis of prosthetic clinical experiences. valve endocarditis (PVE) 4. Zimmerli W. Experimental models in the investigation of device-related infections. J Antimicrob Chemother 1993;31 Heart: first published as 10.1136/hrt.85.5.590 on 1 May 2001. Downloaded from (suppl D):97–102. x Duration of treatment for PVE is usually 5. Horstkotte D, Weist K, Rueden H. Better understanding longer than for native valve endocarditis of the pathogenesis of prosthetic valve endocarditis—recent perspectives for prevention strategies. J Heart Valve Dis Antibiotic sterilisation of coagulase 1998;7:313–15. x • Editorial on the recent perspectives for prevention of PVE. 593 negative staphylococci or enterococci PVE 6. Hyde JAJ, Darouiche RO, Costeron JW. Strategies for is extremely diYcult prophylaxis against prosthetic valve endocarditis. A review article. J Heart Valve Dis 1998;7:316–26. Early surgical intervention is necessary in • This article gives the historic background to the prevention x of PVE and discusses the current state of research in this most cases to prevent secondary area. complications 7. Edmunds LH, Clark RE, Cohn LH, et al. Guidelines for reporting morbidity and mortality after cardiac valvular x Oral anticoagulation should be replaced by operation. Ann Thorac Surg 1988;46:257–9. intravenous heparin or low molecular 8. Horstkotte D, Follath F, v Graevenitz A, et al. Recommendations for prevention, diagnosis and treatment of weight heparin infective endocarditis. The task force on infective endocarditis of the European Society of Cardiology. Eur x For PVE prophylaxis, antibiotics should be Heart J (in press) • A European consensus status giving the latest guidelines taken one hour before and six hours after for prevention, diagnosis, and treatment of infective the interventional procedure endocarditis. 9. Karchmer AW, Gibbons GW. Infections of prosthetic x In hospitalised patients, antibiotics may be heart valves and vascular grafts. In: Bisno AL, Waldvogel FA, eds. Infections associated with indwelling medical administered intravenously in combination devices. Washington: ASM Press, 1994:213–49. with aminoglycosides 10. Horstkotte D, Piper C, Niehues R, et al. Late prosthetic valve endocarditis. Eur Heart J 1995;16(suppl B):39–47. • Prevalence, sources of infection, microbiology, diagnostic without myocardial failure has a poor progno- approach, and medical as well as surgical treatment of sis. If congestion is not promptly removed by 140 patients with PVE from 4182 consecutive patients medical treatment, surgical intervention is operated on in Düsseldorf between 1970 and 1992. mandatory. Allograft aortic root replacement is 11. Pedersen WR, Walker M, Olseon JD, et al. Value of transesophageal echocardiography in evaluation of native a valuable technique in the complex setting of and prosthetic valve endocarditis. Chest 1991;100:351–6. PVE with involvement of the periannular • This study reported better visualisation of valvar 16 vegetations in native as well as in prosthetic valve region. endocarditis with transoesophageal (TOE) than with transthoracic (TTE) echocardiography. TTE was positive in only 5 of 10 patients with infective endocarditis, while TOE Prophylaxis not only yielded abnormal findings in all 10 patients but also revealed additional information in 4 of 5 patients. http://heart.bmj.com/ 12. O’Brien K, Barnes D, Martin RH, et al. Gallium-SPECT in the detection of prosthetic valve endocarditis and aortic As the risk for an infection is much higher in ring abscess. J Nucl Med 1991;32:1791–3. patients with prosthetic heart valves than in 13. Wilson WR, Geraci JE, Danielson GU, et al. patients with valvar heart disease, more inten- Anticoagulant therapy and central nervous system complications in patients with prosthetic valve endocarditis. sive prophylaxis is needed in these patients. In Circulation 1978;57:1004–7. patients with prosthetic valves, the antibiotics • Adequate anticoagulant treatment in patients with PVE resulted in fewer central nervous system complications should be taken one hour before the interven- than inadequate anticoagulation.

tional procedure and a repeat but reduced dos- 14. Wilson WR, Karchmer AW, Dajani AS, et al. Antibiotic on September 25, 2021 by guest. Protected copyright. age administered six hours after the procedure. treatment of adults with infective endocarditis due to streptococci, enterococci, staphylococci, and HACEK If patients are hospitalised, the antibiotics may microorganisms. JAMA 1995;274:1706–13. be applied intravenously in combination with • Consensus opinion regarding management of PVE caused by commonly encountered microorganisms and aminoglycosides one hour before and six hours those cases resulting from infrequent causes of PVE. after the procedure (table 3). 15. Horstkotte D, Piper C, Wiemer M, et al. Dringlicher Herzklappenersatz nach akuter Hirnembolie während florider 1. Blackstone EH, Kirklin JW. Death and other Endokarditis. Med Klinik 1998;93:284-93. time-related events after valve replacement. Circulation • The analysis of an urgent surgical intervention after 1985;72:753–67. embolic cerebral infarction in 22 patients compared to 27 • Experience with 1533 patients undergoing valve surgery medically treated patients revealed that removing the between 1975 and 1979 revealed that PVE occurs source of infection and embolic hazard seems to be uncommonly after original valve replacement surgery beneficial and that surgery should be performed within 72 (4.4% in five years) but with a high mortality of 63%. hours to prevent secondary cerebral haemorrhage. 2. Kloster FE. Complications of artificial heart valves. JAMA 16. Dossche KM, Defauw JJ, Ernst SM, et al. Allograft 1979;241:2201–3. aortic root replacement in prosthetic aortic valve endocarditis: a review of 32 patients. Ann Thorac Surg 3. Vlessis AA, Khaki A, Grunkemeier GL, et al. Risk, 1997;63:1644–9. diagnosis and management of prosthetic valve endocarditis: • This review of 32 patients showed that allograft aortic root a review. J Heart Valve Dis 1997;6:443–65. replacement is a valuable technique in the complex setting • This article reviews the current understanding of PVE and of PVE with involvement of the periannular region with low provides an outline for diagnosis and treatment based on perioperative mortality and morbidity.