Evaluation of Conventional Weaning Criteria in Patients with Acute Respiratory Failure
Kazufumi OKAMOTO, Hiroko IWAMASA* 1 Hiroshi DOGOMORI** and Tohru MORIOKA
We evaluated the reliability of conventional weaning criteria from a ventila- tor during 33 weaning trials on 25 patients with acute respiratory failure (ARF). Of 13 criteria, a. ratio of maximal voluntary ventilation to minute ventilation 1 (MV) > 2, a vital capacity > 12 mlkg- , a spontaneous respiratory rate < 1 1 25 breaths·min- , and a MV < 10 /·min- appeared to be useful for predict- ing successful weaning outcome. However, even using those criteria, there were many falsely-negative cases. The alveolar-arterial Po, gradient < 350 mmHg at an FlO, 1.0 was not useful as a predictor of weaning outcome. The present study demonstrates that conventional criteria are frequently inaccurate for pre- dicting weaning outcomes and suggests that the use of some of these criteria may unnecessarily prolong the length of ventilator support. Since ventilation of most patients with poor oxygenation can be successfully discontinued by pla.cing them on a continuous positive airway pressure system, these results suggest that the improvement of oxygenation is not an indispensable prerequisite [or weaning from mechanical ventilators. (Key words: acute respiratory failure, mechanical ventilation, oxygenation, weaning, weaning criteria) (Okamoto K, Iwamasa H, Dogomori H, et al.: Evaluation of conventional weaning criteria in patients with acute respiratory failure. J Anesth 4: 213-218, 1990)
Mechanical ventilatory support (MVS) is A number of weaning criteria have been 1 7 17 associated with many complications - :l and proposed (listed in table 1 - ), but there thus should be discontinued as soon as a exists disagreement regarding the most ac- patient can maintain sufficient spontaneous curate criterion for predicting weaning out- respiration. However, if weaning trials from comes. The purpose of this prospective study mechanical ventilators are initiated too early, was to evaluate the reliability of conven- they may lead to severe respiratory dis- tional weaning criteria in patients with acute tress or acute left ventricular dysfunctiont". respiratory failure (ARF). Thus, timing is a crucial component for suc- Methods cessful weaning from a ventilator, Twenty-five patients were allotted for this Department of Anestbesiology, Kumamoto Uni- study. All patients were hemodynamically versity MecfjcaI School, Kumamoto, Japan stable and recovering from ARF of medical "'Department of Anesthesiology, Kumamoto Mu- and/or surgical etiology that had necessi- nicipal Hospital, Kumamoto, Japan HDepartment of Anesthesiology, Kagoshima tated MVS. Ten females and 15 males had University School of ./i,ledicine, Kagoshima, Japan a mean age of 58 years. The duration of Address reprint requests to Dr. Okamoto: De- MVS prior to the study ranged from 1 to partment of Anesthesiology, Kumamoto University 12 days. No patient had chronic obstructive Medical School, 1-1-1 Honjo, Kumamoto, 860 Japan lung disease or evidence of diaphragmatic or
J Anesth 4:213-218, 1990 214 Okamoto et al J An esth 1990
Table 1. Criteria proposed for ventilator weaning
Vital capacity (VC) > 10 ml.kg-· l 7,8 Vital capacity (VC) > 12 ml.kg-1 7 Vital capacity (VC) > 15 m]·kg- 1 9 Vital capacity (VC) > I lID Tidal volume (VT) > 300 mIll Tidal volume (VT) > s ml.kg-1 8,12 Minute ventilation (MV) < 10 [.min-1 8,13,14 Minute ventilation (MV) < 180 ml.kg-l.min-1 8,15 Maximal voluntary ventilation to minute ventilation (MVV/MV) > 2 timcs8,13,14
Respiratory rate (RR) < 30 breaths-miu t ' 16 1 Respiratory rate (RR) < 25 breaths.min 11 Alveolar-arterial P02 gradient with an FlO, of 1.0 < 350 mmHg8,17 Ratio of the arterial P02 to the Flo2 (Pao2 /FIo2)
neurologic injury. a Wright respirometer (Medishield, U.K.) Each patient was evaluated at our routine during a brief interruption of the ventila- morning conference with regard to undergo- tor therapy. A spontaneous respiratory rate ing a weaning trial, The decision whether to (RR) was measured concomitantly with MV. try weaning or not was primarily based on Tidal volume (VT) was calculated from MV clinical assessment and the patient's ability and RR. The MVV to MV ratio (MVV/MV) to generate a vital capacity (VC) of at least was obtained by calculation. 10 ml-kg "! body weight or greater. In some After all measurements were made, it was of the patients, VC was a little bit lower attempted to wean the patients from MVS. than 10 ml-kg " ' body weight, but a weaning Of the 25 patients studied, 4 failed to wean. trial was still judged clinically appropriate In these 4 patients, another weaning trial if a patient's general physical condition was was repeated on a different day. Conse- good. All patients consented to this study. quently, a total of 33 weaning trials were The study was conducted in a semire- attempted. Of 33 weaning trials, 18 were cumbent position. The lead II of ECG and made by using T-piece breathing circuit with arterial blood pressure were monitored con- a large 20 I balloon reservoir". Three to 5 tinuously. The radial or dorsalis pedis artery cmH20 of continuous positive airway pres- was cannulated to obtain arterial blood sam- sure (CPAP) were applied if it seemed to ples and all blood gas tensions and pH be clinically preferable. In the remaining were measured with a Corning 168 pH/blood 15 trials, intermittent mandatory ventilation gas analyzer. Prior to each weaning trial, (nvIV) or pressure support ventilation (PSV) arterial blood gases under MVS with an was used as a weaning process. However, ordinary FlO, were measured for calculat- TMV or PSV was not used longer than 4 ing the ratio of arterial oxygen tension to hr. These patients were also placed on the the fraction of inspired oxygen (Pao, /Flo2). T-picce breathing circuit for the final pro- Subsequently, the patients were mechanically cess of weaning. The weaning process was ventilated with pure oxygen to calculate considered to be completed when endotra- the alveolar-arterial oxygen tension differ- cheal tube was extubated or when CPAP was ence (A-aD02 with an Flo2 of 1.0). Minute discontinued in tracheostomized patients. ventilation (MV), VC and maximal volun- The weaning was judged a "failure" if the tary ventilation (MVV) were measured using patient complained of severe dyspnea and/or Vol 4, No 3 Evaluation of weaning criteria in ARF 215
Table 2. The predictive ability of weaning outcome by simple measurement of the conventional weaning criteria
false positive false negative P
-~------~- MVV/MV > 2 times 4.8% ( 1/21) 33.3% (4/12) < 0.01 VC > 12 ml·kg-1 9.1% (2/22) 36.4% (4/11 ) < 0.01 RR < 25 breaths'min-1 12.5% (3/24) 33.3% (3/9) < am RR < 30 breaths·min 1 18.5% (5/27) 33.3% (2/6) < 0.05 MV < 101·min-1 13.6% (3/22) 45.5% (5/11) < 0.05 VC > 15 ml-kg- 1 11.1% (2/18) 53.3% (8/15) < 0.05 MV < 180 ml.kg-1 .min- 1 21.4% (6/28) 40.0% (2/5) N.S. VC > 10 ml·kg- 1 21.4% (6/28) 40.0% (2/5) N.S. VC > 1 I 7.7% (1/13) 60.0% (12/20) N.S. VT > 300 ml 30.8% (8/26) 14.3% (1/7) N.S. 1 VT > 5 ml-kg- 26.1% (6/23) 70.0% (7/10) N.S. A-aD02 < 350 mmHg 33.3% (6/18) 80.0% (12/15) N.S. Pao2/FI02 > 200 32.0% (8/25) 87.5% (7/8) N.S. false positive: predicted success but actual failure false negative: predicted failure but actual success P values were by two-tailed Fisher's exact test developed anxiety, agitation or diaphoresis, VC > ll, a VT > 300 ml and a VT > 5 and was placed back on MVS within 24 ml-kg"! were inconsistent with weaning out- hr. Statistical analysis of the data was per- come and surprisingly, A-aDOz < 350 mmHg formed using two-tailed Fisher's exact test or and Pao, /F102 > 200 were also unreliable chi-square analysis. A P < 0.05 was consid- in predicting weaning outcomes. For exam- ered significant. ple, among the 15 patients receiving MVS for poor oxygenation (A-aDo > 350 mrnHg}, Results z 12 (80%) were successfully discontinued from Of 33 weaning trials, 24 were successful MVS and were able to sustain spontaneous and 9 failures. In the latter group, all pa- respiration on a CPAP system with a large tients required reinstitution of MVS within 9 balloon reservoir. All these patients placed hr due to progressive respiratory distress. No OIl a CPAP system were successfully extu- complications occurred from these trials. bated within 2 to 52 hr. Table 2 shows the predictive ability of Table 3 and 4 show the predictive value each criterion for weaning outcome. Of the of weaning outcome through combinations of 13 criteria, a MVV/MV > 2, a VC > 12 the 3 or 4 criteria. All 17 patients who sat- mlkg"! and a RR < 25 breaths-min t ' were isfied 3 or 4 of the criteria were successfully most consistent with a successful weaning weaned. The less the criteria were satisfied, outcome (P < 0.01). However, these were the smaller the success rate was (P < 0.01). also associated with a significant number of However, it sllould be noted that one patient false-negatives (predicted failure but actual who failed to satisfy all of these 4 criteria success). A RR < 30 breaths-minr", a MV was also successfully weaned. 1 10 l'min- and a VC 15 ml-kg"! also < > Discussion had predictive value for weaning outcome (P < 0.05). However, a MV < 10 l'min-1 MVS is associated with numerous com- and a VC > 15 ml-kg"! were associated plications, including pneumonia, pneumoth- with many more cases of false-negatives than orax, depression of cardiac output, fluid re- the 3 above mentioned criteria. A MV < tention, gastric bleeding due to stress ulcer, 180 ml·kg-1·min-1, a VC > 10 ml-kg" ", a etcl - 3 . Therefore, weaning from MVS should 216 Okamoto et aj .J Anesth 1990
Table 3. The predictive ability of weaning Table 4. The predictive ability of weaning outcome by the combination of a outcome by the combination of a MVV/MV > 2 times, a RR < RR < 25 breaths-mini", a MV 25 breaths-min-1, a MV < 10 < 10 [·min-1 and a VC > 12 1 1 l·min- and s vc > 12 ml.kg- ml-kg"!
No. of the No. of No. of No. of the No. of No. of Success Success criteria weaning weaning criteria weaning weaning rate rate satisfied success failure satisfied success failure 4 17 0 100.0% 3 17 0 100.0% 3 2 1 66.7% 2 3 2 60.0% 2 2 1 66.7% 1 3 3 50.0% 1 2 3 40.0% 0 1 4 20.0% 0 1 4 20.0% P < 0.01 by using chi-square analysis P < 0.01 by using chi-square analysis
disappointing results, as with MVV /MV, be undertaken as soon as any illness re- may be partly due to its dependence on quiring MVS is resolved. If MVS can be patient cooperation. Moreover, VC is as- discontinued earlier, its complications may sociated with respiratory muscle strength be lessened, length of ICU stay decreased, rather than endurancel". Respiratory muscle medication requirements diminished, and pa- fatigue, an important cause of weaning fail- tient comfort improved. However, it is often ure, is associated with endurance rather than difficult to determine when weaning trials strength19. should be initiated. Premature discontinua- Sahn et al. 8,14 have recommended using tion of MVS may lead to severe respiratory peak negative pressure (PNP) < -30 cmUzO distress and left ventricular dysfunction'"". to determine weaning outcome and in a se- Therefore, it is imperative to have reliable ries of 100 patients found that all generated predictors for weaning outcome. values less than -30 crnHj O were success- A MVV/ MV > 2 has been recommended fully weaned, while none with values more as a specific predictor of a successful wean- than - 20 cmRz0 could sustain spontaneous ing outcome8.13,14. However, Tahvanainen et respiration. In contrast, Tahvanainen et a1. 4 al. 4 have recently questioned its predictive recently reported that a value of -30 cmHj O ability. In the present study, the MVV/MV was falsely negative in 100% of patients and > 2 criterion would have resulted in a sig- falsely positive in 26%. We did not mea- nificant number of false-negatives (33.3%) sure PNP in this study due to several of its suggesting that this criterion may unnec- significant drawbacks. Similar to VC, PNP essarily prolong the length of MVS. The requires the patient's cooperation and pro- drawback of MVV / MV as a criteria is its vides limited information about respiratory dependence on patient's effort and good co- muscle endurance-". In addition, PNP does 19 operation. Furthermore, the MVV maneuver not take into account thoracic compliance • is an exhausting test and must be avoided In ARF patients with low lung compliance, in patients with coronary artery disease. This high PNP values may be produced by rel- criterion will not be appropriate as a routine atively small increases in lung volume and test in patients of advanced age. the associated increase in the work of breath- VC is one of the most commonly-used ing to sustain spontaneous respiration may predictors'
RR and VT may be useful in distinguish- The use of this CPAP system may facilitate ing ventilator-dependent patients from those weaning. who can be weaned easily. They found that Finally, we examined the predictive ability patients who exhibited rapid and shallow- of weaning outcome by combination of the breathing patterns immediately following 3 or 4 criteria (tables 3 and 4). The pre- discontinuation of MVS failed the weaning dictive ability of the combinations of the 3 trial. We also observed similar phenomena or 4 criteria was almost similar. The results in RR in most of the present patients who suggest that if a patient at least satisfies RR 1 failed to wean. However, unlike the study of < 25 breaths'min- , MV < 10 i-min 1 and Tobin et alY, we were unable to find any VC > 12 ml-kg"! he or she will successfully predictive value in VT, though this discrep- wean. Still, patients not satisfying some or ancy may be due to the differences of under- all of the 3 criteria may also be successfully lyjng diseases in the patients. In the present weaned. study, a RR < 25 breaths-min-1 was more In conclusion, among the 13 conventional reliable than a RR < 30 breaths-min"! in criteria, a MVV /MV > 2 times, a RR < 25 distinguishing patients who could and could breaths-rninr", a MV < 10 l·min-- 1 and a not be successfully weaned. The results were VC> 12 ml-kg"! appeared to be most useful consistent with those of Gravelyn et al. 20 as predictors of weaning outcome in patients A high MV often implies the presence with ARF. However, these were also highly of increased dead space related to resid- falsely-negative as predictors of weaning out- ual pulmonary disease and/or increased CO 2 come. Therefore, these criteria may prolong productiorr'". If a high MV is required to the length of ventilator support. Since most maintain adequate ventilation, the associated of the patients with poor oxygenation were increase in the work of breathing may lead able to be discontinued from ventilators by to respiratory muscle fatigue and consequent placing them on a CPAP system with a weaning failure. In the present study, a MV large balloon reservoir, the improvement of < 10 /·min-1 was significant in predict- oxygenation is not prerequisite for ventilator ing weaning outcomes while a MV < 180 weaning in patients with ARF. ml.kg-1·min-1 was not, and therefore, we (Received Sep. 20, 1989, accepted for publica- recommend using the former as opposed to tion Nov. 15, 1989) the latter criteria. Unlike MVV/MV, VC and References PNP, MV is independent of the patient's effort: a distinct advantage. 1. Grum CM, Morganroth ML: Initiating me- In a 1970 paper, Pontoppidan et al, rec- chanical ventilation. J Intensive Care Med 3:6-20, 1988 ommended an A-aD02 < 350 mmHg as an additional useful predictor of weaning out- 2. Craven DE, Kunches LM, Kilinsky V, comes. However, we could not find any pre- Lichtenberg DA, Make BJ, Mccabe WR: Risk factors for pneumonia and fatality dictive value for an A-aD0 350 mmHg 2 < in patients receiving continuous mechanical and Pao2/Flo2 > 200. Eighty. percent of ventilation. Am Rev Respir Dis 133:792- patients with poor oxygenation were success- 796, 1986 fully discontinued from MVS. They sustained 3. Tsuno K, Sakanashi Y, Kishi Y, Urata K, spontaneous respiration on a CPAP system Tanoue T, Higashi K, Yano T, Terasaki with a large balloon reservotr" and were ex- H, Morioka T: Acute respiratory failure tubated within 2 to 52 hOUTS without prob- induced by mechanical pulmonary ventila- lerns. Thus, improvements in oxygenation in tion at a peak inspiratory pressure of 40 patients with ARF is not a prerequisite for CmH20. J Anesth 2:176~183, 1988 ventilator weaning. A patient with poor oxy- 4. Tahvanainen J, Salmenpera M, Nikki P: genation may be placed on a CPAP system Extubation criteria after weaning from in- termittent mandatory ventilation and con- with a large balloon reservoir-" if the pa- tinuous positive airway pressure. Crit Care tient's ventilation appears to be sufficient. Med 11:702-707, 1983 218 Olalmoto et al J Anestb 1990
5. Demting RH, Read T, Lind LJ, Flanagan use of a nomogram to estimate proper HL: Incidence and morbidity of extubation ventilation during artificial respiration. N failure in surgical intensive care patients. Engl J Med 251:877-884, 1954 Crit Care Med 16:573-577, 1988 13. Stetson JB: Prolonged Tracheal Intubation. 6. Lemaire F, Teboul J-L, Cinotti L, Giotto G, Boston, Little Brown & Co Inc., 1970, vol. Abrouk F, Steg G, Macquin-Mavier I, Za- 8, pp 767-779 pol WN: Acute left ventricular dysfunction 14. Sahn SA, Lakshminarayan S: Bedside crite- during unsuccessful weaning from mechan- ria for discontinuation of mechanical venti- ical ventilation. Anesthesiology 69:171-179, lation. Chest 63:1002-1005, 1973 1988 15. Lecky JH, Ominsky AJ: Postoperative res- 7. Bendixen HH, Egbert LD, Hedley-Whyt.e piratory management. Chest 62:50S-575, J, Laver MB, Pontoppidan H: Management 1972 of patients undergoing prolonged artificial 16. Kubota T, Ohtake K: Weaning from me- ventilation. Respiratory Care. St. Louis: chanical ventilator. Criteria, method and C.V. Mosby Co., 1965, pp 149-150 respiratory monitor for weaning (Japanese). 8. Sahn SA, Lakshminarayan S, Petty TL: Acute Medicine 11:341-351, 1987 Weaning from mechanical ventilation. 17. Pontoppidan H, Laver MB, Geffin B: Acute JA,MA 235:2208-2212, 1976. respiratory failure in the surgical patient. 9. Feeley TW, Hedley-Whyte J: Weaning Advances in Surgery. Edited by Welch, from controlled ventilation and supplemen- C.E., Chicago, Year Book Medical Publish- tal oxygen. N Engl J Med 292:903-906, ers, 1970, vol. 4, pp 163-254 1975 18. Bshouty ZH, Roeseler J, Reynaert MS, Ro- 10. Safar P, Kunkel HG: Prolonged artificial denstein D: The importance of the balloon ventilation. Respiratory Therapy. Edited by reservoir volume of a CPAP system in Safar, P. Philadelphia, F.A. Davis Co., reducing the work of breathing. Intensive 1969, pp 126-128 Care Med 12:153-156, 1986 11. Tobin MJ, Perez W, Guenther SM, Semmes 19. Tobin MJ: Respiratory monitoring in the BJ, Mador MJ, Allen SJ, Lodato RF, intensive care unit. Am Rev Respir Dis Dantzker DR: The pattern of breathing 138:1625-1642, 1988 during successful and unsuccessful trials of 20. Gravelyn TR, Weg JG: Respiratory rate as weaning from mechanical ventilation. Am an indicator of acute respiratory dysfunc- Rev Respir Dis 134:1111-1118, 1986 tion. JAMA 244:1123 ..1125, 1980 12. Radford EP, Ferris BG, Kriete BC: Clinical