Failed Extubation a Source of Increased Mortality?

Abstract & Commentary

Synopsis: Medical ICU patients requiring reintubation following weaning from mechanical ventilation experienced an almost 4-fold increase in mortality (43% vs 12%) as compared to patients who were initially successfully extubated. Increased ICU and hospital length-of-stays were also seen in the reintubated group.

Source: Epstein SK, et al. Chest 1997;112:186-192.

Reintubation is associated with increased mortality. Epstein and colleagues prospectively studied 404 consecutive patients who were mechanically ventilated for six hours or more and compared the 247 who remained extubated (ES = Extubation success group) to the 42 who required reintubation (EF = Extubation failure group) either within 72 hours or within seven days if the patient remained continuously in the ICU. One-hundred fifteen of the initial group were never extubated (104 died intubated, and 11 went to long-term care facilities with an artificial airway). Age, cause of respiratory failure, other organ failures, days of mechanical ventilation prior to extubation, and severity of illness at intubation and prior to extubation (APACHE II Score) were compared between the groups.

EF differed form ES in mortality (43% vs 12%), ICU length-of-stay (LOS) (21 vs 4.5 days), hospital LOS (30 vs 16 days), and need for transfer to a long-term care facility (38% vs 21%). EF patients were older (64 vs 55 years), had higher APACHE II scores at weaning (12 vs 10), and were more likely to have a cardiac cause for their respiratory failure. On multiple regression analysis, reintubation was an independent predictor of death and need for long-term care. The presence of one or more additional organ failure and the need for dialysis correlated with increased mortality. These same factors independently predicted the need for an ICU stay exceeding seven or 14 days.

The role of the act of reintubation in increasing mortality and morbidity is commented on by Epstein et al. None of the 40 reintubated patients died during the procedure. They speculate that the period of extubation may have increased the stress on the reintubated patients directly causing the increased mortality.


This is an interesting paper that identifies a large group of patients requiring reintubation during treatment of severe medical illness. The important question of the contribution of the actual reintubation (or premature extubation) to mortality is not answered. One area of weakness of this study is lack of consistency of ventilation modes and extubation criteria—individual clinicians were making ventilation and extubation decisions, and many different modes of ventilation were combined in this study. Differences in clinician-associated outcome were not reported or tested. These could have contributed significantly to the results observed.

Epstein et al suggest that either the need for reintubation is a marker of more significant disease or that the intubation (or period of inappropriate extubation) is the direct cause of the increased mortality observed. Statistically, reintubation was an independent predictor of increased mortality; however, the entire ventilated group was not used for comparison. One-hundred fifteen patients in the original cohort were never extubated; most of these patients died (90%), while the others were sent to long-term care facilities. The overall mortality was 32%, whereas 43% of those extubated and reintubated died. Only 12% of the successfully extubated patients died (most were reintubated prior to death, but after 7 days of extubation). When viewed in this light, attempted but failed extubation increased survival (43% mortality) as compared to never attempting extubation (90%) in the first place. Obviously, the patients who were never extubated were more ill than those who met extubation criteria. The role of reintubation per se cannot be established by the methods used in this paper.

In a study of ours (Durbin CG, Kopel RF. A case-control study of patients readmitted to the intensive care unit. Crit Care Med 1993;21:1547-1553), we identified that ICU readmission was associated with increased mortality, and that readmission for respiratory failure was particularly lethal. Our control group was matched for diagnosis, age, sex, and ICU (medical or surgical) to help eliminate the severity of illness differences. We found that patients developing a new problem had a similar mortality to patients admitted to an ICU only once for such a problem, while those readmitted with a recurrence of their original problem experienced significantly increased mortality. This is not surprising as it indicates a more persistent disease and failure to resolve the initial problem. I believe that this is what Epstein et al are demonstrating with their study. (Incidentally, as in this study, we also found that dialysis was a predictor of death).

The contribution of premature extubation to morbidity and mortality is an important question to be answered. However, this study does not provide additional information to answer this question.