Long-Term Outcome Poor After Prolonged Ventilator Weaning

Abstract & Commentary

Synopsis: This large, single-center observational study found that 5-year survival of patients requiring prolonged mechanical ventilation and care at an in-patient hospital-based weaning unit was only 19%.

Source: Stoller JK, et al. Long-term outcomes for patients discharged from a long-term hospital-based weaning unit. Chest. 2003;24:1892-1899.

Only limited data are available on long-term outcomes after prolonged mechanical ventilation. The primary aim of this observational study was to report 5-year survival rates of such patients. The setting was the Respiratory Special Care Unit (ReSCU) at Cleveland Clinic Foundation (CCF) in Cleveland, Ohio. This 6-bed weaning unit accepts patients only from intensive care units (ICUs) within CCF. It offers 24-hour respiratory therapy services, nurses trained in pulmonary care and rehabilitation, daily sessions with other support staff (physical and occupational therapists, dieticians, other), and noninvasive monitoring. Eligible patients must be hemodynamically stable and free of acute arrhythmias requiring telemetry. They must be good candidates for successful weaning as judged by the attending physician. Patients who are not likely to wean may only be admitted for training for home mechanical ventilation. ReSCU discharge criteria include ventilator independence for > 48 hours, consistent inability to sustain spontaneous breathing (failed trials for > 3 days), or hemodynamic instability requiring transfer to an ICU. Discharge is to home, rehabilitation facility, acute care hospital unit, or long-term care facility for continued ventilatory support without further weaning.

All patients admitted to the ReSCU between August 22, 1993, and August 22, 1996, were included in this study. Causes of respiratory failure were categorized into 5 groups: COPD, nonsurgical (ARDS or chronic lung disease other than COPD), respiratory failure complicating surgical interventions (including ARDS associated with surgical interventions), neuromuscular disease, and other.

Over the 3-year study period, 162 patients (with total 204 admissions) were cared for in the ReSCU. Median hospital length of stay prior to admission to the ReSCU was 29 days (interquartile range, 18-45 days). Fifty-nine percent of patients were female and the mean age was 65 years. Using Stoller and colleagues’ categories, these patients most commonly had respiratory failure complicating surgical interventions (50%). Twenty-seven patients (17%) died before discharge from the ReSCU. (Mortality rate among the few patients readmitted to the ReSCU was similar.) Median length of stay in the ReSCU is not presented. Discharge was most commonly to a skilled nursing facility (63%); it is not clear how many patients were still mechanically ventilated at this time. Only 28% of patients were discharged directly to home; 15% of these continued to require mechanical ventilation for at least part of the day.

Five-year follow-up data were available for 94% of patients. Kaplan-Meier survival rates were: 43% (95% confidence interval [CI], 35-51%) at 1 year, 27% (95% CI, 20-34%) at 3 years and 19% (95% CI, 13-25%) at 5 years. The highest risk of death appeared to be within the first 2 years following discharge from the ReSCU. There was a significant association between survival and year of admission to the ReSCU; survival improved slightly with each year between 1993 and 1996. Increasing patient age was noted to be associated with worse survival after adjusting for variables identified a priori (gender, year of ReSCU admission, and category of respiratory failure). Risk ratio for death per 10 years above 65 years was 1.3.

Stoller et al conclude that long-term survival in this patient population is poor. They suggest further investigation of potential factors for poor survival so that we may affect future outcomes

Comment by Saadia R. Akhtar, MD, MSC

Stoller et al’s study has a straightforward, simple observational design and is presented clearly and concisely. The 5-year follow-up is nearly complete, and thus the overall survival data are very useful. Other outcomes though were not assessed. Time-to-liberation from mechanical ventilation and ultimate time-to-home are also not reported. Stoller et al do note an association between increasing age and poor outcome, as has been reported previously. They also attempt to assess the effect of cause of respiratory failure on survival. Unfortunately, their method of categorizing causes and their patient numbers do not allow for separation of important causes that may influence outcome. Thus this information adds little to their report. Finally, although they note that the greatest risk of death appeared to be within the first 2 years after discharge from the ReSCU, they do not provide information about these patients’ characteristics or causes of death; such data may have been useful for generating hypotheses about factors predictive of outcome. What we are presented with then is an important but limited report of long-term survival in this patient population.

It is quite clear from this and 6 prior observational studies (nicely summarized in the discussion) that long-term survival after prolonged mechanical ventilation is poor. Further support of this alone is unnecessary. These data can help to guide patients and families in their decision-making, but they leave many questions unanswered.

Future studies must evaluate outcomes other than survival: formal long-term quality-of-life measurement is vital and may be what is most important to patients. (This has been done only in a very limited way in 2 of the prior observational studies; Carson et al assessed whether patients are independent and ambulatory at 1 year1 and Nasraway et al provide results of patients’ own rating of their overall health.2) Other outcomes such as time to liberation from ventilation, time to discharge to home, rehospitalization (to acute care facilities), and cost analyses are due. The next investigations must also move further toward identifying additional factors associated with outcome and, more importantly, prospectively evaluating these to see whether they are indeed predictive of specific outcomes. Multicenter studies would supply the large numbers of patients required to address these questions in a reasonable time period and would allay some of the issues of generalizability raised by single-center reports. (At least 1 multi-center study is currently on-going.3)

There are also inadequate data available on the long-term outcomes of patients requiring prolonged mechanical ventilation but being cared for in an acute care ICU rather than a specialized weaning unit. In some regions of the country, this is the only alternative for care for patients requiring prolonged ventilation. It is imperative to determine whether outcomes differ between these 2 settings. This may help to guide future resource allocation and may affect development of facilities for chronic patient care.

Stoller et al’s report supports and strengthens the existing observational data on long-term outcomes after prolonged mechanical ventilation. More importantly, though, their work serves as a reminder of the data that we are lacking. I hope this study will inspire and drive us to carefully consider the most appropriate future directions for this area of research. I hope it will be the impetus for developing and carrying out those investigations!

Dr. Akhtar, Pulmonary and Critical Care Medicine Yale University School of Medicine


1. Carson SS, et al. Outcomes after long term acute care: An analysis of 133 mechanically ventilated patients. Am J Resp Crit Care Med. 1999;159(5):1569-1577.

2. Nasraway SA, et al. Survivors of catastrophic illness: Outcome of direct transfer from intensive care to extended care facilities. Crit Care Med. 2000;28(1): 19-25.

3. Scheinhorn DJ, et al. Liberation from prolonged mechanical ventilation. Crit Care Clin. 2002; 18(3):569-595.