ADVERTISEMENT

Outcomes After Prolonged ICU Stays in Postoperative Cardiac Surgery Patients

Federal Practitioner. 2022 November;39(5)s:S6-S11 | doi:10.12788/fp.0300
November 15, 2022|Federal Practitioner
Thomas F. Curran, MD, MBA;Bipin Sunkara, MD;Aleda Leis;Adrian Lim, MD, PharmD;Jonathan Haft, MD;Milo Engoren, MD
Author and Disclosure Information

Thomas F. Curran, MD, MBAa,b; Bipin Sunkara, MDc; Aleda Leisa; Adrian Lim, MD, PharmDd; Jonathan Haft, MDa,b; Milo Engoren, MDa
Correspondence: Thomas Curran (thcurran@med.umich.edu)

aUniversity of Michigan, Ann Arbor, Michigan
bVeterans Affairs Ann Arbor Healthcare System, Ann Arbor, Michigan
cMedical College of Wisconsin, Wauwatosa, Wisconsin
dPromedica Toledo Hospital, Toledo, Ohio

Author disclosures

The authors report no actual or potential conflicts of interest or outside sources of funding with regard to this article.

Disclaimer

The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the US Government, or any of its agencies.

Ethics and consent

This study was approved by the University of Michigan Health System Institutional Review Board (HUM00086820 5/20/2014), which waived informed consent.

Background: Prolonged postoperative intensive care unit (ICU) stays are common after cardiac surgery and are associated with poor outcomes. There are few studies evaluating how risk factors associated with mortality may change during prolonged ICU stays or how mortality may vary with length of stay. We evaluated operative and long-term mortality in postcardiac surgery patients after prolonged ICU stays at 7, 14, 21, and 28 days and factors associated with mortality.

Methods: We included University of Michigan Medical Center cardiac surgery patients with ≥ 7 postoperative days in the ICU. We determined factors associated with hospital mortality at 7, 14, 21, and 28 days of ICU stay using logistic regression, and among hospital survivors, we determined the factors associated with long-term mortality using Cox regression.

Results: Of 8309 ICU admissions from cardiac surgery, 1174 (14%) had ICU stays > 7 days. Operative mortality was 11%, 18%, 22%, and 35% for the 7-, 14-, 21-, and 28-day groups, respectively. Mechanical ventilation on the day of assessment was associated with increased odds ratios of operative mortality in all models. Of the 1049 (89%) hospital survivors, 420 (40%) died by late follow-up. Median (IQR) Cox model survival was 10.7 (0.7) years. Longer ICU stays, postoperative pneumonia, and elevated discharge blood urea nitrogen were associated with increased hazard of dying; whereas higher discharge platelet count and cardiac transplant were protective.

Conclusions: Both operative and late mortality increased as the duration of a ICU stay increased after cardiac surgery.

Using multivariable logistic regression to adjust for factors associated with mortality, we found that receiving mechanical ventilation on the day of analysis was associated with increased operative mortality with AOR increasing from 3.35 (95% CI, 2.82-3.98) for

day 7, to 4.19 (95% CI, 3.25-5.41) for day 14 , to 6.06 (95% CI, 4.25-8.62) for day 21, to 15.68 (95% CI, 8.11-30.13) for day 28; all P values < .001 (Figure 1). Use of vasopressors was associated with an increased operative mortality only for the day 7 group, AOR 2.15 (95% CI, 1.17-2.70), P < .001. For days 7, 14, and 28, severe or moderate chronic lung disease was associated with increased AOR of operative mortality: 2.19 (95% CI, 1.52-3.14; P < .001) for day 7,2.73 (95% CI, 1.99-3.75; P < .001) for day 14, and 37.02 (95% CI, 13.57-100.99; P < .001) for day 28 (Table 3). Of the 1049 (89%) hospital survivors, 420 (40%) died by late follow-up (Figure 2). Median (IQR) Cox model survival was 10.7 (0.7) years for all hospital survivors; however, long-term survival varied by ICU length of stay (Figure 3). Longer ICU stays were associated with higher late mortality: 36% for ≥ 7 days, 41% for ≥ 14 days, 48% for 21 days, and 51% for ≥ 28 days (P < .001). Univariate analysis demonstrated that survivors were less likely to have comorbidities or to be ever smokers. Survivors were younger and less likely to have a coronary artery bypass graft and more likely to have transplant surgery compared with patients who died.

After multivariable Cox regression to adjust for confounders, we found that each postoperative week was associated with a 7% higher hazard of dying (HR, 1.07; 95% CI, 1.07-1.07; P < .001). Postoperative pneumonia was also associated with increased hazard of dying (HR, 1.59; 95% CI, 1.27-1.99; P < .001),

as was elevated blood urea nitrogen. In contrast higher discharge platelet count and cardiac transplant were protective factors (Table 4).

Discussion

We found that operative mortality increased the longer the patient stayed in the ICU, ranging from 11% for ≥ 7 days to 35% for ≥ 28 days. We further found that in ICU survivors, median (IQR) survival was 10.7 (0.7) years. While previous studies have evaluated prolonged ICU stays, they have been limited by studying limited subpopulations, such as patients who are dependent on dialysis or octogenarians, or used a single cutoff to define prolonged ICU stays, variably defined from > 48 hours to > 14 days.2-7,9-12,22 Our study is similar to others that used ≥ 2 cutoffs.1,8 However, our study was novel by providing 4 cutoffs to improve temporal prediction of hospital outcomes. Unlike a study by Ryan and colleagues, which found no increase in mortality with longer stay (43.5% for ≥ 14 days and 45% for ≥ 28 days), our study findings are similar to those of Yu and colleagues (11.1% mortality for prolonged ICU stays of 1 to 2 weeks, 26.6% for 2 to 4 weeks, and 31% for > 4 weeks) and others (8%, 3 to 14 days; 40%, >14 days; 10%, 1 to 2 weeks; 25.7% > 2 weeks) in finding a progressively increased hospital mortality with longer ICU stays.1,4,5,8 These differences may be related to different ICU populations or to improvements in care since Ryan and colleagues study was conducted.

Fewer studies have evaluated factors associated with mortality in cardiac surgery prolonged ICU stay patients. Our study is similar to other studies that evaluated risk factors by finding associations between a variety of comorbidities and process of care associated with both operative and long-term mortality; however, comparison between these studies is limited by the varying factors analyzed.1,3,5,6,8,9,11 We found that mechanical ventilation on days 7, 14, 21, and 28 was strongly associated with operative mortality, similar to noncardiac surgery patients and cardiac surgery patients.6,23,24 While we found several processes of care, such as catecholamine use and transfusions to be associated with mortality, which is similar to other studies, notably, we did not find an association between renal replacement therapy and mortality.1,25 While there is an association between renal replacement therapy and mortality in ICU patients, its status in cardiac surgery patients with prolonged ICU stays is less clear.26 While Ryan and colleagues found an association between renal replacement therapy and hospital mortality in patients staying ≥ 14 days, they did not find it in patients staying ≥.

28 days.1 Other studies of prolonged ICU stays for cardiac surgery patients have also failed to find an association between renal replacement therapy and mortality.5,6,9 Importantly, practice that expedites liberation from mechanical ventilation, such as fast tracking, daily spontaneous breathing trials, extubation to noninvasive respiratory support, and pulmonary rehabilitation may all have potential to limit mechanical ventilation duration and improve hospital survival and deserve further study.27-29Median (IQR) survival in hospital survivors was 10.7 (0.7) years, which is generally better than previously reported, but similar to that reported by Silberman and colleagues.2,4,6,8,11,12 Differences between these studies may relate to different patient populations within the cardiac surgery ICUs, definitions of prolonged ICU stays, or eras of care. Further study is needed to clarify these discrepancies. We found that cardiac transplantation and obesity were associated with the least risk of dying, while smoking, lung disease, and postoperative pneumonia were independently associated with increased hazard of dying. The obesity paradox, where obesity is protective, has been previously observed in cardiac surgery patients.30

Strengths and Limitations

There are several limitations of this study. This is a single center study, and our patient population and processes of care may differ from other centers, limiting its generalizability. Notably, we do fewer coronary bypass operations and more aortic reconstructions and ventricular assist device insertions than do many other centers. Second, we did not have laboratory values for about one-third of patients (preceded EHR implementation). However, we were able to compensate for this by binning values and including missing data as an extra bin.20,21

The main strength of this study is that we were able to combine disparate records to assess a large number of potential factors associated with both operative and long-term mortality. This produced models that had good to very good discrimination. By producing models at 7, 14, 21, and 28 days to predict operative mortality and a model at discharge, it may help to provide objective data to facilitate conversations with patients and their families. However, further studies to externally validate these models should be conducted.

Conclusions

We found that longer prolonged ICU stays are associated with both operative and late mortality. Receiving mechanical ventilation on days 7, 14, 21, or 28 was strongly associated with operative mortality.

ADVERTISEMENT
ADVERTISEMENT
ADVERTISEMENT