- Email: [email protected]
Very Prolonged Stay in the Intensive Care Unit After Cardiac Operations: Early Results and Late Survival
CARDIOTHORACIC ANESTHESIOLOGY: The Annals of Thoracic Surgery CME Program is located online at http://www.annalsthoracicsurgery. org/cme/home. To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal.
Very Prolonged Stay in the Intensive Care Unit After Cardiac Operations: Early Results and Late Survival Shuli Silberman, MD, Daniel Bitran, MD, Daniel Fink, MD, Rachel Tauber, MPH, and Ofer Merin, MD Department of Cardiothoracic Surgery, Shaare Zedek Medical Center affiliated with the Hebrew University of Jerusalem, Jerusalem, Israel
Background. Prolonged intensive care unit (ICU) stay is a surrogate for advanced morbidity or perioperative complications, and resource utilization may become an issue. It is our policy to continue full life support in the ICU, even for patients with a seemingly grim outlook. We examined the effect of duration of ICU stay on early outcomes and late survival. Methods. Between 1993 and 2011, 6,385 patients were admitted to the ICU after cardiac surgery. Patients were grouped according to length of stay in the ICU: group 1, 2 days or less (n [ 4,631; 73%); group 2, 3 to 14 days (n [ 1,423; 22%); group 3, more than 14 days (n [ 331; 5%). Length of stay in ICU for group 3 patients was 38 ± 24 days (range, 15 to 160; median 31). Clinical profile and outcomes were compared between groups. Results. Patients requiring prolonged ICU stay were older, underwent more complex surgery, had greater
comorbidity, and a higher predicted operative mortality (p < 0.0001). They had a higher incidence of adverse events and increased mortality (p < 0.0001). Of the 331 group 3 patients, 60% were discharged: survival of these patients at 1, 3, and 5 years was 78%, 65%, and 52%, respectively. Operative mortality as well as late survival of discharged patients was proportional to duration of ICU stay. Conclusions. Current technology enables keeping sick patients alive for extended periods of time. Nearly two thirds of patients requiring prolonged ICU leave hospital, and of these, 50% attain 5-year survival. These data support offering full and continued support even for patients requiring very prolonged ICU stay.
I
We analyzed our patient data in an attempt to determine the clinical profile of patients requiring prolonged stay in the ICU after cardiac operations, identify predictors for prolonged ICU stay, identify predictors for operative mortality, and determine long-term outcomes of these patients. Primary endpoints were operative and late survival; secondary endpoints were adverse events.
n recent years great advances have been made in the field of invasive cardiology. Therefore, patients undergoing heart surgery are more complex and operative risk is higher. These patients have more advanced cardiac disease as well as associated noncardiac comorbidity. As a result, postoperative complications are greater, and a greater number of patients require longer stay in the intensive care unit (ICU). With highly advanced technology at our disposal, these patients can be kept alive for prolonged periods of time, although not all will survive to be discharged. Furthermore, many who do not survive utilize the greater proportion of available resources [1]. Recent reports [2] show that these patients have worse outcomes and suggest using these data to better inform patients and families about realistic expectations.
Accepted for publication Jan 11, 2013. Presented at the Forty-ninth Annual Meeting of The Society of Thoracic Surgeons, Los Angeles, CA, Jan 26–30, 2013. Address correspondence to Dr Silberman, Department of Cardiothoracic Surgery, Shaare Zedek Medical Center, PO Box 3235, Jerusalem 91031, Israel; e-mail: [email protected].
Ó 2013 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2013;96:15–22) Ó 2013 by The Society of Thoracic Surgeons
Patients and Methods Patients All patients admitted to the ICU after cardiac surgery between 1993 and 2011 (n ¼ 6,385) were included. To facilitate data analysis, patients were divided into three groups according to length of stay in the ICU: group 1 (n ¼ 4,631; 73%) consisted of patients with ICU stay of 2 days or less; group 2 (n ¼ 1,423; 22%), with ICU stay 3 to 14 days; and group 3 (n ¼ 331; 5%), with ICU stay more than 14 days. The quoted length of stay in the ICU is from admission until transfer to the regular ward and does not include readmission to ICU. The clinical profile, operative data, and postoperative events were compared between groups. Problems encountered during the course of 0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2013.01.103
ADULT CARDIAC
ORIGINAL ARTICLES: ADULT CARDIAC
16 ADULT CARDIAC
SILBERMAN ET AL PROLONGED ICU AFTER CARDIAC SURGERY
Ann Thorac Surg 2013;96:15–22
surgery cannot always be identified. As a marker for intraoperative difficulties, we examined the length of time on bypass in excess of aortic clamp time. Discharge from ICU within less than 24 hours was defined as 1 day. Our Institutional Review Board approved this study and waived the need for individual patient consent. Operative mortality was defined as death within 30 days of surgery or within the same hospitalization. Congestive heart failure was defined as New York Heart Association functional class III to IV. Renal failure was defined as serum creatinine greater than 1.5 mg/dL. Postoperative renal deterioration was defined as a twofold increase in baseline serum creatinine. Follow-up for late survival was obtained from the Ministry of the Interior and was 94% complete.
Data Analysis Preoperative, operative, and postoperative data were collected prospectively on our departmental database (Summit Medical, Europe). The data were imported and analyzed using JMP software (SAS Institute, Cary, NC). Continuous variables were compared using Student’s t test and presented as mean SD. Nominal and categorical values were compared using the c2 likelihood ratio or Fisher’s exact test. Multivariate logistic regression was used to identify predictors of outcomes. Univariate and multivariate Cox proportional hazard functions were used to determine predictors for late mortality.
Results Patients requiring prolonged stay in the ICU were older, had greater baseline comorbidity, and had a higher
predicted operative mortality as determined by the European System for Cardiac Operative Risk Evaluation (EuroSCORE [p < 0.0001]). A larger number underwent non–isolated coronary artery bypass graft surgery (CABG), had nonelective surgery, and reoperation (p < 0.0001). Postoperative adverse events were also increased: more patients had stroke, sepsis, or renal failure (p < 0.0001). Patient clinical data are shown in Table 1, operative data are shown in Table 2, and postoperative events are shown in Table 3. By multivariate analysis, we found numerous predictors for prolonged ICU stay (Table 4). These included baseline parameters such as age (p < 0.0001) and chronic pulmonary disease (p ¼ 0.0001). Postoperative events included low cardiac output, sepsis, stroke, and renal deterioration (p < 0.0001). Non– isolated CABG as well as complexity of surgery also emerged as predictors (p < 0.0001). Neither urgency nor perioperative myocardial infarction were predictive of lengthy stay. Operative mortality was in correlation with length of ICU stay (Fig 1). Mortality was lower than predicted in the short stay group and above predicted with longer ICU stay (Table 5). The major cause of death was infection (39%) in the long-term ICU groups and cardiac (63%) in the shortterm group (Table 3). Of the 121 cardiac deaths, 109 (90%) were within 2 weeks of surgery. By multivariate analysis, predictors for operative mortality included baseline characteristics, operative events, and postoperative events (Table 4). Length of stay in ICU emerged as an independent predictor for operative mortality (p ¼ 0.001). Of note, for 3 patients, cause of death is
Table 1. Comparison of Preoperative Demographic, Clinical, and Echocardiographic Data Among the Three Groups Preoperative Data Number Male Age, years Hypertension Diabetes mellitus COPD Renal failure Pulmonary hypertension Stroke Peripheral vascular disease Atrial fibrillation CHFa Shock EuroSCOREb Echocardiography LV dysfunction MR grade III–IV TI gradient a
Group 1 <2 Days
Group 2 3–14 Days
Group 3 >14 Days
p Value
4,631 3,394 (73%) 62 12 2,585 (56%) 1,496 (32%) 326 (7%) 289 (6%) 739 (16%) 363 (8%) 554 (12%) 446 (10%) 1,303 (28%) 50 (1%) 67
1,423 888 (62%) 68 11 976 (69%) 563 (40%) 155 (11%) 275 (19%) 555 (39%) 163 (11%) 197 (14%) 267 (19%) 836 (59%) 79 (6%) 14 15
331 180 (54%) 70 10 244 (74%) 171 (52%) 50 (15%) 99 (30%) 164 (50%) 41 (12%) 54 (16%) 86 (26%) 241 (73%) 41 (12%) 22 19
<0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.02 <0.0001 <0.0001 <0.0001 <0.0001
826 (18%) 988 (21%) 33 14
400 (28%) 632 (44%) 40 15
129 (39%) 199 (60%) 45 34
<0.0001 <0.0001 <0.0001
New York Heart Association functional class III–IV.
b
Percent predicted mortality.
CHF ¼ congestive heart failure; COPD ¼ chronic obstructive pulmonary disease; EuroSCORE ¼ European System for Cardiac Operative Risk Evaluation; LV ¼ left ventricle; MR ¼ mitral regurgitation; TI ¼ tricuspid incompetence.
17
SILBERMAN ET AL PROLONGED ICU AFTER CARDIAC SURGERY
Table 2. Comparison of Operative Data Among the Three Groups Operative Data
Group 1 <2 Days
Group 2 3–14 Days
Group 3 >14 Days
p Value
Isolated CABG Urgent Reoperation BPT, min XCT, min BPT-XCT, min
3,286 (71%) 1,047 (23%) 213 (5%) 93 40 61 30 33 22
555 (39%) 448 (31%) 147 (10%) 123 55 83 38 41 35
82 (25%) 118 (36%) 62 (19%) 147 62 94 46 55 45
< < < < < <
BPT ¼ bypass time; CABG ¼ coronary artery bypass grafting; aortic cross-clamp time.
0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 XCT ¼
unknown; these are patients who died within 30 days of surgery but after discharge from our department. Of 1,754 patients requiring more than 2 days in ICU, 1,514 (86%) were discharged from hospital. Of the 331 patients with more than 14 days in ICU, 200 (60%) were discharged from hospital at 32 20 days (range, 15 to 155; median 25) after surgery. There were 170 patients in the subgroup of patients with more than 30 days of ICU stay; 84 (49%) were discharged from hospital. In that group, 1-year and 5-year survival was 64% and 40%, respectively. Overall late survival was proportional to length of ICU stay (Fig 2). Predictors for late survival among
discharged patients, by Cox regression, included baseline comorbidity as well as postoperative events (Table 6). Of note, length of ICU stay also emerged as a predictor of late survival (p ¼ 0.0002).
Comment Prolonged ICU stay after cardiac surgery is correlated with poor outcomes and major utilization of resources. We present our data for prolonged stay in the ICU, namely, 331 patients stayed longer than 14 days. To our knowledge this is the largest reported series in the English literature. The size of our large series is due mainly to three factors. (1) Our ICU is staffed by our own cardiac surgical staff and is dedicated solely to cardiac surgery patients. Thus we have full autonomy over bed allocation and sole responsibility for patient care. (2) We do not have backup and do not transfer patients to other intensive care units. (3) National as well as institutional policies do not allow discontinuation of life support, even for patients whose outlook appears to be bleak. The duration of stay in the ICU is dictated by comorbidity, intraoperative, and postoperative events. The criteria for “prolonged” ICU stay are not well defined, and in numerous reports range from 2 to 14 days after
Table 3. Early Postoperative Outcomes, Overall Mortality, and Major Cause of Death for Each Group Outcomes ICU stay, days Mean Range Median Ventilation >24 h Perioperative MI Stroke Sepsis Renal deteriorationa Low CO Pacemaker Mortality Cardiac Infection Stroke Other Multiorgan failure Bowel ischemia Pulmonary Bleeding Cirrhosis Gastrointestinal Pulmonary embolism HIT Unknown a
Group 1 <2 days
Group 2 3–14 days
Group 3 >14 days
1.3 0.5 1–2 1 40 (1%) 59 (1%) 31 (0.7%) 42 (1%) 59 (1%) 207 (4%) 39 (1%) 92 (2%) 58 (63%) 6 (7%) 9 (10%) 18 (21%) 6 (7%) 2 (2%) 2 (2%) 2 (2%) 2 (2%) 1 (1%) 1 (1%) . 2 (2%)
5.5 3 3–14 5 499 (35%) 51 (4%) 26 (2%) 63 (4%) 93 (7%) 287 (20%) 68 (5%) 109 (8%) 51 (47%) 23 (21%) 8 (7%) 27 (25%) 6 (6%) 8 (7%) 2 (2%) 2 (2%) 8 (7%) . . . 1 (1%)
38 24 15–160 31 242 (73%) 18 (5%) 50 (15%) 134 (40%) 101 (31%) 146 (44%) 13 (4%) 131 (40%) 12 (9%) 71 (54%) 18 (14%) 30 (24%) 17 (13%) 1 (0.8%) 6 (5%) 1 (0.8%) . 3 (2%) . 2 (2%) .
Twofold rise in baseline creatinine.
CO ¼ cardiac output;
HIT ¼ heparin-induced thrombocytopenia;
ICU ¼ intensive care unit;
MI ¼ myocardial infarction.
p Value < 0.0001
< < < < < < < < < <
0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.3 0.8
ADULT CARDIAC
Ann Thorac Surg 2013;96:15–22
18 ADULT CARDIAC
SILBERMAN ET AL PROLONGED ICU AFTER CARDIAC SURGERY
Ann Thorac Surg 2013;96:15–22
Table 4. Baseline, Operative, and Postoperative Predictors for Intensive Care Unit Stay Longer Than 14 Days and for Operative Mortality, by Multivariate Regression Analysis Variable Baseline Age Female Diabetes mellitus COPD Renal failure EuroSCOREa MR grade III–IV Urgency Shock Operative Non–pure CABG BPT-XCT Postoperative Perioperative MI Low output Sepsis Stroke Renal deteriorationb ICU stay a
Percent predicted operative mortality.
ICU >14 Days OR (95% CI)
p Value
Operative Mortality OR (95% CI)
(1.02–1.06) (1.17–1.62) (1.15–2.08) (1.46–3.24) . 1.01 (1.00–1.02) 1.48 (1.05–2.08) . 2.27 (1.28–4.03)
< 0.0001 0.04 0.004 < 0.0001
0.005
1.03 (1.01–1.05) 2.82 (2.73–2.95) . . 1.58 (1.08–2.3) 1.02 (1.01–1.03) . 1.78 (1.28–2.47) 2.30 (1.32–4.01)
2.31 (1.57–3.38) 1.008 (1.004–1.012)
< 0.0001 < 0.0001
. 1.015 (1.011–1.018)
< < < <
3.73 2.57 8.59 9.89 4.18 1.015
1.04 1.36 1.54 2.17
3.00 13.29 9.13 6.24
b
. (2.18–4.13) (9.21–19.7) (5.14–16.23) (4.23–9.21) .
0.01 0.03
0.0001 0.0001 0.0001 0.0001
(2.03–6.83) (1.85–3.55) (5.68–12.98) (5.60–17.45) (2.77–6.32) (1.006–1.023)
p Value < 0.0001 < 0.0001
0.02 0.001 0.001 0.003
< 0.0001 < < < < <
0.0001 0.0001 0.0001 0.0001 0.0001 0.001
Twofold rise in baseline creatinine.
BPT ¼ cardiopulmonary bypass time; CABG ¼ coronary artery bypass graft surgery; CI ¼ confidence interval; COPD ¼ chronic obstructive pulmonary disease; EuroSCORE ¼ European System for Cardiac Operative Risk Evaluation; ICU ¼ intensive care unit; MI ¼ myocardial infarction; MR ¼ mitral regurgitation; OR ¼ odds ratio; XCT ¼ aortic cross-clamp time.
surgery [1–13]. The traumatic effects of cardiopulmonary bypass usually subside within 2 days and most patients undergoing uneventful cardiac surgery will indeed be discharged from the ICU within that time. Cocker and associates [14] presented a risk model for prediction of a prolonged ICU stay based on 12 preoperative
Fig 1. Operative mortality as a function of length of stay in the intensive care unit (ICU). (d ¼ day; w ¼ week.)
variables. These are in accord with those preoperative predictors in our study. However, their model does not take into account operative and postoperative events and does not address short-term and long-term outcomes. Our data show that patients with an unequivocally uneventful course do leave the ICU within
SILBERMAN ET AL PROLONGED ICU AFTER CARDIAC SURGERY
Table 5. Predicted Versus Observed Mortality Observed Mortality Predicted Mortality 0% to 10% 11% to 20% 21% to 30% 31% to 40% >40%
Group 1 <2 Days
Group 2 3–14 Days
Group 3 >14 Days
1% 5% 13% 16% 36%
6% 9% 8% 10% 19%
31% 41% 40% 32% 63%
Preoperative predicted mortality (logistic EuroSCORE) is shown in the left column. The observed mortality is shown for each group. EuroSCORE ¼ European System for Cardiac Operative Risk Evaluation.
2 days whereas patients remaining in the ICU longer than 2 days have worse outcomes overall. Indeed, the only clear cutoff associated with poor outcomes was at 2 days. Beyond 2 days, the incidence of adverse events was proportional to the duration of ICU stay. In our study, prolonged ICU stay patients had more severe baseline morbidity, both cardiac and noncardiac. In addition, they had more postoperative events. Some—such as prolonged ventilation, stroke, or deterioration in renal function—may be the leading events for prolongation of stay. Other events such as sepsis may be the result and not the cause of prolonged ICU stay, although this cannot always be determined. One cannot avoid the fact that postoperative events are very often dictated by the intraoperative course. Despite surgery that may seem uneventful, there may be “hidden variables” that contribute to poor outcomes. We did find a correlation between what appeared to be
19
difficulties during the course of surgery, as reflected by bypass time in excess of ischemic time, with operative mortality (p < 0.0001). Of note, although preoperative hemodynamic compromise was a strong predictor for operative mortality, it was only marginally correlated with prolonged ICU stay. That is most probably because either these patients died at an early stage or survived and were discharged from the ICU within a short time.
Survival In the very long term group, cause of death was mostly noncardiac. These are patients in whom develop unforeseen events such as stroke or sepsis, events that keep them in the ICU for a prolonged period. For patients with as many as 2 weeks in ICU, the main cause of death was cardiac. The fact that the majority of cardiac deaths were within the first 2 weeks after surgery alludes to the fact that patients with inadequate surgical results do not survive. Patients who died of cardiac causes did so at an early interval and therefore did not reach the inclusion criteria for the prolonged group. That finding is in accord with the report by Hein and colleagues [15] who found that ICU stay of more than 14 days was associated with noncardiac morbidity. Although we could identify a number of predictors for operative mortality, in real time it is difficult to apply them for decisions regarding treatment policy. In reality, these were numerous and we could not identify any one in particular. Prolonged ICU patients are sicker, and although length of stay emerged as an independent predictor for adverse outcomes, it is, in essence, a surrogate for other problems, mostly operative and postoperative events. Indeed, several studies do correlate
Fig 2. Late survival of discharged patients as a function of length of stay in the intensive care unit (ICU). Overall, survival was reduced in correlation with length of intensive care unit stay. (d ¼ day; w ¼ week.)
ADULT CARDIAC
Ann Thorac Surg 2013;96:15–22
20 ADULT CARDIAC
SILBERMAN ET AL PROLONGED ICU AFTER CARDIAC SURGERY
Ann Thorac Surg 2013;96:15–22
Table 6. Baseline, Operative, and Postoperative Predictors for Late Survival, by Cox Multivariate Regression Analysis Variable Baseline Age Female Diabetes mellitus COPD Renal failure Pulmonary HTN CHF MR grade III–IV Shock Operative Non–pure CABG BPT-XCT Postoperative Perioperative MI Low output Sepsis Stroke Renal deteriorationa ICU stay a
HR (95% CI)
1.06 1.07 1.31 1.30 1.34 1.08 1.20 1.08 1.04
(1.05–1.06) (1.01–1.12) (1.25–1.37) (1.21–1.40) (1.26–1.44) (1.01–1.15) (1.13–1.27) (1.02–1.15) (0.89–1.22)
1.06 (1–1.1) 1.004 (1.002–1.006) 1.37 1.08 1.77 1.31 1.20 1.25
(1.20–1.57) (1.0–1.16) (1.04–1.34) (1.10–1.53) (1.08–1.34) (1.12–1.39)
p Value < 0.0001 0.01 < 0.0001 < 0.0001 < 0.0001 0.05 < 0.0001 0.007 0.7 0.05 < 0.0001 < 0.0001 0.07 0.02 < 0.0001 < 0.0001 0.0002
Twofold rise in baseline creatinine.
BPT ¼ cardiopulmonary bypass time; CABG ¼ coronary artery bypass graft; CHF ¼ congestive heart failure; CI ¼ confidence interval; COPD ¼ chronic obstructive pulmonary disease; HR ¼ hazard ratio; HTN ¼ hypertension; ICU ¼ intensive care unit; MI ¼ myocardial infarction; MR ¼ mitral regurgitation; XCT ¼ aortic cross-clamp time.
between operative risk models and prolonged ICU stay after cardiac surgery [12, 16–19]. However, our study does show that a significant number of patients with prolonged ICU stay can eventually be discharged, and this group does enjoy a reasonable long-term survival. Patients with more than 14 days in the ICU attained an overall survival of 50% at 5 years after discharge. Even in the most extreme group with more than 30 days in the ICU, 50% do get discharged; of these, 64% will survive 1 year and 40% will attain 5-year survival. After discharge, the highest rate of mortality is within the first year. Although this is tempered after 1 year, late survival is still poorer as ICU stay is longer. That may be due to greater comorbidity in these patients or possibly because prolonged ICU stay in itself, for whatever reasons, does leave a mark. Indeed, we found ICU stay itself to be an independent predictor of late survival. Despite lack of uniformity in the definition of prolonged ICU stay, all reports show increased operative mortality, ranging between 29% and 53% [1, 2, 5, 6, 9, 10] as well as reduced late survival [1, 3, 5–7] associated with prolonged ICU stay. Our results show the correlation between length of stay in the ICU and overall late survival, and this has not been previously reported.
Quality of Life We do not have an assessment of quality of life for hospital survivors, and that is a limitation of our study.
Some researchers report poor quality of life among patients surviving after prolonged ICU stay [3, 6, 7, 9] whereas others report good quality of life [1, 5, 10, 20]. Hellgren and Stahle [3] emphasize, for a group of valve surgery patients, that the comparison of quality of life should be made in similar patients if for some reason they would not have undergone surgery. The alternative would have been medical therapy, which is associated with poor prognosis for symptomatic valvular patients. In their study [3], although 20% did not experience an improvement, none of the prolonged ICU group regretted the surgery. This finding could be due to the awareness of these patients of the poor prognosis associated with medical treatment. With lack of uniformity in definition of prolonged ICU stay and the somewhat subjective perception of quality of life, it is difficult to apply such conclusions to clinical practice.
Clinical Implications Most reports on the topic of prolonged ICU stay relate to short-term outcomes and resource utilization as well as patient and family expectations once the postoperative course is extended. Williams and associates [1] report 49 patients who stayed more than 14 days in the ICU; these patients comprised only 3.8% of the patient volume but utilized 28% of the total ICU beds. These issues do raise certain ethical dilemmas. With increasing costs, limited resources, and reduced expected survival, while at the same time withholding services for other patients, what is the right decision? Should there be a cutoff point for ICU stay? If so, who is to decide when? Our data do not give a clear answer but do give some understanding of in-hospital outcomes, and in particular, long-term survival. Indeed, a significant number of patients live many years after discharge. Whether prolonged ICU stay is a reflection of extreme attempts in keeping these patients alive or is a measure of patients’ viability is difficult to determine. Furthermore, we cannot determine what the outcomes of such patients might be had they been transferred to a non–intensive care facility at an earlier stage. The challenge is to prospectively predict which patients will not survive despite ICU care. Different models attempt to predict inevitable death in patients after cardiac surgery requiring prolonged ICU stay [1, 12, 21]; however, it would be difficult to implement these models to aid decisions about withdrawal of treatment [22]. For patients with a protracted ICU stay, risk assessment becomes a “moving target” that may vary on a daily, if not hourly, basis. Better prediction of outcomes in real time requires very detailed data collected prospectively. This is beyond the scope of this study but is a topic for further investigation and possible establishment of an "evolving" score. Our study is unique in a number of aspects. The series includes a large number of patients with lengthy followup. Despite poorer results after prolonged stay in the ICU, a significant number of patients do get discharged from hospital. Lastly, whereas other investigators determined a range of cutoff points for lengthy stay, we have
shown a continuous correlation between length of ICU stay after surgery and long-term survival. In conclusion, prolonged stay in the ICU after cardiac surgery is a surrogate for advanced morbidity and predicts reduced survival. This is in correlation with duration of stay. Despite this, a significant number of these patients will survive 10 years or more. Our data shed more light with regard to expectations, both short term and long term. We wish to thank Dr Bernard S. Goldman for his valuable assistance in the preparation of this manuscript.
References 1. Williams MR, Wellner RB, Hartnett EA, et al. Long-term survival and quality of life in cardiac surgical patients with prolonged intensive care unit length of stay. Ann Thorac Surg 2002;73:1472–8. 2. Hassan A, Anderson C, Kypson A, et al. Clinical outcomes in patients with prolonged intensive care unit length of stay after cardiac surgical procedures. Ann Thorac Surg 2012;93: 565–9. 3. Hellgren L, Stahle E. Quality of life after heart valve surgery with prolonged intensive care. Ann Thorac Surg 2005;80: 1693–8. 4. Hein OV, Birnbaum J, Wernecke K, England M, Konertz W, Spies C. Prolonged intensive care unit stay in cardiac surgery: risk factors and long-term survival. Ann Thorac Surg 2006;81:880–5. 5. Lagercranz E, Lindblom D, Sartipy U. Survival and quality of life in cardiac surgery patients with prolonged intensive care. Ann Thorac Surg 2010;89:490–6. 6. Gaudin o M, Girola F, Piscitelli M, et al. Long-term survival and quality of life of patients with prolonged postoperative intensive care unit stay: unmasking an apparent success. J Thorac Cardiovasc Surg 2007;134:465–9. 7. Bapat V, Allen D, Young C, Roxburgh J, Ibrahim M. Survival and quality of life after cardiac surgery complicated by prolonged intensive care. J Cardiac Surg 2005;20:212–7. 8. Heimrath OP, Buth KJ, Legare JF. Long-term outcomes in patients requiring stay of more than 48 hours in the intensive care unit following coronary bypass surgery. J Crit Care 2007;22:153–8.
SILBERMAN ET AL PROLONGED ICU AFTER CARDIAC SURGERY
21
9. Bashour CA, Yared JP, Ryan TA, et al. Long-term survival and functional capacity in cardiac surgery patients after prolonged intensive care. Crit Care Med 2000;28:3847–53. 10. Isgro F, Skuras JA, Keissling AH, Lehmann A. Survival and quality of life after long-term intensive care stay. Thorac Cardiovasc Surg 2002;50:95–9. 11. Ghotkar SV, Grayson AD, Fabri BM, Dihmis WC, Pullan DM. Preoperative calculation of risk for prolonged intensive care unit stay following coronary artery bypass grafting. J Cardiothorac Surgery 2006;1:14. 12. Tu JV, Jaglal SB, Naylor CD. Multicenter validation of a risk index for mortality, intensive care unit stay, and overall hospital length of stay after cardiac surgery. Circulation 1995;91:677–84. 13. Mazzoni M, De Maria R, Bortone F, et al. Long-term outcome of survivors of prolonged intensive care treatment after cardiac surgery. Ann Thorac Surg 2006;82:2080–8. 14. De Cocker J, Messaoudi N, Stockman BA, et al. Preoperative prediction of intensive care unit stay following cardiac surgery. Eur J Cardiothorac Surg 2011;39:60–7. 15. Hein OV, Birnbaum J, Wernecke KD, Konertz W, Spies C. Intensive care unit stay of more than 14 days after cardiac surgery is associated with non-cardiac organ failure. J Int Med Res 2006;34:695–703. 16. Nilsson J, Algotsson L, Hoglund P, Luhrs C, Brandt J. EuroSCORE predicts intensive care unit stay and costs of open heart surgery. Ann Thorac Surg 2004;78:1528–35. 17. Lawrence DR, Valencia O, Smith EEJ, Murday A, Treasure T. Parsonnet score is a good predictor of the duration of intensive care unit stay following cardiac surgery. Heart 2000;83:429–32. 18. Ettema RGA, Peelen LM, Schuurmans MJ, Nierich AP, Kalkman CJ, Moons KGM. Prediction models for prolonged intensive care unit stay after cardiac surgery: systematic review and validation study. Circulation 2010;122:682–9. 19. Van Caenegem O, Jacquet LM, Goenen M. Outcome of cardiac surgery patients with complicated intensive care unit stay. Curr Opin Crit Care 2002;8:404–10. 20. Trouillet JL, Scheimberg A, Vuagnat A, Fagon JY, Chastre J, Gibert C. Long-term outcome and quality of life of patients requiring multidisciplinary intensive care unit admission after cardiac operations. J Thorac Cardiovasc Surg 1996;112: 926–34. 21. Holmes L, Loughead K, Treasure T, Gallivan S. Which patients will not benefit from further intensive care after cardiac surgery? Lancet 1994;344:1200–2. 22. Barnato AE, Angus DC. Value and role of intensive care unit outcome prediction models in end-of-life decision making. Crit Care Clin 2004;20:345–62.
DISCUSSION DR KEVIN W. LOBDELL (Charlotte, NC): You addressed most of the things that I thought about in advance, but maybe you could elaborate. We are all challenged with patients of this nature. There are existing risk models for predicting what happens once a patient is in our unit and they have a set of problems. The difficulty lies in predicting outcomes in a single patient. So, in other words, we might calculate using the CASUS score and determine that the patient has a 95% predicted mortality, but what does one do? DR SILBERMAN: First of all, there are many models predicting operative mortality, but they are preoperative. There are also models predicting risk of long stay in the intensive care unit (ICU). The problem is to predict which patient in the ICU has a chance to survive and leave the hospital. These assessments may change not only on a daily basis but even on an hourly basis. We can count how many blood units patients received, how
many perfusion pumps they require, infection, no infection, and so on, and these will vary even within a single day. If you do a score one day, it will change by the next. So this is a problem. Our policy is just to maintain treatment. For various reasons, be it national, religious, ethical reasons, we just do not withhold treatment. For a patient in extreme condition, with seemingly no chance of survival, common sense may dictate not to start cardiopulmonary resuscitation if the patient arrests. But we will not withhold treatment and we will maintain full support for as long as necessary, and indeed, many of these do go home. Of all patients with more than a month in the ICU, 20% will be alive at 5 years. DR GLENN WHITMAN (Baltimore, MD): It is interesting that you looked at it the way you did, because I think your conclusion, in my opinion, simply creates more dilemmas for us. What so frequently happens in the ICU is that after a couple of weeks
ADULT CARDIAC
Ann Thorac Surg 2013;96:15–22
22 ADULT CARDIAC
SILBERMAN ET AL PROLONGED ICU AFTER CARDIAC SURGERY
Ann Thorac Surg 2013;96:15–22
with multiorgan system failure, we talk to the families about what the chances are of their loved one surviving. This may lead to end-of-life discussions. Your message, however, seems to advocate to continue to be aggressive, because even after 4 weeks, your loved one has a 50% chance of being alive 5 years from now. What would benefit everyone would be to know just which patients do survive, and which do not. Armed with that, we could better advise families and better utilize resources. In my opinion, this paper unrealistically gives us optimism without accounting for the limited resources with which we all will need to deal.
about chances of survival. Have you found that you have more discussions with relatives at time frames since you have seen this or do you still continue doing as you were before?
DR SILBERMAN: We do not tell families that their loved one has a 50% chance of surviving so and so many months or years. Again, it’s a policy to maintain treatment. We are very realistic about the outcomes, and for patients in the ICU for 3 or 4 weeks or beyond, we relate this to the family; we tell them the facts. We just don’t withhold treatment. Moreover, the families don’t have the legal rights to ask to withhold treatment either. I’m not sure of any existing model that can predict which patient will go home and survive a year or 2 or 5. Our data are retrospective, and we did see that a significant number of patients did survive. To create a model for real-time calculation is a bit more difficult, it has to be prospective, and to take into account countless parameters that have to be collected.
DR DAVID A. FULLERTON (Aurora, CO): One of the other variables that could increase the survival after discharge from the hospital is the patient’s age and destination upon discharge, whether they went home or to a skilled nursing facility. Did you have the opportunity to see if any other variables might have affected the survival curve?
DR RICHARD ENGELMAN (Springfield, MA): This is a study that has gone on since 1993 until almost the present, and I would ask, have you looked at a time frame, because perhaps technology has changed over that 8- or 9-year period, and has any change in technology done anything to the data? DR SILBERMAN: The question is a good one. No, we didn’t break it down to time frames. DR ADRIAN JEREMY LEVINE (Stoke-on-Trent, UK): We did a bit of work like this seven, eight years ago. Our findings were not quite as optimistic as yours. What it did do in our group was it enabled us to discuss with relatives far more meaningfully
DR SILBERMAN: Again, we are not blinded by these results. We are still living in the realistic world and we are aware of outcomes. We know that the patient with a long stay in the ICU, for various reasons, has a smaller chance of survival, and we relate this to the families. We discuss with families from day one, and we keep them informed on a daily basis. If the situation looks bad, we tell them; we don’t raise any false hopes.
DR SILBERMAN: Well, age did emerge as a predictor, of course. We are talking about patients who were discharged. Some patients may go through a care facility such as a nursing home, but the majority eventually do go home. We do not have data regarding patients who were chronically in a home and died over there. Information from the Ministry of Interior tells us who died and when, and who is still alive, but we don’t have all the information as to whether they were hospitalized since their discharge or not. DR LOBDELL: Just one comment. I think that there is rich opportunity to understand this better than we do, as evidenced by the discussion. It has always been interesting to me that the STS database risk model for mortality takes into account preoperative and intraoperative variables, but nothing postoperatively. You could take great care of patients and my care may be poor, but the risk model would predict that our outcomes should be the same. So I think that there is a body of information within the database that we could use to understand this better than we ever have.
Very Prolonged Stay in the Intensive Care Unit After Cardiac Operations: Early Results and Late Survival Shuli Silberman, MD, Daniel Bitran, MD, Daniel Fink, MD, Rachel Tauber, MPH, and Ofer Merin, MD Department of Cardiothoracic Surgery, Shaare Zedek Medical Center affiliated with the Hebrew University of Jerusalem, Jerusalem, Israel
Background. Prolonged intensive care unit (ICU) stay is a surrogate for advanced morbidity or perioperative complications, and resource utilization may become an issue. It is our policy to continue full life support in the ICU, even for patients with a seemingly grim outlook. We examined the effect of duration of ICU stay on early outcomes and late survival. Methods. Between 1993 and 2011, 6,385 patients were admitted to the ICU after cardiac surgery. Patients were grouped according to length of stay in the ICU: group 1, 2 days or less (n [ 4,631; 73%); group 2, 3 to 14 days (n [ 1,423; 22%); group 3, more than 14 days (n [ 331; 5%). Length of stay in ICU for group 3 patients was 38 ± 24 days (range, 15 to 160; median 31). Clinical profile and outcomes were compared between groups. Results. Patients requiring prolonged ICU stay were older, underwent more complex surgery, had greater
comorbidity, and a higher predicted operative mortality (p < 0.0001). They had a higher incidence of adverse events and increased mortality (p < 0.0001). Of the 331 group 3 patients, 60% were discharged: survival of these patients at 1, 3, and 5 years was 78%, 65%, and 52%, respectively. Operative mortality as well as late survival of discharged patients was proportional to duration of ICU stay. Conclusions. Current technology enables keeping sick patients alive for extended periods of time. Nearly two thirds of patients requiring prolonged ICU leave hospital, and of these, 50% attain 5-year survival. These data support offering full and continued support even for patients requiring very prolonged ICU stay.
I
We analyzed our patient data in an attempt to determine the clinical profile of patients requiring prolonged stay in the ICU after cardiac operations, identify predictors for prolonged ICU stay, identify predictors for operative mortality, and determine long-term outcomes of these patients. Primary endpoints were operative and late survival; secondary endpoints were adverse events.
n recent years great advances have been made in the field of invasive cardiology. Therefore, patients undergoing heart surgery are more complex and operative risk is higher. These patients have more advanced cardiac disease as well as associated noncardiac comorbidity. As a result, postoperative complications are greater, and a greater number of patients require longer stay in the intensive care unit (ICU). With highly advanced technology at our disposal, these patients can be kept alive for prolonged periods of time, although not all will survive to be discharged. Furthermore, many who do not survive utilize the greater proportion of available resources [1]. Recent reports [2] show that these patients have worse outcomes and suggest using these data to better inform patients and families about realistic expectations.
Accepted for publication Jan 11, 2013. Presented at the Forty-ninth Annual Meeting of The Society of Thoracic Surgeons, Los Angeles, CA, Jan 26–30, 2013. Address correspondence to Dr Silberman, Department of Cardiothoracic Surgery, Shaare Zedek Medical Center, PO Box 3235, Jerusalem 91031, Israel; e-mail: [email protected].
Ó 2013 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2013;96:15–22) Ó 2013 by The Society of Thoracic Surgeons
Patients and Methods Patients All patients admitted to the ICU after cardiac surgery between 1993 and 2011 (n ¼ 6,385) were included. To facilitate data analysis, patients were divided into three groups according to length of stay in the ICU: group 1 (n ¼ 4,631; 73%) consisted of patients with ICU stay of 2 days or less; group 2 (n ¼ 1,423; 22%), with ICU stay 3 to 14 days; and group 3 (n ¼ 331; 5%), with ICU stay more than 14 days. The quoted length of stay in the ICU is from admission until transfer to the regular ward and does not include readmission to ICU. The clinical profile, operative data, and postoperative events were compared between groups. Problems encountered during the course of 0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2013.01.103
ADULT CARDIAC
ORIGINAL ARTICLES: ADULT CARDIAC
16 ADULT CARDIAC
SILBERMAN ET AL PROLONGED ICU AFTER CARDIAC SURGERY
Ann Thorac Surg 2013;96:15–22
surgery cannot always be identified. As a marker for intraoperative difficulties, we examined the length of time on bypass in excess of aortic clamp time. Discharge from ICU within less than 24 hours was defined as 1 day. Our Institutional Review Board approved this study and waived the need for individual patient consent. Operative mortality was defined as death within 30 days of surgery or within the same hospitalization. Congestive heart failure was defined as New York Heart Association functional class III to IV. Renal failure was defined as serum creatinine greater than 1.5 mg/dL. Postoperative renal deterioration was defined as a twofold increase in baseline serum creatinine. Follow-up for late survival was obtained from the Ministry of the Interior and was 94% complete.
Data Analysis Preoperative, operative, and postoperative data were collected prospectively on our departmental database (Summit Medical, Europe). The data were imported and analyzed using JMP software (SAS Institute, Cary, NC). Continuous variables were compared using Student’s t test and presented as mean SD. Nominal and categorical values were compared using the c2 likelihood ratio or Fisher’s exact test. Multivariate logistic regression was used to identify predictors of outcomes. Univariate and multivariate Cox proportional hazard functions were used to determine predictors for late mortality.
Results Patients requiring prolonged stay in the ICU were older, had greater baseline comorbidity, and had a higher
predicted operative mortality as determined by the European System for Cardiac Operative Risk Evaluation (EuroSCORE [p < 0.0001]). A larger number underwent non–isolated coronary artery bypass graft surgery (CABG), had nonelective surgery, and reoperation (p < 0.0001). Postoperative adverse events were also increased: more patients had stroke, sepsis, or renal failure (p < 0.0001). Patient clinical data are shown in Table 1, operative data are shown in Table 2, and postoperative events are shown in Table 3. By multivariate analysis, we found numerous predictors for prolonged ICU stay (Table 4). These included baseline parameters such as age (p < 0.0001) and chronic pulmonary disease (p ¼ 0.0001). Postoperative events included low cardiac output, sepsis, stroke, and renal deterioration (p < 0.0001). Non– isolated CABG as well as complexity of surgery also emerged as predictors (p < 0.0001). Neither urgency nor perioperative myocardial infarction were predictive of lengthy stay. Operative mortality was in correlation with length of ICU stay (Fig 1). Mortality was lower than predicted in the short stay group and above predicted with longer ICU stay (Table 5). The major cause of death was infection (39%) in the long-term ICU groups and cardiac (63%) in the shortterm group (Table 3). Of the 121 cardiac deaths, 109 (90%) were within 2 weeks of surgery. By multivariate analysis, predictors for operative mortality included baseline characteristics, operative events, and postoperative events (Table 4). Length of stay in ICU emerged as an independent predictor for operative mortality (p ¼ 0.001). Of note, for 3 patients, cause of death is
Table 1. Comparison of Preoperative Demographic, Clinical, and Echocardiographic Data Among the Three Groups Preoperative Data Number Male Age, years Hypertension Diabetes mellitus COPD Renal failure Pulmonary hypertension Stroke Peripheral vascular disease Atrial fibrillation CHFa Shock EuroSCOREb Echocardiography LV dysfunction MR grade III–IV TI gradient a
Group 1 <2 Days
Group 2 3–14 Days
Group 3 >14 Days
p Value
4,631 3,394 (73%) 62 12 2,585 (56%) 1,496 (32%) 326 (7%) 289 (6%) 739 (16%) 363 (8%) 554 (12%) 446 (10%) 1,303 (28%) 50 (1%) 67
1,423 888 (62%) 68 11 976 (69%) 563 (40%) 155 (11%) 275 (19%) 555 (39%) 163 (11%) 197 (14%) 267 (19%) 836 (59%) 79 (6%) 14 15
331 180 (54%) 70 10 244 (74%) 171 (52%) 50 (15%) 99 (30%) 164 (50%) 41 (12%) 54 (16%) 86 (26%) 241 (73%) 41 (12%) 22 19
<0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.02 <0.0001 <0.0001 <0.0001 <0.0001
826 (18%) 988 (21%) 33 14
400 (28%) 632 (44%) 40 15
129 (39%) 199 (60%) 45 34
<0.0001 <0.0001 <0.0001
New York Heart Association functional class III–IV.
b
Percent predicted mortality.
CHF ¼ congestive heart failure; COPD ¼ chronic obstructive pulmonary disease; EuroSCORE ¼ European System for Cardiac Operative Risk Evaluation; LV ¼ left ventricle; MR ¼ mitral regurgitation; TI ¼ tricuspid incompetence.
17
SILBERMAN ET AL PROLONGED ICU AFTER CARDIAC SURGERY
Table 2. Comparison of Operative Data Among the Three Groups Operative Data
Group 1 <2 Days
Group 2 3–14 Days
Group 3 >14 Days
p Value
Isolated CABG Urgent Reoperation BPT, min XCT, min BPT-XCT, min
3,286 (71%) 1,047 (23%) 213 (5%) 93 40 61 30 33 22
555 (39%) 448 (31%) 147 (10%) 123 55 83 38 41 35
82 (25%) 118 (36%) 62 (19%) 147 62 94 46 55 45
< < < < < <
BPT ¼ bypass time; CABG ¼ coronary artery bypass grafting; aortic cross-clamp time.
0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 XCT ¼
unknown; these are patients who died within 30 days of surgery but after discharge from our department. Of 1,754 patients requiring more than 2 days in ICU, 1,514 (86%) were discharged from hospital. Of the 331 patients with more than 14 days in ICU, 200 (60%) were discharged from hospital at 32 20 days (range, 15 to 155; median 25) after surgery. There were 170 patients in the subgroup of patients with more than 30 days of ICU stay; 84 (49%) were discharged from hospital. In that group, 1-year and 5-year survival was 64% and 40%, respectively. Overall late survival was proportional to length of ICU stay (Fig 2). Predictors for late survival among
discharged patients, by Cox regression, included baseline comorbidity as well as postoperative events (Table 6). Of note, length of ICU stay also emerged as a predictor of late survival (p ¼ 0.0002).
Comment Prolonged ICU stay after cardiac surgery is correlated with poor outcomes and major utilization of resources. We present our data for prolonged stay in the ICU, namely, 331 patients stayed longer than 14 days. To our knowledge this is the largest reported series in the English literature. The size of our large series is due mainly to three factors. (1) Our ICU is staffed by our own cardiac surgical staff and is dedicated solely to cardiac surgery patients. Thus we have full autonomy over bed allocation and sole responsibility for patient care. (2) We do not have backup and do not transfer patients to other intensive care units. (3) National as well as institutional policies do not allow discontinuation of life support, even for patients whose outlook appears to be bleak. The duration of stay in the ICU is dictated by comorbidity, intraoperative, and postoperative events. The criteria for “prolonged” ICU stay are not well defined, and in numerous reports range from 2 to 14 days after
Table 3. Early Postoperative Outcomes, Overall Mortality, and Major Cause of Death for Each Group Outcomes ICU stay, days Mean Range Median Ventilation >24 h Perioperative MI Stroke Sepsis Renal deteriorationa Low CO Pacemaker Mortality Cardiac Infection Stroke Other Multiorgan failure Bowel ischemia Pulmonary Bleeding Cirrhosis Gastrointestinal Pulmonary embolism HIT Unknown a
Group 1 <2 days
Group 2 3–14 days
Group 3 >14 days
1.3 0.5 1–2 1 40 (1%) 59 (1%) 31 (0.7%) 42 (1%) 59 (1%) 207 (4%) 39 (1%) 92 (2%) 58 (63%) 6 (7%) 9 (10%) 18 (21%) 6 (7%) 2 (2%) 2 (2%) 2 (2%) 2 (2%) 1 (1%) 1 (1%) . 2 (2%)
5.5 3 3–14 5 499 (35%) 51 (4%) 26 (2%) 63 (4%) 93 (7%) 287 (20%) 68 (5%) 109 (8%) 51 (47%) 23 (21%) 8 (7%) 27 (25%) 6 (6%) 8 (7%) 2 (2%) 2 (2%) 8 (7%) . . . 1 (1%)
38 24 15–160 31 242 (73%) 18 (5%) 50 (15%) 134 (40%) 101 (31%) 146 (44%) 13 (4%) 131 (40%) 12 (9%) 71 (54%) 18 (14%) 30 (24%) 17 (13%) 1 (0.8%) 6 (5%) 1 (0.8%) . 3 (2%) . 2 (2%) .
Twofold rise in baseline creatinine.
CO ¼ cardiac output;
HIT ¼ heparin-induced thrombocytopenia;
ICU ¼ intensive care unit;
MI ¼ myocardial infarction.
p Value < 0.0001
< < < < < < < < < <
0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.3 0.8
ADULT CARDIAC
Ann Thorac Surg 2013;96:15–22
18 ADULT CARDIAC
SILBERMAN ET AL PROLONGED ICU AFTER CARDIAC SURGERY
Ann Thorac Surg 2013;96:15–22
Table 4. Baseline, Operative, and Postoperative Predictors for Intensive Care Unit Stay Longer Than 14 Days and for Operative Mortality, by Multivariate Regression Analysis Variable Baseline Age Female Diabetes mellitus COPD Renal failure EuroSCOREa MR grade III–IV Urgency Shock Operative Non–pure CABG BPT-XCT Postoperative Perioperative MI Low output Sepsis Stroke Renal deteriorationb ICU stay a
Percent predicted operative mortality.
ICU >14 Days OR (95% CI)
p Value
Operative Mortality OR (95% CI)
(1.02–1.06) (1.17–1.62) (1.15–2.08) (1.46–3.24) . 1.01 (1.00–1.02) 1.48 (1.05–2.08) . 2.27 (1.28–4.03)
< 0.0001 0.04 0.004 < 0.0001
0.005
1.03 (1.01–1.05) 2.82 (2.73–2.95) . . 1.58 (1.08–2.3) 1.02 (1.01–1.03) . 1.78 (1.28–2.47) 2.30 (1.32–4.01)
2.31 (1.57–3.38) 1.008 (1.004–1.012)
< 0.0001 < 0.0001
. 1.015 (1.011–1.018)
< < < <
3.73 2.57 8.59 9.89 4.18 1.015
1.04 1.36 1.54 2.17
3.00 13.29 9.13 6.24
b
. (2.18–4.13) (9.21–19.7) (5.14–16.23) (4.23–9.21) .
0.01 0.03
0.0001 0.0001 0.0001 0.0001
(2.03–6.83) (1.85–3.55) (5.68–12.98) (5.60–17.45) (2.77–6.32) (1.006–1.023)
p Value < 0.0001 < 0.0001
0.02 0.001 0.001 0.003
< 0.0001 < < < < <
0.0001 0.0001 0.0001 0.0001 0.0001 0.001
Twofold rise in baseline creatinine.
BPT ¼ cardiopulmonary bypass time; CABG ¼ coronary artery bypass graft surgery; CI ¼ confidence interval; COPD ¼ chronic obstructive pulmonary disease; EuroSCORE ¼ European System for Cardiac Operative Risk Evaluation; ICU ¼ intensive care unit; MI ¼ myocardial infarction; MR ¼ mitral regurgitation; OR ¼ odds ratio; XCT ¼ aortic cross-clamp time.
surgery [1–13]. The traumatic effects of cardiopulmonary bypass usually subside within 2 days and most patients undergoing uneventful cardiac surgery will indeed be discharged from the ICU within that time. Cocker and associates [14] presented a risk model for prediction of a prolonged ICU stay based on 12 preoperative
Fig 1. Operative mortality as a function of length of stay in the intensive care unit (ICU). (d ¼ day; w ¼ week.)
variables. These are in accord with those preoperative predictors in our study. However, their model does not take into account operative and postoperative events and does not address short-term and long-term outcomes. Our data show that patients with an unequivocally uneventful course do leave the ICU within
SILBERMAN ET AL PROLONGED ICU AFTER CARDIAC SURGERY
Table 5. Predicted Versus Observed Mortality Observed Mortality Predicted Mortality 0% to 10% 11% to 20% 21% to 30% 31% to 40% >40%
Group 1 <2 Days
Group 2 3–14 Days
Group 3 >14 Days
1% 5% 13% 16% 36%
6% 9% 8% 10% 19%
31% 41% 40% 32% 63%
Preoperative predicted mortality (logistic EuroSCORE) is shown in the left column. The observed mortality is shown for each group. EuroSCORE ¼ European System for Cardiac Operative Risk Evaluation.
2 days whereas patients remaining in the ICU longer than 2 days have worse outcomes overall. Indeed, the only clear cutoff associated with poor outcomes was at 2 days. Beyond 2 days, the incidence of adverse events was proportional to the duration of ICU stay. In our study, prolonged ICU stay patients had more severe baseline morbidity, both cardiac and noncardiac. In addition, they had more postoperative events. Some—such as prolonged ventilation, stroke, or deterioration in renal function—may be the leading events for prolongation of stay. Other events such as sepsis may be the result and not the cause of prolonged ICU stay, although this cannot always be determined. One cannot avoid the fact that postoperative events are very often dictated by the intraoperative course. Despite surgery that may seem uneventful, there may be “hidden variables” that contribute to poor outcomes. We did find a correlation between what appeared to be
19
difficulties during the course of surgery, as reflected by bypass time in excess of ischemic time, with operative mortality (p < 0.0001). Of note, although preoperative hemodynamic compromise was a strong predictor for operative mortality, it was only marginally correlated with prolonged ICU stay. That is most probably because either these patients died at an early stage or survived and were discharged from the ICU within a short time.
Survival In the very long term group, cause of death was mostly noncardiac. These are patients in whom develop unforeseen events such as stroke or sepsis, events that keep them in the ICU for a prolonged period. For patients with as many as 2 weeks in ICU, the main cause of death was cardiac. The fact that the majority of cardiac deaths were within the first 2 weeks after surgery alludes to the fact that patients with inadequate surgical results do not survive. Patients who died of cardiac causes did so at an early interval and therefore did not reach the inclusion criteria for the prolonged group. That finding is in accord with the report by Hein and colleagues [15] who found that ICU stay of more than 14 days was associated with noncardiac morbidity. Although we could identify a number of predictors for operative mortality, in real time it is difficult to apply them for decisions regarding treatment policy. In reality, these were numerous and we could not identify any one in particular. Prolonged ICU patients are sicker, and although length of stay emerged as an independent predictor for adverse outcomes, it is, in essence, a surrogate for other problems, mostly operative and postoperative events. Indeed, several studies do correlate
Fig 2. Late survival of discharged patients as a function of length of stay in the intensive care unit (ICU). Overall, survival was reduced in correlation with length of intensive care unit stay. (d ¼ day; w ¼ week.)
ADULT CARDIAC
Ann Thorac Surg 2013;96:15–22
20 ADULT CARDIAC
SILBERMAN ET AL PROLONGED ICU AFTER CARDIAC SURGERY
Ann Thorac Surg 2013;96:15–22
Table 6. Baseline, Operative, and Postoperative Predictors for Late Survival, by Cox Multivariate Regression Analysis Variable Baseline Age Female Diabetes mellitus COPD Renal failure Pulmonary HTN CHF MR grade III–IV Shock Operative Non–pure CABG BPT-XCT Postoperative Perioperative MI Low output Sepsis Stroke Renal deteriorationa ICU stay a
HR (95% CI)
1.06 1.07 1.31 1.30 1.34 1.08 1.20 1.08 1.04
(1.05–1.06) (1.01–1.12) (1.25–1.37) (1.21–1.40) (1.26–1.44) (1.01–1.15) (1.13–1.27) (1.02–1.15) (0.89–1.22)
1.06 (1–1.1) 1.004 (1.002–1.006) 1.37 1.08 1.77 1.31 1.20 1.25
(1.20–1.57) (1.0–1.16) (1.04–1.34) (1.10–1.53) (1.08–1.34) (1.12–1.39)
p Value < 0.0001 0.01 < 0.0001 < 0.0001 < 0.0001 0.05 < 0.0001 0.007 0.7 0.05 < 0.0001 < 0.0001 0.07 0.02 < 0.0001 < 0.0001 0.0002
Twofold rise in baseline creatinine.
BPT ¼ cardiopulmonary bypass time; CABG ¼ coronary artery bypass graft; CHF ¼ congestive heart failure; CI ¼ confidence interval; COPD ¼ chronic obstructive pulmonary disease; HR ¼ hazard ratio; HTN ¼ hypertension; ICU ¼ intensive care unit; MI ¼ myocardial infarction; MR ¼ mitral regurgitation; XCT ¼ aortic cross-clamp time.
between operative risk models and prolonged ICU stay after cardiac surgery [12, 16–19]. However, our study does show that a significant number of patients with prolonged ICU stay can eventually be discharged, and this group does enjoy a reasonable long-term survival. Patients with more than 14 days in the ICU attained an overall survival of 50% at 5 years after discharge. Even in the most extreme group with more than 30 days in the ICU, 50% do get discharged; of these, 64% will survive 1 year and 40% will attain 5-year survival. After discharge, the highest rate of mortality is within the first year. Although this is tempered after 1 year, late survival is still poorer as ICU stay is longer. That may be due to greater comorbidity in these patients or possibly because prolonged ICU stay in itself, for whatever reasons, does leave a mark. Indeed, we found ICU stay itself to be an independent predictor of late survival. Despite lack of uniformity in the definition of prolonged ICU stay, all reports show increased operative mortality, ranging between 29% and 53% [1, 2, 5, 6, 9, 10] as well as reduced late survival [1, 3, 5–7] associated with prolonged ICU stay. Our results show the correlation between length of stay in the ICU and overall late survival, and this has not been previously reported.
Quality of Life We do not have an assessment of quality of life for hospital survivors, and that is a limitation of our study.
Some researchers report poor quality of life among patients surviving after prolonged ICU stay [3, 6, 7, 9] whereas others report good quality of life [1, 5, 10, 20]. Hellgren and Stahle [3] emphasize, for a group of valve surgery patients, that the comparison of quality of life should be made in similar patients if for some reason they would not have undergone surgery. The alternative would have been medical therapy, which is associated with poor prognosis for symptomatic valvular patients. In their study [3], although 20% did not experience an improvement, none of the prolonged ICU group regretted the surgery. This finding could be due to the awareness of these patients of the poor prognosis associated with medical treatment. With lack of uniformity in definition of prolonged ICU stay and the somewhat subjective perception of quality of life, it is difficult to apply such conclusions to clinical practice.
Clinical Implications Most reports on the topic of prolonged ICU stay relate to short-term outcomes and resource utilization as well as patient and family expectations once the postoperative course is extended. Williams and associates [1] report 49 patients who stayed more than 14 days in the ICU; these patients comprised only 3.8% of the patient volume but utilized 28% of the total ICU beds. These issues do raise certain ethical dilemmas. With increasing costs, limited resources, and reduced expected survival, while at the same time withholding services for other patients, what is the right decision? Should there be a cutoff point for ICU stay? If so, who is to decide when? Our data do not give a clear answer but do give some understanding of in-hospital outcomes, and in particular, long-term survival. Indeed, a significant number of patients live many years after discharge. Whether prolonged ICU stay is a reflection of extreme attempts in keeping these patients alive or is a measure of patients’ viability is difficult to determine. Furthermore, we cannot determine what the outcomes of such patients might be had they been transferred to a non–intensive care facility at an earlier stage. The challenge is to prospectively predict which patients will not survive despite ICU care. Different models attempt to predict inevitable death in patients after cardiac surgery requiring prolonged ICU stay [1, 12, 21]; however, it would be difficult to implement these models to aid decisions about withdrawal of treatment [22]. For patients with a protracted ICU stay, risk assessment becomes a “moving target” that may vary on a daily, if not hourly, basis. Better prediction of outcomes in real time requires very detailed data collected prospectively. This is beyond the scope of this study but is a topic for further investigation and possible establishment of an "evolving" score. Our study is unique in a number of aspects. The series includes a large number of patients with lengthy followup. Despite poorer results after prolonged stay in the ICU, a significant number of patients do get discharged from hospital. Lastly, whereas other investigators determined a range of cutoff points for lengthy stay, we have
shown a continuous correlation between length of ICU stay after surgery and long-term survival. In conclusion, prolonged stay in the ICU after cardiac surgery is a surrogate for advanced morbidity and predicts reduced survival. This is in correlation with duration of stay. Despite this, a significant number of these patients will survive 10 years or more. Our data shed more light with regard to expectations, both short term and long term. We wish to thank Dr Bernard S. Goldman for his valuable assistance in the preparation of this manuscript.
References 1. Williams MR, Wellner RB, Hartnett EA, et al. Long-term survival and quality of life in cardiac surgical patients with prolonged intensive care unit length of stay. Ann Thorac Surg 2002;73:1472–8. 2. Hassan A, Anderson C, Kypson A, et al. Clinical outcomes in patients with prolonged intensive care unit length of stay after cardiac surgical procedures. Ann Thorac Surg 2012;93: 565–9. 3. Hellgren L, Stahle E. Quality of life after heart valve surgery with prolonged intensive care. Ann Thorac Surg 2005;80: 1693–8. 4. Hein OV, Birnbaum J, Wernecke K, England M, Konertz W, Spies C. Prolonged intensive care unit stay in cardiac surgery: risk factors and long-term survival. Ann Thorac Surg 2006;81:880–5. 5. Lagercranz E, Lindblom D, Sartipy U. Survival and quality of life in cardiac surgery patients with prolonged intensive care. Ann Thorac Surg 2010;89:490–6. 6. Gaudin o M, Girola F, Piscitelli M, et al. Long-term survival and quality of life of patients with prolonged postoperative intensive care unit stay: unmasking an apparent success. J Thorac Cardiovasc Surg 2007;134:465–9. 7. Bapat V, Allen D, Young C, Roxburgh J, Ibrahim M. Survival and quality of life after cardiac surgery complicated by prolonged intensive care. J Cardiac Surg 2005;20:212–7. 8. Heimrath OP, Buth KJ, Legare JF. Long-term outcomes in patients requiring stay of more than 48 hours in the intensive care unit following coronary bypass surgery. J Crit Care 2007;22:153–8.
SILBERMAN ET AL PROLONGED ICU AFTER CARDIAC SURGERY
21
9. Bashour CA, Yared JP, Ryan TA, et al. Long-term survival and functional capacity in cardiac surgery patients after prolonged intensive care. Crit Care Med 2000;28:3847–53. 10. Isgro F, Skuras JA, Keissling AH, Lehmann A. Survival and quality of life after long-term intensive care stay. Thorac Cardiovasc Surg 2002;50:95–9. 11. Ghotkar SV, Grayson AD, Fabri BM, Dihmis WC, Pullan DM. Preoperative calculation of risk for prolonged intensive care unit stay following coronary artery bypass grafting. J Cardiothorac Surgery 2006;1:14. 12. Tu JV, Jaglal SB, Naylor CD. Multicenter validation of a risk index for mortality, intensive care unit stay, and overall hospital length of stay after cardiac surgery. Circulation 1995;91:677–84. 13. Mazzoni M, De Maria R, Bortone F, et al. Long-term outcome of survivors of prolonged intensive care treatment after cardiac surgery. Ann Thorac Surg 2006;82:2080–8. 14. De Cocker J, Messaoudi N, Stockman BA, et al. Preoperative prediction of intensive care unit stay following cardiac surgery. Eur J Cardiothorac Surg 2011;39:60–7. 15. Hein OV, Birnbaum J, Wernecke KD, Konertz W, Spies C. Intensive care unit stay of more than 14 days after cardiac surgery is associated with non-cardiac organ failure. J Int Med Res 2006;34:695–703. 16. Nilsson J, Algotsson L, Hoglund P, Luhrs C, Brandt J. EuroSCORE predicts intensive care unit stay and costs of open heart surgery. Ann Thorac Surg 2004;78:1528–35. 17. Lawrence DR, Valencia O, Smith EEJ, Murday A, Treasure T. Parsonnet score is a good predictor of the duration of intensive care unit stay following cardiac surgery. Heart 2000;83:429–32. 18. Ettema RGA, Peelen LM, Schuurmans MJ, Nierich AP, Kalkman CJ, Moons KGM. Prediction models for prolonged intensive care unit stay after cardiac surgery: systematic review and validation study. Circulation 2010;122:682–9. 19. Van Caenegem O, Jacquet LM, Goenen M. Outcome of cardiac surgery patients with complicated intensive care unit stay. Curr Opin Crit Care 2002;8:404–10. 20. Trouillet JL, Scheimberg A, Vuagnat A, Fagon JY, Chastre J, Gibert C. Long-term outcome and quality of life of patients requiring multidisciplinary intensive care unit admission after cardiac operations. J Thorac Cardiovasc Surg 1996;112: 926–34. 21. Holmes L, Loughead K, Treasure T, Gallivan S. Which patients will not benefit from further intensive care after cardiac surgery? Lancet 1994;344:1200–2. 22. Barnato AE, Angus DC. Value and role of intensive care unit outcome prediction models in end-of-life decision making. Crit Care Clin 2004;20:345–62.
DISCUSSION DR KEVIN W. LOBDELL (Charlotte, NC): You addressed most of the things that I thought about in advance, but maybe you could elaborate. We are all challenged with patients of this nature. There are existing risk models for predicting what happens once a patient is in our unit and they have a set of problems. The difficulty lies in predicting outcomes in a single patient. So, in other words, we might calculate using the CASUS score and determine that the patient has a 95% predicted mortality, but what does one do? DR SILBERMAN: First of all, there are many models predicting operative mortality, but they are preoperative. There are also models predicting risk of long stay in the intensive care unit (ICU). The problem is to predict which patient in the ICU has a chance to survive and leave the hospital. These assessments may change not only on a daily basis but even on an hourly basis. We can count how many blood units patients received, how
many perfusion pumps they require, infection, no infection, and so on, and these will vary even within a single day. If you do a score one day, it will change by the next. So this is a problem. Our policy is just to maintain treatment. For various reasons, be it national, religious, ethical reasons, we just do not withhold treatment. For a patient in extreme condition, with seemingly no chance of survival, common sense may dictate not to start cardiopulmonary resuscitation if the patient arrests. But we will not withhold treatment and we will maintain full support for as long as necessary, and indeed, many of these do go home. Of all patients with more than a month in the ICU, 20% will be alive at 5 years. DR GLENN WHITMAN (Baltimore, MD): It is interesting that you looked at it the way you did, because I think your conclusion, in my opinion, simply creates more dilemmas for us. What so frequently happens in the ICU is that after a couple of weeks
ADULT CARDIAC
Ann Thorac Surg 2013;96:15–22
22 ADULT CARDIAC
SILBERMAN ET AL PROLONGED ICU AFTER CARDIAC SURGERY
Ann Thorac Surg 2013;96:15–22
with multiorgan system failure, we talk to the families about what the chances are of their loved one surviving. This may lead to end-of-life discussions. Your message, however, seems to advocate to continue to be aggressive, because even after 4 weeks, your loved one has a 50% chance of being alive 5 years from now. What would benefit everyone would be to know just which patients do survive, and which do not. Armed with that, we could better advise families and better utilize resources. In my opinion, this paper unrealistically gives us optimism without accounting for the limited resources with which we all will need to deal.
about chances of survival. Have you found that you have more discussions with relatives at time frames since you have seen this or do you still continue doing as you were before?
DR SILBERMAN: We do not tell families that their loved one has a 50% chance of surviving so and so many months or years. Again, it’s a policy to maintain treatment. We are very realistic about the outcomes, and for patients in the ICU for 3 or 4 weeks or beyond, we relate this to the family; we tell them the facts. We just don’t withhold treatment. Moreover, the families don’t have the legal rights to ask to withhold treatment either. I’m not sure of any existing model that can predict which patient will go home and survive a year or 2 or 5. Our data are retrospective, and we did see that a significant number of patients did survive. To create a model for real-time calculation is a bit more difficult, it has to be prospective, and to take into account countless parameters that have to be collected.
DR DAVID A. FULLERTON (Aurora, CO): One of the other variables that could increase the survival after discharge from the hospital is the patient’s age and destination upon discharge, whether they went home or to a skilled nursing facility. Did you have the opportunity to see if any other variables might have affected the survival curve?
DR RICHARD ENGELMAN (Springfield, MA): This is a study that has gone on since 1993 until almost the present, and I would ask, have you looked at a time frame, because perhaps technology has changed over that 8- or 9-year period, and has any change in technology done anything to the data? DR SILBERMAN: The question is a good one. No, we didn’t break it down to time frames. DR ADRIAN JEREMY LEVINE (Stoke-on-Trent, UK): We did a bit of work like this seven, eight years ago. Our findings were not quite as optimistic as yours. What it did do in our group was it enabled us to discuss with relatives far more meaningfully
DR SILBERMAN: Again, we are not blinded by these results. We are still living in the realistic world and we are aware of outcomes. We know that the patient with a long stay in the ICU, for various reasons, has a smaller chance of survival, and we relate this to the families. We discuss with families from day one, and we keep them informed on a daily basis. If the situation looks bad, we tell them; we don’t raise any false hopes.
DR SILBERMAN: Well, age did emerge as a predictor, of course. We are talking about patients who were discharged. Some patients may go through a care facility such as a nursing home, but the majority eventually do go home. We do not have data regarding patients who were chronically in a home and died over there. Information from the Ministry of Interior tells us who died and when, and who is still alive, but we don’t have all the information as to whether they were hospitalized since their discharge or not. DR LOBDELL: Just one comment. I think that there is rich opportunity to understand this better than we do, as evidenced by the discussion. It has always been interesting to me that the STS database risk model for mortality takes into account preoperative and intraoperative variables, but nothing postoperatively. You could take great care of patients and my care may be poor, but the risk model would predict that our outcomes should be the same. So I think that there is a body of information within the database that we could use to understand this better than we ever have.