| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2005;79:1976-1986
© 2005 The Society of Thoracic Surgeons
a Division of Cardiovascular Surgery, St. Vincent Mercy Medical Center, Toledo, Ohio
b Division of Cardiovascular Surgery, St. Luke’s Hospital, Maumee, Ohio
c Department of Medicine, Medical College of Ohio, Toledo, Ohio
d Department of Surgery, Medical College of Ohio, Toledo, Ohio
Accepted for publication November 17, 2004.
* Address reprint requests to Dr Habib, Cardiopulmonary Research, St. Vincent Mercy Medical Center, 2213 Cherry St, ACC Bldg, Suite 309, Toledo, OH43608 (E-mail: robert_habib{at}mhsnr.org).
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentReferences |
|---|
METHODS: Primary isolated CABG patients were grouped based on body surface area (BSA, m2) and body mass index (BMI, kg/m2) as follows: 611 very small (BSA 
1.70); 933 slightly small (1.70 < BSA 1.85); 945 moderately obese (32 < BMI <36); 594 very obese (BMI 
36); and 3,018 normal (BSA >1.85; BMI = 22 to 32). Subcohorts of very small (371 pairs, 61%), slightly small (717, 77%), moderately obese (874, 92%), and very obese (516, 87%) patients were propensity-matched to normal.
RESULTS: Compared with normal, very small had more transfusions (46% versus 32%; p < 0.001), reoperation for bleeding (3.2% versus 0.3%; p = 0.002), and pulmonary edema (2.4% versus 0.5%; p = 0.033). For slightly small, transfusion (41% versus 29%; p < 0.001) and bleeding (2.5% versus 1.0%; p = 0.04) were increased. For moderately obese, sternal wound infections (1.9% versus 0.8%; p = 0.04) were greater. Complications were most frequent in very obese: reoperation (5.2% versus 1.6%; p < 0.001), sternal wound infections (3.5% versus 0.2%; p < 0.001), pulmonary edema (2.9% versus 1.2%; p = 0.047), renal failure (6.0% versus 2.3%; p = 0.003), atrial fibrillation (20% versus 12%; p = 0.001), gastrointestinal problems (3.7% versus 1.6%; p = 0.032), and postoperative stay (8.0 versus 6.4 days; p = 0.003). When slightly small and very small are considered together, operative mortality was significantly greater (3.22% versus 1.65%; p = 0.026). Both very small (risk ratio [RR] = 1.39; p = 0.044) and very obese (RR = 1.44; p = 0.020) were independent predictors of worse 0- to 12-year mortality.
CONCLUSIONS: Large deviations from normal body size in either direction—particularly extreme obesity—are associated with increased postoperative morbidity and worse long-term survival.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
Our current understanding of the role of body size on short-term CABG results remains incomplete, while that of long-term outcomes is practically absent. Nearly all studies have been focused on the immediate postoperative period [3, 4, 7–11], and the data suggest a greater propensity for complications, resource utilization, and mortality at body size extremes relative to intermediate size (normal) [3, 12]. Yet, the evidence for an independent association between size and adverse outcomes is inconclusive, and the actual magnitudes of these effects are not well defined.
Body size is a given characteristic that precludes prospective trials to address body size and outcome questions. Also, patients at extremes of body habitus are characterized by substantial risk factor differences compared with normal. This means that conventional multivariate analyses in unmatched cohorts are susceptible to residual confounding. We thus utilized nonparsimonious propensity modeling to obtain matched small (and obese) patients to normal counterparts characterized by similar demographics (other than size), risk factors, and operative variables [13]. Next, we compared operative and 12-year outcomes in matched groups.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
1.70); (2) 933 slightly small (1.70 < BSA 1.85); (3) 945 moderately obese (32 < BMI <36); (4) 594 very obese (BMI 36); and (5) normal (BSA >1.85 and BMI = 22 to 32). A small number of short stature patients (n = 66 [1.1%]; height: 1.27 to 1.55 m) were included both as small (1.45 to 1.85 m2) and obese (32 to 42 kg/m2). Cardiopulmonary bypass (CPB) using standard techniques was utilized in approximately 95% of patients with predominant normothermia (98%) [3, 5]. Demographics, risk factors, cardiac status, medications, coronary disease and grafting, and operative and postoperative data were prospectively collected in the cardiac surgery database according to The Society of Thoracic Surgeons (STS) definitions. Postoperative data included all complications, resource utilization, operative death, and hospital readmission. Long-term survival data were secured from our service follow-up and verified from individual queries of the United States Social Security Death Index database (conducted in February 2004). Allowing for a 3-month lag, this corresponds to a follow-up of 5 to 149 months.
Propensity Matching
Two nonparsimonious propensity models were used to match body habitus extremes patients to normal: 1) small-model derived from all small plus normal patients, and 2) obese-model, from all obese plus normal. Excluding all size variables, 64 patient variables (demographics, coronary disease and grafting, medications and operative) were entered into logistic models irrespective of their significance level. The score derived for each patient from the small model and obese model represented, respectively, the probability that the patient was either small or obese. A custom-made computer algorithm was next used to obtain one-to-one (or greedy) propensity score matching. To maximize the number of matched patients, individual patients belonging to each of the study groups (very small, slightly small, moderately obese, and very obese) were separately matched to the closest possible normal score (matches) to within ±2% score difference (more than 95% were <1%).
Statistical Methods
Continuous data were expressed as mean ± SD. Univariate comparisons were done with the 
2 test or Fisher’s exact test for categorical variables and the unpaired t test or the nonparametric Mann-Whitney rank sum test for continuous variables. Kaplan-Meier 12-year survival data were compared by the log-rank test. These were modeled using Blackstone’s multiphase (early, constant, and late) time-varying hazard model, and relative death hazard functions were derived for the matched comparisons [14]. Effects of explanatory variables on survival were defined by risk ratios (RR) derived by Cox regression analysis. Model selection was done with backward elimination (Wald statistic, confirmed using forward and stepwise selection), and variables with p less than 0.05 were retained (SPSS version 10.0; SPSS, Chicago, IL).
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
|
|
Patient data for all four matched group pairs are summarized in Table 1. Compared with normal, slightly small and moderately obese study patients had no significant differences in any of these variables. Few minor differences persisted after matching for very small and very obese versus their respective normal matches: for very small, insulin dependence (11.6% versus 17.0%; p = 0.036) was less frequent whereas all-arterial grafting (11.9% versus 7.3%; p = 0.034) was more frequent; and for very obese, preoperative aspirin use was greater (68% versus 61%; p = 0.016) whereas right internal thoracic artery use (0.8% versus 3.3%; p = 0.004) and bilateral internal thoracic artery use (0.8% versus 2.7%; p = 0.017) were less.
|
|
Long-Term Survival
Unadjusted Kaplan-Meier survival comparisons are shown in Figure 3, and only very obese were worse than normal (p = 0.04). The corresponding relative hazard ratios are plotted in Figure 4. These indicated that slightly small and very small were associated with systematically worse early-phase death hazard up to about 4 to 5 months after CABG. The very small death hazard was also systematically greater between 6 and 12 years after CABG. No substantial death hazard differences were observed in case of moderately obese. Lastly, very obese did not systematically differ from normal during year 1, but were associated with noticeably worse intermediate-term death hazard (1 to 6 years). After year 8, very obese showed a relatively lower death hazard.
|
|
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
These divergent results on how patient size affects CABG operative outcomes coupled with a scarcity of data on its potential long-term effects motivated this investigation. This is increasingly pertinent given the ongoing demographic changes in adults including an obesity epidemic that affects both the frequency and outcomes of CABG [1, 2]. Mokdad and coworkers [2] used BMI of 30 kg/m2 or more to define obesity and reported a 20.9% national incidence in 2001, representing a 69% increase compared with 1991 (12.0%). In this 1991 to 2003 CABG series, 38% of patients had a BMI of 30 kg/m2 or greater, which demonstrates the important role of obesity in coronary disease. Moreover, significant obesity (BMI >32 kg/m2) increased substantially over this period from 21% to 28% (Fig 1C), and this change comes almost entirely from a rise in the very obese (6% to 13%) while normal patients were less. Our goal was to elucidate the independent effects of body size in CABG free of covariate confounding. To accomplish this, we matched patients by one-to-one (no group size bias), greedy (no repeated patients bias) propensity modeling of the four extreme size groups (see Methods) [13]. The derived propensity models had good-to-excellent discriminative power, and we were able to match a large majority of extreme body size patients to normal size patients.
Postoperative CABG Outcome and Body Size
We report that slightly small and moderately obese patients are associated with a limited increase in morbidity, whereas morbidity was noticeably worse for the very small and very obese cohorts. Compared with normal (Table 2), transfusion and bleeding problems requiring reexploration were more common for both small groups. This is similar to previous reports, and can be linked to greater hemodilution during CPB in small versus larger patients [3, 5, 6]. Higher incidence of pulmonary edema and pneumonia in very small patients may also be related to CPB hemodilutional anemia, which can lead to greater fluid retention including extravascular lung water [3]. On-pump hemodilution is usually greater in small patients since the asanguinous circuit prime represents a greater fraction of the circulating blood-prime mixture. This, in turn, increases the oncotic gradient favoring fluid extravasation particularly when CPB inflammatory endothelial injury is induced [24, 25].
Like prior studies [26, 27], sternal wound infections were more prevalent with increasing obesity probably due to poor healing of under-perfused adipose chest wall tissues. This was the only increased complication for moderately obese (Table 2). For very obese, noncardiac reoperations, pulmonary edema/pneumonia, acute renal failure, atrial fibrillation, and gastrointestinal complications were all increased. Also, consequently, postoperative hospital stays in very obese were longer (8.0 versus 6.4 days; p = 0.003).
None of the individual group comparisons showed significant operative mortality differences. But operative mortality in very small (3.23% versus 1.62%; p = 0.15) and slightly small (3.21% versus 1.67%; p = 0.06) were both twice that of normal, and the lack of significance is likely due to the low operative mortality rates and limited number of patients available. When slightly small (n = 717) and very small (n = 371) are combined, their operative mortality is significantly worse than in normal (18 of 1,088 [1.65%] versus 35 of 1,088 [3.22%]; p = 0.026). That was not true when we combined both obese groups. Similar findings of worse operative mortality in small patients but not large BSA was previously reported in unmatched groups [3, 6, 21].
Long-Term CABG Outcome and Body Size
Surprisingly little is known about the potential impact of body size at the time of CABG on intermediate to longer term survival [3, 28]. Our analysis showed that 12-year CABG survival was significantly worse for very small and very obese (both independent predictors; Fig 5). For very small, the increased death hazard was evident early during year 1 and late after year 8. In contrast, excluding early deaths increases the very obese adjusted RR from 1.44 to 1.65 (Fig 5). Moreover, very obese showed particularly greater death hazard between 1 and 6 years.
It is not obvious why very small and very obese do worse than normal in the long-term even after multivariate adjustment. Follow-up data after discharge were limited to patient death only. Yet, one can speculate about possible reasons for these results. First, the two extreme size groups are associated with a greater propensity of post-CABG complications affecting a number of vital organs (Table 2). Such injury (eg, renal failure) may have sustained long-term effects on organ function and hence on survival. Second, small patients—particularly very small—are susceptible to increased CPB hemodilutional anemia, and subsequently require more transfusions. Increased hemodilutional anemia is associated with vital organ injury with possible long-term effects [5, 6]. Both excessive hemodilution [5] and transfusion [16] are linked to worse long-term mortality. Fortunately, these risks in small patients are potentially modifiable by changes in CPB practice so that excessive hemodilution and transfusions are minimized, for example, use of smaller bypass circuits, retrograde autologous priming, or off-pump surgery. Third, compared with comorbidity-matched normal patients, very small and very obese may be associated with an increased likelihood for developing new comorbidities or exacerbation of already existing comorbidities. For example, very obese are probably more likely to become diabetic or develop hypertension at a given interval after CABG.
We did not find appreciable differences in survival and death hazard for the slightly small and moderately obese groups when compared with matched normal patients. This is in contrast to two recent population studies that indicated that even mild obesity is associated with increased years of life lost [29, 30]. Indeed, Peeters and coworkers [29] showed that the free of cardiovascular disease Framingham Study subpopulation exhibited a significant loss in life span.
Long-term CABG outcome may depend significantly on the utilized grafting strategy, and particularly use of arterial as opposed to vein conduits [31–34]. Grafting data can differ among different size cohorts. However, the frequency of vein, internal thoracic artery and radial artery conduits use in very small, slightly small, and moderately obese study groups was similar to that of their corresponding normal matches (Table 1). Only the very obese versus normal comparison included small but significant differences in right and bilateral internal thoracic artery use. To confirm that this did not bias our findings, we repeated the Cox regression analysis where the small subgroup of patients with right or bilateral internal thoracic artery use and their matches were removed. Here, too, very obese was an independent predictor with a similar risk ratio (1.68; p = 0.006). Worse out-of-hospital CABG outcomes among women—up to 1 year after surgery—have been reported [35,36] and is currently a topic of significant debate. Our data showed that women are disproportionately represented in the extreme size CABG cohorts, which may have played an important role in these differences. Guru and colleagues [36] compared early and late CABG outcomes in men versus women for the entire 1991 to 2000 Ontario, Canada, experience. Their analysis of 54,425 CABG cases, without accounting for body size, resulted in significantly greater adjusted early mortality in women up to 1 year after CABG. Yet, their BSA-adjusted secondary analysis in 13,921 patients suggested no sex differences in early or late mortality (RR = 1.04; p = 0.72). Koch and associates [37] convincingly demonstrated that female sex per se is not a cause of worse operative mortality. Their analysis of the Cleveland Clinic CABG experience revealed that early CABG mortality is essentially identical in males versus females (2.3% versus 2.1%; n = 945 each; p = 0.76) when patients are rigorously matched for demographics including body habitus, risk factors, medications, coronary disease, and cardiopulmonary bypass [37]. Our study is not specifically designed to address the issue of sex. However, the results from the slightly small, moderately obese and very obese versus normal multivariate analyses (Tables 3 and 4) showed that women were in fact associated with less rather than worse long-term mortality (adjusted RR = 0.63 to 0.77). The above indicates that the sex and size effects on CABG outcomes are intricately related and should be addressed by appropriate study designs in the future.
Limitations
Our study was done on a retrospective series from a single institution. Reproducibility of our findings should be confirmed in other patient series. Validity of retrospective analyses to future series is often a topic of debate. However, as size cannot be randomly assigned, the role of body size in CABG cannot be addressed in prospective randomized fashion. Also, there is evidence that when conducted in sufficiently large populations and with the appropriate multivariate analyses, results of retrospective series closely predict results of prospective randomized trials [38–40].
Like most retrospective studies, the long-term survival data are based on all-cause mortality. Autopsy data are rarely available, and the accuracy of cause of death data listed on death certificates is often questionable. Also, since our results were based on matched group comparisons, we expect that significant between-group differences in noncardiac related deaths to be unlikely. Our analysis did not account for the possible effects of changes in patient body habitus after CABG due to lifestyle and diet modifications. These are likely to occur with increased patient awareness in a manner that reduces health risks. For example, normal patients are less likely to become obese and obese patients may be encouraged to lower their BMI toward normal. The latter may have played a role in the lack of worse long-term outcomes in moderately obese.
Lastly, the potential effects of incomplete group matching cannot be totally discounted. In fact, we know that a few variable differences remained after matching in case of very obese and very small comparisons—which are the groups with worse outcomes. Insulin dependence was less frequent (11.6% versus 17.0%) and all-arterial grafting more frequent (11.9% versus 7.3%) in very small relative to normal. Insulin dependence was a powerful predictor of worse outcome in all paired group comparisons (Tables 3 and 4). More arterial grafting is associated with better outcomes [14, 31–33]. Thus, if anything, these residual differences are expected to reduce rather than exaggerate differences between very small and normal. In case of very obese versus normal, small residual differences in right (0.8% versus 3.3%) and bilateral (0.8% versus 2.7%) internal thoracic artery use did not bias our findings as discussed above. Left internal thoracic artery use is an established standard of care [33, 34], and we used it in all patients unless of poor quality or owing to the specific diseased coronary anatomy. Alternatively, bilateral internal thoracic artery dissection is avoided in morbidly obese patients to avoid compromising the blood supply to the sternal tissues given their propensity for sternal wound infections [26, 27]. Finally, it is unlikely that greater preoperative aspirin use in very obese (68.2% versus 61%; p = 0.016) could have affected our findings, particularly since their outcome was similar to normal up to at least 1 year post-CABG and since aspirin is equally prescribed postoperatively.
Conclusion
Even after rigorous matching of body size cohorts to normal size counterparts, we found that large deviations from normal (very small and very obese) are independently associated with increased postoperative morbidity and worse long-term survival. Small patients also exhibited greater operative death, which is perhaps linked to increased on-pump hemodilution, transfusions, and associated complications [5, 6, 15, 16]. We suggest that these effects in small patients are potentially modifiable by changes in current CPB practice. Current CABG demographic trends for our practice show that very obese is the only growing CABG subpopulation and this occurs at the expense of the normal cohort. The very obese cohort was also associated—even after accounting for other comorbidities—with the most perioperative morbidity and the highest relative death hazard after the first year (RR = 1.65). Since prevalence of comorbid factors is itself increased by morbid obesity [1], then the risks we report for very obese status may underestimate its true effects. These data in CABG outcomes are consistent with the effects of obesity in the general adult population [29, 30] and underscore the importance of reversing the well-documented obesity epidemic from a coronary artery disease perspective [2, 23].
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  L. K. Ti, B.-L. Goh, P.-S. Wong, P. Ong, S.-G. Goh, and C.-N. Lee Comparison of Mini-Cardiopulmonary Bypass System With Air-Purge Device to Conventional Bypass System Ann. Thorac. Surg., March 1, 2008; 85(3): 994 - 1000. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Ranucci, A. Pazzaglia, C. Bianchini, G. Bozzetti, and G. Isgro Body Size, Gender, and Transfusions as Determinants of Outcome After Coronary Operations Ann. Thorac. Surg., February 1, 2008; 85(2): 481 - 486. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. A. Albert, N. Halevy, and E. M. Antman Preoperative Evaluation for Cardiac Surgery Card. Surg. Adult, January 1, 2008; 3(2008): 261 - 280. [Full Text]
|
|
|
|
 |
|
 P. C. Austin Propensity-score matching in the cardiovascular surgery literature from 2004 to 2006: a systematic review and suggestions for improvement. J. Thorac. Cardiovasc. Surg., November 1, 2007; 134(5): 1128 - 1135. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. H. Habib and A. Zacharias Body Size and the Early Mortality Gender Gap in Coronary Artery Bypass Grafting Surgery J. Am. Coll. Cardiol., September 11, 2007; 50(11): 1095 - 1095. [Full Text] [PDF]
|
|
|
|
 |
|
 J. D. Puskas, P. D. Kilgo, M. Kutner, S. V. Pusca, O. Lattouf, and R. A. Guyton Off-Pump Techniques Disproportionately Benefit Women and Narrow the Gender Disparity in Outcomes After Coronary Artery Bypass Surgery Circulation, September 11, 2007; 116(11_suppl): I-192 - I-199. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Filardo, C. Hamilton, B. Hamman, and P. Grayburn Obesity and Stroke After Cardiac Surgery: The Impact of Grouping Body Mass Index Ann. Thorac. Surg., September 1, 2007; 84(3): 720 - 722. [Full Text] [PDF]
|
|
|
|
|
|
G. H. Tyson III, E. Rodriguez, O. C. Elci, T. C. Koutlas, W. R. Chitwood Jr, T. B. Ferguson, and A. P. Kypson Cardiac Procedures in Patients With a Body Mass Index Exceeding 45: Outcomes and Long-Term Results Ann. Thorac. Surg., July 1, 2007; 84(1): 3 - 9. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. D. Wagner, G. K. Grunwald, J. S. Rumsfeld, J. O. Hill, P. M. Ho, H. R. Wyatt, and A. L. W. Shroyer Relationship of Body Mass Index With Outcomes After Coronary Artery Bypass Graft Surgery Ann. Thorac. Surg., July 1, 2007; 84(1): 10 - 16. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Zacharias, T. A. Schwann, C. J. Riordan, S. J. Durham, A. S. Shah, and R. H. Habib Obesity and Risk of New-Onset Atrial Fibrillation After Cardiac Surgery Circulation, November 22, 2005; 112(21): 3247 - 3255. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
