Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

A gender-based analysis of the obesity paradox in cardiac surgery: height for women, weight for men?

A gender-based analysis of the obesity paradox in cardiac surgery: height for women, weight for men? Abstract View largeDownload slide View largeDownload slide OBJECTIVES In cardiac surgery, obesity is associated with a lower mortality risk. This study aims to investigate the association between body mass index (BMI) and operative mortality separately in female patients and male patients undergoing cardiac surgery and to separate the effects of weight and height in each gender-based cohort of patients. METHODS A retrospective cohort study including 7939 consecutive patients who underwent cardiac surgery was conducted. The outcome measure was the operative mortality. RESULTS In men, there was a U-shaped relationship between the BMI and the operative mortality, with the lower mortality rate at a BMI of 35 kg/m2. In women, the relationship is J-shaped, with the lower mortality at a BMI of 22 kg/m2. Female patients with obesity class II–III had a relative risk for operative mortality of 2.6 [95% confidence interval (CI) 1.37–4.81, P = 0.002]. The relationship between weight and mortality rate is a U-shaped bot in men and women, with the lower mortality rate at 100 kg for men and 70 kg for women. Height was linearly and inversely associated with the operative mortality in men and women. After correction for the potential confounders, height, but not weight, was independently associated with operative mortality in women (odds ratio 0.949, 95% CI 0.915–0.983; P = 0.004); conversely, in men, this association exists for weight (odds ratio 1.017, 95% CI 1.001–1.032; P = 0.034), but not height. CONCLUSIONS Contrary to men, in women obesity does not reduce the operative mortality in cardiac surgery, whereas the height seems to be associated with a lower mortality. Adult, Cardiac surgical procedures, Gender, Obesity, Mortality INTRODUCTION A recent large retrospective cohort study [1] found that obesity is associated with lower morbidity and mortality after cardiac surgery, confirming previous relatively smaller studies where this association was identified [2–4]. These results concur in the identification of the so-called ‘obesity paradox’, where an important risk factor for cardiovascular disease and cardiovascular death [5] is conversely a protective factor in patients with acute coronary syndromes [6], heart failure [7] and those undergoing cardiac surgery. Many possible explanations for this paradox in the cardiac surgery population have been raised, including collider bias (where the association in favour of the obese patients actually reflects a higher mortality in frail, cachectic, underweight patients) or selection bias (obese patients are admitted to cardiac surgery only if without other risk factors) or an indirect age effect (obese patients are usually younger) or finally a lower degree of haemodilution and less need for transfusions in patients with large body mass index (BMI) [8, 9]. In 2008, a study from our group [4] raised the hypothesis that BMI and body surface area might affect the outcome of cardiac surgery differently in women and men. Additionally, previous studies addressing the association between obesity and outcomes in cardiac surgery used the BMI to stratify the patients into categories, in accordance with the World Health Organization [10]. However, the BMI is calculated from weight and height, and no previous studies in this field tried to separate the effects of these 2 variables. The present single-centre, retrospective, cohort study aims to investigate the association between BMI and operative mortality separately in female patients and male patients undergoing cardiac surgery and to separate the effects of weight and height in each gender-based cohort of patients. METHODS This study is based on a retrospective, single-centre cohort study of patients undergoing cardiac surgery in the period January 2010 to November 2017. All data were retrieved from the institutional database of the IRCCS Policlinico San Donato. The study design was approved by the local ethics committee (IRCCS San Raffaele Hospital) and the need for a written informed consent was waived, providing adequate anonymization of all the data utilized. All the patients gave a written consent for the scientific treatment of their data in an anonymous form at the admission in the Hospital. Patients During the period considered, 11 374 patients underwent heart surgery at our institution. Patients aged <18 years and adult patients undergoing congenital heart surgery were excluded from the analysis, reaching a final patient population of 7939 adult patients undergoing cardiac surgery. Data collection and definitions Data collection is derived from the institutional cardiac surgery database, which includes demographics, comorbidities, heart function and laboratory data, conditions at surgery, type of surgery, intraoperative details and outcome data. The BMI was calculated as the weight in kilograms divided by the square of the height in metres [11]. It was further categorized into 5 classes according to the World Health Organization classification [10]: underweight (BMI <18.5 kg/m2), normal weight (BMI 18.5 to <25 kg/m2), overweight (BMI 25 to <30 kg/m2), obese class I (BMI 30 to <35 kg/m2) and obese class II–III (BMI ≥35 kg/m2). Obesity classes II and III were pooled together because of the poor number of cases in obesity class III. Operative mortality was defined as mortality occurring during hospital stay or within 30 days from surgery after discharge. Statistical analysis Data are presented as number (%) for categorical data and median (interquartile range) for continuous data. The association between BMI classes and operative mortality was tested using the Pearson’s χ2 analysis, and interclass differences and cubic spline functions in figures were calculated using post hoc analysis. The associations between BMI, weight and height (continuous variables) and the operative mortality were assessed separately for women and men with regression analyses (linear and quadratic) choosing the equation with the best fit. When the best fit was found for a quadratic equation, the value of BMI and weight associated with the lowest mortality risk was identified at the nadir of the parabolic relationship. The potential sources of bias were addressed as follows: the independent role of BMI, weight and height as determinants of the operative mortality was tested with logistic regression analyses, after correction for the age, the EuroSCORE II (for operative risk stratification) and the cardiopulmonary bypass (CPB) duration (as an index of intraoperative complexity). For values being linearly associated with the operative mortality, the absolute value was included as an independent variable; for values showing a quadratic relationship with the operative mortality, the negative or positive offset from the value associated with the lowest mortality rate was considered and included as an absolute value in the logistic regression analysis. Odds ratios and 95% confidence intervals (CIs) were derived from the logistic regression analyses. Multivariable logistic regression models provided an estimate of operative mortality with a confounder-adjusted marginal standardization. Additional analyses included a propensity matching of the population based on 3 BMI-based categories. Propensity matching was based on the EuroSCORE II, the age and the CPB duration. All tests were 2-sided; a P-value of <0.05 was considered significant. All the statistical analyses were performed with computerized packages (SPSS 20.0, IBM, Chicago, IL, USA and GraphPad Prism 5.0, GraphPad, San Diego, CA, USA). RESULTS The general characteristics of the patient population are summarized in Table 1. Overall, there were 2658 (33.5%) women and 5281 (66.5%) men. One hundred sixty-five (2.1%) patients were classified as underweight, 3302 (41.6%) as normal weight, 3183 (40.1%) as overweight, 1014 (12.8%) as obese class I and 275 (3.5%) as obese class II–III. The distribution of weight and height in women and men is shown in the Supplementary Material, Figs S1 and S2. Table 1: General characteristics of the patient population Variables Overall (N = 7939) Women (N = 2685) Men (N = 5281) Age (years) 69 (60–76) 72 (63–78) 68 (59–75) Weight (kg) 73 (64–82) 64 (56–72) 77 (70–85) Height (cm) 169 (160–174) 160 (156–165) 170 (168–176) Body mass index (kg/m2) 25.7 (23.4–28.4) 24.9 (22–28.1) 26 (23.9–28.6) BMI class  Underweight 165 (2.1) 110 (4.1) 55 (1.0)  Normal weight 3302 (41.6) 1279 (48.1) 2023 (38.3)  Overweight 3183 (40.1) 843 (31.7) 2340 (44.3)  Obesity class I 1014 (12.8) 298 (11.2) 716 (13.6)  Obesity class II–III 275 (3.5) 128 (4.8) 147 (2.8) Hypertension 5080 (64) 1695 (63) 3385 (64) Left ventricular ejection fraction (%) 55 (49–60) 56 (50–62) 55 (46–60) Congestive heart failure 646 (8.1) 265 (10) 381 (7.2) Diabetes on medication 1459 (18.4) 411 (15.5) 1048 (19.8) Serum creatinine (mg/dl) 1.0 (0.8–1.2) 0.85 (0.7–1.1) 1.0 (0.9–1.2) Haematocrit (%) 39 (35.7–41.9) 37.4 (34.1–40) 40 (36.6–42.6) COPD 581 (7.3) 164 (6.2) 417 (7.9) Previous CVA 250 (3.1) 73 (2.7) 177 (3.4) Conditions of surgery  Elective 6674 (84.1) 2290 (86.2) 4384 (83)  Urgent 1120 (14.1) 319 (12) 809 (15.2)  Emergency 145 (1.8) 49 (1.8) 96 (1.8) Redo surgery 605 (7.6) 245 (9.2) 360 (6.8) EuroSCORE II 2.2 (1.2–4.7) 2.5 (1.4–5.6) 2 (1.2–4.4) Type of surgery  Isolated CABG 2446 (30.8) 419 (15.8) 2027 (38.4)  Isolated valve 2447 (30.8) 1040 (39.1) 1407 (26.6)  CABG + valve 1282 (16.1) 407 (15.3) 875 (16.6)  Ascending aorta 606 (7.6) 214 (8.1) 392 (7.4)  Others 1158 (14.6) 605 (22.5) 580 (11) CPB duration (min) 76 (58–104) 78 (59–104) 76 (58–104) Variables Overall (N = 7939) Women (N = 2685) Men (N = 5281) Age (years) 69 (60–76) 72 (63–78) 68 (59–75) Weight (kg) 73 (64–82) 64 (56–72) 77 (70–85) Height (cm) 169 (160–174) 160 (156–165) 170 (168–176) Body mass index (kg/m2) 25.7 (23.4–28.4) 24.9 (22–28.1) 26 (23.9–28.6) BMI class  Underweight 165 (2.1) 110 (4.1) 55 (1.0)  Normal weight 3302 (41.6) 1279 (48.1) 2023 (38.3)  Overweight 3183 (40.1) 843 (31.7) 2340 (44.3)  Obesity class I 1014 (12.8) 298 (11.2) 716 (13.6)  Obesity class II–III 275 (3.5) 128 (4.8) 147 (2.8) Hypertension 5080 (64) 1695 (63) 3385 (64) Left ventricular ejection fraction (%) 55 (49–60) 56 (50–62) 55 (46–60) Congestive heart failure 646 (8.1) 265 (10) 381 (7.2) Diabetes on medication 1459 (18.4) 411 (15.5) 1048 (19.8) Serum creatinine (mg/dl) 1.0 (0.8–1.2) 0.85 (0.7–1.1) 1.0 (0.9–1.2) Haematocrit (%) 39 (35.7–41.9) 37.4 (34.1–40) 40 (36.6–42.6) COPD 581 (7.3) 164 (6.2) 417 (7.9) Previous CVA 250 (3.1) 73 (2.7) 177 (3.4) Conditions of surgery  Elective 6674 (84.1) 2290 (86.2) 4384 (83)  Urgent 1120 (14.1) 319 (12) 809 (15.2)  Emergency 145 (1.8) 49 (1.8) 96 (1.8) Redo surgery 605 (7.6) 245 (9.2) 360 (6.8) EuroSCORE II 2.2 (1.2–4.7) 2.5 (1.4–5.6) 2 (1.2–4.4) Type of surgery  Isolated CABG 2446 (30.8) 419 (15.8) 2027 (38.4)  Isolated valve 2447 (30.8) 1040 (39.1) 1407 (26.6)  CABG + valve 1282 (16.1) 407 (15.3) 875 (16.6)  Ascending aorta 606 (7.6) 214 (8.1) 392 (7.4)  Others 1158 (14.6) 605 (22.5) 580 (11) CPB duration (min) 76 (58–104) 78 (59–104) 76 (58–104) Data are presented as n (%) or median (interquartile range). BMI: body mass index; CABG: coronary artery bypass graft; COPD: chronic obstructive pulmonary disease; CPB: cardiopulmonary bypass; CVA: cerebrovascular accident. View Large Table 1: General characteristics of the patient population Variables Overall (N = 7939) Women (N = 2685) Men (N = 5281) Age (years) 69 (60–76) 72 (63–78) 68 (59–75) Weight (kg) 73 (64–82) 64 (56–72) 77 (70–85) Height (cm) 169 (160–174) 160 (156–165) 170 (168–176) Body mass index (kg/m2) 25.7 (23.4–28.4) 24.9 (22–28.1) 26 (23.9–28.6) BMI class  Underweight 165 (2.1) 110 (4.1) 55 (1.0)  Normal weight 3302 (41.6) 1279 (48.1) 2023 (38.3)  Overweight 3183 (40.1) 843 (31.7) 2340 (44.3)  Obesity class I 1014 (12.8) 298 (11.2) 716 (13.6)  Obesity class II–III 275 (3.5) 128 (4.8) 147 (2.8) Hypertension 5080 (64) 1695 (63) 3385 (64) Left ventricular ejection fraction (%) 55 (49–60) 56 (50–62) 55 (46–60) Congestive heart failure 646 (8.1) 265 (10) 381 (7.2) Diabetes on medication 1459 (18.4) 411 (15.5) 1048 (19.8) Serum creatinine (mg/dl) 1.0 (0.8–1.2) 0.85 (0.7–1.1) 1.0 (0.9–1.2) Haematocrit (%) 39 (35.7–41.9) 37.4 (34.1–40) 40 (36.6–42.6) COPD 581 (7.3) 164 (6.2) 417 (7.9) Previous CVA 250 (3.1) 73 (2.7) 177 (3.4) Conditions of surgery  Elective 6674 (84.1) 2290 (86.2) 4384 (83)  Urgent 1120 (14.1) 319 (12) 809 (15.2)  Emergency 145 (1.8) 49 (1.8) 96 (1.8) Redo surgery 605 (7.6) 245 (9.2) 360 (6.8) EuroSCORE II 2.2 (1.2–4.7) 2.5 (1.4–5.6) 2 (1.2–4.4) Type of surgery  Isolated CABG 2446 (30.8) 419 (15.8) 2027 (38.4)  Isolated valve 2447 (30.8) 1040 (39.1) 1407 (26.6)  CABG + valve 1282 (16.1) 407 (15.3) 875 (16.6)  Ascending aorta 606 (7.6) 214 (8.1) 392 (7.4)  Others 1158 (14.6) 605 (22.5) 580 (11) CPB duration (min) 76 (58–104) 78 (59–104) 76 (58–104) Variables Overall (N = 7939) Women (N = 2685) Men (N = 5281) Age (years) 69 (60–76) 72 (63–78) 68 (59–75) Weight (kg) 73 (64–82) 64 (56–72) 77 (70–85) Height (cm) 169 (160–174) 160 (156–165) 170 (168–176) Body mass index (kg/m2) 25.7 (23.4–28.4) 24.9 (22–28.1) 26 (23.9–28.6) BMI class  Underweight 165 (2.1) 110 (4.1) 55 (1.0)  Normal weight 3302 (41.6) 1279 (48.1) 2023 (38.3)  Overweight 3183 (40.1) 843 (31.7) 2340 (44.3)  Obesity class I 1014 (12.8) 298 (11.2) 716 (13.6)  Obesity class II–III 275 (3.5) 128 (4.8) 147 (2.8) Hypertension 5080 (64) 1695 (63) 3385 (64) Left ventricular ejection fraction (%) 55 (49–60) 56 (50–62) 55 (46–60) Congestive heart failure 646 (8.1) 265 (10) 381 (7.2) Diabetes on medication 1459 (18.4) 411 (15.5) 1048 (19.8) Serum creatinine (mg/dl) 1.0 (0.8–1.2) 0.85 (0.7–1.1) 1.0 (0.9–1.2) Haematocrit (%) 39 (35.7–41.9) 37.4 (34.1–40) 40 (36.6–42.6) COPD 581 (7.3) 164 (6.2) 417 (7.9) Previous CVA 250 (3.1) 73 (2.7) 177 (3.4) Conditions of surgery  Elective 6674 (84.1) 2290 (86.2) 4384 (83)  Urgent 1120 (14.1) 319 (12) 809 (15.2)  Emergency 145 (1.8) 49 (1.8) 96 (1.8) Redo surgery 605 (7.6) 245 (9.2) 360 (6.8) EuroSCORE II 2.2 (1.2–4.7) 2.5 (1.4–5.6) 2 (1.2–4.4) Type of surgery  Isolated CABG 2446 (30.8) 419 (15.8) 2027 (38.4)  Isolated valve 2447 (30.8) 1040 (39.1) 1407 (26.6)  CABG + valve 1282 (16.1) 407 (15.3) 875 (16.6)  Ascending aorta 606 (7.6) 214 (8.1) 392 (7.4)  Others 1158 (14.6) 605 (22.5) 580 (11) CPB duration (min) 76 (58–104) 78 (59–104) 76 (58–104) Data are presented as n (%) or median (interquartile range). BMI: body mass index; CABG: coronary artery bypass graft; COPD: chronic obstructive pulmonary disease; CPB: cardiopulmonary bypass; CVA: cerebrovascular accident. View Large Operative mortality was 3.6% in the overall patient population (4.1% in women and 3.5% in men). The crude association between BMI class and operative mortality is shown in Fig. 1. In the overall patient population, the ‘U-shaped’ association between BMI class and mortality is confirmed with the lowest mortality rate for obese class I patients (2.6%) and the highest values at the 2 extremities (4.8% for underweight and 6.5% for obese class II–III). The Pearson’s χ2 analysis confirmed a significant (P = 0.012) impact of the BMI class on operative mortality. To adjust for possible confounders differently stratified in the various BMI classes, a propensity matching was performed. Three BMI-based categories were identified (underweight-normal, overweight-obesity class I and obesity class II–III). The smaller group was obesity class II–III (275 patients). Propensity-matched groups were extracted from the general population on a 2:1 ratio with the obesity class II–III group, with a propensity score based on the age, the expected operative mortality risk (EuroSCORE II) and the surgical complexity (CPB duration). The propensity-matched group was comparable with the exception of conditions intrinsically dependent on BMI (weight, height, diabetes and chronic obstructive pulmonary disease). Within these propensity-matched groups, the BMI-based differences in operative mortality were confirmed; and the ‘U-shaped’ relationship was confirmed with a mortality rate significantly (P = 0.015) higher in underweight-normal patients (4.1%) and obesity class II–III patients (6.5%) than overweight-obesity class I patients (2.4%) (Supplementary Material, Table S1). Figure 1: View largeDownload slide Relationship between the BMI classes and observed operative mortality in the overall population and separately for women and men. Bars are 95% confidence interval. BMI: body mass index. Figure 1: View largeDownload slide Relationship between the BMI classes and observed operative mortality in the overall population and separately for women and men. Bars are 95% confidence interval. BMI: body mass index. There were gender-based differences in BMI–mortality association. In women, the relationship appears ‘J-shaped’ with higher mortality values for obese class II–III patients (9.4%). The relative risk of mortality for obese class II–III patients with respect to the other weight classes is 2.6 (95% CI 1.37–4.81, P = 0.002). In men, the lowest mortality rate (2.5%) was observed in obese class I patients; however, the difference with respect to the other groups is significant only for normal weight patients (relative risk 0.57, 95% CI 0.34–0.95; P = 0.029). A higher mortality rate was observed for underweight patients (5.5%); however, this group is under-represented (55 cases only) and the differences in mortality rate with respect to the other groups are not significant. When considering the BMI as a continuous variable, the association with the operative mortality rate again behaves differently in women and men. The second-order polynomial regression showed the best fit for both the populations (Fig. 2), and the lowest mortality rate was identified at a BMI of 22 kg/m2 for women and 35 kg/m2 for men. The increase in mortality at low values of BMI is similar in men and women; conversely, patterns of high BMI appear to increase the mortality rate to a greater extent in women than in men. Figure 2: View largeDownload slide Unadjusted association (quadratic regression analysis) between the BMI and the observed operative mortality rate of women and men. Dashed lines delimitate the 95% confidence area. BMI: body mass index. Figure 2: View largeDownload slide Unadjusted association (quadratic regression analysis) between the BMI and the observed operative mortality rate of women and men. Dashed lines delimitate the 95% confidence area. BMI: body mass index. A separate analysis was conducted for isolated coronary artery bypass grafting (CABG) surgery and aortic valve replacement (Supplementary Material, Figs S3 and S4). In isolated CABG (N = 2446, women = 419, men = 2027), the effects detected in the overall population were confirmed at an even higher degree, with an increased mortality in men with a low BMI and no increase for high BMI (J-shaped relationship), and a confirmed U-shaped relationship in women. However, in obese women the mortality is greatly increased with values >10% for a BMI >35 kg/m2. In isolated aortic valve replacement (N = 1446, women = 591, men = 875), the relationship is U-shaped for both men and women. Separate analyses for the components of the BMI (weight and height) are presented in Figs 3 and 4. The association between weight and operative mortality (Fig. 3) was again defined by the second-order polynomial regression functions, with the lowest mortality rate identified at 70 kg for women and 100 kg for men. The relationship mirrors that of the BMI with a similar increase in mortality in women and men of low weight, whereas there was a marginal increase in mortality risk in men and a large increase in women with increasing weight. Figure 3: View largeDownload slide Unadjusted association (quadratic regression analysis) between the weight and the observed operative mortality rate of women and men. Dashed lines delimitate the 95% confidence area. Figure 3: View largeDownload slide Unadjusted association (quadratic regression analysis) between the weight and the observed operative mortality rate of women and men. Dashed lines delimitate the 95% confidence area. Figure 4: View largeDownload slide Unadjusted association (linear regression analysis) between the height and the observed operative mortality rate of women and men. Dashed lines delimitate the 95% confidence area. Figure 4: View largeDownload slide Unadjusted association (linear regression analysis) between the height and the observed operative mortality rate of women and men. Dashed lines delimitate the 95% confidence area. The impact of height on the mortality rate (Fig. 4) was defined using linear regression analyses that demonstrated to better fit the association than polynomial regressions (P-values for association 0.002 for women and 0.018 for men), with a mortality rate that decreased with increasing height, which was more pronounced in women. The impact of BMI, weight and height in determining the mortality risk was checked after correction for the age, the EuroSCORE II and the CPB duration separately for women (Table 2) and men (Table 3). The impact of BMI and weight was assessed in terms of absolute (negative or positive) offset from the best (nadir) values identified by the second-order functions of Figs 2 and 3; height was considered as the absolute value. Logistic regression models for operative mortality were applied. Table 2: Crude and adjusted operative mortality risk analysis as a function of body mass index, weight and height in female patients (N = 2658) Factors Regression coefficient Odds ratio 95% CI P-value BMI (crude)a 0.043 1.043 1.000–1.089 0.050 BMI (adjusted)a 0.080 1.084 1.029–1.141 0.002 Age (years) 0.050 1.051 1.025–1.078 0.001 EuroSCORE II 0.079 1.082 1.061–1.104 0.001 CPB duration (min) 0.015 1.016 1.012–1.019 0.001 Weight (crude)b 0.024 1.024 1.001–1.047 0.038 Weight (adjusted)b 0.021 1.021 0.994–1.049 0.133 Age (years) 0.046 1.047 1.022–1.072 0.001 EuroSCORE II 0.076 1.079 1.058–1.101 0.001 CPB duration (min) 0.015 1.015 1.012–1.019 0.001 Height (crude)c −0.047 0.954 0.926–0.984 0.002 Height (adjusted)c −0.053 0.949 0.915–0.983 0.004 Age (years) 0.043 1.044 1.018–1.070 0.001 EuroSCORE II 0.078 1.081 1.060–1.102 0.001 CPB duration (min) 0.015 1.015 1.012–1.019 0.001 Factors Regression coefficient Odds ratio 95% CI P-value BMI (crude)a 0.043 1.043 1.000–1.089 0.050 BMI (adjusted)a 0.080 1.084 1.029–1.141 0.002 Age (years) 0.050 1.051 1.025–1.078 0.001 EuroSCORE II 0.079 1.082 1.061–1.104 0.001 CPB duration (min) 0.015 1.016 1.012–1.019 0.001 Weight (crude)b 0.024 1.024 1.001–1.047 0.038 Weight (adjusted)b 0.021 1.021 0.994–1.049 0.133 Age (years) 0.046 1.047 1.022–1.072 0.001 EuroSCORE II 0.076 1.079 1.058–1.101 0.001 CPB duration (min) 0.015 1.015 1.012–1.019 0.001 Height (crude)c −0.047 0.954 0.926–0.984 0.002 Height (adjusted)c −0.053 0.949 0.915–0.983 0.004 Age (years) 0.043 1.044 1.018–1.070 0.001 EuroSCORE II 0.078 1.081 1.060–1.102 0.001 CPB duration (min) 0.015 1.015 1.012–1.019 0.001 a Units of negative/positive offset from the value of 22 kg/m2. b Negative/positive offset (kg) from the value of 70 kg. c Absolute value (cm). BMI: body mass index; CI: confidence interval; CPB: cardiopulmonary bypass. View Large Table 2: Crude and adjusted operative mortality risk analysis as a function of body mass index, weight and height in female patients (N = 2658) Factors Regression coefficient Odds ratio 95% CI P-value BMI (crude)a 0.043 1.043 1.000–1.089 0.050 BMI (adjusted)a 0.080 1.084 1.029–1.141 0.002 Age (years) 0.050 1.051 1.025–1.078 0.001 EuroSCORE II 0.079 1.082 1.061–1.104 0.001 CPB duration (min) 0.015 1.016 1.012–1.019 0.001 Weight (crude)b 0.024 1.024 1.001–1.047 0.038 Weight (adjusted)b 0.021 1.021 0.994–1.049 0.133 Age (years) 0.046 1.047 1.022–1.072 0.001 EuroSCORE II 0.076 1.079 1.058–1.101 0.001 CPB duration (min) 0.015 1.015 1.012–1.019 0.001 Height (crude)c −0.047 0.954 0.926–0.984 0.002 Height (adjusted)c −0.053 0.949 0.915–0.983 0.004 Age (years) 0.043 1.044 1.018–1.070 0.001 EuroSCORE II 0.078 1.081 1.060–1.102 0.001 CPB duration (min) 0.015 1.015 1.012–1.019 0.001 Factors Regression coefficient Odds ratio 95% CI P-value BMI (crude)a 0.043 1.043 1.000–1.089 0.050 BMI (adjusted)a 0.080 1.084 1.029–1.141 0.002 Age (years) 0.050 1.051 1.025–1.078 0.001 EuroSCORE II 0.079 1.082 1.061–1.104 0.001 CPB duration (min) 0.015 1.016 1.012–1.019 0.001 Weight (crude)b 0.024 1.024 1.001–1.047 0.038 Weight (adjusted)b 0.021 1.021 0.994–1.049 0.133 Age (years) 0.046 1.047 1.022–1.072 0.001 EuroSCORE II 0.076 1.079 1.058–1.101 0.001 CPB duration (min) 0.015 1.015 1.012–1.019 0.001 Height (crude)c −0.047 0.954 0.926–0.984 0.002 Height (adjusted)c −0.053 0.949 0.915–0.983 0.004 Age (years) 0.043 1.044 1.018–1.070 0.001 EuroSCORE II 0.078 1.081 1.060–1.102 0.001 CPB duration (min) 0.015 1.015 1.012–1.019 0.001 a Units of negative/positive offset from the value of 22 kg/m2. b Negative/positive offset (kg) from the value of 70 kg. c Absolute value (cm). BMI: body mass index; CI: confidence interval; CPB: cardiopulmonary bypass. View Large Table 3: Crude and adjusted operative mortality risk analysis as a function of body mass index, weight and height in male patients (N = 5281) Factors Regression coefficient Odds ratio 95% CI P-value BMI (crude)a 0.052 1.054 1.010–1.099 0.015 BMI (adjusted)a 0.032 1.033 0.985–1.083 0.185 Age (years) 0.053 1.054 1.037–1.072 0.001 EuroSCORE II 0.058 1.059 1.044–1.075 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Weight (crude)b 0.024 1.025 1.011–1.038 0.001 Weight (adjusted)b 0.017 1.017 1.001–1.032 0.034 Age (years) 0.051 1.053 1.035–1.071 0.001 EuroSCORE II 0.058 1.059 1.044–1.075 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Height (crude)c −0.025 0.975 0.955–0.996 0.018 Height (adjusted)c −0.015 0.985 0.961–1.010 0.232 Age (years) 0.051 1.053 1.035–1.071 0.001 EuroSCORE II 0.059 1.060 1.045–1.076 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Factors Regression coefficient Odds ratio 95% CI P-value BMI (crude)a 0.052 1.054 1.010–1.099 0.015 BMI (adjusted)a 0.032 1.033 0.985–1.083 0.185 Age (years) 0.053 1.054 1.037–1.072 0.001 EuroSCORE II 0.058 1.059 1.044–1.075 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Weight (crude)b 0.024 1.025 1.011–1.038 0.001 Weight (adjusted)b 0.017 1.017 1.001–1.032 0.034 Age (years) 0.051 1.053 1.035–1.071 0.001 EuroSCORE II 0.058 1.059 1.044–1.075 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Height (crude)c −0.025 0.975 0.955–0.996 0.018 Height (adjusted)c −0.015 0.985 0.961–1.010 0.232 Age (years) 0.051 1.053 1.035–1.071 0.001 EuroSCORE II 0.059 1.060 1.045–1.076 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 a Units of negative/positive offset from the value of 35 kg/m2. b Negative/positive offset (kg) from the value of 100 kg. c Absolute value (cm). BMI: body mass index; CI: confidence interval; CPB: cardiopulmonary bypass. View Large Table 3: Crude and adjusted operative mortality risk analysis as a function of body mass index, weight and height in male patients (N = 5281) Factors Regression coefficient Odds ratio 95% CI P-value BMI (crude)a 0.052 1.054 1.010–1.099 0.015 BMI (adjusted)a 0.032 1.033 0.985–1.083 0.185 Age (years) 0.053 1.054 1.037–1.072 0.001 EuroSCORE II 0.058 1.059 1.044–1.075 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Weight (crude)b 0.024 1.025 1.011–1.038 0.001 Weight (adjusted)b 0.017 1.017 1.001–1.032 0.034 Age (years) 0.051 1.053 1.035–1.071 0.001 EuroSCORE II 0.058 1.059 1.044–1.075 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Height (crude)c −0.025 0.975 0.955–0.996 0.018 Height (adjusted)c −0.015 0.985 0.961–1.010 0.232 Age (years) 0.051 1.053 1.035–1.071 0.001 EuroSCORE II 0.059 1.060 1.045–1.076 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Factors Regression coefficient Odds ratio 95% CI P-value BMI (crude)a 0.052 1.054 1.010–1.099 0.015 BMI (adjusted)a 0.032 1.033 0.985–1.083 0.185 Age (years) 0.053 1.054 1.037–1.072 0.001 EuroSCORE II 0.058 1.059 1.044–1.075 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Weight (crude)b 0.024 1.025 1.011–1.038 0.001 Weight (adjusted)b 0.017 1.017 1.001–1.032 0.034 Age (years) 0.051 1.053 1.035–1.071 0.001 EuroSCORE II 0.058 1.059 1.044–1.075 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Height (crude)c −0.025 0.975 0.955–0.996 0.018 Height (adjusted)c −0.015 0.985 0.961–1.010 0.232 Age (years) 0.051 1.053 1.035–1.071 0.001 EuroSCORE II 0.059 1.060 1.045–1.076 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 a Units of negative/positive offset from the value of 35 kg/m2. b Negative/positive offset (kg) from the value of 100 kg. c Absolute value (cm). BMI: body mass index; CI: confidence interval; CPB: cardiopulmonary bypass. View Large In women, after adjustment for the confounders, the offset from the best BMI remained independently associated with operative mortality (relative risk increase 8.4% per each unit of negative/positive offset from 22 kg/m2); negative/positive offset from the best weight of 70 kg lost significance; and the height remained independently associated with operative mortality (relative risk decrease of 5.1% per each centimetre increase in height). In men, after adjustment for the confounders, the offset from the best BMI lost significance; the offset from the best weight remained independently associated with operative mortality (relative risk increase of 1.7% per each unit of negative/positive offset from 100 kg); and the height lost significance for association with the operative mortality. The predicted mortality rate in logistic regression adjusted models as a function of height and offset from the best weight is reported in the Supplementary Material, Figs S5 and S6. The predicted mortality declines for increasing the height to a larger extent in women than in men (Supplementary Material, Fig. S5), whereas it increases for increasing the weight offset to a larger extent in men than in women (Supplementary Material, Fig. S6). On the basis of the previous results, a sensitivity analysis was conducted. Tall women were considered as those in the upper quartile of the distribution (≥165 cm), and moderately heavy men as those in the third quartile of distribution (weight 77–95 kg) (Table 4). After correction for the age, the weight, the EuroSCORE II and the CPB duration, tall women had an odds ratio for operative mortality of 0.544 (95% CI 0.309–0.956, P = 0.034), with weight being independently associated with operative mortality. Moderately heavy men had an odds ratio for operative mortality of 0.552 (95% CI 0.378–0.804, P = 0.002), and the height was not an independent factor for operative mortality. Table 4: Operative mortality risk analysis for tall women and moderately heavy men Factors Regression coefficient Odds ratio 95% CI P-value Women (N = 2685)  Height ≥165 cm −0.609 0.544 0.309–0.956 0.034  Weight (kg) 0.019 1.019 1.001–1.038 0.039  Age (years) 0.048 1.049 1.023–1.076 0.001  EuroSCORE II 0.079 1.082 1.061–1.103 0.001  CPB duration (min) 0.015 1.016 1.012–1.019 0.001 Men (N = 5281)  Weight 77–95 kg −0.594 0.552 0.379–0.804 0.002  Height (cm) −0.004 0.996 0.971–1.021 0.741  Age (years) 0.052 1.053 1.035–1.071 0.001  EuroSCORE II 0.058 1.060 1.044–1.076 0.001  CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Factors Regression coefficient Odds ratio 95% CI P-value Women (N = 2685)  Height ≥165 cm −0.609 0.544 0.309–0.956 0.034  Weight (kg) 0.019 1.019 1.001–1.038 0.039  Age (years) 0.048 1.049 1.023–1.076 0.001  EuroSCORE II 0.079 1.082 1.061–1.103 0.001  CPB duration (min) 0.015 1.016 1.012–1.019 0.001 Men (N = 5281)  Weight 77–95 kg −0.594 0.552 0.379–0.804 0.002  Height (cm) −0.004 0.996 0.971–1.021 0.741  Age (years) 0.052 1.053 1.035–1.071 0.001  EuroSCORE II 0.058 1.060 1.044–1.076 0.001  CPB duration (min) 0.017 1.017 1.015–1.020 0.001 CI: confidence interval; CPB: cardiopulmonary bypass. View Large Table 4: Operative mortality risk analysis for tall women and moderately heavy men Factors Regression coefficient Odds ratio 95% CI P-value Women (N = 2685)  Height ≥165 cm −0.609 0.544 0.309–0.956 0.034  Weight (kg) 0.019 1.019 1.001–1.038 0.039  Age (years) 0.048 1.049 1.023–1.076 0.001  EuroSCORE II 0.079 1.082 1.061–1.103 0.001  CPB duration (min) 0.015 1.016 1.012–1.019 0.001 Men (N = 5281)  Weight 77–95 kg −0.594 0.552 0.379–0.804 0.002  Height (cm) −0.004 0.996 0.971–1.021 0.741  Age (years) 0.052 1.053 1.035–1.071 0.001  EuroSCORE II 0.058 1.060 1.044–1.076 0.001  CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Factors Regression coefficient Odds ratio 95% CI P-value Women (N = 2685)  Height ≥165 cm −0.609 0.544 0.309–0.956 0.034  Weight (kg) 0.019 1.019 1.001–1.038 0.039  Age (years) 0.048 1.049 1.023–1.076 0.001  EuroSCORE II 0.079 1.082 1.061–1.103 0.001  CPB duration (min) 0.015 1.016 1.012–1.019 0.001 Men (N = 5281)  Weight 77–95 kg −0.594 0.552 0.379–0.804 0.002  Height (cm) −0.004 0.996 0.971–1.021 0.741  Age (years) 0.052 1.053 1.035–1.071 0.001  EuroSCORE II 0.058 1.060 1.044–1.076 0.001  CPB duration (min) 0.017 1.017 1.015–1.020 0.001 CI: confidence interval; CPB: cardiopulmonary bypass. View Large DISCUSSION Our study confirms that the relationship between BMI and operative mortality in cardiac surgery is U-shaped with the lowest mortality rate at a BMI of about 35 kg/m2. The U-shaped relationship was confirmed after propensity matching, with a significantly lower mortality rate for overweight-obesity class I patients with respect to underweight-normal patients and obesity class II–III patients. This result is in agreement with a very large population study [1], where the mortality rate was very close to the one registered at our institution (3.3% in-hospital vs 3.6% in-hospital/30 days from surgery). Notably, the obesity rate (any class) was much lower in our study (16.3%) than in the UK series (30%). When analysed separately for different genders, the results offered new insights into the matter. Basically, the U-shaped relationship seems to be formed by the combination of 2 J-shaped relationships, where in women the high mortality risk in obese class II–III accounts for the right branch, and in men the high mortality risk in underweight accounts for the left branch of the U-shaped relationship. Additionally, the values of BMI associated with the lowest mortality rate are distinctly different in women (22 kg/m2) and in men (35 kg/m2). This combination of a different gender-based, J-shaped relationship is strongly confirmed for isolated CABG surgery. Overall, the effect of BMI on the outcome is therefore highly different in women versus men. In women, no significant differences in mortality rates are observed in the groups underweight to obese class I, and an important increase is observed for obese class II–III. Therefore, it seems difficult to apply the concept of the ‘obesity paradox’ to female patients: the lowest mortality in women corresponds to a BMI (22 kg/m2) and a weight (70 kg) that certainly do not fit with the definition of obesity. Conversely, in men, this paradox is maintained with the best results obtained at a BMI (35 kg/m2) and a weight (100 kg) that suit well with an obesity pattern. When looking at the single components of the BMI, tall women (≥165 cm) and moderately heavy men (77–95 kg) have the best outcome, after correction for the weight (in women), the height (in men), age, EuroSCORE II and CPB duration, with an operative mortality risk about half that of the other patients. There are other examples, in different settings, of a different gender-based relationship between obesity and outcomes. In a recent study, Vest et al. [12] investigated the obesity paradox in heart failure patients separately for women and men. They confirmed the obesity paradox in the total patient population; while this behaviour was confirmed in women, the hazard ratio for mortality was higher in overweight–obese male patients. These results were partially confirmed in terms of heart failure symptoms in a recent study [13], and again in the medical setting, the obesity paradox was more pronounced in women than in men in a population of patients with myocardial infarction [14]. These results apparently conflict with our findings, where the U-shaped relationship is confirmed in men, but not in women. However, the clinical setting is totally different, and the acute impact of surgery (in a population with only 8% of heart failure patients) is certainly a determinant of these differences. The existing studies exploring the obesity paradox in cardiac surgery do not take into account the effect of gender, and our observation introduces a new perspective into this topic. A recent study [15] on 15 134 patients undergoing cardiac surgery found that, once adjusted for other confounders, the BMI was not a determinant of hospital mortality, whereas the female gender remained an independent risk factor. This suggests that an interaction between gender and BMI as determinants of mortality exists, as confirmed by our specific analysis. There are many possible explanations underlying the observation that the 2 extremes of the BMI distribution are associated with bad outcomes. Basically, a very low BMI may be an expression of frailty; moreover, this condition is associated with a greater impact of haemodilution on CPB, which is a determinant of bad outcomes [16], and with a larger use of allogeneic blood products [4]. However, in our series this applies to men only; it is therefore likely that a low BMI may be associated with a frailer condition more often seen in men than in women. The higher mortality risk in severe (class II–III) obesity is usually attributed to postoperative pulmonary complications, surgical site infections and technical surgical problems (prosthesis–patient mismatch in valve surgery). This severe obesity-related mortality risk appears more pronounced in women than in men, in our series, but we are lacking a clear interpretation of this finding. In our analysis, we could identify a different role of height and weight in women versus men. This introduces the very well-known concept that BMI, per se, is not a good discriminant between an elevated body weight because of the high levels of lean versus fat body mass. More adequate measures include waist circumference, waist–hip ratio, skinfold estimates of % body fat and bioelectrical impedance analysis of body composition [17]. All these measures are lacking in our study (as in previous studies in the setting of cardiac surgery), but the finding of a protective effect of height in women (and not in men) is worth some consideration. Height is included in the waist circumference/height ratio that is considered a good estimate of intra-abdominal fat and cardiometabolic risk [18]. Therefore, it is reasonable to assume that height is inversely associated with cardiovascular risk. In our patient population, height is inversely associated with operative mortality, but this effect is more pronounced and independently associated with mortality in women only. This gender-related behaviour of height was found in other studies. In a large population study [19], the decrease in BMI with increasing height was more pronounced in women than in men: in the tallest compared with the shortest height quartile, it was 0.77 in men (95% CI 0.69–0.86) and 1.98 in women (95% CI 1.89–2.08). The role of height in the calculation of lean body mass, fat mass and % fat mass was recently highlighted in a study based on the National Health and Nutrition Examination Survey [20]. Highly predictive equations for body composition were developed based on age, race, height, weight and waist circumference. However, waist circumference was a strong predictor in men but not in women. The role of height in determining a lower fat mass is, therefore, more relevant in women than in men. The above reported data help in understanding our results which found that height plays a role in determining fat mass and its distribution, and which appeared to be more relevant in women than in men. This role defines height as a protective factor in cardiac surgery, independently from weight in women. From a practical perspective, our data may offer some possible insights. Underweight male patients may probably benefit from specific nutritional and exercise training programmes before cardiac surgery, whenever feasible. Conversely, underweight women may probably not necessarily benefit from this approach. Weight loss programmes before cardiac surgery are indicated for both women and men when belonging to obesity class II–III; however, in women ≥ 165 cm, weight loss is probably less relevant. Limitations There are limitations in our study. The most important is the lack of relevant measures of body fat distribution (waist circumference, waist–hip ratio, skinfold estimates of % body fat and bioelectrical impedance analysis of body composition). Additionally, a frailty score assessment to identify underweight patients with or without frailty was not conducted. CONCLUSION In conclusion, our study demonstrates that there is a gender-based difference for BMI, weight and height as determinants of operative mortality in patients undergoing cardiac surgery. Further studies including more specific measures of fat distribution are warranted. Funding This work was supported by the IRCCS Policlinico San Donato, a Clinical Research Hospital recognized and partially supported by the Italian Ministry of Health. Conflict of interest: none declared. REFERENCES 1 Mariscalco G , Wozniak MJ , Dawson AG , Serraino GF , Porter R , Nath M et al. Body mass index and mortality among adults undergoing cardiac surgery: a nationwide study with a systematic review and meta-analysis . Circulation 2017 ; 135 : 850 – 63 . Google Scholar Crossref Search ADS PubMed 2 Potapov EV , Loebe M , Anker S , Stein J , Bondy S , Nasseri BA et al. Impact of body mass index on outcome in patients after coronary artery bypass grafting with and without valve surgery . Eur Heart J 2003 ; 24 : 1933 – 41 . Google Scholar Crossref Search ADS PubMed 3 Stamou SC , Nussbaum M , Stiegel RM , Reames MK , Skipper ER , Robicsek F et al. Effect of body mass index on outcomes after cardiac surgery: is there an obesity paradox? Ann Thorac Surg 2011 ; 91 : 42 – 7 . Google Scholar Crossref Search ADS PubMed 4 Ranucci M , Pazzaglia A , Bianchini C , Bozzetti G , Isgrò G. Body size, gender, and transfusions as determinants of outcome after coronary operations . Ann Thorac Surg 2008 ; 85 : 481 – 6 . Google Scholar Crossref Search ADS PubMed 5 Flegal KM , Kit BK , Orpana H , Graubard BI. Association of all-cause mortality with overweight and obesity using standard body mass index categories: a systematic review and meta-analysis . JAMA 2013 ; 309 : 71 – 82 . Google Scholar Crossref Search ADS PubMed 6 Angerås O , Albertsson P , Karason K , Råmunddal T , Matejka G , James S et al. Evidence for obesity paradox in patients with acute coronary syndromes: a report from the Swedish Coronary Angiography and Angioplasty Registry . Eur Heart J 2013 ; 34 : 345 – 53 . Google Scholar Crossref Search ADS PubMed 7 Oreopoulos A , Padwal R , Kalantar-Zadeh K , Fonarow GC , Norris CM , McAlister FA. Body mass index and mortality in heart failure: a meta-analysis . Am Heart J 2008 ; 156 : 13 – 22 . Google Scholar Crossref Search ADS PubMed 8 Ho KM , Bertenshaw C , Same S , Schneider M , Williams KA , Godsell T et al. Differential associations between body mass index and outcomes after elective adult cardiac surgery: a linked data cohort study . Anaesth Intensive Care 2013 ; 41 : 573 – 83 . Google Scholar Crossref Search ADS PubMed 9 Habib RH , Zacharias A , Schwann TA , Riordan CJ , Durham SJ , Shah A. Effects of obesity and small body size on operative and long-term outcomes of coronary artery bypass surgery: a propensity-matched analysis . Ann Thorac Surg 2005 ; 79 : 1976 – 86 . Google Scholar Crossref Search ADS PubMed 10 WHO. Physical Status: The Use and Interpretation of Anthropometry: Report of a WHO Expert Committee . WHO Technical Report Series 854. Geneva, Switzerland : World Health Organization , 1995 . 11 Criqui MH , Klauber MR , Barrett-Connor E , Holdbrook MJ , Suarez L , Wingard DL. Adjustment for obesity in studies of cardiovascular disease . Am J Epidemiol 1982 ; 116 : 685 – 91 . Google Scholar Crossref Search ADS PubMed 12 Vest AR , Wu Y , Hachamovitch R , Young JB , Cho L. The heart failure overweight/obesity survival paradox . JACC Heart Fail 2015 ; 3 : 917 – 26 . Google Scholar Crossref Search ADS PubMed 13 Heo S , Moser DK , Pressler SJ , Dunbar SB , Lee KS , Kim J et al. Association between obesity and heart failure symptoms in male and female patients . Clin Obes 2017 ; 7 : 77 – 85 . Google Scholar Crossref Search ADS PubMed 14 Keller K , Munzel T , Ostad MA. Sex-specific differences in mortality and the obesity paradox of patients with myocardial infarction ages >70 y . Nutrition 2018 ; 46 : 124 – 30 . Google Scholar Crossref Search ADS PubMed 15 Hartrumpf M , Kuehnel R-U , Albes JM. The obesity paradox is still there: a risk analysis of over 15000 cardiosurgical patients based on body mass index . Interact CardioVasc Thorac Surg 2017 ; 25 : 18 – 24 . Google Scholar Crossref Search ADS PubMed 16 Habib RH , Zacharias A , Schwann TA , Riordan CJ , Durham SJ , Shah A. Adverse effects of low hematocrit during cardiopulmonary bypass in the adult: should current practice be changed? J Thorac Cardiovasc Surg 2003 ; 125 : 1438 – 50 . Google Scholar Crossref Search ADS PubMed 17 Clark AL , Fonarow GC , Horwich TB. Obesity and the obesity paradox in heart failure . Prog Cardiovasc Dis 2014 ; 56 : 409 – 14 . Google Scholar Crossref Search ADS PubMed 18 Ashwell M , Cole TJ , Dixon AK. Ratio of waist circumference to height is strong predictor of intra-abdominal fat . BMJ 1996 ; 313 : 555 – 60 . Google Scholar Crossref Search ADS 19 Sperrin M , Marshall AD , Higgins V , Renehan AG , Buchan IE. Body mass index relates weight to height differently in women and older adults: serial cross-sectional surveys in England (1992–2011) . J Public Health 2016 ; 38 : 607 – 13 . Google Scholar Crossref Search ADS 20 Lee DH , Keum N , Hu FB , Orav EJ , Rimm EB , Sun Q et al. Development and validation of anthropometric prediction equations for lean body mass, fat mass and percent fat in adults using the National Health and Nutrition Examination Survey (NHANES) 1999-2006 . Br J Nutr 2017 ; 118 : 858 – 66 . Google Scholar Crossref Search ADS PubMed © The Author(s) 2019. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png European Journal of Cardio-Thoracic Surgery Oxford University Press

A gender-based analysis of the obesity paradox in cardiac surgery: height for women, weight for men?

Download PDF
7 pages

Loading next page...
 
/lp/oxford-university-press/a-gender-based-analysis-of-the-obesity-paradox-in-cardiac-surgery-M657j0SDhs

References (23)

Publisher
Oxford University Press
Copyright
© The Author(s) 2019. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
ISSN
1010-7940
eISSN
1873-734X
DOI
10.1093/ejcts/ezy454
Publisher site
See Article on Publisher Site

Abstract

Abstract View largeDownload slide View largeDownload slide OBJECTIVES In cardiac surgery, obesity is associated with a lower mortality risk. This study aims to investigate the association between body mass index (BMI) and operative mortality separately in female patients and male patients undergoing cardiac surgery and to separate the effects of weight and height in each gender-based cohort of patients. METHODS A retrospective cohort study including 7939 consecutive patients who underwent cardiac surgery was conducted. The outcome measure was the operative mortality. RESULTS In men, there was a U-shaped relationship between the BMI and the operative mortality, with the lower mortality rate at a BMI of 35 kg/m2. In women, the relationship is J-shaped, with the lower mortality at a BMI of 22 kg/m2. Female patients with obesity class II–III had a relative risk for operative mortality of 2.6 [95% confidence interval (CI) 1.37–4.81, P = 0.002]. The relationship between weight and mortality rate is a U-shaped bot in men and women, with the lower mortality rate at 100 kg for men and 70 kg for women. Height was linearly and inversely associated with the operative mortality in men and women. After correction for the potential confounders, height, but not weight, was independently associated with operative mortality in women (odds ratio 0.949, 95% CI 0.915–0.983; P = 0.004); conversely, in men, this association exists for weight (odds ratio 1.017, 95% CI 1.001–1.032; P = 0.034), but not height. CONCLUSIONS Contrary to men, in women obesity does not reduce the operative mortality in cardiac surgery, whereas the height seems to be associated with a lower mortality. Adult, Cardiac surgical procedures, Gender, Obesity, Mortality INTRODUCTION A recent large retrospective cohort study [1] found that obesity is associated with lower morbidity and mortality after cardiac surgery, confirming previous relatively smaller studies where this association was identified [2–4]. These results concur in the identification of the so-called ‘obesity paradox’, where an important risk factor for cardiovascular disease and cardiovascular death [5] is conversely a protective factor in patients with acute coronary syndromes [6], heart failure [7] and those undergoing cardiac surgery. Many possible explanations for this paradox in the cardiac surgery population have been raised, including collider bias (where the association in favour of the obese patients actually reflects a higher mortality in frail, cachectic, underweight patients) or selection bias (obese patients are admitted to cardiac surgery only if without other risk factors) or an indirect age effect (obese patients are usually younger) or finally a lower degree of haemodilution and less need for transfusions in patients with large body mass index (BMI) [8, 9]. In 2008, a study from our group [4] raised the hypothesis that BMI and body surface area might affect the outcome of cardiac surgery differently in women and men. Additionally, previous studies addressing the association between obesity and outcomes in cardiac surgery used the BMI to stratify the patients into categories, in accordance with the World Health Organization [10]. However, the BMI is calculated from weight and height, and no previous studies in this field tried to separate the effects of these 2 variables. The present single-centre, retrospective, cohort study aims to investigate the association between BMI and operative mortality separately in female patients and male patients undergoing cardiac surgery and to separate the effects of weight and height in each gender-based cohort of patients. METHODS This study is based on a retrospective, single-centre cohort study of patients undergoing cardiac surgery in the period January 2010 to November 2017. All data were retrieved from the institutional database of the IRCCS Policlinico San Donato. The study design was approved by the local ethics committee (IRCCS San Raffaele Hospital) and the need for a written informed consent was waived, providing adequate anonymization of all the data utilized. All the patients gave a written consent for the scientific treatment of their data in an anonymous form at the admission in the Hospital. Patients During the period considered, 11 374 patients underwent heart surgery at our institution. Patients aged <18 years and adult patients undergoing congenital heart surgery were excluded from the analysis, reaching a final patient population of 7939 adult patients undergoing cardiac surgery. Data collection and definitions Data collection is derived from the institutional cardiac surgery database, which includes demographics, comorbidities, heart function and laboratory data, conditions at surgery, type of surgery, intraoperative details and outcome data. The BMI was calculated as the weight in kilograms divided by the square of the height in metres [11]. It was further categorized into 5 classes according to the World Health Organization classification [10]: underweight (BMI <18.5 kg/m2), normal weight (BMI 18.5 to <25 kg/m2), overweight (BMI 25 to <30 kg/m2), obese class I (BMI 30 to <35 kg/m2) and obese class II–III (BMI ≥35 kg/m2). Obesity classes II and III were pooled together because of the poor number of cases in obesity class III. Operative mortality was defined as mortality occurring during hospital stay or within 30 days from surgery after discharge. Statistical analysis Data are presented as number (%) for categorical data and median (interquartile range) for continuous data. The association between BMI classes and operative mortality was tested using the Pearson’s χ2 analysis, and interclass differences and cubic spline functions in figures were calculated using post hoc analysis. The associations between BMI, weight and height (continuous variables) and the operative mortality were assessed separately for women and men with regression analyses (linear and quadratic) choosing the equation with the best fit. When the best fit was found for a quadratic equation, the value of BMI and weight associated with the lowest mortality risk was identified at the nadir of the parabolic relationship. The potential sources of bias were addressed as follows: the independent role of BMI, weight and height as determinants of the operative mortality was tested with logistic regression analyses, after correction for the age, the EuroSCORE II (for operative risk stratification) and the cardiopulmonary bypass (CPB) duration (as an index of intraoperative complexity). For values being linearly associated with the operative mortality, the absolute value was included as an independent variable; for values showing a quadratic relationship with the operative mortality, the negative or positive offset from the value associated with the lowest mortality rate was considered and included as an absolute value in the logistic regression analysis. Odds ratios and 95% confidence intervals (CIs) were derived from the logistic regression analyses. Multivariable logistic regression models provided an estimate of operative mortality with a confounder-adjusted marginal standardization. Additional analyses included a propensity matching of the population based on 3 BMI-based categories. Propensity matching was based on the EuroSCORE II, the age and the CPB duration. All tests were 2-sided; a P-value of <0.05 was considered significant. All the statistical analyses were performed with computerized packages (SPSS 20.0, IBM, Chicago, IL, USA and GraphPad Prism 5.0, GraphPad, San Diego, CA, USA). RESULTS The general characteristics of the patient population are summarized in Table 1. Overall, there were 2658 (33.5%) women and 5281 (66.5%) men. One hundred sixty-five (2.1%) patients were classified as underweight, 3302 (41.6%) as normal weight, 3183 (40.1%) as overweight, 1014 (12.8%) as obese class I and 275 (3.5%) as obese class II–III. The distribution of weight and height in women and men is shown in the Supplementary Material, Figs S1 and S2. Table 1: General characteristics of the patient population Variables Overall (N = 7939) Women (N = 2685) Men (N = 5281) Age (years) 69 (60–76) 72 (63–78) 68 (59–75) Weight (kg) 73 (64–82) 64 (56–72) 77 (70–85) Height (cm) 169 (160–174) 160 (156–165) 170 (168–176) Body mass index (kg/m2) 25.7 (23.4–28.4) 24.9 (22–28.1) 26 (23.9–28.6) BMI class  Underweight 165 (2.1) 110 (4.1) 55 (1.0)  Normal weight 3302 (41.6) 1279 (48.1) 2023 (38.3)  Overweight 3183 (40.1) 843 (31.7) 2340 (44.3)  Obesity class I 1014 (12.8) 298 (11.2) 716 (13.6)  Obesity class II–III 275 (3.5) 128 (4.8) 147 (2.8) Hypertension 5080 (64) 1695 (63) 3385 (64) Left ventricular ejection fraction (%) 55 (49–60) 56 (50–62) 55 (46–60) Congestive heart failure 646 (8.1) 265 (10) 381 (7.2) Diabetes on medication 1459 (18.4) 411 (15.5) 1048 (19.8) Serum creatinine (mg/dl) 1.0 (0.8–1.2) 0.85 (0.7–1.1) 1.0 (0.9–1.2) Haematocrit (%) 39 (35.7–41.9) 37.4 (34.1–40) 40 (36.6–42.6) COPD 581 (7.3) 164 (6.2) 417 (7.9) Previous CVA 250 (3.1) 73 (2.7) 177 (3.4) Conditions of surgery  Elective 6674 (84.1) 2290 (86.2) 4384 (83)  Urgent 1120 (14.1) 319 (12) 809 (15.2)  Emergency 145 (1.8) 49 (1.8) 96 (1.8) Redo surgery 605 (7.6) 245 (9.2) 360 (6.8) EuroSCORE II 2.2 (1.2–4.7) 2.5 (1.4–5.6) 2 (1.2–4.4) Type of surgery  Isolated CABG 2446 (30.8) 419 (15.8) 2027 (38.4)  Isolated valve 2447 (30.8) 1040 (39.1) 1407 (26.6)  CABG + valve 1282 (16.1) 407 (15.3) 875 (16.6)  Ascending aorta 606 (7.6) 214 (8.1) 392 (7.4)  Others 1158 (14.6) 605 (22.5) 580 (11) CPB duration (min) 76 (58–104) 78 (59–104) 76 (58–104) Variables Overall (N = 7939) Women (N = 2685) Men (N = 5281) Age (years) 69 (60–76) 72 (63–78) 68 (59–75) Weight (kg) 73 (64–82) 64 (56–72) 77 (70–85) Height (cm) 169 (160–174) 160 (156–165) 170 (168–176) Body mass index (kg/m2) 25.7 (23.4–28.4) 24.9 (22–28.1) 26 (23.9–28.6) BMI class  Underweight 165 (2.1) 110 (4.1) 55 (1.0)  Normal weight 3302 (41.6) 1279 (48.1) 2023 (38.3)  Overweight 3183 (40.1) 843 (31.7) 2340 (44.3)  Obesity class I 1014 (12.8) 298 (11.2) 716 (13.6)  Obesity class II–III 275 (3.5) 128 (4.8) 147 (2.8) Hypertension 5080 (64) 1695 (63) 3385 (64) Left ventricular ejection fraction (%) 55 (49–60) 56 (50–62) 55 (46–60) Congestive heart failure 646 (8.1) 265 (10) 381 (7.2) Diabetes on medication 1459 (18.4) 411 (15.5) 1048 (19.8) Serum creatinine (mg/dl) 1.0 (0.8–1.2) 0.85 (0.7–1.1) 1.0 (0.9–1.2) Haematocrit (%) 39 (35.7–41.9) 37.4 (34.1–40) 40 (36.6–42.6) COPD 581 (7.3) 164 (6.2) 417 (7.9) Previous CVA 250 (3.1) 73 (2.7) 177 (3.4) Conditions of surgery  Elective 6674 (84.1) 2290 (86.2) 4384 (83)  Urgent 1120 (14.1) 319 (12) 809 (15.2)  Emergency 145 (1.8) 49 (1.8) 96 (1.8) Redo surgery 605 (7.6) 245 (9.2) 360 (6.8) EuroSCORE II 2.2 (1.2–4.7) 2.5 (1.4–5.6) 2 (1.2–4.4) Type of surgery  Isolated CABG 2446 (30.8) 419 (15.8) 2027 (38.4)  Isolated valve 2447 (30.8) 1040 (39.1) 1407 (26.6)  CABG + valve 1282 (16.1) 407 (15.3) 875 (16.6)  Ascending aorta 606 (7.6) 214 (8.1) 392 (7.4)  Others 1158 (14.6) 605 (22.5) 580 (11) CPB duration (min) 76 (58–104) 78 (59–104) 76 (58–104) Data are presented as n (%) or median (interquartile range). BMI: body mass index; CABG: coronary artery bypass graft; COPD: chronic obstructive pulmonary disease; CPB: cardiopulmonary bypass; CVA: cerebrovascular accident. View Large Table 1: General characteristics of the patient population Variables Overall (N = 7939) Women (N = 2685) Men (N = 5281) Age (years) 69 (60–76) 72 (63–78) 68 (59–75) Weight (kg) 73 (64–82) 64 (56–72) 77 (70–85) Height (cm) 169 (160–174) 160 (156–165) 170 (168–176) Body mass index (kg/m2) 25.7 (23.4–28.4) 24.9 (22–28.1) 26 (23.9–28.6) BMI class  Underweight 165 (2.1) 110 (4.1) 55 (1.0)  Normal weight 3302 (41.6) 1279 (48.1) 2023 (38.3)  Overweight 3183 (40.1) 843 (31.7) 2340 (44.3)  Obesity class I 1014 (12.8) 298 (11.2) 716 (13.6)  Obesity class II–III 275 (3.5) 128 (4.8) 147 (2.8) Hypertension 5080 (64) 1695 (63) 3385 (64) Left ventricular ejection fraction (%) 55 (49–60) 56 (50–62) 55 (46–60) Congestive heart failure 646 (8.1) 265 (10) 381 (7.2) Diabetes on medication 1459 (18.4) 411 (15.5) 1048 (19.8) Serum creatinine (mg/dl) 1.0 (0.8–1.2) 0.85 (0.7–1.1) 1.0 (0.9–1.2) Haematocrit (%) 39 (35.7–41.9) 37.4 (34.1–40) 40 (36.6–42.6) COPD 581 (7.3) 164 (6.2) 417 (7.9) Previous CVA 250 (3.1) 73 (2.7) 177 (3.4) Conditions of surgery  Elective 6674 (84.1) 2290 (86.2) 4384 (83)  Urgent 1120 (14.1) 319 (12) 809 (15.2)  Emergency 145 (1.8) 49 (1.8) 96 (1.8) Redo surgery 605 (7.6) 245 (9.2) 360 (6.8) EuroSCORE II 2.2 (1.2–4.7) 2.5 (1.4–5.6) 2 (1.2–4.4) Type of surgery  Isolated CABG 2446 (30.8) 419 (15.8) 2027 (38.4)  Isolated valve 2447 (30.8) 1040 (39.1) 1407 (26.6)  CABG + valve 1282 (16.1) 407 (15.3) 875 (16.6)  Ascending aorta 606 (7.6) 214 (8.1) 392 (7.4)  Others 1158 (14.6) 605 (22.5) 580 (11) CPB duration (min) 76 (58–104) 78 (59–104) 76 (58–104) Variables Overall (N = 7939) Women (N = 2685) Men (N = 5281) Age (years) 69 (60–76) 72 (63–78) 68 (59–75) Weight (kg) 73 (64–82) 64 (56–72) 77 (70–85) Height (cm) 169 (160–174) 160 (156–165) 170 (168–176) Body mass index (kg/m2) 25.7 (23.4–28.4) 24.9 (22–28.1) 26 (23.9–28.6) BMI class  Underweight 165 (2.1) 110 (4.1) 55 (1.0)  Normal weight 3302 (41.6) 1279 (48.1) 2023 (38.3)  Overweight 3183 (40.1) 843 (31.7) 2340 (44.3)  Obesity class I 1014 (12.8) 298 (11.2) 716 (13.6)  Obesity class II–III 275 (3.5) 128 (4.8) 147 (2.8) Hypertension 5080 (64) 1695 (63) 3385 (64) Left ventricular ejection fraction (%) 55 (49–60) 56 (50–62) 55 (46–60) Congestive heart failure 646 (8.1) 265 (10) 381 (7.2) Diabetes on medication 1459 (18.4) 411 (15.5) 1048 (19.8) Serum creatinine (mg/dl) 1.0 (0.8–1.2) 0.85 (0.7–1.1) 1.0 (0.9–1.2) Haematocrit (%) 39 (35.7–41.9) 37.4 (34.1–40) 40 (36.6–42.6) COPD 581 (7.3) 164 (6.2) 417 (7.9) Previous CVA 250 (3.1) 73 (2.7) 177 (3.4) Conditions of surgery  Elective 6674 (84.1) 2290 (86.2) 4384 (83)  Urgent 1120 (14.1) 319 (12) 809 (15.2)  Emergency 145 (1.8) 49 (1.8) 96 (1.8) Redo surgery 605 (7.6) 245 (9.2) 360 (6.8) EuroSCORE II 2.2 (1.2–4.7) 2.5 (1.4–5.6) 2 (1.2–4.4) Type of surgery  Isolated CABG 2446 (30.8) 419 (15.8) 2027 (38.4)  Isolated valve 2447 (30.8) 1040 (39.1) 1407 (26.6)  CABG + valve 1282 (16.1) 407 (15.3) 875 (16.6)  Ascending aorta 606 (7.6) 214 (8.1) 392 (7.4)  Others 1158 (14.6) 605 (22.5) 580 (11) CPB duration (min) 76 (58–104) 78 (59–104) 76 (58–104) Data are presented as n (%) or median (interquartile range). BMI: body mass index; CABG: coronary artery bypass graft; COPD: chronic obstructive pulmonary disease; CPB: cardiopulmonary bypass; CVA: cerebrovascular accident. View Large Operative mortality was 3.6% in the overall patient population (4.1% in women and 3.5% in men). The crude association between BMI class and operative mortality is shown in Fig. 1. In the overall patient population, the ‘U-shaped’ association between BMI class and mortality is confirmed with the lowest mortality rate for obese class I patients (2.6%) and the highest values at the 2 extremities (4.8% for underweight and 6.5% for obese class II–III). The Pearson’s χ2 analysis confirmed a significant (P = 0.012) impact of the BMI class on operative mortality. To adjust for possible confounders differently stratified in the various BMI classes, a propensity matching was performed. Three BMI-based categories were identified (underweight-normal, overweight-obesity class I and obesity class II–III). The smaller group was obesity class II–III (275 patients). Propensity-matched groups were extracted from the general population on a 2:1 ratio with the obesity class II–III group, with a propensity score based on the age, the expected operative mortality risk (EuroSCORE II) and the surgical complexity (CPB duration). The propensity-matched group was comparable with the exception of conditions intrinsically dependent on BMI (weight, height, diabetes and chronic obstructive pulmonary disease). Within these propensity-matched groups, the BMI-based differences in operative mortality were confirmed; and the ‘U-shaped’ relationship was confirmed with a mortality rate significantly (P = 0.015) higher in underweight-normal patients (4.1%) and obesity class II–III patients (6.5%) than overweight-obesity class I patients (2.4%) (Supplementary Material, Table S1). Figure 1: View largeDownload slide Relationship between the BMI classes and observed operative mortality in the overall population and separately for women and men. Bars are 95% confidence interval. BMI: body mass index. Figure 1: View largeDownload slide Relationship between the BMI classes and observed operative mortality in the overall population and separately for women and men. Bars are 95% confidence interval. BMI: body mass index. There were gender-based differences in BMI–mortality association. In women, the relationship appears ‘J-shaped’ with higher mortality values for obese class II–III patients (9.4%). The relative risk of mortality for obese class II–III patients with respect to the other weight classes is 2.6 (95% CI 1.37–4.81, P = 0.002). In men, the lowest mortality rate (2.5%) was observed in obese class I patients; however, the difference with respect to the other groups is significant only for normal weight patients (relative risk 0.57, 95% CI 0.34–0.95; P = 0.029). A higher mortality rate was observed for underweight patients (5.5%); however, this group is under-represented (55 cases only) and the differences in mortality rate with respect to the other groups are not significant. When considering the BMI as a continuous variable, the association with the operative mortality rate again behaves differently in women and men. The second-order polynomial regression showed the best fit for both the populations (Fig. 2), and the lowest mortality rate was identified at a BMI of 22 kg/m2 for women and 35 kg/m2 for men. The increase in mortality at low values of BMI is similar in men and women; conversely, patterns of high BMI appear to increase the mortality rate to a greater extent in women than in men. Figure 2: View largeDownload slide Unadjusted association (quadratic regression analysis) between the BMI and the observed operative mortality rate of women and men. Dashed lines delimitate the 95% confidence area. BMI: body mass index. Figure 2: View largeDownload slide Unadjusted association (quadratic regression analysis) between the BMI and the observed operative mortality rate of women and men. Dashed lines delimitate the 95% confidence area. BMI: body mass index. A separate analysis was conducted for isolated coronary artery bypass grafting (CABG) surgery and aortic valve replacement (Supplementary Material, Figs S3 and S4). In isolated CABG (N = 2446, women = 419, men = 2027), the effects detected in the overall population were confirmed at an even higher degree, with an increased mortality in men with a low BMI and no increase for high BMI (J-shaped relationship), and a confirmed U-shaped relationship in women. However, in obese women the mortality is greatly increased with values >10% for a BMI >35 kg/m2. In isolated aortic valve replacement (N = 1446, women = 591, men = 875), the relationship is U-shaped for both men and women. Separate analyses for the components of the BMI (weight and height) are presented in Figs 3 and 4. The association between weight and operative mortality (Fig. 3) was again defined by the second-order polynomial regression functions, with the lowest mortality rate identified at 70 kg for women and 100 kg for men. The relationship mirrors that of the BMI with a similar increase in mortality in women and men of low weight, whereas there was a marginal increase in mortality risk in men and a large increase in women with increasing weight. Figure 3: View largeDownload slide Unadjusted association (quadratic regression analysis) between the weight and the observed operative mortality rate of women and men. Dashed lines delimitate the 95% confidence area. Figure 3: View largeDownload slide Unadjusted association (quadratic regression analysis) between the weight and the observed operative mortality rate of women and men. Dashed lines delimitate the 95% confidence area. Figure 4: View largeDownload slide Unadjusted association (linear regression analysis) between the height and the observed operative mortality rate of women and men. Dashed lines delimitate the 95% confidence area. Figure 4: View largeDownload slide Unadjusted association (linear regression analysis) between the height and the observed operative mortality rate of women and men. Dashed lines delimitate the 95% confidence area. The impact of height on the mortality rate (Fig. 4) was defined using linear regression analyses that demonstrated to better fit the association than polynomial regressions (P-values for association 0.002 for women and 0.018 for men), with a mortality rate that decreased with increasing height, which was more pronounced in women. The impact of BMI, weight and height in determining the mortality risk was checked after correction for the age, the EuroSCORE II and the CPB duration separately for women (Table 2) and men (Table 3). The impact of BMI and weight was assessed in terms of absolute (negative or positive) offset from the best (nadir) values identified by the second-order functions of Figs 2 and 3; height was considered as the absolute value. Logistic regression models for operative mortality were applied. Table 2: Crude and adjusted operative mortality risk analysis as a function of body mass index, weight and height in female patients (N = 2658) Factors Regression coefficient Odds ratio 95% CI P-value BMI (crude)a 0.043 1.043 1.000–1.089 0.050 BMI (adjusted)a 0.080 1.084 1.029–1.141 0.002 Age (years) 0.050 1.051 1.025–1.078 0.001 EuroSCORE II 0.079 1.082 1.061–1.104 0.001 CPB duration (min) 0.015 1.016 1.012–1.019 0.001 Weight (crude)b 0.024 1.024 1.001–1.047 0.038 Weight (adjusted)b 0.021 1.021 0.994–1.049 0.133 Age (years) 0.046 1.047 1.022–1.072 0.001 EuroSCORE II 0.076 1.079 1.058–1.101 0.001 CPB duration (min) 0.015 1.015 1.012–1.019 0.001 Height (crude)c −0.047 0.954 0.926–0.984 0.002 Height (adjusted)c −0.053 0.949 0.915–0.983 0.004 Age (years) 0.043 1.044 1.018–1.070 0.001 EuroSCORE II 0.078 1.081 1.060–1.102 0.001 CPB duration (min) 0.015 1.015 1.012–1.019 0.001 Factors Regression coefficient Odds ratio 95% CI P-value BMI (crude)a 0.043 1.043 1.000–1.089 0.050 BMI (adjusted)a 0.080 1.084 1.029–1.141 0.002 Age (years) 0.050 1.051 1.025–1.078 0.001 EuroSCORE II 0.079 1.082 1.061–1.104 0.001 CPB duration (min) 0.015 1.016 1.012–1.019 0.001 Weight (crude)b 0.024 1.024 1.001–1.047 0.038 Weight (adjusted)b 0.021 1.021 0.994–1.049 0.133 Age (years) 0.046 1.047 1.022–1.072 0.001 EuroSCORE II 0.076 1.079 1.058–1.101 0.001 CPB duration (min) 0.015 1.015 1.012–1.019 0.001 Height (crude)c −0.047 0.954 0.926–0.984 0.002 Height (adjusted)c −0.053 0.949 0.915–0.983 0.004 Age (years) 0.043 1.044 1.018–1.070 0.001 EuroSCORE II 0.078 1.081 1.060–1.102 0.001 CPB duration (min) 0.015 1.015 1.012–1.019 0.001 a Units of negative/positive offset from the value of 22 kg/m2. b Negative/positive offset (kg) from the value of 70 kg. c Absolute value (cm). BMI: body mass index; CI: confidence interval; CPB: cardiopulmonary bypass. View Large Table 2: Crude and adjusted operative mortality risk analysis as a function of body mass index, weight and height in female patients (N = 2658) Factors Regression coefficient Odds ratio 95% CI P-value BMI (crude)a 0.043 1.043 1.000–1.089 0.050 BMI (adjusted)a 0.080 1.084 1.029–1.141 0.002 Age (years) 0.050 1.051 1.025–1.078 0.001 EuroSCORE II 0.079 1.082 1.061–1.104 0.001 CPB duration (min) 0.015 1.016 1.012–1.019 0.001 Weight (crude)b 0.024 1.024 1.001–1.047 0.038 Weight (adjusted)b 0.021 1.021 0.994–1.049 0.133 Age (years) 0.046 1.047 1.022–1.072 0.001 EuroSCORE II 0.076 1.079 1.058–1.101 0.001 CPB duration (min) 0.015 1.015 1.012–1.019 0.001 Height (crude)c −0.047 0.954 0.926–0.984 0.002 Height (adjusted)c −0.053 0.949 0.915–0.983 0.004 Age (years) 0.043 1.044 1.018–1.070 0.001 EuroSCORE II 0.078 1.081 1.060–1.102 0.001 CPB duration (min) 0.015 1.015 1.012–1.019 0.001 Factors Regression coefficient Odds ratio 95% CI P-value BMI (crude)a 0.043 1.043 1.000–1.089 0.050 BMI (adjusted)a 0.080 1.084 1.029–1.141 0.002 Age (years) 0.050 1.051 1.025–1.078 0.001 EuroSCORE II 0.079 1.082 1.061–1.104 0.001 CPB duration (min) 0.015 1.016 1.012–1.019 0.001 Weight (crude)b 0.024 1.024 1.001–1.047 0.038 Weight (adjusted)b 0.021 1.021 0.994–1.049 0.133 Age (years) 0.046 1.047 1.022–1.072 0.001 EuroSCORE II 0.076 1.079 1.058–1.101 0.001 CPB duration (min) 0.015 1.015 1.012–1.019 0.001 Height (crude)c −0.047 0.954 0.926–0.984 0.002 Height (adjusted)c −0.053 0.949 0.915–0.983 0.004 Age (years) 0.043 1.044 1.018–1.070 0.001 EuroSCORE II 0.078 1.081 1.060–1.102 0.001 CPB duration (min) 0.015 1.015 1.012–1.019 0.001 a Units of negative/positive offset from the value of 22 kg/m2. b Negative/positive offset (kg) from the value of 70 kg. c Absolute value (cm). BMI: body mass index; CI: confidence interval; CPB: cardiopulmonary bypass. View Large Table 3: Crude and adjusted operative mortality risk analysis as a function of body mass index, weight and height in male patients (N = 5281) Factors Regression coefficient Odds ratio 95% CI P-value BMI (crude)a 0.052 1.054 1.010–1.099 0.015 BMI (adjusted)a 0.032 1.033 0.985–1.083 0.185 Age (years) 0.053 1.054 1.037–1.072 0.001 EuroSCORE II 0.058 1.059 1.044–1.075 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Weight (crude)b 0.024 1.025 1.011–1.038 0.001 Weight (adjusted)b 0.017 1.017 1.001–1.032 0.034 Age (years) 0.051 1.053 1.035–1.071 0.001 EuroSCORE II 0.058 1.059 1.044–1.075 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Height (crude)c −0.025 0.975 0.955–0.996 0.018 Height (adjusted)c −0.015 0.985 0.961–1.010 0.232 Age (years) 0.051 1.053 1.035–1.071 0.001 EuroSCORE II 0.059 1.060 1.045–1.076 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Factors Regression coefficient Odds ratio 95% CI P-value BMI (crude)a 0.052 1.054 1.010–1.099 0.015 BMI (adjusted)a 0.032 1.033 0.985–1.083 0.185 Age (years) 0.053 1.054 1.037–1.072 0.001 EuroSCORE II 0.058 1.059 1.044–1.075 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Weight (crude)b 0.024 1.025 1.011–1.038 0.001 Weight (adjusted)b 0.017 1.017 1.001–1.032 0.034 Age (years) 0.051 1.053 1.035–1.071 0.001 EuroSCORE II 0.058 1.059 1.044–1.075 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Height (crude)c −0.025 0.975 0.955–0.996 0.018 Height (adjusted)c −0.015 0.985 0.961–1.010 0.232 Age (years) 0.051 1.053 1.035–1.071 0.001 EuroSCORE II 0.059 1.060 1.045–1.076 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 a Units of negative/positive offset from the value of 35 kg/m2. b Negative/positive offset (kg) from the value of 100 kg. c Absolute value (cm). BMI: body mass index; CI: confidence interval; CPB: cardiopulmonary bypass. View Large Table 3: Crude and adjusted operative mortality risk analysis as a function of body mass index, weight and height in male patients (N = 5281) Factors Regression coefficient Odds ratio 95% CI P-value BMI (crude)a 0.052 1.054 1.010–1.099 0.015 BMI (adjusted)a 0.032 1.033 0.985–1.083 0.185 Age (years) 0.053 1.054 1.037–1.072 0.001 EuroSCORE II 0.058 1.059 1.044–1.075 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Weight (crude)b 0.024 1.025 1.011–1.038 0.001 Weight (adjusted)b 0.017 1.017 1.001–1.032 0.034 Age (years) 0.051 1.053 1.035–1.071 0.001 EuroSCORE II 0.058 1.059 1.044–1.075 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Height (crude)c −0.025 0.975 0.955–0.996 0.018 Height (adjusted)c −0.015 0.985 0.961–1.010 0.232 Age (years) 0.051 1.053 1.035–1.071 0.001 EuroSCORE II 0.059 1.060 1.045–1.076 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Factors Regression coefficient Odds ratio 95% CI P-value BMI (crude)a 0.052 1.054 1.010–1.099 0.015 BMI (adjusted)a 0.032 1.033 0.985–1.083 0.185 Age (years) 0.053 1.054 1.037–1.072 0.001 EuroSCORE II 0.058 1.059 1.044–1.075 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Weight (crude)b 0.024 1.025 1.011–1.038 0.001 Weight (adjusted)b 0.017 1.017 1.001–1.032 0.034 Age (years) 0.051 1.053 1.035–1.071 0.001 EuroSCORE II 0.058 1.059 1.044–1.075 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Height (crude)c −0.025 0.975 0.955–0.996 0.018 Height (adjusted)c −0.015 0.985 0.961–1.010 0.232 Age (years) 0.051 1.053 1.035–1.071 0.001 EuroSCORE II 0.059 1.060 1.045–1.076 0.001 CPB duration (min) 0.017 1.017 1.015–1.020 0.001 a Units of negative/positive offset from the value of 35 kg/m2. b Negative/positive offset (kg) from the value of 100 kg. c Absolute value (cm). BMI: body mass index; CI: confidence interval; CPB: cardiopulmonary bypass. View Large In women, after adjustment for the confounders, the offset from the best BMI remained independently associated with operative mortality (relative risk increase 8.4% per each unit of negative/positive offset from 22 kg/m2); negative/positive offset from the best weight of 70 kg lost significance; and the height remained independently associated with operative mortality (relative risk decrease of 5.1% per each centimetre increase in height). In men, after adjustment for the confounders, the offset from the best BMI lost significance; the offset from the best weight remained independently associated with operative mortality (relative risk increase of 1.7% per each unit of negative/positive offset from 100 kg); and the height lost significance for association with the operative mortality. The predicted mortality rate in logistic regression adjusted models as a function of height and offset from the best weight is reported in the Supplementary Material, Figs S5 and S6. The predicted mortality declines for increasing the height to a larger extent in women than in men (Supplementary Material, Fig. S5), whereas it increases for increasing the weight offset to a larger extent in men than in women (Supplementary Material, Fig. S6). On the basis of the previous results, a sensitivity analysis was conducted. Tall women were considered as those in the upper quartile of the distribution (≥165 cm), and moderately heavy men as those in the third quartile of distribution (weight 77–95 kg) (Table 4). After correction for the age, the weight, the EuroSCORE II and the CPB duration, tall women had an odds ratio for operative mortality of 0.544 (95% CI 0.309–0.956, P = 0.034), with weight being independently associated with operative mortality. Moderately heavy men had an odds ratio for operative mortality of 0.552 (95% CI 0.378–0.804, P = 0.002), and the height was not an independent factor for operative mortality. Table 4: Operative mortality risk analysis for tall women and moderately heavy men Factors Regression coefficient Odds ratio 95% CI P-value Women (N = 2685)  Height ≥165 cm −0.609 0.544 0.309–0.956 0.034  Weight (kg) 0.019 1.019 1.001–1.038 0.039  Age (years) 0.048 1.049 1.023–1.076 0.001  EuroSCORE II 0.079 1.082 1.061–1.103 0.001  CPB duration (min) 0.015 1.016 1.012–1.019 0.001 Men (N = 5281)  Weight 77–95 kg −0.594 0.552 0.379–0.804 0.002  Height (cm) −0.004 0.996 0.971–1.021 0.741  Age (years) 0.052 1.053 1.035–1.071 0.001  EuroSCORE II 0.058 1.060 1.044–1.076 0.001  CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Factors Regression coefficient Odds ratio 95% CI P-value Women (N = 2685)  Height ≥165 cm −0.609 0.544 0.309–0.956 0.034  Weight (kg) 0.019 1.019 1.001–1.038 0.039  Age (years) 0.048 1.049 1.023–1.076 0.001  EuroSCORE II 0.079 1.082 1.061–1.103 0.001  CPB duration (min) 0.015 1.016 1.012–1.019 0.001 Men (N = 5281)  Weight 77–95 kg −0.594 0.552 0.379–0.804 0.002  Height (cm) −0.004 0.996 0.971–1.021 0.741  Age (years) 0.052 1.053 1.035–1.071 0.001  EuroSCORE II 0.058 1.060 1.044–1.076 0.001  CPB duration (min) 0.017 1.017 1.015–1.020 0.001 CI: confidence interval; CPB: cardiopulmonary bypass. View Large Table 4: Operative mortality risk analysis for tall women and moderately heavy men Factors Regression coefficient Odds ratio 95% CI P-value Women (N = 2685)  Height ≥165 cm −0.609 0.544 0.309–0.956 0.034  Weight (kg) 0.019 1.019 1.001–1.038 0.039  Age (years) 0.048 1.049 1.023–1.076 0.001  EuroSCORE II 0.079 1.082 1.061–1.103 0.001  CPB duration (min) 0.015 1.016 1.012–1.019 0.001 Men (N = 5281)  Weight 77–95 kg −0.594 0.552 0.379–0.804 0.002  Height (cm) −0.004 0.996 0.971–1.021 0.741  Age (years) 0.052 1.053 1.035–1.071 0.001  EuroSCORE II 0.058 1.060 1.044–1.076 0.001  CPB duration (min) 0.017 1.017 1.015–1.020 0.001 Factors Regression coefficient Odds ratio 95% CI P-value Women (N = 2685)  Height ≥165 cm −0.609 0.544 0.309–0.956 0.034  Weight (kg) 0.019 1.019 1.001–1.038 0.039  Age (years) 0.048 1.049 1.023–1.076 0.001  EuroSCORE II 0.079 1.082 1.061–1.103 0.001  CPB duration (min) 0.015 1.016 1.012–1.019 0.001 Men (N = 5281)  Weight 77–95 kg −0.594 0.552 0.379–0.804 0.002  Height (cm) −0.004 0.996 0.971–1.021 0.741  Age (years) 0.052 1.053 1.035–1.071 0.001  EuroSCORE II 0.058 1.060 1.044–1.076 0.001  CPB duration (min) 0.017 1.017 1.015–1.020 0.001 CI: confidence interval; CPB: cardiopulmonary bypass. View Large DISCUSSION Our study confirms that the relationship between BMI and operative mortality in cardiac surgery is U-shaped with the lowest mortality rate at a BMI of about 35 kg/m2. The U-shaped relationship was confirmed after propensity matching, with a significantly lower mortality rate for overweight-obesity class I patients with respect to underweight-normal patients and obesity class II–III patients. This result is in agreement with a very large population study [1], where the mortality rate was very close to the one registered at our institution (3.3% in-hospital vs 3.6% in-hospital/30 days from surgery). Notably, the obesity rate (any class) was much lower in our study (16.3%) than in the UK series (30%). When analysed separately for different genders, the results offered new insights into the matter. Basically, the U-shaped relationship seems to be formed by the combination of 2 J-shaped relationships, where in women the high mortality risk in obese class II–III accounts for the right branch, and in men the high mortality risk in underweight accounts for the left branch of the U-shaped relationship. Additionally, the values of BMI associated with the lowest mortality rate are distinctly different in women (22 kg/m2) and in men (35 kg/m2). This combination of a different gender-based, J-shaped relationship is strongly confirmed for isolated CABG surgery. Overall, the effect of BMI on the outcome is therefore highly different in women versus men. In women, no significant differences in mortality rates are observed in the groups underweight to obese class I, and an important increase is observed for obese class II–III. Therefore, it seems difficult to apply the concept of the ‘obesity paradox’ to female patients: the lowest mortality in women corresponds to a BMI (22 kg/m2) and a weight (70 kg) that certainly do not fit with the definition of obesity. Conversely, in men, this paradox is maintained with the best results obtained at a BMI (35 kg/m2) and a weight (100 kg) that suit well with an obesity pattern. When looking at the single components of the BMI, tall women (≥165 cm) and moderately heavy men (77–95 kg) have the best outcome, after correction for the weight (in women), the height (in men), age, EuroSCORE II and CPB duration, with an operative mortality risk about half that of the other patients. There are other examples, in different settings, of a different gender-based relationship between obesity and outcomes. In a recent study, Vest et al. [12] investigated the obesity paradox in heart failure patients separately for women and men. They confirmed the obesity paradox in the total patient population; while this behaviour was confirmed in women, the hazard ratio for mortality was higher in overweight–obese male patients. These results were partially confirmed in terms of heart failure symptoms in a recent study [13], and again in the medical setting, the obesity paradox was more pronounced in women than in men in a population of patients with myocardial infarction [14]. These results apparently conflict with our findings, where the U-shaped relationship is confirmed in men, but not in women. However, the clinical setting is totally different, and the acute impact of surgery (in a population with only 8% of heart failure patients) is certainly a determinant of these differences. The existing studies exploring the obesity paradox in cardiac surgery do not take into account the effect of gender, and our observation introduces a new perspective into this topic. A recent study [15] on 15 134 patients undergoing cardiac surgery found that, once adjusted for other confounders, the BMI was not a determinant of hospital mortality, whereas the female gender remained an independent risk factor. This suggests that an interaction between gender and BMI as determinants of mortality exists, as confirmed by our specific analysis. There are many possible explanations underlying the observation that the 2 extremes of the BMI distribution are associated with bad outcomes. Basically, a very low BMI may be an expression of frailty; moreover, this condition is associated with a greater impact of haemodilution on CPB, which is a determinant of bad outcomes [16], and with a larger use of allogeneic blood products [4]. However, in our series this applies to men only; it is therefore likely that a low BMI may be associated with a frailer condition more often seen in men than in women. The higher mortality risk in severe (class II–III) obesity is usually attributed to postoperative pulmonary complications, surgical site infections and technical surgical problems (prosthesis–patient mismatch in valve surgery). This severe obesity-related mortality risk appears more pronounced in women than in men, in our series, but we are lacking a clear interpretation of this finding. In our analysis, we could identify a different role of height and weight in women versus men. This introduces the very well-known concept that BMI, per se, is not a good discriminant between an elevated body weight because of the high levels of lean versus fat body mass. More adequate measures include waist circumference, waist–hip ratio, skinfold estimates of % body fat and bioelectrical impedance analysis of body composition [17]. All these measures are lacking in our study (as in previous studies in the setting of cardiac surgery), but the finding of a protective effect of height in women (and not in men) is worth some consideration. Height is included in the waist circumference/height ratio that is considered a good estimate of intra-abdominal fat and cardiometabolic risk [18]. Therefore, it is reasonable to assume that height is inversely associated with cardiovascular risk. In our patient population, height is inversely associated with operative mortality, but this effect is more pronounced and independently associated with mortality in women only. This gender-related behaviour of height was found in other studies. In a large population study [19], the decrease in BMI with increasing height was more pronounced in women than in men: in the tallest compared with the shortest height quartile, it was 0.77 in men (95% CI 0.69–0.86) and 1.98 in women (95% CI 1.89–2.08). The role of height in the calculation of lean body mass, fat mass and % fat mass was recently highlighted in a study based on the National Health and Nutrition Examination Survey [20]. Highly predictive equations for body composition were developed based on age, race, height, weight and waist circumference. However, waist circumference was a strong predictor in men but not in women. The role of height in determining a lower fat mass is, therefore, more relevant in women than in men. The above reported data help in understanding our results which found that height plays a role in determining fat mass and its distribution, and which appeared to be more relevant in women than in men. This role defines height as a protective factor in cardiac surgery, independently from weight in women. From a practical perspective, our data may offer some possible insights. Underweight male patients may probably benefit from specific nutritional and exercise training programmes before cardiac surgery, whenever feasible. Conversely, underweight women may probably not necessarily benefit from this approach. Weight loss programmes before cardiac surgery are indicated for both women and men when belonging to obesity class II–III; however, in women ≥ 165 cm, weight loss is probably less relevant. Limitations There are limitations in our study. The most important is the lack of relevant measures of body fat distribution (waist circumference, waist–hip ratio, skinfold estimates of % body fat and bioelectrical impedance analysis of body composition). Additionally, a frailty score assessment to identify underweight patients with or without frailty was not conducted. CONCLUSION In conclusion, our study demonstrates that there is a gender-based difference for BMI, weight and height as determinants of operative mortality in patients undergoing cardiac surgery. Further studies including more specific measures of fat distribution are warranted. Funding This work was supported by the IRCCS Policlinico San Donato, a Clinical Research Hospital recognized and partially supported by the Italian Ministry of Health. Conflict of interest: none declared. REFERENCES 1 Mariscalco G , Wozniak MJ , Dawson AG , Serraino GF , Porter R , Nath M et al. Body mass index and mortality among adults undergoing cardiac surgery: a nationwide study with a systematic review and meta-analysis . Circulation 2017 ; 135 : 850 – 63 . Google Scholar Crossref Search ADS PubMed 2 Potapov EV , Loebe M , Anker S , Stein J , Bondy S , Nasseri BA et al. Impact of body mass index on outcome in patients after coronary artery bypass grafting with and without valve surgery . Eur Heart J 2003 ; 24 : 1933 – 41 . Google Scholar Crossref Search ADS PubMed 3 Stamou SC , Nussbaum M , Stiegel RM , Reames MK , Skipper ER , Robicsek F et al. Effect of body mass index on outcomes after cardiac surgery: is there an obesity paradox? Ann Thorac Surg 2011 ; 91 : 42 – 7 . Google Scholar Crossref Search ADS PubMed 4 Ranucci M , Pazzaglia A , Bianchini C , Bozzetti G , Isgrò G. Body size, gender, and transfusions as determinants of outcome after coronary operations . Ann Thorac Surg 2008 ; 85 : 481 – 6 . Google Scholar Crossref Search ADS PubMed 5 Flegal KM , Kit BK , Orpana H , Graubard BI. Association of all-cause mortality with overweight and obesity using standard body mass index categories: a systematic review and meta-analysis . JAMA 2013 ; 309 : 71 – 82 . Google Scholar Crossref Search ADS PubMed 6 Angerås O , Albertsson P , Karason K , Råmunddal T , Matejka G , James S et al. Evidence for obesity paradox in patients with acute coronary syndromes: a report from the Swedish Coronary Angiography and Angioplasty Registry . Eur Heart J 2013 ; 34 : 345 – 53 . Google Scholar Crossref Search ADS PubMed 7 Oreopoulos A , Padwal R , Kalantar-Zadeh K , Fonarow GC , Norris CM , McAlister FA. Body mass index and mortality in heart failure: a meta-analysis . Am Heart J 2008 ; 156 : 13 – 22 . Google Scholar Crossref Search ADS PubMed 8 Ho KM , Bertenshaw C , Same S , Schneider M , Williams KA , Godsell T et al. Differential associations between body mass index and outcomes after elective adult cardiac surgery: a linked data cohort study . Anaesth Intensive Care 2013 ; 41 : 573 – 83 . Google Scholar Crossref Search ADS PubMed 9 Habib RH , Zacharias A , Schwann TA , Riordan CJ , Durham SJ , Shah A. Effects of obesity and small body size on operative and long-term outcomes of coronary artery bypass surgery: a propensity-matched analysis . Ann Thorac Surg 2005 ; 79 : 1976 – 86 . Google Scholar Crossref Search ADS PubMed 10 WHO. Physical Status: The Use and Interpretation of Anthropometry: Report of a WHO Expert Committee . WHO Technical Report Series 854. Geneva, Switzerland : World Health Organization , 1995 . 11 Criqui MH , Klauber MR , Barrett-Connor E , Holdbrook MJ , Suarez L , Wingard DL. Adjustment for obesity in studies of cardiovascular disease . Am J Epidemiol 1982 ; 116 : 685 – 91 . Google Scholar Crossref Search ADS PubMed 12 Vest AR , Wu Y , Hachamovitch R , Young JB , Cho L. The heart failure overweight/obesity survival paradox . JACC Heart Fail 2015 ; 3 : 917 – 26 . Google Scholar Crossref Search ADS PubMed 13 Heo S , Moser DK , Pressler SJ , Dunbar SB , Lee KS , Kim J et al. Association between obesity and heart failure symptoms in male and female patients . Clin Obes 2017 ; 7 : 77 – 85 . Google Scholar Crossref Search ADS PubMed 14 Keller K , Munzel T , Ostad MA. Sex-specific differences in mortality and the obesity paradox of patients with myocardial infarction ages >70 y . Nutrition 2018 ; 46 : 124 – 30 . Google Scholar Crossref Search ADS PubMed 15 Hartrumpf M , Kuehnel R-U , Albes JM. The obesity paradox is still there: a risk analysis of over 15000 cardiosurgical patients based on body mass index . Interact CardioVasc Thorac Surg 2017 ; 25 : 18 – 24 . Google Scholar Crossref Search ADS PubMed 16 Habib RH , Zacharias A , Schwann TA , Riordan CJ , Durham SJ , Shah A. Adverse effects of low hematocrit during cardiopulmonary bypass in the adult: should current practice be changed? J Thorac Cardiovasc Surg 2003 ; 125 : 1438 – 50 . Google Scholar Crossref Search ADS PubMed 17 Clark AL , Fonarow GC , Horwich TB. Obesity and the obesity paradox in heart failure . Prog Cardiovasc Dis 2014 ; 56 : 409 – 14 . Google Scholar Crossref Search ADS PubMed 18 Ashwell M , Cole TJ , Dixon AK. Ratio of waist circumference to height is strong predictor of intra-abdominal fat . BMJ 1996 ; 313 : 555 – 60 . Google Scholar Crossref Search ADS 19 Sperrin M , Marshall AD , Higgins V , Renehan AG , Buchan IE. Body mass index relates weight to height differently in women and older adults: serial cross-sectional surveys in England (1992–2011) . J Public Health 2016 ; 38 : 607 – 13 . Google Scholar Crossref Search ADS 20 Lee DH , Keum N , Hu FB , Orav EJ , Rimm EB , Sun Q et al. Development and validation of anthropometric prediction equations for lean body mass, fat mass and percent fat in adults using the National Health and Nutrition Examination Survey (NHANES) 1999-2006 . Br J Nutr 2017 ; 118 : 858 – 66 . Google Scholar Crossref Search ADS PubMed © The Author(s) 2019. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Journal

European Journal of Cardio-Thoracic Surgery – Oxford University Press

Published: Jul 1, 2019

There are no references for this article.