World J Surg (2015) 39:2376–2385 DOI 10.1007/s00268-015-3112-7
ORIGINAL SCIENTIFIC REPORT
The Effect of Body Mass Index on Perioperative Outcomes After Major Surgery: Results from the National Surgical Quality Improvement Program (ACS-NSQIP) 2005–2011 Akshay Sood1,3 • Firas Abdollah1 • Jesse D. Sammon1 • Kaustav Majumder2 • Marianne Schmid3 • James O. Peabody1 • Mark A. Preston3 • Adam S. Kibel3 Mani Menon1 • Quoc-Dien Trinh3
•
Published online: 10 June 2015 Ó Socie´te´ Internationale de Chirurgie 2015
Abstract Background Obesity is associated with poor surgical outcomes and disparity in access-to-care. There is a lack of quality data on the effect of body mass index (BMI) on perioperative outcomes. Accordingly, we sought to determine the procedure specific, independent-effect of BMI on 30-day perioperative outcomes in patients undergoing major surgery. Methods Participants included individuals undergoing one of 16 major surgery (cardiovascular, orthopedic, oncologic; n = 141,802) recorded in the ACS-NSQIP (2005–2011). Outcomes evaluated included complications, blood transfusion, length-of-stay (LOS), re-intervention, readmission, and perioperative mortality. Multivariableregression models assessed the independent-effect of BMI on outcomes. Results Nearly, 74 % of patients had a BMI disturbance; the majority being overweight (35.3 %) or obese (29.8 %). Morbidly obese patients constituted a small but significant proportion of the patients (5.7 %; n = 8067). In adjusted-analyses, morbidly obese patients had significantly increased odds of wound complications in 15 of the examined procedures, of renal complications after 6-procedures, of thromboembolism after 5-procedures, of pulmonary, septic and UTI complications after 2-procedures, and of cardiovascular complications after CABG. Conversely, obese/overweight patients, except for increased odds of wound complications after select procedures, had significantly decreased odds of perioperative mortality, prolonged-LOS and blood transfusion relative to normal BMI patients after 4, 8, and 9 of the examined procedures. Conclusions The prevalence of BMI derangements in surgical patients is high. The effect of BMI on outcomes is procedure specific. Patients with BMI between 18.5 and 40-kg/m2 at time of surgery fare equally well with regard to complications and mortality. However, morbidly obese patients are at-risk for postsurgical complications and targeted preoperative-optimization may improve outcomes and attenuate disparity in access-to-care.
Electronic supplementary material The online version of this article (doi:10.1007/s00268-015-3112-7) contains supplementary material, which is available to authorized users. & Quoc-Dien Trinh [email protected]; [email protected] 1
Center for Outcomes Research, Analytics and Evaluation (VCORE), Vattikuti Urology Institute, Henry Ford Health System, Detroit, MI, USA
123
2
Department of Surgery, University of Minnesota, Minneapolis, MN, USA
3
Division of Urologic Surgery, Center for Surgery and Public Health, Brigham and Women’s Hospital, Harvard Medical School, 45 Francis St, ASB II-3, Boston, MA 02115, USA
World J Surg (2015) 39:2376–2385
2377
Introduction
Methods
Obesity is a growing worldwide epidemic [1, 2]. The prevalence of obesity among adults in the U.S. has been steadily increasing for the past several decades such that today 1 in every 3 adults is obese and 1 in every 20 is morbidly obese [3]. Obesity is associated with an increased risk for developing serious medical conditions, including hypertension, dyslipidemia, diabetes, heart disease, degenerative arthritis, and cancers [4–7]. Many of these conditions and/or their associated complications/sequelae mandate surgical management. However, the decision to operate on these multimorbid patients is not straightforward as perioperative outcomes in patients with body mass index (BMI) derangements are often inferior relative to the general surgical population [8–14]. As such, morbid-obesity is a relative contraindication for many procedures/approaches including organ transplantation [15]. Surgeons today are thus faced with the dilemma of whether to carry on with immediate surgery to avoid disease progression or to delay treatment and gain benefit from preoperative weight optimization. These decisions will only be further complicated by the Centers for Medicare and Medicaid Services (CMS)’ upcoming non-reimbursement policy for hospital-acquired complications, common in patients with BMI derangements [16, 17]. Prior studies evaluating surgical outcomes in patients with weight disorders have shown unfavorable outcomes in morbidly obese and underweight patients, while mild obesity (BMI = 30–35 kg/m2) has been linked with equivocal or sometimes superior outcomes—coined the ‘‘obesity paradox’’ [8–14]. However, many of these studies were single-institution series [12–14], focused on a single surgical specialty [9, 10], or did not assess outcomes in a procedure-specific manner [8, 11]. Moreover, the BMI cutoffs used were not consistent between reports [9]. The most comprehensive report to date, by Mullen et al. [11], evaluated the role of BMI in patients undergoing cancer surgery, but the data reported in that study is now a decade old. On the basis of these considerations, we sought to assess the impact of BMI on 30-day outcomes within a contemporary (2005–2011) and comprehensive group of 16 major surgeries. We utilized data from the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP), a multi-institutional quality improvement initiative. We aimed to address two questions: who are the patients ‘at-risk’ for surgical complications, and thus, may benefit from preoperative optimization, and what may be an acceptable BMI range from a surgical perspective.
Data source The current study relies on the ACS-NSQIP Participant User Files [18]. ACS-NSQIP is an American College of Surgeons led quality improvement initiative; the ACS-NSQIP database contains risk-adjusted surgical patient data from the participant hospitals. Trained Surgical Clinical Reviewers (SCR) prospectively collect the ACS-NSQIP data and validated data from patients’ medical charts allows quantification of 30-day surgical outcomes, including post-discharge information [19]. Rigorous SCR training and regular audits ensure accuracy, completeness, and quality of the data [20, 21]. In 2011, the ACS-NSQIP included data from 315 participant institutions with more than 1.7 million cases having been contributed. Study population We focused on 16 major cancer and non-cancer surgeries. The rationale behind the selection of these procedures was to provide a comprehensive assessment of the most common major cancer and non-cancer surgeries performed in the U.S., to ensure the relevance of this study to the surgical community at large [22]. Patients were identified in the ACS-NSQIP (2005–2011) using Current Procedural Terminology (CPT) codes for cardiovascular-surgery: carotid endarterectomy (CEA), coronary artery bypass grafting (CABG), lower extremity (LE) bypass surgery, abdominal aortic aneurysm (AAA) repair, cardiac valve repair/replacement; orthopedicsurgery: total hip or knee arthroplasty (THA/TKA); and oncologic-surgery: esophagectomy, gastrectomy, pancreatectomy, colectomy, radical prostatectomy, cystectomy, nephrectomy, hysterectomy or pneumonectomy. For patients undergoing oncologic interventions the extraction was limited to a concomitant cancer diagnosis code, and only patients with complete information for baseline parameters were included (eTable 1: CPT and ICD-9 codes). Patients were stratified according to the World Health Organization (WHO) classification of BMI; underweight (\18.5 kg/m2), normal (18.5–24.9 kg/m2), overweight (25–29.9 kg/m2), obese or class I–II obesity (30–39.9 kg/ m2), and morbidly obese or class III obesity (C40 kg/m2). Patients with missing BMI information (n = 542) were excluded from the analysis. Covariates For each patient, age, gender, race, smoking status, alcohol consumption, comorbidities including history of hypertension, diabetes, cardiopulmonary disease, cerebrovascular disease and others (liver disease, peripheral vascular disease,
123
2378
and chronic kidney disease), preoperative hematocrit (Hct), serum creatinine (SCr), American Society of Anesthesiologists (ASA) score, operative time, and anesthesia type were recorded. Furthermore, approach (open vs. minimally invasive) was assessed wherever relevant (see below). Endpoints Postoperative complications were grouped according to previously reported methodology: [23] cardiovascular (postoperative cardiac arrest and myocardial infarction), pulmonary (pneumonia, postoperative reintubation, and ventilatory support [48-h), neurological (coma [24-h and cerebrovascular accident), thromboembolic [deep venous thrombosis (DVT) and pulmonary embolism], septic (sepsis and septic shock), renal (acute renal failure and progressive renal insufficiency), urinary tract infections (UTI) and wound complications (superficial, deep, and organ-space surgical site infections, and wound dehiscence). Additional outcomes examined were the need for transfusion, re-intervention, prolonged length-of-stay (pLOS), readmission, and perioperative mortality. Procedure-specific pLOS was defined as a length-of-stay greater than the 75th percentile (eTable 2). 30-day readmission data were only available starting January 1st 2011; therefore, only 1 year of data (n = 21,719) were analyzed for this outcome. Perioperative mortality was defined as death within 30-days of surgery. Statistical analyses Descriptive statistics of categorical variables focused on frequencies and proportions. Medians and interquartile ranges were reported for continuously coded variables. Procedurespecific multivariable-regression models were used to analyze the association between BMI and outcomes. Models controlled for age, gender, race, smoking status, alcohol intake, comorbidities, pre-operative SCr and Hct, ASA scre, anesthesia type and operative time. For colectomy, nephrectomy, prostatectomy and hysterectomy, additional adjustment was performed for open vs. minimally invasive approach. All statistical analyses were performed using the R-statistical package (R Foundation, Vienna, Austria), with a two-sided significance level set at p \ 0.05. An IRB-waiver was obtained prior to conducting this study, in accordance with institutional regulation when dealing with de-identified administrative data.
Results Baseline characteristics Table 1 summarizes the descriptive characteristics in 141,802 patients who underwent one of 16 major surgical
123
World J Surg (2015) 39:2376–2385
procedures, stratified according to BMI (cardiovascular: 49.4 %, orthopedic: 17.8 %, oncologic: 32.8 %). Nearly, 74 % of all patients had a BMI disturbance with the majority being either overweight (35.3 %) or obese (29.8 %). Morbidly obese and underweight patients represented a small but significant proportion of the patients at 5.7 and 2.8 %, respectively. Perioperative outcomes Univariable outcomes stratified by BMI are shown in Table 2. Underweight and morbidly obese patients experienced more perioperative complications when compared to patients with normal BMI (18.9 and 16 % vs. 14.4 %, respectively, p \ 0.001). Specifically, underweight patients experienced more cardiovascular (1.9 vs. 1.5 %, p \ 0.001), pulmonary (8.8 vs. 5.4 %, p \ 0.001), neurologic (1.4 vs. 1 %, p \ 0.001), and septic complications (6.0 vs. 4.1 %, p \ 0.001) when compared to normal weight patients. Underweight patients also experienced more re-interventions (8.3 vs. 6.1 %, p \ 0.001) and readmissions (11.8 vs. 9.6 %, p \ 0.001). Conversely, morbidly obese patients experienced more wound (8.3 vs. 5.2 %, p \ 0.001), renal (1.8 vs. 1.1 %, p \ 0.001) and thromboembolic complications (1.8 vs. 1.2 %, p \ 0.001) in comparison to normal weight patients. Finally, overweight and obese patients did not fare poorly relative to normal BMI patients; both groups rather experienced lower complication, re-intervention and readmission rates (p \ 0.001 for all). Multivariable analyses eTables 3 and 4 show multivariable procedure-specific regression analysis. Forest plots (for underweight and morbidly obese only) illustrating the odds of experiencing an adverse outcome are shown in Fig. 1a–c for oncological procedures, and in Fig. 2a, b for non-oncological procedures. In adjusted-analyses, underweight patients were more likely to develop pulmonary complications after cardiovascular or thoracic procedures (CEA OR 1.53, p = 0.036; CABG OR 2.37, p = 0.010; valve repair/replacement OR 2.24, p = 0.021; LE Bypass OR 1.46, p = 0.007; AAA repair OR 1.5, p = 0.021; Pneumonectomy OR 1.84, p = 0.047). These patients undergoing CEA, LE bypass, THA, TKA, and gastrectomy had increased odds of experiencing pLOS (p \ 0.001 for all). The odds of 30-day perioperative mortality in these patients were also increased in a procedure-specific manner; CABG (OR 3.76, p = 0.009), LE bypass (OR 1.94, p = 0.001) and esophagectomy (OR 4.6, p = 0.001). Morbidly obese patients had significantly higher odds of developing a complication after 9 out of 16 procedures
World J Surg (2015) 39:2376–2385
2379
Table 1 Descriptive characteristics of 141,802 patients ([16 years old) undergoing one of the 16 major surgical procedures (cancer or noncancer) stratified according to BMI; National Surgical Quality Improvement Program (NSQIP) Database 2005–2011 Variables No of patients ( %) Age; years, median (IQR)
b
Overall
\18.5
18.5–24.9
25.0–29.9
30–39.9
C40
p
141,802 (100)
3,971 (2.8)
37,422 (26.4)
50,422 (35.3)
42,319 (29.8)
8067 (5.7)
–
68.0 (16)
71.0 (18)
71.0 (17)
69.0 (16)
67.0 (15)
63.0 (13)
ORIGINAL SCIENTIFIC REPORT
The Effect of Body Mass Index on Perioperative Outcomes After Major Surgery: Results from the National Surgical Quality Improvement Program (ACS-NSQIP) 2005–2011 Akshay Sood1,3 • Firas Abdollah1 • Jesse D. Sammon1 • Kaustav Majumder2 • Marianne Schmid3 • James O. Peabody1 • Mark A. Preston3 • Adam S. Kibel3 Mani Menon1 • Quoc-Dien Trinh3
•
Published online: 10 June 2015 Ó Socie´te´ Internationale de Chirurgie 2015
Abstract Background Obesity is associated with poor surgical outcomes and disparity in access-to-care. There is a lack of quality data on the effect of body mass index (BMI) on perioperative outcomes. Accordingly, we sought to determine the procedure specific, independent-effect of BMI on 30-day perioperative outcomes in patients undergoing major surgery. Methods Participants included individuals undergoing one of 16 major surgery (cardiovascular, orthopedic, oncologic; n = 141,802) recorded in the ACS-NSQIP (2005–2011). Outcomes evaluated included complications, blood transfusion, length-of-stay (LOS), re-intervention, readmission, and perioperative mortality. Multivariableregression models assessed the independent-effect of BMI on outcomes. Results Nearly, 74 % of patients had a BMI disturbance; the majority being overweight (35.3 %) or obese (29.8 %). Morbidly obese patients constituted a small but significant proportion of the patients (5.7 %; n = 8067). In adjusted-analyses, morbidly obese patients had significantly increased odds of wound complications in 15 of the examined procedures, of renal complications after 6-procedures, of thromboembolism after 5-procedures, of pulmonary, septic and UTI complications after 2-procedures, and of cardiovascular complications after CABG. Conversely, obese/overweight patients, except for increased odds of wound complications after select procedures, had significantly decreased odds of perioperative mortality, prolonged-LOS and blood transfusion relative to normal BMI patients after 4, 8, and 9 of the examined procedures. Conclusions The prevalence of BMI derangements in surgical patients is high. The effect of BMI on outcomes is procedure specific. Patients with BMI between 18.5 and 40-kg/m2 at time of surgery fare equally well with regard to complications and mortality. However, morbidly obese patients are at-risk for postsurgical complications and targeted preoperative-optimization may improve outcomes and attenuate disparity in access-to-care.
Electronic supplementary material The online version of this article (doi:10.1007/s00268-015-3112-7) contains supplementary material, which is available to authorized users. & Quoc-Dien Trinh [email protected]; [email protected] 1
Center for Outcomes Research, Analytics and Evaluation (VCORE), Vattikuti Urology Institute, Henry Ford Health System, Detroit, MI, USA
123
2
Department of Surgery, University of Minnesota, Minneapolis, MN, USA
3
Division of Urologic Surgery, Center for Surgery and Public Health, Brigham and Women’s Hospital, Harvard Medical School, 45 Francis St, ASB II-3, Boston, MA 02115, USA
World J Surg (2015) 39:2376–2385
2377
Introduction
Methods
Obesity is a growing worldwide epidemic [1, 2]. The prevalence of obesity among adults in the U.S. has been steadily increasing for the past several decades such that today 1 in every 3 adults is obese and 1 in every 20 is morbidly obese [3]. Obesity is associated with an increased risk for developing serious medical conditions, including hypertension, dyslipidemia, diabetes, heart disease, degenerative arthritis, and cancers [4–7]. Many of these conditions and/or their associated complications/sequelae mandate surgical management. However, the decision to operate on these multimorbid patients is not straightforward as perioperative outcomes in patients with body mass index (BMI) derangements are often inferior relative to the general surgical population [8–14]. As such, morbid-obesity is a relative contraindication for many procedures/approaches including organ transplantation [15]. Surgeons today are thus faced with the dilemma of whether to carry on with immediate surgery to avoid disease progression or to delay treatment and gain benefit from preoperative weight optimization. These decisions will only be further complicated by the Centers for Medicare and Medicaid Services (CMS)’ upcoming non-reimbursement policy for hospital-acquired complications, common in patients with BMI derangements [16, 17]. Prior studies evaluating surgical outcomes in patients with weight disorders have shown unfavorable outcomes in morbidly obese and underweight patients, while mild obesity (BMI = 30–35 kg/m2) has been linked with equivocal or sometimes superior outcomes—coined the ‘‘obesity paradox’’ [8–14]. However, many of these studies were single-institution series [12–14], focused on a single surgical specialty [9, 10], or did not assess outcomes in a procedure-specific manner [8, 11]. Moreover, the BMI cutoffs used were not consistent between reports [9]. The most comprehensive report to date, by Mullen et al. [11], evaluated the role of BMI in patients undergoing cancer surgery, but the data reported in that study is now a decade old. On the basis of these considerations, we sought to assess the impact of BMI on 30-day outcomes within a contemporary (2005–2011) and comprehensive group of 16 major surgeries. We utilized data from the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP), a multi-institutional quality improvement initiative. We aimed to address two questions: who are the patients ‘at-risk’ for surgical complications, and thus, may benefit from preoperative optimization, and what may be an acceptable BMI range from a surgical perspective.
Data source The current study relies on the ACS-NSQIP Participant User Files [18]. ACS-NSQIP is an American College of Surgeons led quality improvement initiative; the ACS-NSQIP database contains risk-adjusted surgical patient data from the participant hospitals. Trained Surgical Clinical Reviewers (SCR) prospectively collect the ACS-NSQIP data and validated data from patients’ medical charts allows quantification of 30-day surgical outcomes, including post-discharge information [19]. Rigorous SCR training and regular audits ensure accuracy, completeness, and quality of the data [20, 21]. In 2011, the ACS-NSQIP included data from 315 participant institutions with more than 1.7 million cases having been contributed. Study population We focused on 16 major cancer and non-cancer surgeries. The rationale behind the selection of these procedures was to provide a comprehensive assessment of the most common major cancer and non-cancer surgeries performed in the U.S., to ensure the relevance of this study to the surgical community at large [22]. Patients were identified in the ACS-NSQIP (2005–2011) using Current Procedural Terminology (CPT) codes for cardiovascular-surgery: carotid endarterectomy (CEA), coronary artery bypass grafting (CABG), lower extremity (LE) bypass surgery, abdominal aortic aneurysm (AAA) repair, cardiac valve repair/replacement; orthopedicsurgery: total hip or knee arthroplasty (THA/TKA); and oncologic-surgery: esophagectomy, gastrectomy, pancreatectomy, colectomy, radical prostatectomy, cystectomy, nephrectomy, hysterectomy or pneumonectomy. For patients undergoing oncologic interventions the extraction was limited to a concomitant cancer diagnosis code, and only patients with complete information for baseline parameters were included (eTable 1: CPT and ICD-9 codes). Patients were stratified according to the World Health Organization (WHO) classification of BMI; underweight (\18.5 kg/m2), normal (18.5–24.9 kg/m2), overweight (25–29.9 kg/m2), obese or class I–II obesity (30–39.9 kg/ m2), and morbidly obese or class III obesity (C40 kg/m2). Patients with missing BMI information (n = 542) were excluded from the analysis. Covariates For each patient, age, gender, race, smoking status, alcohol consumption, comorbidities including history of hypertension, diabetes, cardiopulmonary disease, cerebrovascular disease and others (liver disease, peripheral vascular disease,
123
2378
and chronic kidney disease), preoperative hematocrit (Hct), serum creatinine (SCr), American Society of Anesthesiologists (ASA) score, operative time, and anesthesia type were recorded. Furthermore, approach (open vs. minimally invasive) was assessed wherever relevant (see below). Endpoints Postoperative complications were grouped according to previously reported methodology: [23] cardiovascular (postoperative cardiac arrest and myocardial infarction), pulmonary (pneumonia, postoperative reintubation, and ventilatory support [48-h), neurological (coma [24-h and cerebrovascular accident), thromboembolic [deep venous thrombosis (DVT) and pulmonary embolism], septic (sepsis and septic shock), renal (acute renal failure and progressive renal insufficiency), urinary tract infections (UTI) and wound complications (superficial, deep, and organ-space surgical site infections, and wound dehiscence). Additional outcomes examined were the need for transfusion, re-intervention, prolonged length-of-stay (pLOS), readmission, and perioperative mortality. Procedure-specific pLOS was defined as a length-of-stay greater than the 75th percentile (eTable 2). 30-day readmission data were only available starting January 1st 2011; therefore, only 1 year of data (n = 21,719) were analyzed for this outcome. Perioperative mortality was defined as death within 30-days of surgery. Statistical analyses Descriptive statistics of categorical variables focused on frequencies and proportions. Medians and interquartile ranges were reported for continuously coded variables. Procedurespecific multivariable-regression models were used to analyze the association between BMI and outcomes. Models controlled for age, gender, race, smoking status, alcohol intake, comorbidities, pre-operative SCr and Hct, ASA scre, anesthesia type and operative time. For colectomy, nephrectomy, prostatectomy and hysterectomy, additional adjustment was performed for open vs. minimally invasive approach. All statistical analyses were performed using the R-statistical package (R Foundation, Vienna, Austria), with a two-sided significance level set at p \ 0.05. An IRB-waiver was obtained prior to conducting this study, in accordance with institutional regulation when dealing with de-identified administrative data.
Results Baseline characteristics Table 1 summarizes the descriptive characteristics in 141,802 patients who underwent one of 16 major surgical
123
World J Surg (2015) 39:2376–2385
procedures, stratified according to BMI (cardiovascular: 49.4 %, orthopedic: 17.8 %, oncologic: 32.8 %). Nearly, 74 % of all patients had a BMI disturbance with the majority being either overweight (35.3 %) or obese (29.8 %). Morbidly obese and underweight patients represented a small but significant proportion of the patients at 5.7 and 2.8 %, respectively. Perioperative outcomes Univariable outcomes stratified by BMI are shown in Table 2. Underweight and morbidly obese patients experienced more perioperative complications when compared to patients with normal BMI (18.9 and 16 % vs. 14.4 %, respectively, p \ 0.001). Specifically, underweight patients experienced more cardiovascular (1.9 vs. 1.5 %, p \ 0.001), pulmonary (8.8 vs. 5.4 %, p \ 0.001), neurologic (1.4 vs. 1 %, p \ 0.001), and septic complications (6.0 vs. 4.1 %, p \ 0.001) when compared to normal weight patients. Underweight patients also experienced more re-interventions (8.3 vs. 6.1 %, p \ 0.001) and readmissions (11.8 vs. 9.6 %, p \ 0.001). Conversely, morbidly obese patients experienced more wound (8.3 vs. 5.2 %, p \ 0.001), renal (1.8 vs. 1.1 %, p \ 0.001) and thromboembolic complications (1.8 vs. 1.2 %, p \ 0.001) in comparison to normal weight patients. Finally, overweight and obese patients did not fare poorly relative to normal BMI patients; both groups rather experienced lower complication, re-intervention and readmission rates (p \ 0.001 for all). Multivariable analyses eTables 3 and 4 show multivariable procedure-specific regression analysis. Forest plots (for underweight and morbidly obese only) illustrating the odds of experiencing an adverse outcome are shown in Fig. 1a–c for oncological procedures, and in Fig. 2a, b for non-oncological procedures. In adjusted-analyses, underweight patients were more likely to develop pulmonary complications after cardiovascular or thoracic procedures (CEA OR 1.53, p = 0.036; CABG OR 2.37, p = 0.010; valve repair/replacement OR 2.24, p = 0.021; LE Bypass OR 1.46, p = 0.007; AAA repair OR 1.5, p = 0.021; Pneumonectomy OR 1.84, p = 0.047). These patients undergoing CEA, LE bypass, THA, TKA, and gastrectomy had increased odds of experiencing pLOS (p \ 0.001 for all). The odds of 30-day perioperative mortality in these patients were also increased in a procedure-specific manner; CABG (OR 3.76, p = 0.009), LE bypass (OR 1.94, p = 0.001) and esophagectomy (OR 4.6, p = 0.001). Morbidly obese patients had significantly higher odds of developing a complication after 9 out of 16 procedures
World J Surg (2015) 39:2376–2385
2379
Table 1 Descriptive characteristics of 141,802 patients ([16 years old) undergoing one of the 16 major surgical procedures (cancer or noncancer) stratified according to BMI; National Surgical Quality Improvement Program (NSQIP) Database 2005–2011 Variables No of patients ( %) Age; years, median (IQR)
b
Overall
\18.5
18.5–24.9
25.0–29.9
30–39.9
C40
p
141,802 (100)
3,971 (2.8)
37,422 (26.4)
50,422 (35.3)
42,319 (29.8)
8067 (5.7)
–
68.0 (16)
71.0 (18)
71.0 (17)
69.0 (16)
67.0 (15)
63.0 (13)
