Objective
The objective of the study was to assess the impact of body mass index (BMI) on 30 day perioperative morbidity following abdominal hysterectomy.
Study Design
The 2006-2010 National Surgical Quality Improvement Program data registry was retrospectively reviewed for patients undergoing abdominal hysterectomy. Logistic regression was used to investigate the relationship between BMI and postoperative complications.
Results
A total of 9917 patients were captured, of which, 2219 were of an ideal BMI, 2765 were overweight, and 4933 were obese. Complications occurred in 11.3% of the procedures, with obese patients experiencing significantly higher rates of morbidity compared with overweight and nonobese patients (13.2%, 9.7%, and 9.0%, respectively; P < .001). Surgical complications were rare; however, a significant step-wise progression was observed with increasing BMI ( P < .001). The rate of reoperations and overall medical complication did not differ among cohorts, although the incidence of deep vein thromboses (DVTs) was notably elevated in obese and overweight patients ( P = .032). Adjusted odds ratios (ORs) found both overweight and obese patients to be at a significantly higher risk of surgical complications (OR, 1.6 and 3.0, respectively) and wound infections (OR, 1.7 and 3.0, respectively). Overweight patients were also at higher risk for DVTs (OR, 4.6) and obese patients for overall morbidity (OR, 1.4) and wound disruption (OR, 3.6).
Conclusion
Obese and overweight patients demonstrated an increased risk for periorperative morbidity following abdominal hysterectomies.
With more than 600,000 procedures performed each year in the United States alone, the hysterectomy is the single most common gynecological operation in the world. Traditionally, the abdominal approach to hysterectomy has been the most common, accounting for nearly two-thirds of the procedures in 2003. Recently the less invasive vaginal and laparoscopic approaches have gained popularity and are now the recommended approaches for treating benign diseases by the American Congress of Obstetricians and Gynecologists. The abdominal approach, however, still remains an important option, especially in patients contraindicated to vaginal and laparoscopic hysterectomies.
With the increasing prevalence of obesity in the United States, it is becoming increasingly important to characterize the effects of body mass index (BMI) on morbidity following hysterectomy. Although an elevated BMI is anecdotally regarded by many gynecologists as a surgical risk factor, studies have repeatedly shown that obesity is actually not a contraindication to the laparoscopic approach. Attempts at examining the relationship in the abdominal approach, however, have yielded inconsistent results. Previous studies have been limited by many factors including the method of data collection and small cohorts. Furthermore, these studies often collected data from a signal surgical center, severely limiting the generalizability of their results.
The National Surgical Quality Improvement Program (NSQIP) was established by the American College of Surgeons (ACS) as a means for quantifying and improving surgical outcomes. Data were prospectively collected by trained surgical nurses at more than 250 institutions throughout the United States from randomly selected patients. Utilizing the comprehensive nature of the NSQIP registry, the objective of this study was to evaluate the role of BMI in a large cohort of patients undergoing an abdominal hysterectomy.
Materials and Methods
Data acquisition and patient selection
Data from the 2006-2010 ACS NSQIP participant use files were retrospectively analyzed. Data collection methods for the NSQIP registry have been previously described in detail. In brief, 240 variables, including patient demographics, comorbidities, preoperative laboratory values, and perioperative details, are prospectively collected for each patient. Postoperative outcomes are tracked for 30 days following the primary operation. Data are independently abstracted by trained surgical nurses and are subject to random audits providing high-quality, standardized data that have demonstrated a disagreement rate of less than 1.8%.
To identify patients undergoing abdominal hysterectomy, the NSQIP registry was queried using the primary Current Procedural Terminology codes pertaining to the abdominal hysterectomy procedure (58150, 58152, 58180, 58200, 58210, and 58240). Patients coded with multiple hysterectomy types and those with an unknown BMI were excluded.
Outcomes
Primary outcomes of interest were overall morbidity, surgical complications, medical complications, and reoperations within 30 days. Overall morbidity was defined as the presence of 1 or more of the following complications: superficial, deep, and organ/space surgical site infections (SSIs), wound dehiscence, deep-vein thrombosis (DVT), pulmonary embolism (PE), unplanned reintubation, ventilator dependence longer than 48 hours, renal insufficiency, acute renal failure, coma, stroke, cardiac arrest, myocardial infarction, peripheral nerve injury, pneumonia, urinary tract infection (UTI), bleeding requiring transfusion, and sepsis/septic shock. Surgical complications consisted of superficial, deep, and organ/space SSI and wound dehiscence. Medical complications consisted of DVT, PE, unplanned reintubation, ventilator dependence longer than 48 hours, renal insufficiency, acute renal failure, coma, stroke, cardiac arrest, myocardial infarction, peripheral nerve injury, pneumonia, urinary tract infection, bleeding requiring transfusion, and sepsis/septic shock. All outcomes were tracked for 30 days postoperatively and used as defined in the NSQIP user guide.
Risk adjustments
Variables for risk adjustment included patient demographics, comorbidities, and total operative length. Patient demographics included age, BMI, race, outpatient status (as per each participating hospital’s own definition of inpatient and outpatient status), smoking (defined as active smoking over the year prior to the index operation), steroid use, radiotherapy within 90 days of the operation, chemotherapy within 30 days of the operation, and prior operations within 30 days (defined as any major surgical procedure performed within 30 days prior to the assessed operation utilizing general, spinal, or epidural anesthesia).
In accordance with the World Health Organization’s stratification, overweight was defined as a BMI from 25 to 29.9 kg/m 2 and obese was defined as a BMI of 30 kg/m 2 or greater. Comorbidities of interest included diabetes, dyspnea, hypertension, chronic obstructive pulmonary disorder (COPD), congestive heart failure, bleeding disorders, previous percutaneous coronary intervention (PCI) or cardiac surgery, and previous stroke or transient ischemic attack (TIA). Additionally, the sum of the relative value units for additional procedures were also used to adjust for added complexity and risk of concurrent procedures, as has been described previously.
Statistical analysis
Descriptive statistics and complication profiles were calculated for the study population using Pearson’s χ 2 and analysis of variance tests for categorical and quantitative variables, respectively. Multivariate regression models were used to control for potential confounding variables and to identify independent risk factors for postoperative complications. This analysis was used to quantify the association between different BMI categories and overall morbidity, surgical complications, wound infections, wound disruptions, medical complications, DVT, and reoperations, comparing both the overweight and obese cohorts to patients with an ideal BMI for each one.
Preoperative variables underwent bivariate screening using Pearson’s χ 2 and independent Student t tests for categorical and continuous variables, respectively, and variables with a significance value of P < .2 were included in the regression models. To improve model precision, preoperative variables with fewer than 10 events were excluded from the models. Hosmer-Lemmeshow (HL) and C statistics were computed to assess the models’ calibration and discrimination.
Results
Study population
A total of 9917 patients underwent an abdominal hysterectomy during the study period and met the inclusion criteria, of which, 2219 had an ideal BMI, 2765 were overweight, and 4933 were obese. A comparison of patient characteristics is summarized in Table 1 . There were no significant differences in the distribution of age among cohorts; however, smoking status and neoadjuvant chemotherapy both demonstrated a stepwise decrease in prevalence from ideal to overweight to obese ( P < .001). Of the analyzed comorbidities, a significant difference in the distribution of patients was observed only for hypertension, dyspnea, and diabetes ( P < .001). There were no also significant differences in the distributions of COPD, congestive heart failure, bleeding disorders, previous PCI or cardiac surgery, and previous stroke or TIA ( Table 1 ).
| Demographic | Ideal | Overweight | Obese | P value |
|---|---|---|---|---|
| BMI <25 kg/m 2 | BMI ≤25 kg/m 2 to <30 kg/m 2 | BMI ≤30 kg/m 2 | ||
| (n = 2219) | (n = 2765) | (n = 4933) | ||
| Age, y | 48.8 ± 11.7 | 49.3 ± 11.3 | 49.0 ± 11.0 | .353 |
| BMI, kg/m 2 | 22.29 ± 1.98 | 27.53 ± 1.42 | 37.67 ± 6.93 | < .001 |
| Race, % | < .001 | |||
| White | 69.58 | 67.88 | 66.55 | |
| Black | 11.31 | 16.49 | 20.94 | |
| Asian | 6.58 | 2.39 | 0.77 | |
| Other | 12.53 | 13.24 | 11.74 | |
| Clinical characteristics, % | ||||
| Outpatient | 2.79 | 1.95 | 2.29 | .144 |
| Active smoker | 22.08 | 19.31 | 17.09 | < .001 |
| Steroid use | 1.08 | 1.12 | 0.97 | .807 |
| Radiotherapy <90 d | 0.27 | 0.25 | 0.14 | .415 |
| Chemotherapy <30 d | 0.90 | 0.58 | 0.20 | < .001 |
| Previous OP <30 d | 0.81 | 1.19 | 0.79 | .174 |
| Comorbidities, % | ||||
| Diabetes | 3.24 | 6.44 | 13.99 | < .001 |
| Dyspnea | 3.11 | 3.62 | 6.91 | < .001 |
| Hypertension | 19.24 | 27.16 | 43.16 | < .001 |
| COPD | 1.53 | 1.12 | 1.66 | .165 |
| Congestive heart failure | 0.23 | 0.11 | 0.16 | .593 |
| Bleeding disorders | 1.17 | 1.59 | 1.28 | .379 |
| Previous PCI/cardiac surgery | 1.53 | 1.74 | 2.01 | .351 |
| Pervious stroke/TIA | 1.80 | 1.99 | 2.05 | .788 |
Complication profile
Overall morbidity differed significantly among cohorts, with ideal BMI, overweight, and obese patients experiencing morbidity rates of 9.0%, 9.7%, and 13.2%, respectively ( P <0.001). Few surgical complications were observed; however, obese and overweight patients experienced significantly higher rates of wound infections, wound disruptions, and overall surgical complications ( Table 2 ).
| Variable | Ideal | Overweight | Obese | P value | |||
|---|---|---|---|---|---|---|---|
| BMI <25 kg/m 2 | BMI ≤25 kg/m 2 to <30 kg/m 2 | BMI ≤30 kg/m 2 | |||||
| (n = 2219) | (n = 2765) | (n = 4933) | |||||
| Overall complications | 199 | 8.97% | 267 | 9.66% | 653 | 13.24% | < .001 |
| Surgical complications | 44 | 1.98% | 90 | 3.25% | 326 | 6.61% | < .001 |
| Wound infection | 38 | 1.71% | 80 | 2.89% | 283 | 5.74% | < .001 |
| Wound disruption | 7 | 0.32% | 13 | 0.47% | 63 | 1.28% | < .001 |
| Medical complications | 170 | 7.66% | 200 | 7.23% | 398 | 8.07% | .416 |
| DVT | 3 | 0.14% | 17 | 0.61% | 21 | 0.43% | .032 |
| Pulmonary embolism | 8 | 0.36% | 15 | 0.54% | 28 | 0.57% | .511 |
| Unplanned reintubation | 5 | 0.23% | 8 | 0.29% | 17 | 0.34% | .689 |
| Ventilation longer than 48 h | 5 | 0.23% | 7 | 0.25% | 20 | 0.41% | .347 |
| Renal insufficiency | 1 | 0.05% | 5 | 0.18% | 11 | 0.22% | .24 |
| Acute renal failure | 1 | 0.05% | 2 | 0.07% | 10 | 0.20% | .141 |
| Coma | 0 | 0.00% | 0 | 0.00% | 2 | 0.04% | .364 |
| Stroke | 0 | 0.00% | 1 | 0.04% | 4 | 0.08% | .341 |
| Cardiac arrest | 1 | 0.05% | 1 | 0.04% | 8 | 0.16% | .159 |
| Myocardial infarction | 6 | 0.27% | 2 | 0.07% | 7 | 0.14% | .196 |
| Peripheral nerve injury | 1 | 0.05% | 3 | 0.11% | 5 | 0.10% | .716 |
| Pneumonia | 13 | 0.59% | 8 | 0.29% | 27 | 0.55% | .216 |
| UTI | 65 | 2.93% | 77 | 2.78% | 137 | 2.78% | .932 |
| Transfusion | 76 | 3.42% | 75 | 2.71% | 140 | 2.84% | .285 |
| Sepsis/septic shock | 23 | 1.04% | 33 | 1.19% | 54 | 1.09% | .863 |
| Reoperation | 44 | 1.98% | 51 | 1.84% | 91 | 1.84% | .923 |
| Length of stay, d | 3.05 ± 2.57 | 3.00 ± 2.94 | 3.06 ± 2.96 | .671 | |||
| Length of surgery, min | 109.73 ± 67.79 | 112.70 ± 66.21 | 127.37 ± 69.33 | < .001 | |||
Medical complication rates did not differ significantly among BMI cohorts ( P = .416). When examined individually, rates of DVT were notably elevated in the overweight population as compared with obese and ideal BMI patients (0.6%, 0.4%, and 0.1%, respectively; P = .032). Reoperations within 30 days and length of postoperative stay did not differ significantly among cohorts. Additionally, longer operative times were observed in obese patients, averaging 127.4 minutes compared with 112.7 minutes in overweight patients and 109.7 minutes in ideal BMI patients ( P < .001).
Logistic regression
After adjusting for potential confounders, a high BMI was significantly associated with overall 30 day morbidity, surgical complications, wound infections, wound disruptions, and DVT ( Table 3 ). Overall morbidity was 1.3 times more likely in obese patients ( P = .001) but was not significantly associated with being overweight. Surgical complications and wound infections, however, both demonstrated a significant, stepwise increase in the odds ratio from overweight to obese ( Table 3 ). Wound disruption was also associated with a significantly increased risk in obese patients (odds ratio [OR], 3.6; P = .002) but demonstrated only a nonsignificant trend toward a higher risk in overweight patients. Although overall medical complications were not significantly associated with a higher BMI, individually DVT was 4.5 times as likely in overweight patients ( P = .016). A high BMI was not found to be a risk factor for reoperation ( Table 3 ).
| Complication | Overweight | Obese | HL | C statistic | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| OR | 95% CI | P value | OR | 95% CI | P value | |||||
| Overall morbidity | 1.050 | 0.861 | 1.282 | .628 | 1.338 | 1.120 | 1.599 | .001 | 0.502 | 0.670 |
| Surgical complications | 1.622 | 1.123 | 2.344 | .010 | 3.014 | 2.173 | 4.182 | .001 | 0.726 | 0.706 |
| Wound infection | 1.661 | 1.121 | 2.461 | .011 | 2.954 | 2.079 | 4.196 | .001 | 0.666 | 0.700 |
| Wound disruption | 1.472 | 0.584 | 3.708 | .412 | 3.556 | 1.601 | 7.896 | .002 | 0.276 | 0.747 |
| Medical complications | 0.905 | 0.726 | 1.128 | .375 | 0.906 | 0.742 | 1.107 | .336 | 0.348 | 0.674 |
| DVT | 4.548 | 1.329 | 15.565 | .016 | 2.789 | 0.828 | 9.389 | .098 | 0.381 | 0.735 |
| Reoperation | 0.920 | 0.611 | 1.386 | .690 | 0.837 | 0.574 | 1.222 | .357 | 0.914 | 0.620 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
