Abstract

Introduction

Over 40% of adults in the United States are diagnosed with obesity which is associated with serious comorbidities, such as diabetes mellitus, cardiovascular diseases, cancers, and renal diseases [1-3]. Body mass index (BMI) measures weight in relation to height and is often used to diagnose obesity. A normal BMI range is between 18.5-24.9 kg/m², and an overweight BMI range is between 25-29.9 kg/m². Obesity is defined as a BMI greater than or equal to 30 kg/m². It can be further broken down into obesity class 1 (30-34.9 kg/m²), class 2 (35-39.9 kg/m²), and severe obesity (greater than or equal to 40 kg/m²). BMI equal to or greater than 50 kg/m² is also known as extreme obesity. With the growing prevalence of obesity in the general population, there is an increasing number of patients with obesity who undergo elective surgery. Perioperatively, elevated BMI is associated with an increase in operating time, risk of infection, and length of hospitalization [4].

Acute renal failure (ARF) is defined as a rapid fall in the rate of glomerular filtration, which manifests clinically as an abrupt and sustained increase in the serum levels of urea and creatinine that may or may not require dialysis [5]. It is associated with poor clinical outcomes, such as increased length of hospital stay, subsequent development of chronic kidney disease, and increased in-hospital mortality [6,7]. Previous work has shown that elevated BMI is an independent risk factor for acute kidney injury in critically ill patients [8], and this has been reproduced in various postsurgical populations [9,10]. However, many of these studies were performed on small datasets. Further, there is a lack of empirical evidence on the relationship between elevated BMI and postoperative ARF following general surgery. Thus, the objective of this study was to determine the association between increasing BMI and the development of ARF within 30 days of an elective general surgery procedure using a large, multicenter dataset of surgical patients from across the United States and Canada.

Materials and methods

Study design

The study was conducted at the Department of Surgery, University of Toronto, Toronto, Canada. We performed a retrospective cohort study of adult patients who underwent elective general surgery procedures and were accounted for in the American College of Surgeons National Surgical Quality Improvement Program® (ACS NSQIP®) database [11] to determine the relationship between BMI and postoperative ARF. Multivariable regression analyses were performed to determine the relationship between BMI categories and postoperative ARF after adjusting for pertinent patient- and surgical-level confounders. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines.

Data sources

The ACS NSQIP® Participant Use Data Files (PUF) from January 1, 2015, to December 31, 2019, were used for this study [11]. Trained and calibrated surgical clinical reviewers collect data from patient charts at each participating hospital using standardized definitions [12,13]. The PUFs contain aggregate, high-quality, standardized clinical data on preoperative, intraoperative, and postoperative variables from over 700 hospitals in the United States and Canada [13]. The PUFs do not identify hospitals, healthcare providers, or patients in compliance with the Health Insurance Portability and Accessibility Act (HIPAA). The Research Ethics Board at Unity Health (Toronto, ON) approved this study (REB ID 00042883).

Study population

All adult (age ≥18) patients who underwent elective general surgery operations as designated by the ACS NSQIP® were included in this study. Exclusion criteria included emergency surgery, outpatient surgery, presence of preoperative acute renal failure or preoperative use of dialysis, age above 80, and cases with missing data. Patients with a length of stay of less than two days were excluded due to the likelihood of a missed diagnosis of ARF during relatively short inpatient admissions.

Outcomes

The primary outcome of the study was the presence of postoperative ARF within 30 days after surgery. We used two variables in the NSQIP database to define this outcome: progressive renal insufficiency (using a variable named RENAINSF) and acute renal failure requiring dialysis (a variable named OPRENAFL). Progressive renal insufficiency is defined as a decreased kidney function as indicated by an increase in serum creatinine by >2 mg/dl from preoperative levels with no requirement for dialysis. Acute renal failure requiring dialysis is defined as the development of renal failure requiring hemodialysis, peritoneal dialysis, hemofiltration, hemodiafiltration, or ultrafiltration in a patient who did not require preoperative dialysis [13]. Secondary outcomes included 30-day readmission (using variable READMISSION), 30-day unplanned reoperation (using variable REOPERATION1), and 30-day all-cause mortality (using variable DOpertoD).

Covariates

Covariates considered in the multivariable analyses included patient characteristics such as BMI, sex, race (American Indian or Alaska Native, Asian, Black or African American, Native Hawaiian or Pacific Islander, White and Unknown or Not Reported), Hispanic ethnicity, age, and health-related variables such as smoking, dyspnea, functional health status, ventilator dependence, severe chronic obstructive pulmonary disease, ascites, congestive heart failure, hypertension requiring medication, disseminated cancer, immunosuppressant used for a chronic condition, greater than 10% loss of body weight in the six months before surgery, and bleeding disorders. Preoperative laboratory variables of serum creatinine and blood urea nitrogen were included. Perioperative variables, such as red blood cell (RBC) transfusion within 72 hours before surgery, wound classification, American Society of Anesthesiologists (ASA) classification, wound closure, sepsis present at the time of surgery, deep incisional surgical site infection (SSI) present at the time of surgery, organ/space SSI present at the time of surgery, pneumonia present at the time of surgery, on ventilator greater than 48 hours present at the time of surgery, sepsis within 48 hours before surgery, septic shock present at the time of surgery, principal anesthesia technique, quarter of admission, year of operation, work relative value unit, and total operation time were also considered in this study. Information about hospital and surgeon-specific variables is not reported in the NSQIP database and was therefore not included in the analysis.

Statistical analysis

Distributions of data points in continuous variables were described with mean (standard deviation, (SD)) or median (interquartile range (IQR)) for normal and non-normal distributions, respectively. Frequency analyses (%) were used to summarize data points in categorical variables. Several pre-specified associations were investigated between patient- and surgical-level variables and the outcomes. The accepted statistical practice of considering no more than one explanatory variable for every 10 patients who experienced an outcome was employed for variable selection [14]. Univariate logistic regression was used to investigate the relationship between the explanatory variables and postoperative ARF. All pairwise associations between each explanatory variable were studied to ensure none was overly correlated. BMI categories and explanatory variables with p≤0.05 on univariate analyses were entered into a multivariable logistic model using forward selection. The same processes were repeated to develop multivariable models predicting secondary outcomes. Goodness-of-fit was examined for all multivariable models using c-statistics and five-fold cross-validation with mean square error. Confidence intervals and p-values reported reflect a two-tailed α level of 0.05. Statistical analyses were conducted using R software version 3.6.3 (R Foundation for Statistical Computing, Vienna, Austria, https://www.R-project.org/)

Results

Characteristics of study patients

Out of the 5,011,560 patients who were accounted for in the NSQIP database between January 1, 2015, and December 31, 2019, 2,172,095 patients underwent a general surgery procedure. After exclusion, a total of 424,527 patients were included in the final analysis, with 3638 patients who developed postoperative ARF within 30 days (Figure ​(Figure1).1). Of the patients who developed ARF after surgery, 63.2% had progressive renal insufficiency, and 37.6% developed acute renal failure requiring dialysis.

Figure 1

Open in a separate window

Study sample

NSQIP: National Surgical Quality Improvement Program

Unadjusted comparisons between patients with ARF and patients without ARF

Patients with ARF were generally older (median age of 64 (IQR 56-71) years vs. 58 (IQR 48-67) years, p<0.001), less likely to be female (38.9% vs. 57.1%, p<0.001), and more likely to be a current smoker within one year (21.9% vs. 16.7%, p<0.001) than those without ARF (Table ​(Table1).1). Patients with ARF had a higher median BMI (30.9 (IQR 26.4-32.3) kg/m² vs. 30.1 (IQR 25.8-31.4) kg/m², p<0.001). Further, they were more likely to have diabetes (33.0% vs. 19.1%, p<0.001), severe chronic obstructive pulmonary disease (8.7% vs. 4.3%, p<0.001), bleeding disorders (5.4% vs. 2.5%, p<0.001), hypertension requiring medication (71.7% vs. 46.7%, p<0.001), congestive heart failure (2.3% vs. 0.6%, p<0.001), disseminated cancer (10.9% vs. 6.9%, p<0.001), recent significant weight loss (6.0% vs. 3.5%, p<0.001), and use of immunosuppressants (8.2% vs. 6.2%, p<0.001) (Table ​(Table1).1). Patients who developed ARF also had higher preoperative blood urea nitrogen (BUN) values (mean of 19.1 (SD 11.9) mg/dL vs. 15.2 (SD 7.5) mg/dL, p<0.001) and preoperative creatinine values (mean of 1.1 (SD 0.66) mg/dL vs. 0.9 (SD 0.4) mg/dL, p<0.001). Patients who developed ARF were more likely to have developed sepsis within 48 hours prior to surgery (2.8% vs. 1.4%, p<0.001), and more likely to have contaminated or dirty/infected wounds (20.3% vs. 14.0%, p<0.001) (Table ​(Table1).1). Patients who developed ARF were also more likely to have had either no layers surgically closed (1.2% vs. 0.4%, p<0.001) or only deep layers closed (1.7% vs. 1.0%, p<0.001) (Table ​(Table1).1). As well, patients who developed ARF generally had longer surgeries (mean of 255.3 (SD 147.8) minutes vs. 190.0 (SD 119.6) minutes, p<0.001) (Table ​(Table1).1). In unadjusted comparison analyses of postoperative outcomes, patients who developed ARF were also more likely to require a return to the operating room (30.7% vs. 4.0%, p<0.001) and had a longer length of stay in the hospital (mean 13.3 (SD 12.7) days vs. 5.2 (SD 5.0) days, p<0.001) (Table ​(Table2).2). As well, they were much more likely to require readmission to the hospital within 30 days (43.6% vs. 9.5%, p<0.001) (Table ​(Table22).

Multivariable models predicting ARF 

Multivariable regression analyses demonstrated that there was a stepwise increase in odds of postoperative ARF with increasing BMI categories compared to normal BMI category (overweight: OR 1.11 (95% CI 1.0-1.23), class 1 and 2 obesity: OR 1.32 (95%-CI 1.2-1.46), severe obesity: OR 1.45 (95%-CI 1.27-1.66), and extreme obesity: OR 1.78 (95%-CI 1.47-2.15)) (Figure ​(Figure2).2). There was no difference in the odds of postoperative ARF in the underweight BMI compared to the normal BMI. Other independent predictors of postoperative ARF included male sex (OR 1.60 (95%-CI 1.49-1.72)), increasing age (OR 1.02 (95%-CI 1.02-1.02)), current smoker (OR 1.46 (95%-CI 1.34-1.59)), hypertension requiring medication (OR 1.75 (95%-CI 1.61-1.90)), ascites (OR 3.48 (95%-CI 2.55-4.77)), and immunosuppressive use (OR 1.44 (95%-CI 1.28-1.63)) (Table ​(Table33).

Figure 2

Open in a separate window

Multivariable regression analyses to assess the relationship between BMI categories and outcomes.

Multivariable regression analyses to assess the relationship between BMI categories and the following postoperative outcomes: A) acute renal failure; B) mortality; C) readmission; and D) reoperation. Normal BMI (18.5-24.9 kg/m2) was used as the reference. Odds ratios and 95% confidence intervals are shown on the x-axis, with BMI categories on the y-axis.

BMI: body mass index

Underweight BMI was independently associated with increased odds of death (OR 1.67 (95% CI 1.36-2.06)) (Tables ​(Tables4,4, ​,5,5, Appendices 1, 2), and overweight BMI was associated with decreased odds of death (OR 0.84 (95% CI 0.75-0.94)). Obesity was not associated with postoperative death compared to the normal BMI. Underweight BMI was independently associated with an increased risk of readmission (OR 1.12 (95% CI 1.04-1.20)) (Tables ​(Tables6,6, ​,7,7, Appendices 3, 4). Increasing BMI compared to normal BMI was associated with lower odds of readmission (overweight: OR 0.94 (95% CI 0.91-0.97), class 1 and 2 obesity: OR 0.91 (95% CI 0.88-0.94), severe obesity: OR 0.76 (95% CI 0.73-0.79), and extreme obesity: OR 0.72 (95% CI 0.68-0.77)) (Tables ​(Tables6,6, ​,7).7). Similarly, underweight BMI was independently associated with higher odds of reoperation (OR 1.23 (95% CI 1.12-1.36)) and increasing BMI was associated with lower odds of reoperation (overweight: OR 0.87 (95% CI 0.83-0.90), class 1 and 2 obesity: OR 0.88 (95% CI 0.84-0.92), severe obesity: OR 0.85 (95% CI 0.80-0.91), and extreme obesity: OR 0.77 (95% CI 0.70-0.85)) (Tables ​(Tables8,8, ​,9,9, Appendices 5, 6).

Model validation

The calculated mean square error values and c-statistics from five-fold cross-validation on all our multivariable regression models demonstrated model consistency. Mean square error and c-statistic values are reported in Appendix 7.

Discussion

We performed a retrospective cohort study of 424,527 patients in the 2015-2019 NSQIP database who underwent elective general surgery procedures to demonstrate that increasing BMI was independently associated with postoperative ARF compared to normal BMI. We also showed that underweight BMI was associated with increased odds of postoperative death, readmission, and reoperation, while increasing BMI was associated with lower odds of readmission and reoperation. Our study represents one of the largest contemporary analyses to determine the relationship between BMI and postoperative ARF after elective general surgery.

Our study found a stepwise increase in the odds of developing postoperative ARF with increasing BMI in patients who underwent general surgery procedures. This finding is in line with previous work that found that elevated BMI was associated with increased risk of postoperative ARF in several types of surgical procedures, such as cardiac surgery [15,16], tracheostomies [17], pancreaticoduodenectomies [18], surgery for diverticular disease [19,20], and hip fractures [21]. Our study conducted robust risk-adjusted analyses on the relationship between BMI and postoperative ARF using a contemporary, multicenter, and multiyear database and focused on general surgery. Thus, the findings from our study add to the growing body of evidence linking obesity to postoperative ARF. The underlying pathophysiology behind the relationship between obesity and ARF is thought to involve multiple mechanisms, including chronic inflammation, oxidative stress, and endothelial dysfunction [10,22]. Obesity often leads to dyslipidemia, and the resultant lipid accumulation in the kidneys can promote inflammation, oxidative stress, and fibrosis, contributing to renal dysfunction [23]. Adipose tissue also releases pro-inflammatory cytokines and chemokines, leading to chronic low-grade inflammation and oxidative stress [24]. These harmful effects can contribute to renal damage. Elevated BMI can raise the glomerular filtration rate due to increased metabolic demand and renal blood flow, which can contribute to kidney damage over time [22]. Obesity is also linked to the activation of the renin-angiotensin-aldosterone system, resulting in increased renal vascular resistance, salt retention, and hypertension, all of which can contribute to kidney injury [22].

The relationship between underweight BMI and mortality has been studied in the past and demonstrated an increased risk of death among those with lower than normal BMI [25,26]. Our finding of increased morbidity and mortality for underweight patients is in line with the scientific literature. This supports the theory that most of the patients with underweight BMI are undernourished and more likely to have weakened immune systems, dietary deficits, and decreased physiologic reserve. Consequently, they are more likely to experience postoperative deaths, readmissions, and reoperations. A previous literature review indicated that underweight patients had higher crude mortality following coronary artery bypass graft surgery, indicating that undernutrition should be investigated further as a risk factor [27]. The low fat-free mass index was also linked to an increased likelihood of negative outcomes following cardiac surgery, stressing the importance of optimizing nutritional status for undernourished patients [28]. Monitoring the health of underweight patients is critical, as is providing adequate nutritional assistance to help mitigate this risk.

Our finding that increasing BMI was protective for readmission and reoperation has been previously described in studies and is known as the “obesity paradox” [29]. The obesity paradox challenges the paradigm of obesity as a perioperative risk factor. Although obesity is associated with various comorbidities, in patients who undergo acute stress, such as surgery and critical illness, increased BMI has been found to be a protective factor linked to lower morbidity and mortality [29]. This obesity paradox has been demonstrated in various study samples [29,30]. In critically ill patients, obesity has been linked to increased morbidity but not mortality, with one study showing that obesity was protective against mortality in these individuals. Our findings emphasize the complex relationship of obesity with health outcomes. More work is needed to fully understand the mechanisms underlying the obesity paradox and to determine the circumstances under which obesity may confer a survival benefit.

Decreasing the risk of postoperative ARF in patients with elevated BMI requires a comprehensive approach that targets the underlying processes and risk factors for both conditions. Losing weight through lifestyle modifications such as dietary changes and increased physical exercise has shown promise in improving renal function. In addition, our study identified diabetes requiring medication as a protective factor against ARF. This may be attributed to the effective management of diabetes and monitoring of renal function in these patients, which could mitigate the detrimental effects of obesity. This finding further supports the theory that managing comorbid chronic conditions such as hypertension and diabetes with appropriate medication and monitoring might help reduce the risk of ARF. Further, there may be an important role in educating patients with obesity on their increased risk of renal complications when they undergo general surgery procedures. Close monitoring of renal function in these patients is critical to identify and manage modifiable risk factors early. As well, patients may benefit from a multidisciplinary approach with involvement from physicians, nurses, dietitians, and fitness experts. Further consideration should be given to renal-sparing strategies, including careful drug choices and intraoperative fluid resuscitation.

There are limitations in this study. First, hospital-specific variables, such as geographic location or institution type, and surgeon-specific variables were not included in the analysis as the NSQIP does not report these variables. Second, despite adjusting for key confounders, the possibility of residual confounding remains, particularly from unmeasured factors such as lifestyle, socioeconomic status, or detailed comorbidity history, which could influence the observed associations. Third, we were not able to analyze outcomes that occurred beyond the 30 days after surgery because they were not reported in the NSQIP database. As a result, any delayed or chronic renal dysfunction may not have been captured. Fourth, while BMI was used as the primary measure of obesity, it does not capture other important aspects such as fat distribution or muscle mass, which may provide additional insights into the relationship between obesity and ARF. Lastly, patients with missing values for diabetes status, dyspnea, functional status, surgical wound closure, preoperative BUN values, and preoperative creatinine values were excluded; however, we were unable to probe further into why some patients did not have these values collected.

Conclusions

Our study demonstrated that obesity was independently associated with increasing odds of developing ARF after elective general surgery procedures compared to the normal BMI. These findings emphasize the need for targeted interventions such as weight management, renal monitoring, and optimized perioperative care to reduce the incidence of ARF and improve outcomes in this population.

Appendices

Appendix 1 

Appendix 2

Appendix 3

Appendix 4

Appendix 5

Appendix 6

Appendix 7

Disclosures

Author Contributions

References