Predictors of preoperative hypoalbuminemia and the association with 30-day morbidity and mortality after surgery for endometrial cancer Pubblico
Smith, Christopher (Summer 2022)
Abstract
Background: Endometrial cancer (EC) is the most common gynecologic malignancy in the United States and remains the fourth most common cancer in women. Once a diagnosis of EC is made, the principal management of most patients is with surgical staging. Current clinical guidelines recommend obtaining liver function tests as part of an initial evaluation of a patient with newly diagnosed EC of which serum albumin concentration levels are a part. Hypoalbuminemia, a surrogate for poor nutritional status, has been associated with adverse surgical outcomes in patients undergoing various surgical procedures.
Objective: The primary purpose of this study is to investigate preoperative predictors of low preoperative serum albumin (hypoalbuminemia, HA), and the relationship between HA and 30-day postoperative morbidity and mortality for women with endometrial cancer undergoing a surgical staging procedure.
Methods: Data analysis was conducted as a retrospective cohort study utilizing the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database from 2012-2016. Patients undergoing elective, same-day surgery were included in the analysis. Patient variables and complications were defined and abstracted according to the ACS-NSQIP methodology. The primary exposure, preoperative serum albumin was classified as normal (≥ 3.5 g/dL) or HA (< 3.5 g/dL). Associations of preoperative risk factors with preoperative HA were identified using logistic regression modeling. Logistic regression assessed the crude and adjusted association of HA with 30-day postoperative morbidity and mortality. Cox proportional hazards regression modeling was used to analyze the effect of HA on the probability of 30-day survival after surgery.
Results: The study population included 17058 patients; 1330 (7.80%) were classified as HA. Preoperative risk factors for HA were Black race (adjusted odds ratio, aOR 1.43, p = 0.0067), dyspnea with moderate exertion (aOR 1.48, p = 0.0136), history of congestive heart failure (aOR 2.55, p = 0.0172), ascites (aOR 2.94, p = 0.0004), disseminated cancer (aOR 1.61, p = 0.0045), weight loss > 10% (aOR 2.20, p = 0.0048), liver dysfunction (aOR 2.64, p <0.0001), renal dysfunction (aOR 1.93, p <0.0001), anemia (aOR 4.29, p <0.0001), thrombocytosis (aOR 2.86, p <0.0001), preoperative blood transfusion (aOR 3.30, p = 0.0002) and worsening American Society of Anesthesiologists Classification (Class 3: aOR 1.49, p = 0.0005; Class 4: aOR 2.55, p <0.0001). Adjusted odds ratios for 30-day postoperative composite morbidity (aOR 1.55, 95% Confidence Interval [CI] 1.28 – 1.87) and all-cause mortality (aOR 3.45, 95% CI 1.75 – 6.63) were significantly higher in patients with HA. Adjusting for covariates, HA was associated with significantly lower probability of 30-day survival after surgery (adjusted hazard ratio, aHR = 3.81, 95% CI 2.17 – 6.63).
Conclusions: Preoperative HA in women undergoing surgery for EC is associated with increased postoperative morbidity and mortality. By identifying those at risk for HA, nutritional optimization could be used to improve postoperative outcomes in such a high-risk surgical population.
Table of Contents
Background 1
Methods 3
Data Source 3
Statistical Analysis 5
Descriptive Statistics 5
Multiple Logistic Regression 6
Survival Analysis 8
Kaplan-Meier Approach to Survival Probability & Log-Rank Test 8
Cox Proportional Hazards Model 9
Proportional Hazards Model Selection 9
Cox Proportional Hazards Model Assessment 11
Results 14
Univariate Analysis of Preoperative Characteristics 14
Univariate Analysis of Intraoperative Characteristics 15
Univariate Analysis of Postoperative Complications 16
Multivariable Logistic Regression 17
Predictors of Preoperative Hypoalbuminemia 17
Association of Preoperative Hypoalbuminemia and 30-day Postoperative Morbidity and Mortality 18
Association of Preoperative Hypoalbuminemia and 30-day Survival 20
Discussion 24
References 30
Tables & Figures 37
Table 1. Preoperative characteristics of patients undergoing surgery for endometrial cancer stratified by serum albumin concentration 37
Table 2. Intraoperative characteristics of patients undergoing surgery for endometrial cancer stratified by serum albumin concentration 38
Table 3. Thirty-day Postoperative characteristics of patients undergoing surgery for endometrial cancer stratified by serum albumin concentration 39
Table 4. Predictors of preoperative hypoalbuminemia for patients undergoing surgery for endometrial cancer 40
Table 5. Association of hypoalbuminemia with 30-day postoperative outcomes for patients undergoing surgery for endometrial cancer 41
Table 6. Association of decreasing serum albumin by 0.25 g/dL with 30-day postoperative outcomes for patients undergoing surgery for endometrial cancer 42
Table 7. Association of perioperative variables with 30-day survival probability following surgery for endometrial cancer 43
Figure 1. Preoperative predictors for preoperative hypoalbuminemia among women undergoing surgery for endometrial cancer 44
Figure 2. The association of a -0.25 g/dL decrease in preoperative serum albumin concentration with 30-day postoperative morbidity and mortality among women undergoing surgery for endometrial cancer 45
Figure 3. Kaplan-Meier Curves for the unadjusted effect of preoperative hypoalbuminemia on 30-day survival probability following surgery for endometrial cancer 46
Figure 4. Plot of likelihood displacement statistic and martingale residuals to identify any influential observations 47
Figure 5. Univariable log-log plots testing for proportional hazards assumption violations for the survival model 48
Figure 6. Univariable observed vs. expected plots evaluating for violations of the proportional hazards assumption 49
Figure 7. Plot of empirical score process based on 5000 martingale residuals simulations of the univariable 30-day survival model 50
Figure 8. Plot of empirical score process based on 5000 martingale residuals simulations of the multivariable 30-day survival model 51
Figure 9. Plots of weighted Schoenfeld residuals versus time to assess the proportional hazards assumption of the univariable model for probability of 30-day survival 52
Figure 10. Plots of weighted Schoenfeld residuals versus time to assess the proportional hazards assumption of the multivariable model for probability of 30-day survival 53
Figure 11. Effect of preoperative hypoalbuminemia on 30-day survival following surgery for endometrial cancer 54
Figure 12. The effect of a 0.25 g/dL decrease in preoperative serum albumin concentration on the 30-day survival of women undergoing surgery for endometrial cancer 55
Appendices 56
Appendix A. Primary current procedure terminology codes included in the study 56
Appendix B. International statistical classification of diseases and related health problems codes included
in the study 57
About this Master's Thesis
| School | |
|---|---|
| Department | |
| Subfield / Discipline | |
| Degree | |
| Submission | |
| Language |
|
| Research Field | |
| Parola chiave | |
| Committee Chair / Thesis Advisor | |
| Committee Members |
Primary PDF
| Thumbnail | Title | Date Uploaded | Actions |
|---|---|---|---|
|
|
Predictors of preoperative hypoalbuminemia and the association with 30-day morbidity and mortality after surgery for endometrial cancer () | 2022-07-23 19:09:08 -0400 |
|
Supplemental Files
| Thumbnail | Title | Date Uploaded | Actions |
|---|---|---|---|
|
Original Word-Processing File (Original Word-Processing File) | 2022-07-23 19:09:21 -0400 |
|