ORIGINAL STUDY
Racial Disparity in 30-Day Morbidity and Mortality After Surgery for Ovarian Cancer Haider Mahdi, MD,* Amelia Jernigan, MD,* David Lockhart, BA,Þ Mehdi Moslemi-Kebria, MD,* and Peter G. Rose, MD*
Background: The improved survival observed in recent years for women with ovarian cancer (OC) has not been realized among African American (AA) compared with white (W) women. The contribution of immediate postoperative morbidity and mortality to this survival disparity remains unclear. This study aims to examine disparities in postoperative 30-day morbidity and mortality between AA and W women with OC. Materials and Methods: Patients with OC were identified from the American College of Surgeons (ACS) National Surgical Quality Improvement Program (NSQIP) 2005 to 2011. African American and subgroups were studied. Multivariable logistic regression models were performed. Results: Of 1649 women, 1510 (92%) were W and 139 (8%) were AA. The rate of 30-day postoperative complications and mortality among the entire cohort were 30% and 2%, respectively. No differences in postoperative complications were noted between AA and W women (33% vs 30%, P = 0.47) including surgical (29% vs 26%, P = 0.40) and nonsurgical (10% vs 9%, P = 0.75) complications. The mean length of hospital stay was longer in AA women, but there was no difference in surgical re-exploration and operative time. No difference in 30-day mortality was found between AA and W women (3% vs 2%, P = 0.45). African Americans were younger and more likely to be obese, have diabetes, hypertension, preoperative weight loss, higher serum creatinine level greater than or equal to 2 mg/dL, hypoalbuminemia, and anemia. After adjusting for surgical complexity and associated comorbidities, AA race was not an independent predictor of 30-day postoperative complications (odds ratio, 0.99; 95% confidence interval [CI], 0.65Y1.5; P = 0.96) or mortality (odds ratio, 0.89; 95% confidence interval, 0.25Y2.43; P = 0.83). Conclusions: African American race was not an independent predictor of poor 30-day outcomes. Interestingly, AAs with OC are underrepresented in quality-seeking hospitals. Efforts to minimize this racial disparity should target optimization of comorbidities and improving access to high-volume centers for AA women. Key Words: Race, Surgery, Ovarian cancer, Perioperative outcome, African American, White Received August 8, 2014, and in revised form October 5, 2014. Accepted for publication October 6, 2014. (Int J Gynecol Cancer 2015;25: 55Y62)
*Gynecologic Oncology Division, Ob/Gyn and Women’s Health Institute, Cleveland Clinic, Cleveland, OH; and †Department of Biostatistics, University of Washington, Seattle, WA. Address correspondence and reprint requests to Haider Mahdi, MD, Ob/Gyn and Women’s Health Institute, Copyright * 2014 by IGCS and ESGO ISSN: 1048-891X DOI: 10.1097/IGC.0000000000000324 International Journal of Gynecological Cancer
9500 Euclid Ave, Cleveland, OH 44195. E-mail: [email protected]. The authors declare no conflicts of interest. Additional contributions: For this study, the linked American College of Surgeons National Surgical Quality Improvement Program (NSQIP) database was used. The interpretation and reporting of these data are the sole responsibility of the authors. We acknowledge the efforts of the American College of Surgeons NSQIP for the creation of the NSQIP database.
& Volume 25, Number 1, January 2015
Copyright © 2014 by IGCS and ESGO. Unauthorized reproduction of this article is prohibited.
55
International Journal of Gynecological Cancer
Mahdi et al
cancer is the most lethal gynecologic malignancy, O varian representing the fifth leading cause of cancer-related
death in the United States. Considerable advances in surgical and medical therapies have resulted in a significant improvement in survival from this disease. Unfortunately, the Surveillance, Epidemiology, and End Results data 1975 to 2005 indicate that these improvements in ovarian cancer (OC) survival were more prominent in white (W) women with lower survival noted for African American (AA) women with OC.1,2 There is great interest in identifying the reasons for these discrepancies so that they may be remedied in an attempt to narrow the gap in survival between W and AA women. In an attempt to explain these racial discrepancies, many studies have focused on socioeconomic factors, tumor biology, and differences in access to care, treatment, and surgical outcomes. African American women are more likely than their W counterparts to present with advanced-stage disease.2 However, it is noteworthy and concerning that multiple studies have demonstrated that AA women with OC are more likely to experience delays in the initiation of adjuvant chemotherapy and are less likely than W women to receive standard-of-care treatment recommended in accordance with guidelines.3Y5 Treatment adherence to the National Comprehensive Cancer Network guidelines results in improved survival outcomes.3,6 Although treatment at highvolume centers by high-volume providers is associated with both guidelines compliance and better survival, race and socioeconomic status have been associated with an increased likelihood of treatment at a low-volume center by a lowvolume provider.6,7 Both the existence of disparate shortterm perioperative outcomes and their impact on racial disparities observed with regards to survival outcomes are important additional pieces of the picture that have not been thoroughly explored. If AA women present with more preoperative medical comorbidities and more advanced disease, it is worth exploring the possibility that they require more extensive surgeries and experience more morbidity and mortality in the perioperative period. Short-term mortality could contribute to differences in overall survival. Furthermore, perioperative morbidity can contribute to treatment delays including time to receive chemotherapy, which can result in worse progression-free and overall survivals.8 Using the ACS-NSQIP database, we detailed the perioperative comorbidities, surgical complexity, and postoperative outcomes of women who underwent surgical treatment for OC. The objective of this study was to analyze the association of race with 30-day postoperative morbidity and mortality.
METHODS Data Source American College of Surgeons NSQIP is a risk-adjusted data collection mechanism that collects and analyzes clinical outcomes data. Participating hospitals use their collected data to develop quality initiatives that improve surgical care and to identify elements in provided health care that can be improved when compared with other institutions. The ACS-NSQIP collects data on 135 variables, including preoperative risk
56
& Volume 25, Number 1, January 2015
factors (including patient demographics, comorbidities, and laboratory values), intraoperative variables, and 30-day postoperative mortality and morbidity outcomes for a systematic and prospective sample of patients undergoing major surgical procedures. Data are collected in a standardized fashion according to strict definitions by dedicated surgical clinical nurse reviewers. Patients are followed throughout their hospital course and after discharge from hospital for up to 30 days postoperatively. A site’s surgical clinical reviewer captures these data using a variety of methods including medical chart extraction, doctor’s office records, 30-day telephone interview with the patients, and other methods. Patients who were diagnosed with OC were identified from the 2005 to 2011 ACS-NSQIP participant use files, which include data collected from 258 academic and community hospitals throughout the United States using the International Classification of Diseases, Ninth Revision codes. Patients with OC were included if they underwent at least salpingo-oophorectomy, debulking, or any of the surgeries listed below the surgical complexity scoring using the Current Procedural Terminology codes. Patients who underwent pelvic exenteration were excluded. Two subgroups were abstracted for comparisonVW and AA patients.
Risk Factors and Outcome All risk factors available in the ACS-NSQIP database were compared between the 2 groups. The primary end points of the study were 30-day mortality, postoperative morbidity, procedure-related complications, surgical re-exploration (return to the operating room within 30 days), and length of hospital stay. Composite end points were created to categorize postoperative complication into few related groups: surgical complications (all surgical site infections, wound disruption, bleeding requiring transfusion, and peripheral nerve injury), renal complications (progressive renal failure, acute renal failure), pulmonary complications (pneumonia, unplanned intubation, respiratory insufficiency requiring ventilation for 48 hours), infectious complications (systematic inflammatory response syndrome, sepsis, septic shock, surgical site infection, and pneumonia), cardiovascular complications (pulmonary embolism, stroke/cerebrovascular event, cardiac arrest, myocardial infarction, deep vein thrombosis requiring therapy), and any nonsurgical complication (any complication except surgical complications). Patients with preoperative sepsis were excluded from the study. Patients who are ventilator dependent, with renal failure, or on dialysis before surgery were excluded from their respective complication category.
Statistical Analysis Associations between categorical covariates were assessed using W2 test. Group differences in means of continuous measures were assessed using the Student t test or the Wilcoxon rank sum test. The preoperative laboratory values were used both as continuous and categorical variables: serum albumin (93 vs e 3 g/dL), hematocrit (G35 vs Q 35%), serum creatinine (Q2 vs G 2 mg/dL), platelets (G350,000 vs Q 350,000 cell/mL), and white blood cell (WBC) are 109/mL. * 2014 IGCS and ESGO
Copyright © 2014 by IGCS and ESGO. Unauthorized reproduction of this article is prohibited.
International Journal of Gynecological Cancer
& Volume 25, Number 1, January 2015
To adjust for surgical complexity, patients were given a specific score for each procedure. Then, based on the number of procedures performed, the sum of these scores was calculated. A score of 1 was given to any of the following procedures: hysterectomy and/or salpingo-oophorectomy, debulking, lymphadenectomy, or omentectomy. A score of 2 was given to any of the following procedures: small or large bowel resection, gastrectomy, hepatectomy, splenectomy, and pancreatectomy. Multivariable logistic regression models were used to assess the association between race and 30-day postoperative complications while controlling for all other confounders. For the creation of the models, we considered all preoperative variables available in the ACS-NSQIP database, including demographics (age and race), preoperative health status and comorbidities, preoperative laboratory values (serum albumin, creatinine, WBC count, platelet count, and hematocrit), and operative factors (operative time, American Society of Anesthesiologists (ASA) class, surgical complexity). The preoperative morbidities, operative risk factors, or demographic variables that were significantly associated with postoperative complications in the univariate logistic regression model with P G 0.05 were included in the multivariable regression model. Another multivariable model for postoperative mortality was created including the preoperative morbidities, operative risk factors, or demographic variables that were significantly associated with 30-day mortality and race in the univariate logistic regression model with P G 0.05. The final logistic regression models were run using the 2 racial groups and all confounders found in this way. All tests of significance were at the P G 0.05 level, and P values were 2-tailed. STATA 10.0 program (College Station, TX) was used for the analysis of the data.
RESULTS
Race and Perioperative Outcome in OC
complications between AA and W patients stratified are listed in Table 2. No difference in the risk of postoperative complications was found between AA and W (33% vs 30%, P = 0.47). Furthermore, no differences in surgical (29% vs 26%, P = 0.40) and no surgical (10% vs 9%, P = 0.75) complications were noted between AA and W, respectively. Among nonsurgical complications, there was no significant difference in the rate of cardiac (4% vs 4%, P = 0.81), pulmonary (4% vs 4%, P = 0.62), infectious (5% vs 4%, P = 0.60), or renal (1% vs 1%, P = 0.85) between AA and W patients, respectively (Table 2). No difference in surgical reexploration was found between AA and W patients (3% vs 4%, P = 0.42). However, AA women were more likely to have longer hospital stay compared with W women with mean hospital stay of 8 versus 6.5 days (SD 6.9 vs 6.3 days), respectively (P G 0.001). In multivariable logistic regression model after adjustment for all confounders including surgical complexity, demographics, and associated morbidities, AA race was not an independent predictor of ‘‘any postoperative complications’’ (odds ratio [OR], 0.99; 95% confidence interval [CI], 0.65Y1.5; P = 0.96; Table 3). Significant predictors of postoperative complications were age, ascites, surgical complexity, operative time, anemia, hypoalbuminemia, and higher ASA class.
Thirty-Day Postoperative Mortality No difference in 30-day mortality was found between AA and W patients (3% vs 2%, P = 0.45, Table 2). In multivariable logistic regression model after adjustment for all confounders including surgical complexity, demographics, and associated morbidities, AA race was not an independent predictor of 30-day postoperative mortality (OR, 0.89; 95% CI, 0.25Y2.43; P = 0.83; Table 4). Significant predictors of postoperative mortality were surgical complexity, preoperative weight loss, and hypoalbuminemia.
Demographics and Clinical Characteristics One thousand six hundred forty-nine women were included in this study, 1510 (92%) were W and 139 (8%) were AA. Demographics and clinical characteristics of each group are shown in Table 1. African American patients with OC were younger (mean 58.6 vs 61.3 years, P G 0.001) and more likely to be obese (P G 0.001), have diabetes (P G 0.001), or hypertension requiring medications (P = 0.006) and weight loss within 6 months before surgery (P = 0.003, Table 1). Furthermore, AA were more likely to have higher serum creatinine level greater than or equal 2 mg/dL (6% vs 2%, P = 0.018), hypoalbuminemia (serum albumin level e3 mg/dL) (40% vs 24%, P G 0.001), and anemia (hematocrit G35 mg/dL) (45% vs 25%, P G 0.001) (Table 1). No difference in surgical complexity was found between the 2 racial groups (P = 0.27, Table 1). However, there was a trend in favor of longer mean operative time for AA compared with W (203 vs 189 minutes, P = 0.065, Table 1).
Postoperative Complications The rate of any complication among the entire cohort was 30%. The rates and differences in the risk of postoperative
DISCUSSION We used the ACS-NSQIP database to describe the disparities in perioperative comorbidities, surgical complexity, and postoperative outcomes between W and AA women who underwent surgery for OC. African American women with OC were more likely to have preoperative morbidities and longer hospital stay, but no difference in surgical complexity and reoperation was found compared with their W counterparts. Furthermore, AA race was not an independent predictor of adverse 30-day postoperative morbidity or mortality after surgery for OC even in multivariate analysis after adjusting for preoperative morbidities and surgical complexity. Therefore, it does not seem that immediate postoperative outcome discrepancies can account for the differences in disease-specific and overall survival that have been reported for AA women with OC. Complete cytoreductive surgery is one of the most important goals of OC care and the extent of surgery needed to achieve this goal varies from patient to patient. Understanding discrepancies in critical surgical outcomes and perioperative events are vital to understand discrepancies in
* 2014 IGCS and ESGO
Copyright © 2014 by IGCS and ESGO. Unauthorized reproduction of this article is prohibited.
57
Mahdi et al
International Journal of Gynecological Cancer
& Volume 25, Number 1, January 2015
TABLE 1. Demographic and clinical characteristics of patients with OC stratified by race Variable
Total Cohort (n = 1649)
Age groups, n (%) G60 y 60Y69 y 70Y79 y Q80 y Body mass index, n (%) G18 18Y29.9 30Y39.9 Q40 ASA class, n (%) 1Y2 3 Q4 Dependent functional status, n (%) Current smoker (within 1 year), n (%) Diabetes, n (%) Hypertension on medications, n (%) Cardiac comorbidities, n (%) Respiratory comorbidities, n (%) Renal comorbidities, n (%) Neurologic comorbidities, n (%) Ascites, n (%) Weight loss, n (%) Blood transfusion before surgery, n (%) Steroid use, n (%) Chemotherapy before surgery Yes Unknown Preoperative anemia, n (%)* Preoperative hypoalbuminemia, n (%)† Preoperative serum creatinine Q2 mg/dL, n (%) Preoperative leukocytosis, n (%)‡ Preoperative thrombocytosis, n (%) Surgical complexity, n (%) Complexity score 1Y3 Complexity score 93 Emergency, n (%) Operative time, n (%) e60 min 61Y120 min 121Y180 min Q180 min
W (n = 1510)
AA (n = 139)
745 468 316 120
(45) (28) (19) (7)
671 (44) 438 (29) 293 (19) 108 (7)
74 30 23 12
(53) (22) (17) (9)
1017 20 451 142
(62) (1) (28) (9)
953 (64) 19 (1) 399 (27) 121 (8)
64 1 52 21
(46) (1) (38) (15)
764 815 70 36 218 171 710 55 51 3 57 359 81 19 35
(46) (49) (4) (2) (13) (10) (43) (3) (3) (0) (3) (22) (5) (1) (2)
710 (47) 739 (49) 61 (4) 33 (2) 198 (13) 141 (9) 635 (42) 51 (3) 45 (3) 2 (0) 49 (3) 334 (22) 67 (4) 14 (1) 32 (2)
54 (39) 76 (55) 9 (6) 3 (8) 2 (20) 30 (22) 75 (54) 4 (8) 6 (4) 1 (1) 8 (6) 25 (18) 14 (10) 5 (4) 3 (2)
104 545 436/1607 300/1204 43/1568 123/1602 486/1600
(6) (33) (27) (25) (3) (8) (30)
92 (6) 512 (34) 374/1471 (25) 259/1107 (24) 35/1432 (2) 111/1466 (8) 445/1465 (30)
1388/1630 (85) 242/1630 (15) 32 (2)
1267/1493 (85) 226/1493 (15) 29 (2)
40/1648 326/1648 488/1648 794/1648
(2) (20) (30) (48)
38/1509 (3) 297/1509 (20) 459/1509 (30) 715/1509 (47)
12 33 62/136 41/103 8/136 12/136 41/135
(9) (24) (46) (40) (6) (8) (30)
121/137 (88) 16/137 (12) 3 (2) 2/139 29/139 29/139 78/139
(1) (21) (21) (57)
P
0.13
G0.001
0.10 0.98 0.67 G0.0001 0.007 0.75 0.38 0.12 0.12 0.25 0.003 0.005 0.97 0.038
G0.001 0.0003 0.02 0.76 0.99 0.27 0.84
0.07
(Continued on next page)
58
* 2014 IGCS and ESGO
Copyright © 2014 by IGCS and ESGO. Unauthorized reproduction of this article is prohibited.
International Journal of Gynecological Cancer
& Volume 25, Number 1, January 2015
Race and Perioperative Outcome in OC
Total Cohort (n = 1649)
W (n = 1510)
AA (n = 139)
189 (89)
203 (89)
TABLE 1. (Continued) Variable Operative time, mean (SD), min§
P 0.065
Statistical significance was determined by a P G 0.05. Preoperative thrombocytosis was defined as a platelet count of more than 350 109/L. Elevated preoperative serum creatinine level was defined as greater than or equal to 2 mg/dL. *Preoperative anemia was defined as a hematocrit less than 35%. †Preoperative hypoalbuminemia was defined as a serum albumin level less than or equal to 3 g/dL. ‡Preoperative leukocytosis was defined as a WBC count greater than 11 109/L. §Mean values are provided, and comparisons are made with Student t test.
long-term OC survival. African American women who are surgically cytoreduced at a tertiary care center are as likely to be left with minimal residual disease with an elimination of the racial discrepancy in overall survival.9 Using NSQIP data, Parsons et al10 looked at short-term survival outcomes after cancer surgery and found, similar to us, that non-W patients have similar short-term operative outcomes but remain hospitalized longer. Notably, they pointed out that ACS-NSQIP hospitals are quality-driven hospitals and that improved access to quality-driven surgical facilities may drive a narrowing in racial disparities in the immediate postoperative period. In OC, improved short- and long-term outcomes as well as improved compliance with recommended treatment guidelines have repeatedly been associated with high-volume, quality-driven, tertiary care facilities; furthermore, treatment at these facilities has been demonstrated to result in an amelioration of racial disparities observed with regards to outcomes and survival.3,6,7,9 In this study, AA patients were more likely to have preoperative comorbidities compared with W patients, specifically higher body mass index, diabetes, hypertension requiring medications, serum creatinine level greater than or equal 2, hypoalbuminemia, and anemia. Some of these preoperative comorbidities were significantly associated with adverse postoperative morbidity and mortality in multivariate analysis such as hypoalbuminemia and anemia. In fact, hypoalbuminemia was an independent predictor of adverse postoperative morbidity and 30-day mortality after surgery for OC. Uppal et al11 recently reported a significant increase in the risk of perioperative complications and 30-day mortality among patients with serum albumin level less than 3 g/dL undergoing open surgery for gynecologic malignancies. Other independent predictors of adverse perioperative complications were surgical complexity, operative time, ascites, higher ASA class, and older age. Most of these factors are not modifiable risk factors and have been reported in prior studies. However, optimal management of perioperative comorbidities and careful selection of the appropriate candidate for primary debulking surgery are very important to minimize the risk of perioperative morbidities and mortality and also to improve the short- and long-term outcome of this patient population. Our results indicate AA women stay in the hospital longer. This is likely multifactorial including higher prevalence of preoperative comorbidities, socioeconomic, and
insurance status. Importantly, if modifiable causes for their prolonged stay can be identified and ameliorated with proper preoperative planning, this identifies a key area for potential cost savings. Although our study provides insight into the short-term operative outcomes across the 2 racial groups in ACS-NSQIP, we acknowledge several limitations. First, we lacked information on patients’ socioeconomic and insurance status, which have been shown to correlate with race. Alternatively, it is also possible that AA patients treated within ACS-NSQIP hospitals have better socioeconomic status when compared with other patients who receive care at lower quality hospitals. Interestingly, AA women with OC were underrepresented in quality-seeking hospitals represented in NSQIP data set. On the other hand, Census data showed that AAs comprise 13.1% of the entire population in the United States for 2013.12 We did not adjust for surgeon or hospital volume, which have been associated with surgical outcomes and survival. Our results were not adjusted for stage at cancer diagnosis, which influenced treatment plans and overall survival outcomes. However, we adjusted for extent of surgical treatment, which is a proxy for burden of disease. Furthermore, only 15% of the patients in this study underwent ultraradical tumor debulking, which could be a surrogate marker of less advanced disease or of low surgical aggressiveness in advanced disease. Finally, much of the data used requires reliance on Current Procedural Terminology and International Classification of Diseases, Ninth Revision codes, which raises concerns regarding accuracy and validity of the data. Notably, other investigators have closely scrutinized the NSQIP database and found it to be highly accurate.13 Despite these limitations, we present evidence gathered from more than 250 quality-seeking hospitals that reveal that AA and W women experience relatively comparable short-term operative outcomes and mortality after cytoreductive surgery for OC. In summary, AA race was not independently associated with a greater risk for postoperative complications or mortality after OC surgery. However, AA women were more likely to have preoperative comorbidities and stay longer in the hospital. Therefore, efforts should be directed to optimally manage the preoperative comorbidities in this patient population to achieve postoperative outcome equivalent to their W counterpart. Identification of factors that contribute to the poor survival of AA women will illuminate ways to improve the care they receive and outcomes they experience. Potential
* 2014 IGCS and ESGO
Copyright © 2014 by IGCS and ESGO. Unauthorized reproduction of this article is prohibited.
59
60
Mortality, n/available data (%) Overall 34/1581 Complexity score 1Y3 29/1331 Complexity score 93 4/236 Surgical complications, n/available data (%) Overall 437/1649 Complexity score 1Y3 289/1388 Complexity score 93 146/242 Nonsurgical complications, n/available data (%) Overall 154/1649 Complexity score 1Y3 105/1388 Complexity score 93 48/242 Renal complications, n/available data (%) Overall 10/1649 Complexity score 1Y3 6/1388 Complexity score 93 4/242 Pulmonary complications, n/available data (%) Overall 59/1649 Complexity score 1Y3 42/1388 Complexity score 93 17/242 Sepsis, n/available data (%) Overall 69/1649 Complexity score 1Y3 41/1388 Complexity score 93 27/242 Cardiac complications, n/available data (%) Overall 65/1649 Complexity score 1Y3 51/1388 Complexity score 93 14/242 Any complication, n/available data (%) Overall 502/1649 Complexity score 1Y3 343/1388 Complexity score 93 157/242 Surgical re-exploration, n/available data (%) Overall 69/1648
Copyright © 2014 by IGCS and ESGO. Unauthorized reproduction of this article is prohibited.
53/1510 (4) 38/1267 (3) 15/226 (7) 62/1510 (4) 35/1267 (3) 26/226 (12) 59/1510 (4) 46/1221 (4) 13/213 (6)
(4) (3) (7) (4) (3) (11) (4) (4) (6)
65/1509 (4)
9/1501 (1) 5/1257 (0) 4/222 (2)
(1) (0) (2)
(4)
140/1510 (9) 94/1267 (7) 45/226 (20)
(9) (8) (20)
456/1510 (30) 309/1267 (24) 145/226 (64)
396/1510 (26) 260/1267 (21) 134/226 (59)
(27) (21) (60)
4/135 (3)
46/139 (33) 34/121 (28) 12/16 (75)
6/139 (4) 5/116 (4) 1/16 (6)
7/132 (5) 6/121 (5) 1/16 (6)
6/139 (4) 4/117 (3) 2/16 (12)
1/138 (1) 1/120 (1) 0/16 (0)
14/139 (10) 11/110 (9) 3/16 (19)
41/139 (29) 29/121 (24) 12/16 (75)
4/131 (3) 4/111 (3) 0/14 (0)
AA (n = 139)
0.42
0.47 0.36 0.38
0.81 0.77 0.93
0.60 0.17 0.51
0.62 0.85 0.37
0.85 0.48 0.59
0.75 0.50 0.91
0.40 0.37 0.21
0.45 0.31 0.61
P
International Journal of Gynecological Cancer
(30) (25) (65)
30/1450 (2) 25/1191 (2) 4/218 (2)
W
(2) (2) (2)
Total Cohort
TABLE 2. Thirty-day outcomes among patients who underwent surgery for OC stratified by race and surgical complexity
Mahdi et al
& Volume 25, Number 1, January 2015
* 2014 IGCS and ESGO
0.298 0.252 0.269
0.008 0.002 0.019
0.95 0.18
International Journal of Gynecological Cancer
& Volume 25, Number 1, January 2015
Race and Perioperative Outcome in OC
TABLE 3. Multivariate regression analysis for any complication within 30 days after surgery for OC
AA Surgical complexity score 3Y4 Surgical complexity score 94 Age 60Y69 y Age 70Y79 y Age Q80 y Respiratory comorbidities Hypertension on medications Ascites Blood transfusion before surgery Chemotherapy before surgery ASA class 3 ASA class 4Y5 Operative time 1Y2 h Operative time 2Y3 h Operative time 93 h Serum creatinine level Q2 Albumin level 93 Hematocrit Q35 High WBC count (911,000) Platelets count Q350,000 cell/mL Statistical significance was determined by a P G 0.05. *Mean values are provided, and comparisons are made with Student t test.
0.29 (0.72) 0.30 (0.60) 0.95 (1.04)
0.46 (0.83) 0.36 (0.69) 1.25 (1.29)
7.99 (6.91) 7.22 (6.04) 14.38 (9.81) 6.59 (6.35) 5.91 (6.01) 9.82 (7.29)
Complexity score 1Y3 Complexity score 93 Length of hospital stay, mean (SD), d* Overall Complexity score 1Y3 Complexity score 93 No. complications, mean (SD)* Overall Complexity score 1Y3 Complexity score 93
47/1388 (3) 22/241 (9)
43/1267 (3) 22/225 (10)
4/117 (3) 0/16 (0)
Variable
OR (95% CI) 0.99 1.63 4.40 1.36 1.44 1.47 1.50 1.08 1.79 1.50
P
(0.65Y1.50) 0.96 (1.26Y2.1) 0.0002 (2.63Y7.46) G0.0001 (1.02Y1.83) 0.038 (1.03Y2.01) 0.032 (0.91Y2.37) 0.11 (0.78Y2.85) 0.21 (0.083Y1.39) 0.55 (1.35Y2.39) 0.0001 (0.53Y4.44) 0.44
1.49 (0.90Y2.43) 1.36 (1.05Y1.75) 1.81 (1.01Y3.21)
0.11 0.019 0.04 0.10 9.17 (1.91Y160.4) 0.03 17.9 (3.78Y322.6) 0.005 1.41 (0.71Y1.14) 0.11 0.70 (0.51Y0.97) 0.037 0.52 (0.40Y0.68) G0.0001 1.43 (0.93Y2.18) 0.10 1.21 (0.92Y1.59) 0.16
targets remain access to high-quality medical care both before their diagnosis to optimize preoperative comorbidities and afterwards to maximize the receipt of timely guidelineadherent cancer treatments plans. It is unlikely that immediate postoperative complications or mortality contribute to the racial discrepancies seen in OC survival. Further research is needed to explore the reasons behind the racial disparities that have been observed, and efforts should be made to ensure women of all races have access to timely and guidelineadherent adjuvant therapy.
TABLE 4. Multivariate regression analysis for 30-day mortality after surgery for OC Variable AA race Surgical complexity score 3Y4 Surgical complexity score 94 Weight loss with 6 months before surgery Albumin level 94
OR (95% CI) 0.89 0.41 0.44 5.23
P
(0.25Y2.43) 0.83 (0.16Y0.92) 0.04 (0.02Y2.16) 0.42 (1.94Y12.68) G0.001
0.14 (0.06Y0.30)
* 2014 IGCS and ESGO
Copyright © 2014 by IGCS and ESGO. Unauthorized reproduction of this article is prohibited.
G0.001
61
International Journal of Gynecological Cancer
Mahdi et al
REFERENCES 1. Albain KS, Unger JM, Crowley JJ, et al. Racial disparities in cancer survival among randomized clinical trials patients of the southwest oncology group. J Natl Cancer Inst. 2009;101:984Y992. 2. Chan JK, Zhang M, Hu JM, et al. Racial disparities in surgical treatment and survival of epithelial ovarian cancer in united states. J Surg Oncol. 2008;97:103Y107. 3. Howell EA, Egorova N, Hayes MP, et al. Racial disparities in the treatment of advanced epithelial ovarian cancer. Obstet Gynecol. 2013;122:1025Y1032. 4. Bristow RE, Powell MA, Al-Hammadi N, et al. Disparities in ovarian cancer care quality and survival according to race and socioeconomic status. J Natl Cancer Inst. 2013;105:823Y832. 5. Wright J, Doan T, McBride R, et al. Variability in chemotherapy delivery for elderly women with advanced stage ovarian cancer and its impact on survival. Br J Cancer. 2008;98:1197Y1203. 6. Bristow RE, Chang J, Ziogas A, et al. Adherence to treatment guidelines for ovarian cancer as a measure of quality care. Obstet Gynecol. 2013;121:1226Y1234. 7. Bristow RE, Chang J, Ziogas A, et al. High-volume ovarian cancer care: survival impact and disparities in access for advanced-stage disease. Gynecol Oncol. 2014;132:403Y410.
62
& Volume 25, Number 1, January 2015
8. Mahner S, Eulenburg C, Staehle A, et al. Prognostic impact of the time interval between surgery and chemotherapy in advanced ovarian cancer: analysis of prospective randomised phase III trials. Eur J Cancer. 2013;49:142Y149. 9. Bristow RE, Ueda S, Gerardi MA, et al. Analysis of racial disparities in stage IIIC epithelial ovarian cancer care and outcomes in a tertiary gynecologic oncology referral center. Gynecol Oncol. 2011;122:319Y323. 10. Parsons HM, Habermann EB, Stain SC, et al. What happens to racial and ethnic minorities after cancer surgery at american college of surgeons national surgical quality improvement program hospitals? J Am Coll Surg. 2012;214:539Y547 discussion 547Y549. 11. Uppal S, Al-Niaimi A, Rice LW, et al. Preoperative hypoalbuminemia is an independent predictor of poor perioperative outcomes in women undergoing open surgery for gynecologic malignancies. Gynecol Oncol. 2013;131:416Y422. 12. U.S. Census Bureau. State and County QuickFacts. http://quickfacts. census.gov/qfd/states/00000.html. Accessed May 5, 2014. 13. Sellers MM, Merkow RP, Halverson A, et al. Validation of new readmission data in the american college of surgeons national surgical quality improvement program. J Am Coll Surg. 2013;216:420Y427.
* 2014 IGCS and ESGO
Copyright © 2014 by IGCS and ESGO. Unauthorized reproduction of this article is prohibited.
Racial Disparity in 30-Day Morbidity and Mortality After Surgery for Ovarian Cancer Haider Mahdi, MD,* Amelia Jernigan, MD,* David Lockhart, BA,Þ Mehdi Moslemi-Kebria, MD,* and Peter G. Rose, MD*
Background: The improved survival observed in recent years for women with ovarian cancer (OC) has not been realized among African American (AA) compared with white (W) women. The contribution of immediate postoperative morbidity and mortality to this survival disparity remains unclear. This study aims to examine disparities in postoperative 30-day morbidity and mortality between AA and W women with OC. Materials and Methods: Patients with OC were identified from the American College of Surgeons (ACS) National Surgical Quality Improvement Program (NSQIP) 2005 to 2011. African American and subgroups were studied. Multivariable logistic regression models were performed. Results: Of 1649 women, 1510 (92%) were W and 139 (8%) were AA. The rate of 30-day postoperative complications and mortality among the entire cohort were 30% and 2%, respectively. No differences in postoperative complications were noted between AA and W women (33% vs 30%, P = 0.47) including surgical (29% vs 26%, P = 0.40) and nonsurgical (10% vs 9%, P = 0.75) complications. The mean length of hospital stay was longer in AA women, but there was no difference in surgical re-exploration and operative time. No difference in 30-day mortality was found between AA and W women (3% vs 2%, P = 0.45). African Americans were younger and more likely to be obese, have diabetes, hypertension, preoperative weight loss, higher serum creatinine level greater than or equal to 2 mg/dL, hypoalbuminemia, and anemia. After adjusting for surgical complexity and associated comorbidities, AA race was not an independent predictor of 30-day postoperative complications (odds ratio, 0.99; 95% confidence interval [CI], 0.65Y1.5; P = 0.96) or mortality (odds ratio, 0.89; 95% confidence interval, 0.25Y2.43; P = 0.83). Conclusions: African American race was not an independent predictor of poor 30-day outcomes. Interestingly, AAs with OC are underrepresented in quality-seeking hospitals. Efforts to minimize this racial disparity should target optimization of comorbidities and improving access to high-volume centers for AA women. Key Words: Race, Surgery, Ovarian cancer, Perioperative outcome, African American, White Received August 8, 2014, and in revised form October 5, 2014. Accepted for publication October 6, 2014. (Int J Gynecol Cancer 2015;25: 55Y62)
*Gynecologic Oncology Division, Ob/Gyn and Women’s Health Institute, Cleveland Clinic, Cleveland, OH; and †Department of Biostatistics, University of Washington, Seattle, WA. Address correspondence and reprint requests to Haider Mahdi, MD, Ob/Gyn and Women’s Health Institute, Copyright * 2014 by IGCS and ESGO ISSN: 1048-891X DOI: 10.1097/IGC.0000000000000324 International Journal of Gynecological Cancer
9500 Euclid Ave, Cleveland, OH 44195. E-mail: [email protected]. The authors declare no conflicts of interest. Additional contributions: For this study, the linked American College of Surgeons National Surgical Quality Improvement Program (NSQIP) database was used. The interpretation and reporting of these data are the sole responsibility of the authors. We acknowledge the efforts of the American College of Surgeons NSQIP for the creation of the NSQIP database.
& Volume 25, Number 1, January 2015
Copyright © 2014 by IGCS and ESGO. Unauthorized reproduction of this article is prohibited.
55
International Journal of Gynecological Cancer
Mahdi et al
cancer is the most lethal gynecologic malignancy, O varian representing the fifth leading cause of cancer-related
death in the United States. Considerable advances in surgical and medical therapies have resulted in a significant improvement in survival from this disease. Unfortunately, the Surveillance, Epidemiology, and End Results data 1975 to 2005 indicate that these improvements in ovarian cancer (OC) survival were more prominent in white (W) women with lower survival noted for African American (AA) women with OC.1,2 There is great interest in identifying the reasons for these discrepancies so that they may be remedied in an attempt to narrow the gap in survival between W and AA women. In an attempt to explain these racial discrepancies, many studies have focused on socioeconomic factors, tumor biology, and differences in access to care, treatment, and surgical outcomes. African American women are more likely than their W counterparts to present with advanced-stage disease.2 However, it is noteworthy and concerning that multiple studies have demonstrated that AA women with OC are more likely to experience delays in the initiation of adjuvant chemotherapy and are less likely than W women to receive standard-of-care treatment recommended in accordance with guidelines.3Y5 Treatment adherence to the National Comprehensive Cancer Network guidelines results in improved survival outcomes.3,6 Although treatment at highvolume centers by high-volume providers is associated with both guidelines compliance and better survival, race and socioeconomic status have been associated with an increased likelihood of treatment at a low-volume center by a lowvolume provider.6,7 Both the existence of disparate shortterm perioperative outcomes and their impact on racial disparities observed with regards to survival outcomes are important additional pieces of the picture that have not been thoroughly explored. If AA women present with more preoperative medical comorbidities and more advanced disease, it is worth exploring the possibility that they require more extensive surgeries and experience more morbidity and mortality in the perioperative period. Short-term mortality could contribute to differences in overall survival. Furthermore, perioperative morbidity can contribute to treatment delays including time to receive chemotherapy, which can result in worse progression-free and overall survivals.8 Using the ACS-NSQIP database, we detailed the perioperative comorbidities, surgical complexity, and postoperative outcomes of women who underwent surgical treatment for OC. The objective of this study was to analyze the association of race with 30-day postoperative morbidity and mortality.
METHODS Data Source American College of Surgeons NSQIP is a risk-adjusted data collection mechanism that collects and analyzes clinical outcomes data. Participating hospitals use their collected data to develop quality initiatives that improve surgical care and to identify elements in provided health care that can be improved when compared with other institutions. The ACS-NSQIP collects data on 135 variables, including preoperative risk
56
& Volume 25, Number 1, January 2015
factors (including patient demographics, comorbidities, and laboratory values), intraoperative variables, and 30-day postoperative mortality and morbidity outcomes for a systematic and prospective sample of patients undergoing major surgical procedures. Data are collected in a standardized fashion according to strict definitions by dedicated surgical clinical nurse reviewers. Patients are followed throughout their hospital course and after discharge from hospital for up to 30 days postoperatively. A site’s surgical clinical reviewer captures these data using a variety of methods including medical chart extraction, doctor’s office records, 30-day telephone interview with the patients, and other methods. Patients who were diagnosed with OC were identified from the 2005 to 2011 ACS-NSQIP participant use files, which include data collected from 258 academic and community hospitals throughout the United States using the International Classification of Diseases, Ninth Revision codes. Patients with OC were included if they underwent at least salpingo-oophorectomy, debulking, or any of the surgeries listed below the surgical complexity scoring using the Current Procedural Terminology codes. Patients who underwent pelvic exenteration were excluded. Two subgroups were abstracted for comparisonVW and AA patients.
Risk Factors and Outcome All risk factors available in the ACS-NSQIP database were compared between the 2 groups. The primary end points of the study were 30-day mortality, postoperative morbidity, procedure-related complications, surgical re-exploration (return to the operating room within 30 days), and length of hospital stay. Composite end points were created to categorize postoperative complication into few related groups: surgical complications (all surgical site infections, wound disruption, bleeding requiring transfusion, and peripheral nerve injury), renal complications (progressive renal failure, acute renal failure), pulmonary complications (pneumonia, unplanned intubation, respiratory insufficiency requiring ventilation for 48 hours), infectious complications (systematic inflammatory response syndrome, sepsis, septic shock, surgical site infection, and pneumonia), cardiovascular complications (pulmonary embolism, stroke/cerebrovascular event, cardiac arrest, myocardial infarction, deep vein thrombosis requiring therapy), and any nonsurgical complication (any complication except surgical complications). Patients with preoperative sepsis were excluded from the study. Patients who are ventilator dependent, with renal failure, or on dialysis before surgery were excluded from their respective complication category.
Statistical Analysis Associations between categorical covariates were assessed using W2 test. Group differences in means of continuous measures were assessed using the Student t test or the Wilcoxon rank sum test. The preoperative laboratory values were used both as continuous and categorical variables: serum albumin (93 vs e 3 g/dL), hematocrit (G35 vs Q 35%), serum creatinine (Q2 vs G 2 mg/dL), platelets (G350,000 vs Q 350,000 cell/mL), and white blood cell (WBC) are 109/mL. * 2014 IGCS and ESGO
Copyright © 2014 by IGCS and ESGO. Unauthorized reproduction of this article is prohibited.
International Journal of Gynecological Cancer
& Volume 25, Number 1, January 2015
To adjust for surgical complexity, patients were given a specific score for each procedure. Then, based on the number of procedures performed, the sum of these scores was calculated. A score of 1 was given to any of the following procedures: hysterectomy and/or salpingo-oophorectomy, debulking, lymphadenectomy, or omentectomy. A score of 2 was given to any of the following procedures: small or large bowel resection, gastrectomy, hepatectomy, splenectomy, and pancreatectomy. Multivariable logistic regression models were used to assess the association between race and 30-day postoperative complications while controlling for all other confounders. For the creation of the models, we considered all preoperative variables available in the ACS-NSQIP database, including demographics (age and race), preoperative health status and comorbidities, preoperative laboratory values (serum albumin, creatinine, WBC count, platelet count, and hematocrit), and operative factors (operative time, American Society of Anesthesiologists (ASA) class, surgical complexity). The preoperative morbidities, operative risk factors, or demographic variables that were significantly associated with postoperative complications in the univariate logistic regression model with P G 0.05 were included in the multivariable regression model. Another multivariable model for postoperative mortality was created including the preoperative morbidities, operative risk factors, or demographic variables that were significantly associated with 30-day mortality and race in the univariate logistic regression model with P G 0.05. The final logistic regression models were run using the 2 racial groups and all confounders found in this way. All tests of significance were at the P G 0.05 level, and P values were 2-tailed. STATA 10.0 program (College Station, TX) was used for the analysis of the data.
RESULTS
Race and Perioperative Outcome in OC
complications between AA and W patients stratified are listed in Table 2. No difference in the risk of postoperative complications was found between AA and W (33% vs 30%, P = 0.47). Furthermore, no differences in surgical (29% vs 26%, P = 0.40) and no surgical (10% vs 9%, P = 0.75) complications were noted between AA and W, respectively. Among nonsurgical complications, there was no significant difference in the rate of cardiac (4% vs 4%, P = 0.81), pulmonary (4% vs 4%, P = 0.62), infectious (5% vs 4%, P = 0.60), or renal (1% vs 1%, P = 0.85) between AA and W patients, respectively (Table 2). No difference in surgical reexploration was found between AA and W patients (3% vs 4%, P = 0.42). However, AA women were more likely to have longer hospital stay compared with W women with mean hospital stay of 8 versus 6.5 days (SD 6.9 vs 6.3 days), respectively (P G 0.001). In multivariable logistic regression model after adjustment for all confounders including surgical complexity, demographics, and associated morbidities, AA race was not an independent predictor of ‘‘any postoperative complications’’ (odds ratio [OR], 0.99; 95% confidence interval [CI], 0.65Y1.5; P = 0.96; Table 3). Significant predictors of postoperative complications were age, ascites, surgical complexity, operative time, anemia, hypoalbuminemia, and higher ASA class.
Thirty-Day Postoperative Mortality No difference in 30-day mortality was found between AA and W patients (3% vs 2%, P = 0.45, Table 2). In multivariable logistic regression model after adjustment for all confounders including surgical complexity, demographics, and associated morbidities, AA race was not an independent predictor of 30-day postoperative mortality (OR, 0.89; 95% CI, 0.25Y2.43; P = 0.83; Table 4). Significant predictors of postoperative mortality were surgical complexity, preoperative weight loss, and hypoalbuminemia.
Demographics and Clinical Characteristics One thousand six hundred forty-nine women were included in this study, 1510 (92%) were W and 139 (8%) were AA. Demographics and clinical characteristics of each group are shown in Table 1. African American patients with OC were younger (mean 58.6 vs 61.3 years, P G 0.001) and more likely to be obese (P G 0.001), have diabetes (P G 0.001), or hypertension requiring medications (P = 0.006) and weight loss within 6 months before surgery (P = 0.003, Table 1). Furthermore, AA were more likely to have higher serum creatinine level greater than or equal 2 mg/dL (6% vs 2%, P = 0.018), hypoalbuminemia (serum albumin level e3 mg/dL) (40% vs 24%, P G 0.001), and anemia (hematocrit G35 mg/dL) (45% vs 25%, P G 0.001) (Table 1). No difference in surgical complexity was found between the 2 racial groups (P = 0.27, Table 1). However, there was a trend in favor of longer mean operative time for AA compared with W (203 vs 189 minutes, P = 0.065, Table 1).
Postoperative Complications The rate of any complication among the entire cohort was 30%. The rates and differences in the risk of postoperative
DISCUSSION We used the ACS-NSQIP database to describe the disparities in perioperative comorbidities, surgical complexity, and postoperative outcomes between W and AA women who underwent surgery for OC. African American women with OC were more likely to have preoperative morbidities and longer hospital stay, but no difference in surgical complexity and reoperation was found compared with their W counterparts. Furthermore, AA race was not an independent predictor of adverse 30-day postoperative morbidity or mortality after surgery for OC even in multivariate analysis after adjusting for preoperative morbidities and surgical complexity. Therefore, it does not seem that immediate postoperative outcome discrepancies can account for the differences in disease-specific and overall survival that have been reported for AA women with OC. Complete cytoreductive surgery is one of the most important goals of OC care and the extent of surgery needed to achieve this goal varies from patient to patient. Understanding discrepancies in critical surgical outcomes and perioperative events are vital to understand discrepancies in
* 2014 IGCS and ESGO
Copyright © 2014 by IGCS and ESGO. Unauthorized reproduction of this article is prohibited.
57
Mahdi et al
International Journal of Gynecological Cancer
& Volume 25, Number 1, January 2015
TABLE 1. Demographic and clinical characteristics of patients with OC stratified by race Variable
Total Cohort (n = 1649)
Age groups, n (%) G60 y 60Y69 y 70Y79 y Q80 y Body mass index, n (%) G18 18Y29.9 30Y39.9 Q40 ASA class, n (%) 1Y2 3 Q4 Dependent functional status, n (%) Current smoker (within 1 year), n (%) Diabetes, n (%) Hypertension on medications, n (%) Cardiac comorbidities, n (%) Respiratory comorbidities, n (%) Renal comorbidities, n (%) Neurologic comorbidities, n (%) Ascites, n (%) Weight loss, n (%) Blood transfusion before surgery, n (%) Steroid use, n (%) Chemotherapy before surgery Yes Unknown Preoperative anemia, n (%)* Preoperative hypoalbuminemia, n (%)† Preoperative serum creatinine Q2 mg/dL, n (%) Preoperative leukocytosis, n (%)‡ Preoperative thrombocytosis, n (%) Surgical complexity, n (%) Complexity score 1Y3 Complexity score 93 Emergency, n (%) Operative time, n (%) e60 min 61Y120 min 121Y180 min Q180 min
W (n = 1510)
AA (n = 139)
745 468 316 120
(45) (28) (19) (7)
671 (44) 438 (29) 293 (19) 108 (7)
74 30 23 12
(53) (22) (17) (9)
1017 20 451 142
(62) (1) (28) (9)
953 (64) 19 (1) 399 (27) 121 (8)
64 1 52 21
(46) (1) (38) (15)
764 815 70 36 218 171 710 55 51 3 57 359 81 19 35
(46) (49) (4) (2) (13) (10) (43) (3) (3) (0) (3) (22) (5) (1) (2)
710 (47) 739 (49) 61 (4) 33 (2) 198 (13) 141 (9) 635 (42) 51 (3) 45 (3) 2 (0) 49 (3) 334 (22) 67 (4) 14 (1) 32 (2)
54 (39) 76 (55) 9 (6) 3 (8) 2 (20) 30 (22) 75 (54) 4 (8) 6 (4) 1 (1) 8 (6) 25 (18) 14 (10) 5 (4) 3 (2)
104 545 436/1607 300/1204 43/1568 123/1602 486/1600
(6) (33) (27) (25) (3) (8) (30)
92 (6) 512 (34) 374/1471 (25) 259/1107 (24) 35/1432 (2) 111/1466 (8) 445/1465 (30)
1388/1630 (85) 242/1630 (15) 32 (2)
1267/1493 (85) 226/1493 (15) 29 (2)
40/1648 326/1648 488/1648 794/1648
(2) (20) (30) (48)
38/1509 (3) 297/1509 (20) 459/1509 (30) 715/1509 (47)
12 33 62/136 41/103 8/136 12/136 41/135
(9) (24) (46) (40) (6) (8) (30)
121/137 (88) 16/137 (12) 3 (2) 2/139 29/139 29/139 78/139
(1) (21) (21) (57)
P
0.13
G0.001
0.10 0.98 0.67 G0.0001 0.007 0.75 0.38 0.12 0.12 0.25 0.003 0.005 0.97 0.038
G0.001 0.0003 0.02 0.76 0.99 0.27 0.84
0.07
(Continued on next page)
58
* 2014 IGCS and ESGO
Copyright © 2014 by IGCS and ESGO. Unauthorized reproduction of this article is prohibited.
International Journal of Gynecological Cancer
& Volume 25, Number 1, January 2015
Race and Perioperative Outcome in OC
Total Cohort (n = 1649)
W (n = 1510)
AA (n = 139)
189 (89)
203 (89)
TABLE 1. (Continued) Variable Operative time, mean (SD), min§
P 0.065
Statistical significance was determined by a P G 0.05. Preoperative thrombocytosis was defined as a platelet count of more than 350 109/L. Elevated preoperative serum creatinine level was defined as greater than or equal to 2 mg/dL. *Preoperative anemia was defined as a hematocrit less than 35%. †Preoperative hypoalbuminemia was defined as a serum albumin level less than or equal to 3 g/dL. ‡Preoperative leukocytosis was defined as a WBC count greater than 11 109/L. §Mean values are provided, and comparisons are made with Student t test.
long-term OC survival. African American women who are surgically cytoreduced at a tertiary care center are as likely to be left with minimal residual disease with an elimination of the racial discrepancy in overall survival.9 Using NSQIP data, Parsons et al10 looked at short-term survival outcomes after cancer surgery and found, similar to us, that non-W patients have similar short-term operative outcomes but remain hospitalized longer. Notably, they pointed out that ACS-NSQIP hospitals are quality-driven hospitals and that improved access to quality-driven surgical facilities may drive a narrowing in racial disparities in the immediate postoperative period. In OC, improved short- and long-term outcomes as well as improved compliance with recommended treatment guidelines have repeatedly been associated with high-volume, quality-driven, tertiary care facilities; furthermore, treatment at these facilities has been demonstrated to result in an amelioration of racial disparities observed with regards to outcomes and survival.3,6,7,9 In this study, AA patients were more likely to have preoperative comorbidities compared with W patients, specifically higher body mass index, diabetes, hypertension requiring medications, serum creatinine level greater than or equal 2, hypoalbuminemia, and anemia. Some of these preoperative comorbidities were significantly associated with adverse postoperative morbidity and mortality in multivariate analysis such as hypoalbuminemia and anemia. In fact, hypoalbuminemia was an independent predictor of adverse postoperative morbidity and 30-day mortality after surgery for OC. Uppal et al11 recently reported a significant increase in the risk of perioperative complications and 30-day mortality among patients with serum albumin level less than 3 g/dL undergoing open surgery for gynecologic malignancies. Other independent predictors of adverse perioperative complications were surgical complexity, operative time, ascites, higher ASA class, and older age. Most of these factors are not modifiable risk factors and have been reported in prior studies. However, optimal management of perioperative comorbidities and careful selection of the appropriate candidate for primary debulking surgery are very important to minimize the risk of perioperative morbidities and mortality and also to improve the short- and long-term outcome of this patient population. Our results indicate AA women stay in the hospital longer. This is likely multifactorial including higher prevalence of preoperative comorbidities, socioeconomic, and
insurance status. Importantly, if modifiable causes for their prolonged stay can be identified and ameliorated with proper preoperative planning, this identifies a key area for potential cost savings. Although our study provides insight into the short-term operative outcomes across the 2 racial groups in ACS-NSQIP, we acknowledge several limitations. First, we lacked information on patients’ socioeconomic and insurance status, which have been shown to correlate with race. Alternatively, it is also possible that AA patients treated within ACS-NSQIP hospitals have better socioeconomic status when compared with other patients who receive care at lower quality hospitals. Interestingly, AA women with OC were underrepresented in quality-seeking hospitals represented in NSQIP data set. On the other hand, Census data showed that AAs comprise 13.1% of the entire population in the United States for 2013.12 We did not adjust for surgeon or hospital volume, which have been associated with surgical outcomes and survival. Our results were not adjusted for stage at cancer diagnosis, which influenced treatment plans and overall survival outcomes. However, we adjusted for extent of surgical treatment, which is a proxy for burden of disease. Furthermore, only 15% of the patients in this study underwent ultraradical tumor debulking, which could be a surrogate marker of less advanced disease or of low surgical aggressiveness in advanced disease. Finally, much of the data used requires reliance on Current Procedural Terminology and International Classification of Diseases, Ninth Revision codes, which raises concerns regarding accuracy and validity of the data. Notably, other investigators have closely scrutinized the NSQIP database and found it to be highly accurate.13 Despite these limitations, we present evidence gathered from more than 250 quality-seeking hospitals that reveal that AA and W women experience relatively comparable short-term operative outcomes and mortality after cytoreductive surgery for OC. In summary, AA race was not independently associated with a greater risk for postoperative complications or mortality after OC surgery. However, AA women were more likely to have preoperative comorbidities and stay longer in the hospital. Therefore, efforts should be directed to optimally manage the preoperative comorbidities in this patient population to achieve postoperative outcome equivalent to their W counterpart. Identification of factors that contribute to the poor survival of AA women will illuminate ways to improve the care they receive and outcomes they experience. Potential
* 2014 IGCS and ESGO
Copyright © 2014 by IGCS and ESGO. Unauthorized reproduction of this article is prohibited.
59
60
Mortality, n/available data (%) Overall 34/1581 Complexity score 1Y3 29/1331 Complexity score 93 4/236 Surgical complications, n/available data (%) Overall 437/1649 Complexity score 1Y3 289/1388 Complexity score 93 146/242 Nonsurgical complications, n/available data (%) Overall 154/1649 Complexity score 1Y3 105/1388 Complexity score 93 48/242 Renal complications, n/available data (%) Overall 10/1649 Complexity score 1Y3 6/1388 Complexity score 93 4/242 Pulmonary complications, n/available data (%) Overall 59/1649 Complexity score 1Y3 42/1388 Complexity score 93 17/242 Sepsis, n/available data (%) Overall 69/1649 Complexity score 1Y3 41/1388 Complexity score 93 27/242 Cardiac complications, n/available data (%) Overall 65/1649 Complexity score 1Y3 51/1388 Complexity score 93 14/242 Any complication, n/available data (%) Overall 502/1649 Complexity score 1Y3 343/1388 Complexity score 93 157/242 Surgical re-exploration, n/available data (%) Overall 69/1648
Copyright © 2014 by IGCS and ESGO. Unauthorized reproduction of this article is prohibited.
53/1510 (4) 38/1267 (3) 15/226 (7) 62/1510 (4) 35/1267 (3) 26/226 (12) 59/1510 (4) 46/1221 (4) 13/213 (6)
(4) (3) (7) (4) (3) (11) (4) (4) (6)
65/1509 (4)
9/1501 (1) 5/1257 (0) 4/222 (2)
(1) (0) (2)
(4)
140/1510 (9) 94/1267 (7) 45/226 (20)
(9) (8) (20)
456/1510 (30) 309/1267 (24) 145/226 (64)
396/1510 (26) 260/1267 (21) 134/226 (59)
(27) (21) (60)
4/135 (3)
46/139 (33) 34/121 (28) 12/16 (75)
6/139 (4) 5/116 (4) 1/16 (6)
7/132 (5) 6/121 (5) 1/16 (6)
6/139 (4) 4/117 (3) 2/16 (12)
1/138 (1) 1/120 (1) 0/16 (0)
14/139 (10) 11/110 (9) 3/16 (19)
41/139 (29) 29/121 (24) 12/16 (75)
4/131 (3) 4/111 (3) 0/14 (0)
AA (n = 139)
0.42
0.47 0.36 0.38
0.81 0.77 0.93
0.60 0.17 0.51
0.62 0.85 0.37
0.85 0.48 0.59
0.75 0.50 0.91
0.40 0.37 0.21
0.45 0.31 0.61
P
International Journal of Gynecological Cancer
(30) (25) (65)
30/1450 (2) 25/1191 (2) 4/218 (2)
W
(2) (2) (2)
Total Cohort
TABLE 2. Thirty-day outcomes among patients who underwent surgery for OC stratified by race and surgical complexity
Mahdi et al
& Volume 25, Number 1, January 2015
* 2014 IGCS and ESGO
0.298 0.252 0.269
0.008 0.002 0.019
0.95 0.18
International Journal of Gynecological Cancer
& Volume 25, Number 1, January 2015
Race and Perioperative Outcome in OC
TABLE 3. Multivariate regression analysis for any complication within 30 days after surgery for OC
AA Surgical complexity score 3Y4 Surgical complexity score 94 Age 60Y69 y Age 70Y79 y Age Q80 y Respiratory comorbidities Hypertension on medications Ascites Blood transfusion before surgery Chemotherapy before surgery ASA class 3 ASA class 4Y5 Operative time 1Y2 h Operative time 2Y3 h Operative time 93 h Serum creatinine level Q2 Albumin level 93 Hematocrit Q35 High WBC count (911,000) Platelets count Q350,000 cell/mL Statistical significance was determined by a P G 0.05. *Mean values are provided, and comparisons are made with Student t test.
0.29 (0.72) 0.30 (0.60) 0.95 (1.04)
0.46 (0.83) 0.36 (0.69) 1.25 (1.29)
7.99 (6.91) 7.22 (6.04) 14.38 (9.81) 6.59 (6.35) 5.91 (6.01) 9.82 (7.29)
Complexity score 1Y3 Complexity score 93 Length of hospital stay, mean (SD), d* Overall Complexity score 1Y3 Complexity score 93 No. complications, mean (SD)* Overall Complexity score 1Y3 Complexity score 93
47/1388 (3) 22/241 (9)
43/1267 (3) 22/225 (10)
4/117 (3) 0/16 (0)
Variable
OR (95% CI) 0.99 1.63 4.40 1.36 1.44 1.47 1.50 1.08 1.79 1.50
P
(0.65Y1.50) 0.96 (1.26Y2.1) 0.0002 (2.63Y7.46) G0.0001 (1.02Y1.83) 0.038 (1.03Y2.01) 0.032 (0.91Y2.37) 0.11 (0.78Y2.85) 0.21 (0.083Y1.39) 0.55 (1.35Y2.39) 0.0001 (0.53Y4.44) 0.44
1.49 (0.90Y2.43) 1.36 (1.05Y1.75) 1.81 (1.01Y3.21)
0.11 0.019 0.04 0.10 9.17 (1.91Y160.4) 0.03 17.9 (3.78Y322.6) 0.005 1.41 (0.71Y1.14) 0.11 0.70 (0.51Y0.97) 0.037 0.52 (0.40Y0.68) G0.0001 1.43 (0.93Y2.18) 0.10 1.21 (0.92Y1.59) 0.16
targets remain access to high-quality medical care both before their diagnosis to optimize preoperative comorbidities and afterwards to maximize the receipt of timely guidelineadherent cancer treatments plans. It is unlikely that immediate postoperative complications or mortality contribute to the racial discrepancies seen in OC survival. Further research is needed to explore the reasons behind the racial disparities that have been observed, and efforts should be made to ensure women of all races have access to timely and guidelineadherent adjuvant therapy.
TABLE 4. Multivariate regression analysis for 30-day mortality after surgery for OC Variable AA race Surgical complexity score 3Y4 Surgical complexity score 94 Weight loss with 6 months before surgery Albumin level 94
OR (95% CI) 0.89 0.41 0.44 5.23
P
(0.25Y2.43) 0.83 (0.16Y0.92) 0.04 (0.02Y2.16) 0.42 (1.94Y12.68) G0.001
0.14 (0.06Y0.30)
* 2014 IGCS and ESGO
Copyright © 2014 by IGCS and ESGO. Unauthorized reproduction of this article is prohibited.
G0.001
61
International Journal of Gynecological Cancer
Mahdi et al
REFERENCES 1. Albain KS, Unger JM, Crowley JJ, et al. Racial disparities in cancer survival among randomized clinical trials patients of the southwest oncology group. J Natl Cancer Inst. 2009;101:984Y992. 2. Chan JK, Zhang M, Hu JM, et al. Racial disparities in surgical treatment and survival of epithelial ovarian cancer in united states. J Surg Oncol. 2008;97:103Y107. 3. Howell EA, Egorova N, Hayes MP, et al. Racial disparities in the treatment of advanced epithelial ovarian cancer. Obstet Gynecol. 2013;122:1025Y1032. 4. Bristow RE, Powell MA, Al-Hammadi N, et al. Disparities in ovarian cancer care quality and survival according to race and socioeconomic status. J Natl Cancer Inst. 2013;105:823Y832. 5. Wright J, Doan T, McBride R, et al. Variability in chemotherapy delivery for elderly women with advanced stage ovarian cancer and its impact on survival. Br J Cancer. 2008;98:1197Y1203. 6. Bristow RE, Chang J, Ziogas A, et al. Adherence to treatment guidelines for ovarian cancer as a measure of quality care. Obstet Gynecol. 2013;121:1226Y1234. 7. Bristow RE, Chang J, Ziogas A, et al. High-volume ovarian cancer care: survival impact and disparities in access for advanced-stage disease. Gynecol Oncol. 2014;132:403Y410.
62
& Volume 25, Number 1, January 2015
8. Mahner S, Eulenburg C, Staehle A, et al. Prognostic impact of the time interval between surgery and chemotherapy in advanced ovarian cancer: analysis of prospective randomised phase III trials. Eur J Cancer. 2013;49:142Y149. 9. Bristow RE, Ueda S, Gerardi MA, et al. Analysis of racial disparities in stage IIIC epithelial ovarian cancer care and outcomes in a tertiary gynecologic oncology referral center. Gynecol Oncol. 2011;122:319Y323. 10. Parsons HM, Habermann EB, Stain SC, et al. What happens to racial and ethnic minorities after cancer surgery at american college of surgeons national surgical quality improvement program hospitals? J Am Coll Surg. 2012;214:539Y547 discussion 547Y549. 11. Uppal S, Al-Niaimi A, Rice LW, et al. Preoperative hypoalbuminemia is an independent predictor of poor perioperative outcomes in women undergoing open surgery for gynecologic malignancies. Gynecol Oncol. 2013;131:416Y422. 12. U.S. Census Bureau. State and County QuickFacts. http://quickfacts. census.gov/qfd/states/00000.html. Accessed May 5, 2014. 13. Sellers MM, Merkow RP, Halverson A, et al. Validation of new readmission data in the american college of surgeons national surgical quality improvement program. J Am Coll Surg. 2013;216:420Y427.
* 2014 IGCS and ESGO
Copyright © 2014 by IGCS and ESGO. Unauthorized reproduction of this article is prohibited.
