Abstract
IMPORTANCE
Optimal chemotherapy dosing in obese patients remains uncertain, with variation in practice. Dose reduction strategies are often used to avoid chemotoxicity, but recent American Society of Clinical Oncology guidelines recommend full dose.
OBJECTIVE
To evaluate the impact of body mass index (BMI) on chemotherapy dosing and of dose reduction on ovarian cancer survival.
DESIGN, SETTING, AND PARTICIPANTS
Cohort study in Kaiser Permanente Northern California (KPNC) health care setting of patients with primary invasive epithelial ovarian cancers diagnosed from January 2000 through March 2013. Analyses focused on 806 patients receiving adjuvant first-line therapy of carboplatin and paclitaxel with curative intent.
MAIN OUTCOMES AND MEASURES
Overall and ovarian cancer–specific mortality. Deaths were identified through the KPNC Mortality Linkage System, with median follow-up of 52.5 months. Hazard ratios (HRs) and 95% CIs were estimated from proportional hazards regression, accounting for prognostic variables including age at diagnosis, race, stage, grade, histologic type, chemotoxic effects, comorbidities, cancer antigen 125 levels, and BMI at diagnosis.
RESULTS
The strongest predictor of dose reduction was a high BMI. Compared with normal-weight women, obese class III women received 38% and 45% lower doses in milligrams per kilogram of body weight of paclitaxel and carboplatin, respectively (P < .001 for each agent). They also received lower relative dose intensity (RDI) for each agent and the combined regimen, calculated as average RDI (ARDI). Mean ARDI was 73.7% for obese class III women and 88.2% for normal-weight women (P < .001). Lower ARDI (<70%) was associated with worse overall (HR, 1.62 [95% CI, 1.10–2.37]) and ovarian cancer–specific survival (HR, 1.69 [95% CI, 1.12–2.55]). Women who were obese at diagnosis appeared to have better survival. In multivariable-adjusted analyses considering joint effects by BMI and ARDI, compared with women with normal weight and no dose reduction, normal-weight women with dose reduction (ARDI < 85%) experienced worse survival (HR, 1.50 [95% CI, 1.02–2.21]). For each BMI category, those with ARDI less than 85% had worse survival than those without dose reduction. The improved survival among obese women was no longer apparent with dose reduction.
CONCLUSIONS AND RELEVANCE
Lower RDI was an independent predictor of ovarian cancer mortality. This finding was strongest among normal-weight women but seen at all levels of BMI. Our results suggest that body size should not be a major factor influencing dose reduction decisions in women with ovarian cancer.
Ovarian cancer is the second most common gynecologic cancer and the leading cause of death from gynecologic malignant neoplasm in the United States.1 Standard treatment for epithelial ovarian cancer (EOC) consists of surgical cytoreduction followed by platinum and taxane–based chemotherapy,2 with the primary adjuvant chemotherapy regimen being combination therapy with carboplatin and paclitaxel.3 However, there is considerable variation in practice and uncertainty regarding optimal chemotherapy dosing of overweight and obese patients, mainly based on concerns that using full dose based on actual weight may lead to increased toxicity.4,5
Body surface area (BSA) has been used for calculation of chemotherapy dose for some agents for more than 40 years. The use of BSA was based on the notion that it is proportionate to metabolism, a theory first proposed by Rubner in 1883, with the most common formula currently used for estimating BSA developed more than 100 years ago and based on a few subjects.6 Because dosing of some chemotherapy drugs, such as paclitaxel, is based on BSA, concerns that use of high doses of drugs in obese patients may result in excess toxicity has resulted in various strategies for dose reduction. For example, dose may be capped at a BSA of 2 m2 or calculated on the basis of ideal rather than actual weight.7 In contrast, carboplatin dosing is based on renal function, measured as glomerular filtration rate, and estimated by calculating creatinine clearance.8,9 Several formulas are used in clinical practice to calculate creatinine clearance, with the commonly used Cockcroft-Gault equation8 also a function of weight. There is uncertainty regarding the use of ideal or actual weight in the formula, with concern that using actual bodyweight may overestimate glomerular filtration rate and, consequently, carboplatin dosing.8,9
Although there is growing evidence that dose reduction decreases cancer survival, several surveys have estimated that up to 40% of obese patients receive reduced doses.5 An expert panel convened by the American Society of Clinical Oncology recommended full dose based on actual weight to treat the obese patient and found no evidence of associated increased short- or long-term toxicity.5 However, these conclusions were mostly based on breast cancer studies. Optimal chemotherapy dosing in overweight and obese patients with ovarian cancer remains uncertain.
This study evaluated the association of body mass index (BMI; calculated as weight in kilograms divided by height in meters squared) and chemotherapy dosing, predictors of dose reduction, and chemotherapy efficacy in improving survival and related toxic effects according to BMI in a large population-based cohort study of patients with ovarian cancer.
Methods
Study Population
The Kaiser Permanente Research on Ovarian Cancer Study (KP-ROCS) is a cohort study of EOC based in the Kaiser Permanente Northern California (KPNC) integrated health care setting. A flowchart describing the KP-ROCS cohort is shown in Figure 1. Primary invasive EOC cases were identified through the KPNC Cancer Registry and were eligible for inclusion if age at diagnosis was 21 years or older, with date of diagnosis between January 2000 and March 2013. Women who were not KPNC members at diagnosis or who disenrolled from the KPNC health plan before full standard chemotherapy treatment could have been completed were also excluded. This resulted in a cohort of 2299 patients with EOC.
Figure 1.
Flowchart for Patient Selection in the Kaiser Permanente Research on Ovarian Cancer Study (KP-ROCS) Cohort and the Chemotherapy Subcohort
Kaiser Permanente Northern California began to include BMI or weight and height data in electronic records in 2002. We defined BMI at diagnosis as the measurement closest to the diagnosis date and within 6 months before diagnosis. If BMI was not available during that period, then the closest measurement to the diagnosis date within 2 months after diagnosis and before treatment was selected. Body mass index was categorized according to World Health Organization classification, as shown in Table 1. As illustrated in Figure 1, 453 cases were excluded because they did not have data on BMI at diagnosis, resulting in a cohort of 1846 EOC cases. Cases without data on BMI were those occurring early in the case identification period (ie, occurring prior to inclusion of BMI in electronic health records). Distributions by major demographic and clinical characteristics for women with and without BMI at diagnosis were similar (data not shown).
Table 1.
Select Demographic and Clinical Characteristics, Kaiser Permanente Research on Ovarian Cancer Study Cohort and Chemotherapy Subcohort
| Characteristic | No. (%) | P Valuea | |
|---|---|---|---|
| Overall Cohort (n = 1846) |
Chemotherapy Subcohort (n = 806) |
||
| Age at diagnosis, y | |||
| 21–39 | 86 (4.7) | 39 (4.8) | <.001 |
| 40–49 | 254 (13.8) | 117 (14.5) | |
| 50–69 | 983 (53.2) | 503 (62.4) | |
| >70 | 523 (28.3) | 147 (18.2) | |
| Race | |||
| White | 1474 (79.9) | 651 (80.8) | .87 |
| African American | 106 (5.7) | 42 (5.2) | |
| Asian | 219 (11.9) | 95 (11.8) | |
| Other | 46 (2.5) | 18 (2.2) | |
| Ethnicity | |||
| Hispanic | 168 (9.1) | 82 (10.2) | .57 |
| Not Hispanic | 1655 (89.6) | 714 (88.6) | |
| AJCC stage | |||
| I | 404 (21.9) | 219 (27.2) | <.001 |
| II | 171 (9.3) | 98 (12.2) | |
| III | 769 (41.7) | 342 (42.4) | |
| IV | 476 (25.8) | 140 (17.4) | |
| Unknown | 26 (1.4) | 7 (0.9) | |
| Grade, SEER definition | |||
| Well differentiated | 125 (6.8) | 50 (6.2) | <.001 |
| Moderately differentiated | 237 (12.8) | 133 (16.5) | |
| Poorly differentiated | 663 (35.9) | 318 (39.4) | |
| Undifferentiated | 311 (16.8) | 133 (16.5) | |
| Unknown | 510 (27.6) | 172 (21.3) | |
| Histologic type | |||
| Serous | 963 (52.2) | 432 (53.6) | <.001 |
| Mucinous | 94 (5.1) | 39 (4.8) | |
| Endometrioid | 190 (10.3) | 107 (13.3) | |
| Clear cell | 133 (7.2) | 76 (9.4) | |
| Other | 466 (25.2) | 152 (18.9) | |
| BMI at diagnosis | |||
| Underweight, <18.50 | 48 (2.6) | 20 (2.5) | .69 |
| Normal, 18.50–24.99 | 685 (37.1) | 297 (36.8) | |
| Overweight, 25.00–29.99 | 566 (30.7) | 248 (30.8) | |
| Obese 1, 30.00–34.99 | 289 (15.7) | 134 (16.6) | |
| Obese 2, 35.00–39.99 | 150 (8.1) | 70 (8.7) | |
| Obese 3, ≥40.00 | 108 (5.8) | 37 (4.6) | |
Abbreviations: American Joint Committee on Cancer; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); SEER, Surveillance, Epidemiology, and End Results.
Based on χ2 goodness-of-fit test.
We included patients who started a standard regimen (surgery followed by chemotherapy administration every 3 weeks) of intravenous adjuvant first-line therapy of carboplatin and paclitaxel within 6 months of diagnosis with curative intent. This regimen was used in approximately 79% of patients receiving chemotherapy. Selecting patients with the same regimen and route of administration minimized variation in dosing parameters and possible adverse effects that may result in dose reduction and delay, thereby facilitating focus on the effects of body size on dose received. Exclusion of an additional 35 women missing dosing variables resulted in an analytic subcohort of 806 patients with EOC (Figure 1).
Data Obtained for Analysis
Data on demographic characteristics (age, race, ethnicity), tumor characteristics (stage, grade, histologic type), treatment information, toxic effects, and comorbidities were obtained from KPNC electronic databases, including its Virtual Data Warehouse (VDW).10,11 Information on race/ethnicity came from the VDW’s Tumor File, which is based on the KPNC Cancer Registry, and was considered as potential covariates. The VDW extracts data into a standardized data model that facilitates research conducted at KPNC and other integrated health care settings that participate in the National Cancer Institute (NCI)–supported Cancer Research Network (http://crn.cancer.gov). Variables for chemotherapy dosing analyses, described in eTable 1 in the Supplement, included actual dose, relative dose intensity (RDI) (actual dose administered per week/expected dose per week), early discontinuation, and treatment delay. The RDI was computed for paclitaxel and carboplatin separately, and for the combination regimen by computing the average RDI (ARDI), as described by others.9,12,13 Expected doses were based on National Comprehensive Cancer Network Guidelines (http://www.nccn.org) and are presented in eTable 1 in the Supplement. Actual weight was used when calculating BSA and creatine clearance to compute expected doses. The area under the curve (AUC) value used for carboplatin dosing was that indicated in the medical record for the first course of therapy and was 6 in 75% and 5 in 21% of the women. Serum creatinine to estimate creatinine clearance was also the value associated with the first course of therapy. Chemotherapy reduction was defined as an RDI for the full regimen of less than either 85%14–16 or 70%.17 Early discontinuation was defined as not completing the full 6 scheduled treatments18 or alternatively, 4 scheduled treatments. Treatment delay was defined as a delay in receiving scheduled chemotherapy treatment of more than 7 days.
Chemotoxic effects considered as potential predictors of chemotherapy dose reduction included myelosuppression (severe neutropenia or thrombocytopenia) and neuropathy before chemotherapy (within 1 month prior to first cycle for neutropenia and thrombocytopenia and within 1 month prior to diagnosis to first cycle for neuropathy) or during chemotherapy (between first and last cycle). The National Cancer Institute’s Common Terminology Criteria for Adverse Events (NCI CTCAE), version 3.0, was used to identify chemotherapy-related toxicity grade 3/4. Use of granulocyte colony-stimulating factor (G-CSF) for prophylaxis or treatment of neutropenia was also obtained from the VDW.
Comorbidities considered were those known to be related to obesity, to impair chemotherapy dosing or survival,19 or to occur commonly among patients with ovarian cancer.20 These included diabetes mellitus, hypertension, cardiovascular disease, and renal disease comprising acute kidney disease during chemotherapy or chronic renal insufficiency occurring at least 1 year before diagnosis throughout the end of follow-up. On the basis of our inclusion criteria, all patients underwent surgery before chemotherapy; cancer antigen 125 (CA125) level after treatment was used as a marker of residual disease.21,22
Outcomes were total and ovarian cancer–specific deaths through December 31, 2013, identified through the KPNC Mortality Linkage System, which includes date of death and International Classification of Diseases, Ninth Revision, Clinical Modification coded underlying cause of death. Through that date, 334 deaths, 291 due to ovarian cancer, were identified in the chemotherapy subcohort, with median (SD) follow-up of 52.5 (38.4) months. The study was approved by the institutional review boards of Kaiser Permanente Northern California, which waived the requirement for informed consent due to the retrospective nature of the study and high fatality rate for this type of cancer, and Rutgers Biomedical and Health Sciences at Rutgers University.
Statistical Analyses
Distributions for demographic and clinical characteristics were compared across categories of BMI using χ2 tests. χ2 Goodness-of-fit tests were used to compare distributions in the overall cohort and the chemotherapy subcohort. Mean values of RDI and ARDI, actual dose of paclitaxel and carboplatin, number of cycles, and treatment duration across BMI categories were compared using analysis of variance. The proportions of women with dose reduction, early discontinuation, and treatment delay across BMI groups were also compared using χ2 tests. Predictors of dose reduction (RDI < 85%) were evaluated by logistic regression analysis with dose reduction as a dichotomous outcome.
The impact of dose reduction on survival was evaluated by comparing survival curves using the Kaplan-Meier method and log-rank test and by estimating HRs and 95% CIs for all-cause and ovarian cancer–specific mortality using Cox proportional hazards regression analysis. Type 3 Wald χ2 tests were conducted to evaluate whether there were significant differences in mortality among exposure categories.P values for trend were computed by including the median in each category as a continuous variable in regression models. Covariates included age at diagnosis, race, American Joint Committee on Cancer stage, grade, histologic type, chemotherapy-related toxic effects, BMI at diagnosis, comorbidities (diabetes, hypertension, cardiovascular disease, and renal disease), and CA125 levels after treatment. All P values were 2 sided, and P < .05 was defined as statistically significant. SAS, version 9.2 (SAS Institute), was used for analyses.
Results
Demographic and clinical characteristics were generally similar in our chemotherapy subcohort and the overall KP-ROCS cohort, with the exception, as expected, of the chemotherapy subcohort having a smaller proportion of women older than 70 years or with advanced disease (Table 1). Approximately 30% of patients were obese and another 31% overweight, whereas fewer than 3% were underweight.
BMI and Chemotherapy Dosing
Dose reduction (RDI < 85%) was significantly more common among obese women, particularly those with BMI at least 35, for both paclitaxel (P = .01) and carboplatin (P < .001) (eTable 2 in the Supplement). There was a significant inverse association between BMI and RDI for each agent and for ARDI, for both the first cycle and all cycles combined. Mean ARDI for the regimen for all cycles was 73.7% for obese class 3 women and 88.2% for normal-weight women (P < .001). We also found a significant inverse association between BMI and actual total dose of paclitaxel and carboplatin administered per kilogram of body weight. Compared with normal-weight women, obese class 3 women received 38% and 45% lower doses in milligrams per kilogram of bodyweight of paclitaxel and carboplatin, respectively (P < .001 for both agents).
Toxic Effects and Comorbidities by BMI at Diagnosis
Aside from cardiovascular disease, comorbidities were more common among obese than normal-weight women (eTable 3 in the Supplement). Neutropenia during chemotherapy and G-CSF use were less common among obese than normal-weight women. Chemotherapy-related neuropathy was unrelated to BMI. Chemotoxic effects were more common among those receiving lower RDI for all categories of BMI (data not shown).
Predictors of Dose Reduction
Body mass index at diagnosis was the strongest predictor of dose reduction (RDI < 85%), with a stronger association for carboplatin than paclitaxel (eTable 4 in the Supplement). Compared with normal-weight women, obese women were substantially more likely (for ARDI, odds ratio, 2.85 [95% CI, 1.79–4.55] for obese class 1, and odds ratio, 19.85 [95% CI, 7.21–54.65] for obese class 3) to have received dose reduction. Advanced stage, chemotoxic effects, and cardiovascular disease were also associated with dose reduction.
Dose Reduction and Overall and Ovarian Cancer–Specific Survival
There was a suggestion that obesity at diagnosis was associated with better survival after adjusting for prognostic factors (Table 2). As shown in Table 2 and Table 3, compared with those with an ARDI of 85% to 100%, lower ARDI (<70%) was a strong predictor of overall (HR, 1.62 [95% CI, 1.10–2.37]) and ovarian cancer–specific mortality (HR, 1.69 [95% CI, 1.12–2.55]). Similar findings are shown in eFigure 1 in the Supplement. Similar associations were observed for paclitaxel RDI and carboplatin RDI, although there was not a clear dose-response association for carboplatin RDI (Table 3).
Table 2.
Association of Selected Factors With Mortality After Ovarian Cancer Diagnosis
| Overall Ovarian Cancer Mortality | Ovarian Cancer–Specific Mortality | ||||||
|---|---|---|---|---|---|---|---|
| Factor | No. | Events | HRa (95% CI) |
Type 3 P Valueb |
Events | HRa (95% CI) |
Type 3 P Valueb |
| Age, continuous | 806 | 334 | 1.03 (1.02–1.04) | … | 291 | 1.03 (1.02–1.04) | … |
| Race | |||||||
| White | 651 | 272 | 1 [Reference] | .18 | 238 | 1 [Reference] | .29 |
| Black | 42 | 23 | 1.61 (1.03–2.52) | 19 | 1.52 (0.93–2.49) | ||
| Asian | 95 | 32 | 0.91 (0.61–1.36) | 27 | 0.90 (0.58–1.39) | ||
| Other | 18 | 7 | 1.28 (0.57–2.87) | 7 | 1.41 (0.62–3.21) | ||
| Stage | |||||||
| I/II | 317 | 50 | 1 [Reference] | <.001 | 39 | 1 [Reference] | <.001 |
| III | 342 | 177 | 4.12 (2.85–5.96) | 158 | 4.48 (2.97–6.76) | ||
| IV | 140 | 103 | 7.45 (4.96–11.19) | 90 | 7.80 (4.98–12.22) | ||
| Grade | |||||||
| Well differentiated | 50 | 5 | 1 [Reference] | .20 | 4 | 1 [Reference] | .29 |
| Moderately differentiated | 133 | 39 | 1.91 (0.73–4.99) | 34 | 2.10 (0.72–6.08) | ||
| Poorly differentiated | 318 | 153 | 2.41 (0.96–6.09) | 134 | 2.61 (0.93–7.32) | ||
| Undifferentiated | 133 | 52 | 2.67 (1.02–6.98) | 45 | 2.80 (0.96–8.16) | ||
| Histologic type | |||||||
| Serous | 432 | 214 | 1 [Reference] | .05 | 191 | 1 [Reference] | .02 |
| Mucinous | 39 | 10 | 2.63 (1.30–5.31) | 10 | 3.16 (1.54–6.47) | ||
| Endometrioid | 107 | 23 | 0.88 (0.56–1.39) | 18 | 0.81 (0.48–1.34) | ||
| Clear cell | 76 | 21 | 1.38 (0.82–2.33) | 16 | 1.25 (0.69–2.24) | ||
| Other | 152 | 66 | 1.11 (0.83–1.49) | 56 | 1.03 (0.75–1.42) | ||
| Toxic effectsc | 309 | 137 | 1.15 (0.90–1.45) | .27 | 119 | 1.08 (0.83–1.41) | .55 |
| G-CSF usec | 182 | 72 | 0.79 (0.59–1.05) | .11 | 64 | 0.81 (0.60–1.10) | .18 |
| Diabetes mellitusc | 142 | 62 | 0.98 (0.72–1.35) | .92 | 53 | 0.96 (0.69–1.36) | .84 |
| Hypertensionc | 435 | 187 | 0.74 (0.57–0.96) | .02 | 160 | 0.73 (0.55–0.97) | .03 |
| Cardiovascular diseasec | 392 | 168 | 0.77 (0.59–0.99) | .05 | 144 | 0.72 (0.55–0.95) | .02 |
| Renal diseasec | 243 | 132 | 1.53 (1.19–1.96) | <.001 | 115 | 1.49 (1.14–1.95) | .003 |
| CA125 | |||||||
| <35 | 678 | 239 | 1 [Reference] | <.001 | 21 | 1 [Reference] | <.001 |
| 35–70 | 37 | 25 | 2.18 (1.41–3.38) | 21 | 2.06 (1.28–3.32) | ||
| >70 | 69 | 53 | 2.94 (2.07–4.17) | 51 | 3.16 (2.19–4.54) | ||
| Unknown | 22 | 17 | 6.44 (3.76–11.02) | 13 | 5.54 (3.02–10.16) | ||
| BMI | |||||||
| Underweight | 20 | 10 | 0.66 (0.33–1.32) | .20 | 10 | 0.78 (0.39–1.56) | .16 |
| Normal | 297 | 127 | 1 [Reference] | 110 | 1 [Reference] | ||
| Overweight | 248 | 101 | 0.86 (0.65–1.13) | 90 | 0.88 (0.65–1.18) | ||
| Obese class 1 | 134 | 54 | 0.75 (0.53–1.06) | 43 | 0.67 (0.46–0.99) | ||
| Obese class 2 | 70 | 23 | 0.55 (0.34–0.91) | 19 | 0.53 (0.31–0.91) | ||
| Obese class 3 | 37 | 19 | 0.76 (0.42–1.36) | 19 | 0.82 (0.45–1.49) | ||
| Average RDI, % | |||||||
| >100 | 79 | 35 | 0.84 (0.57–1.25) | .05 | 29 | 0.78 (0.51–1.20) | .03 |
| 100–85 | 386 | 142 | 1 [Reference] | 122 | 1 [Reference] | ||
| <85–70 | 253 | 113 | 1.16 (0.88–1.52) | 101 | 1.21 (0.90–1.62) | ||
| <70 | 88 | 44 | 1.62 (1.10–2.37) | 39 | 1.69 (1.12–2.55) | ||
Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); CA125, cancer antigen 125; ellipses, not applicable; G-CSF, granulocyte colony-stimulating factor; HR, hazard ratio; RDI, relative dose intensity.
HR and CI mutually adjusted for all other factors in the Table.
Based on type 3 Wald χ2 tests.
Reference category is not having these conditions (or nonuser for G-CSF).
Table 3.
Total and Ovarian Cancer–Specific Mortality Associated With Chemotherapy Dose Reduction
| Overall Ovarian Cancer Mortality | Ovarian Cancer–Specific Mortality | |||||||
|---|---|---|---|---|---|---|---|---|
| Dose Reduction | No. | Events | Median Survival, mo |
HR (95% CI)a |
P Valuec | Events | HR (95% CI)a |
P Value |
| Paclitaxel RDI, % | ||||||||
| Dichotomized variable | ||||||||
| ≥85 | 584 | 227 | 44 | 1 [Reference] | .02 | 195 | 1 [Reference] | .007 |
| <85 | 222 | 107 | 39b | 1.35 (1.05–1.73) | 96 | 1.44 (1.11–1.88) | ||
| Four-level variable | ||||||||
| >100 | 141 | 52 | 41 | 0.95 (0.68–1.32) | .005 | 41 | 0.82 (0.57–1.17) | .003 |
| 100–85 | 443 | 175 | 47 | 1 [Reference] | 154 | 1 [Reference] | ||
| <85–70 | 163 | 74 | 43 | 1.21 (0.90–1.62) | 69 | 1.32 (0.98–1.78) | ||
| <70 | 59 | 33 | 28b | 1.78 (1.18–2.69) | 27 | 1.62 (1.03–2.56) | ||
| Carboplatin RDI, % | ||||||||
| Dichotomized variable | ||||||||
| ≥85 | 339 | 135 | 47 | 1 [Reference] | .05 | 114 | 1 [Reference] | .03 |
| <85 | 467 | 199 | 40 | 1.28 (1.00–1.64) | 177 | 1.35 (1.03–1.76) | ||
| Four-level variable | ||||||||
| >100 | 89 | 41 | 54 | 1.05 (0.71–1.56) | .29 | 34 | 0.98 (0.64–1.51) | .17 |
| 100–85 | 250 | 94 | 46 | 1 [Reference] | 80 | 1 [Reference] | ||
| <85–70 | 262 | 116 | 40 | 1.38 (1.03–1.85) | 103 | 1.42 (1.04–1.94) | ||
| <70 | 205 | 83 | 41 | 1.16 (0.82–1.64) | 74 | 1.19 (0.82–1.73) | ||
| Average RDI, % | ||||||||
| Dichotomized variable | ||||||||
| ≥85 | 465 | 177 | 46 | 1 [Reference] | .04 | 151 | 1 [Reference] | .02 |
| <85 | 341 | 157 | 39b | 1.29 (1.01–1.64) | 140 | 1.36 (1.05–1.77) | ||
| Four-level variable | ||||||||
| >100 | 79 | 35 | 54 | 0.84 (0.57–1.25) | .006 | 29 | 0.78 (0.51–1.20) | .003 |
| 100–85 | 386 | 142 | 44 | 1 [Reference] | 122 | 1 [Reference] | ||
| <85–70 | 253 | 113 | 42 | 1.16 (0.88–1.52) | 101 | 1.21 (0.90–1.62) | ||
| <70 | 88 | 44 | 28b | 1.62 (1.10–2.37) | 39 | 1.69 (1.12–2.55) | ||
Abbreviations: HR, hazard ratio; RDI, relative dose intensity.
Adjusted for age at diagnosis, race, body mass index at diagnosis, stage, grade, histologic type, toxic effects during chemotherapy, granulocyte colony-stimulating factor use, diabetes mellitus, hypertension, cardiovascular disease, renal disease, and cancer antigen 125 levels after treatment.
P value <.05 (log-rank test).
For each agent's RDI and average RDI, P value for dichotomized variables corresponds to type 3 P value. For the 4-level variables, P value for trend is presented. Type 3 P value tests are an overall test of whether there are significance differences in mortality across levels of exposure. P for trend tests whether there is a significant dose-response relationship.
Joint Effects of BMI and Dose Reduction on Survival
Compared with normal-weight women receiving ARDI of at least 85%, women of normal weight receiving dose reduction had worse survival (HR, 1.50 [95% CI, 1.02–2.21]) (Figure 2A). For the 3 body size categories, those with dose reduction had poorer survival; the possible survival advantage for obese women disappeared with dose reduction. The interaction between body size and dose reduction was not statistically significant (P = .36). Similar findings were seen for ovarian cancer–specific survival (Figure 2B).
Figure 2. Hazard Ratios for Overall and Ovarian Cancer–Specific Survival According to Body Mass Index (BMI) Category and Average Relative Dose Intensity (ARDI).
Hazard ratios and 95% CIs are adjusted for age at diagnosis, race, BMI at diagnosis, stage, grade, histologic type, toxic effects during chemotherapy, granulocyte colony-stimulating factor use, diabetes mellitus, hypertension, cardiovascular disease, renal disease, and cancer antigen 125 (CA125) levels after treatment.
Discussion
In this study we found that obese patients with ovarian cancer received considerably less paclitaxel and carboplatin per kilogram of body weight and lower RDI for each chemotherapy agent and for the combined regimen. We also found that the strongest predictor of dose reduction, defined as an RDI below 85%, was a high BMI. Dose reduction was associated with reduced survival time, particularly for normal weight women. This association was apparent even after accounting for diagnostic and prognostic factors such as stage of disease, comorbid conditions, or posttreatment CA125 levels, a marker of residual disease.
Only a few studies have evaluated dose reduction in patients with ovarian cancer. Barrett et al,23 using data from the SCOTROC I clinical trial based in Scotland, in which taxane (docetaxel or paclitaxel) dose was calculated on the basis of BSA with no dose capping, reported no statistical difference in taxane dose intensity among BMI categories—unsurprising because the study used no dose capping. Wright et al,24 using Gynecologic Oncology Group clinical trial data, evaluated the impact of bodyweight on carboplatin-related toxic effects and survival inpatients (n = 387) treated with carboplatin and dosed on the basis of the Jelliffe formula, which includes age and creatinine clearance but does not account for weight. They found that, compared with normal-weight patients, obese patients were less likely to have dose reduction and treatment delay and experienced fewer toxic effects. Dose reduction was defined as at least a 30% decrease in dose after the first cycle. An additional study of ovarian cancer25 (n = 198) found that obese patients were less likely to have dose reduction or treatment delay compared with underweight or normal-weight patients, but the difference was not statistically significant (P = .11).
Chemotherapy dosing decisions in obese women tend to occur at the first cycle (eg, to cap dose at a BSA of 2.0 m2 or to use ideal weight rather than actual weight), and thus dose reduction during the course of chemotherapy may be less common for such women.26 For this reason, using RDI is a better method to evaluate dose reduction. In a recent Australian study,16 there were no differences in dose reduction across BMI categories, but a significantly higher proportion of obese patients received an ARDI less than 85% for carboplatin and paclitaxel, in agreement with our study. Body mass index was also a major predictor of RDI less than 85% in another study.15
In agreement with others,24 we found that obese women were less likely to experience toxic effects, in particular neutropenia. Decreased occurrence of toxic effects may be a consequence of differences in metabolism that are not well understood, such as binding of chemotherapy agents in adipose tissue, or in clearance of these drugs in obese women.26 Regardless of BMI, women who received dose reduction had a higher prevalence of chemotoxic effects. This is likely a reflection of chemotoxic effects resulting in decisions to reduce dose.
Few studies have evaluated the impact of RDI on ovarian cancer survival. Unadjusted analyses in a small study based on participants in clinical trials revealed no impact on survival among those receiving RDI less than 75%.27 In contrast, another small study (n = 138) using retrospective medical record review in an Alabama hospital reported decreased survival for patients with ovarian cancer receiving RDI less than 70%, but no multivariable analyses were conducted. Only 2 studies have presented results adjusted for prognostic variables, with conflicting conclusions, both based on retrospective medical record review and including fewer than 350 patients, and both evaluating an RDI less than 85%.15,16 Hanna et al,15 in a study in North Carolina, found decreased overall survival, similar to our findings. The other study, conducted in Australia,16 found no association with overall survival. Although both studies adjusted for several prognostic factors, neither took into account chemotoxic effects or comorbidities.
To our knowledge, our study is the first to evaluate the impact of reduced RDI on survival across BMI categories. Our findings are in agreement with those of Barrett et al,23 who found that obese women with ovarian cancer participating in the SCOTROC I trial who received dosing based on full weight did not experience worse survival than normal-weight women. Oncologists are often reticent to use full doses in obese women because of concern for toxicity and thus cap the dose. Our study showed that neither survival nor toxicity is worse in obese women given full drug doses of chemotherapy.
In our study, the impact of dose reduction on survival was more apparent for paclitaxel than for carboplatin, for which we did not find a clear dose-response relationship with mortality. Whereas this may be attributed to differences in metabolism and mechanism of action between the 2 drugs, it may also be driven by expected dose calculations. Carboplatin dosing requires selection of an AUC value, most commonly 5 or 6, which may reflect clinical judgment incorporating decisions to decrease dose in those who are obese or have more advanced disease or preexisting conditions. This would result in a lower calculated expected dose and higher RDI among those who tend to have worse survival. To evaluate this possibility, we compared these characteristics by AUC used. Whereas obesity did not appear to be a deciding factor, those with AUC of 5 were more likely to have received a diagnosis at stage IV or to have diabetes, renal disease, or hypertension, with differences in presence of diabetes statistically significant (P = .02). Nevertheless, survival analysis results for carboplatin remained essentially unchanged when analyses were limited to women with AUC of 6. Furthermore, the optimal weight for use in the Cockcroft-Gault formula to calculate expected dose for carboplatin remains uncertain. Studies have shown that in obese women, using ideal body weight underestimates creatinine clearance whereas actual body weight overestimates it.28 Because we used actual weight rather than ideal bodyweight, expected carboplatin dose may have been overestimated in obese women, which would result in underestimation of RDI (ie, the ratio of observed to expected doses).
Our data show that overweight and obesity are nearly as common among patients with ovarian cancer as in the general population.29 Previous studies were based mostly on data from clinical trials, which typically exclude women with obesity-related comorbidities, or observational studies in volunteers that include only patients who were healthy enough and interested in participating. In contrast, we included all cases diagnosed in an integrated health care system that provides a substantial proportion of care in a large geographical area and who received standard treatment with carboplatin-paclitaxel. Cancers were identified from the KPNC Cancer Registry, which follows data quality and reporting standards of the North American Association of Centralized Cancer Registries and the NCI Surveillance, Epidemiology, and End Results (SEER) Program, and contains information on all patients who receive a diagnosis or treatment at a KPNC facility. The distributions of stage at diagnosis30,31(p33) and 5-year relative survival rate by stage31(p35) are similar to SEER estimates, suggesting generalizability of our findings.
Conclusions
Dose reduction for paclitaxel and carboplatin was associated with poorer survival in women with ovarian cancer, and body size was the strongest predictor of dose reduction. This association with poorer survival was apparent regardless of body size, with the effect strongest among patients with normal weight. Our results suggest that BMI should not be a major factor influencing dose reduction decisions in women with ovarian cancer.
Supplementary Material
At a Glance.
The impact of body mass index (BMI) on chemotherapy dosing and dose reduction and ovarian cancer survival was assessed in 806 women in a cohort study.
Body mass index was the strongest predictor of dose reduction, with obese women receiving less paclitaxel and carboplatin per kilogram of body weight, and lower relative dose intensity (RDI) for each agent and the combined regimen.
Lower average RDI for the regimen was associated with worse overall and ovarian cancer–specific survival (HR, 1.62 [95% CI, 1.10–2.37] and HR, 1.69 [95% CI, 1.12–2.55], respectively).
For each BMI category, those with average RDI less than 85% had worse survival than those without dose reduction.
Acknowledgments
Funding/Support: This study was supported by grants from the National Cancer Institute (K22 CA138563, UC2 CA148185, U24 CA171524) and the Kaiser Permanente Center for Effectiveness and Safety Research.
Role of the Funder/Sponsor: The National Cancer Institute and the Kaiser Permanente Center for Effectiveness and Safety Research had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Footnotes
Author Contributions: Ms Lee and Dr Kushi had full access to all data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Bandera, Kushi.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Bandera, Kushi.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Bandera, Lee, Kushi.
Obtained funding: Bandera, Kushi.
Administrative, technical, or material support: Bandera, Lee, Kushi.
Study supervision: Bandera, Kushi.
Conflict of Interest Disclosures: None reported.
REFERENCES
- 1.American Cancer Society. Cancer Facts and Figures 2015. Atlanta, GA: American Cancer Society; 2015. [Google Scholar]
- 2.Nagel CI, Backes FJ, Hade EM, et al. Effect of chemotherapy delays and dose reductions on progression free and overall survival in the treatment of epithelial ovarian cancer. Gynecol Oncol. 2012;124(2):221–224. doi: 10.1016/j.ygyno.2011.10.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Jelovac D, Armstrong DK. Recent progress in the diagnosis and treatment of ovarian cancer. CA Cancer J Clin. 2011;61(3):183–203. doi: 10.3322/caac.20113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Field KM, Kosmider S, Jefford M, et al. Chemotherapy dosing strategies in the obese, elderly, and thin patient: results of a nationwide survey. J Oncol Pract. 2008;4(3):108–113. doi: 10.1200/JOP.0832001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Griggs JJ, Mangu PB, Anderson H, et al. American Society of Clinical Oncology. Appropriate chemotherapy dosing for obese adult patients with cancer: American Society of Clinical Oncology clinical practice guideline. J Clin Oncol. 2012;30(13):1553–1561. doi: 10.1200/JCO.2011.39.9436. [DOI] [PubMed] [Google Scholar]
- 6.Wolin KY, Carson K, Colditz GA. Obesity and cancer. Oncologist. 2010;15(6):556–565. doi: 10.1634/theoncologist.2009-0285. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Modesitt SC, van Nagell JR., Jr The impact of obesity on the incidence and treatment of gynecologic cancers: a review. Obstet Gynecol Surv. 2005;60(10):683–692. doi: 10.1097/01.ogx.0000180866.62409.01. [DOI] [PubMed] [Google Scholar]
- 8.Ekhart C, Rodenhuis S, Schellens JH, Beijnen JH, Huitema AD. Carboplatin dosing in overweight and obese patients with normal renal function, does weight matter? Cancer Chemother Pharmacol. 2009;64(1):115–122. doi: 10.1007/s00280-008-0856-x. [DOI] [PubMed] [Google Scholar]
- 9.Nightingale G, Trovato JA, Lee M, Thompson J. Carboplatin dosing in overweight and obese patients: a single-center experience. J Hematol Oncol Pharm. 2011;1:18–24. [Google Scholar]
- 10.Ross TR, Ng D, Brown JS, et al. The HMO Research Network Virtual Data Warehouse: a public data model to support collaboration. EGEMS (Wash DC) 2014;2(1):1049. doi: 10.13063/2327-9214.1049. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hornbrook MC, Hart G, Ellis JL, et al. Building a virtual cancer research organization. J Natl Cancer Inst Monogr. 2005;(35):12–25. doi: 10.1093/jncimonographs/lgi033. [DOI] [PubMed] [Google Scholar]
- 12.Hryniuk WM, Goodyear M. The calculation of received dose intensity. J Clin Oncol. 1990;8(12):1935–1937. doi: 10.1200/JCO.1990.8.12.1935. [DOI] [PubMed] [Google Scholar]
- 13.Griggs JJ, Sorbero ME, Stark AT, Heininger SE, Dick AW. Racial disparity in the dose and dose intensity of breast cancer adjuvant chemotherapy. Breast Cancer Res Treat. 2003;81(1):21–31. doi: 10.1023/A:1025481505537. [DOI] [PubMed] [Google Scholar]
- 14.Griggs JJ, Culakova E, Sorbero ME, et al. Effect of patient socioeconomic status and body mass index on the quality of breast cancer adjuvant chemotherapy. J Clin Oncol. 2007;25(3):277–284. doi: 10.1200/JCO.2006.08.3063. [DOI] [PubMed] [Google Scholar]
- 15.Hanna RK, Poniewierski MS, Laskey RA, et al. Predictors of reduced relative dose intensity and its relationship to mortality in women receiving multi-agent chemotherapy for epithelial ovarian cancer. Gynecol Oncol. 2013;129(1):74–80. doi: 10.1016/j.ygyno.2012.12.017. [DOI] [PubMed] [Google Scholar]
- 16.Au-Yeung G, Webb PM, DeFazio A, Fereday S, Bressel M, Mileshkin L. Impact of obesity on chemotherapy dosing for women with advanced stage serous ovarian cancer in the Australian Ovarian Cancer Study (AOCS) Gynecol Oncol. 2014;133(1):16–22. doi: 10.1016/j.ygyno.2014.01.030. [DOI] [PubMed] [Google Scholar]
- 17.Fauci JM, Whitworth JM, Schneider KE, et al. Prognostic significance of the relative dose intensity of chemotherapy in primary treatment of epithelial ovarian cancer. Gynecol Oncol. 2011;122(3):532–535. doi: 10.1016/j.ygyno.2011.05.023. [DOI] [PubMed] [Google Scholar]
- 18.Hershman DL, Unger JM, Barlow WE, et al. Treatment quality and outcomes of African American versus white breast cancer patients: retrospective analysis of Southwest Oncology studies S8814/S8897. J Clin Oncol. 2009;27(13):2157–2162. doi: 10.1200/JCO.2008.19.1163. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Stålberg K, Svensson T, Lönn S, Kieler H. The influence of comorbidity on mortality in ovarian cancer patients. Gynecol Oncol. 2014;133(2):298–303. doi: 10.1016/j.ygyno.2014.02.024. [DOI] [PubMed] [Google Scholar]
- 20.Chia VM, O’Malley CD, Danese MD, et al. Prevalence and incidence of comorbidities in elderly women with ovarian cancer. Gynecol Oncol. 2013;129(2):346–352. doi: 10.1016/j.ygyno.2013.02.014. [DOI] [PubMed] [Google Scholar]
- 21.Rodriguez N, Rauh-Hain JA, Shoni M, et al. Changes in serum CA-125 can predict optimal cytoreduction to no gross residual disease in patients with advanced stage ovarian cancer treated with neoadjuvant chemotherapy. Gynecol Oncol. 2012;125(2):362–366. doi: 10.1016/j.ygyno.2012.02.006. [DOI] [PubMed] [Google Scholar]
- 22.Pelissier A, Bonneau C, Chéreau E, et al. CA125 kinetic parameters predict optimal cytoreduction in patients with advanced epithelial ovarian cancer treated with neoadjuvant chemotherapy. Gynecol Oncol. 2014;135(3):542–546. doi: 10.1016/j.ygyno.2014.09.005. [DOI] [PubMed] [Google Scholar]
- 23.Barrett SV, Paul J, Hay A, Vasey PA, Kaye SB, Glasspool RM Scottish Gynaecological Cancer Trials Group. Does body mass index affect progression-free or overall survival in patients with ovarian cancer? results from SCOTROC I trial. Ann Oncol. 2008;19(5):898–902. doi: 10.1093/annonc/mdm606. [DOI] [PubMed] [Google Scholar]
- 24.Wright JD, Tian C, Mutch DG, et al. Carboplatin dosing in obese women with ovarian cancer: a Gynecologic Oncology Group study. Gynecol Oncol. 2008;109(3):353–358. doi: 10.1016/j.ygyno.2008.02.023. [DOI] [PubMed] [Google Scholar]
- 25.Backes FJ, Nagel CI, Bussewitz E, Donner J, Hade E, Salani R. The impact of body weight on ovarian cancer outcomes. Int J Gynecol Cancer. 2011;21(9):1601–1605. doi: 10.1097/IGC.0b013e31822d2aa3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Lyman GH, Sparreboom A. Chemotherapy dosing in overweight and obese patients with cancer. Nat Rev Clin Oncol. 2013;10(8):451–459. doi: 10.1038/nrclinonc.2013.108. [DOI] [PubMed] [Google Scholar]
- 27.Repetto L, Pace M, Mammoliti S, et al. The impact of received dose intensity on the outcome of advanced ovarian cancer. Eur J Cancer. 1993;29A(2):181–184. doi: 10.1016/0959-8049(93)90169-g. [DOI] [PubMed] [Google Scholar]
- 28.Brown DL, Masselink AJ, Lalla CD. Functional range of creatinine clearance for renal drug dosing: a practical solution to the controversy of which weight to use in the Cockcroft-Gault equation. Ann Pharmacother. 2013;47(7–8):1039–1044. doi: 10.1345/aph.1S176. [DOI] [PubMed] [Google Scholar]
- 29.Flegal KM, Carroll MD, Kit BK, Ogden CL. Prevalence of obesity and trends in the distribution of body mass index among US adults, 1999–2010. JAMA. 2012;307(5):491–497. doi: 10.1001/jama.2012.39. [DOI] [PubMed] [Google Scholar]
- 30.Goodman MT, Correa CN, Tung KH, et al. Stage at diagnosis of ovarian cancer in the United States, 1992–1997. Cancer. 2003;97(10) suppl:2648–2659. doi: 10.1002/cncr.11347. [DOI] [PubMed] [Google Scholar]
- 31.Annual Report on Trends, Incidence, and Outcomes: Summarizing Data Reported to the California Cancer Registry and the SEER Program of the National Cancer Institute. Oakland, CA: Kaiser Permanente of Northern California; 2011. Kaiser Permanente Northern California Cancer Registry at the Division of Research. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.