Abstract
Background
The effect of body mass index (BMI) on treatment outcomes for patients with locally advanced cervical carcinoma undergoing definitive chemoradiation is unclear.
Methods
This study cohort included all cervical carcinoma patients (n = 404) with stage IB1 and positive lymph nodes or stage ≥ IB2 treated at our facility from January 1998 to January 2008. Mean follow-up time was 47.2 months. BMI was calculated using the National Institute of Health online calculator. BMI categories were created according to the World Health Organization classification system. Primary outcomes were overall survival, disease free survival, and complication rate. Univariate and multivariate analyses were performed. Kaplan-Meier survival curves were generated and compared using Cox proportional hazard models.
Results
On multivariate analysis, when compared to normal weight subjects (18.5-24.9 kg/m2), a BMI < 18.5 kg/m2 was associated with decreased overall survival (HR 2.37, 95% CI 1.28 - 4.38, p<0.01). The 5-year overall survivals were 33%, 60%, and 68% for a BMI < 18.5 kg/m2, BMI 18.5-24.9 kg/m2, and a BMI > 24.9 kg/m2 respectively. A BMI < 18.5 kg/m2 was associated with increased risk for grade 3 or 4 complications when compared to patients with a BMI > 24.9 kg/m2 (radiation enteritis: 16.7% vs. 13.6 % p= 0.03, fistula: 11.1% vs. 8.8% p= 0.05, bowel obstruction 33.3% vs. 4.4% p< 0.001, lymphedema: 5.6% vs. 1.2% p=0.02).
Conclusions
Underweight patients (BMI < 18.5 kg/m2) with locally advanced cervical cancer have diminished overall survival and more complications than normal weight and obese patients.
Introduction
Obesity is a growing health and economic problem worldwide, and is especially significant within the United States. Over the past twenty years, there has been a dramatic increase in obesity rates in adults and children. In 2006, only four states had obesity rates less than 20%, and twenty-two states had rates greater than 25%. Mississippi, West Virginia, and Alabama reported incidences of obesity greater than 30%.(1) There is growing concern regarding increased health problems associated with obesity and their ensuing costs to society. Obesity has long been recognized as a strong risk factor for many co-morbid conditions such as the metabolic syndrome, diabetes, the obesity-hypoventilation syndrome, and cardiovascular diseases. There is now increasing awareness of the impact obesity may also have on cancer, as new evidence suggests that overeating and obesity may be the largest preventable cause of cancer among non-smokers and may even account for one in five cancer deaths among women.(2, 3)
Breast, colon, ovarian, endometrial, and pancreatic cancers leading causes of cancer mortality in women, are associated with an increased incidence in the obese population.(3, 4) An elevated body mass index (BMI) is also associated with higher death rates as a result of cancer in the esophagus, colon and rectum, liver, pancreas, and non-Hodgkin's lymphoma. A previous prospective study of over 400,000 women found that women in the heaviest cohort (BMI ≥ 40 kg/m2) had a cancer death rate that was 62% higher than that of women with normal weight.(3) Increased death rates for obese women with cancer could reflect an increased incidence of cancer in these women compared with lean women, a worse prognosis for obese women with cancer, or some combination of both.(2)
Among gynecologic malignancies, the impact of BMI on prognosis varies. Significant trends for increasing risk of death in women with an elevated BMI have been reported for cancers of the breast, uterus, cervix, and ovary.(3) Other findings suggest women with a higher BMI have a more favorable outcome especially in ovarian and endometrial cancer patients. The data linking obesity or any differences in BMI to cervical cancer survival outcomes are even less consistent.(5-8) We aimed to further elucidate the role of body mass index on morbidity and survival in patients with locally advanced cervical cancer treated with definitive chemoradiation therapy.
Patients and Methods
A total of 404 subjects formed the study population and included all women with FIGO stage IB1 and positive lymph nodes by positron emission tomography (PET) and FIGO stages IB2-IVA between January 1998 and December 2007. Within the study period, 423 patients were initially identified with 16 subjects excluded due to incomplete medical records and 3 excluded due to missing BMI information. Patient and tumor characteristics are presented in Table 1. All patients underwent standard chemo-radiation therapy for cervical cancer at Barnes-Jewish Hospital/Washington University School of Medicine in Saint Louis, MO. Each subject's clinical stage was verified during manual review of medical records. All pathology reports were previously reviewed by expert pathologists in the field of gynecology at our institution. Demographic, laboratory, surgical, radiotherapy treatment modality, chemotherapy and clinical outcome data were extracted by a single physician from inpatient and outpatient medical records. Additionally, the serum biomarkers albumin, hemoglobin, and lymphocyte count were documented prior to treatment in all subjects. This study was approved by the institutional review board of the Human Research Protection Office at Washington University.
Table 1. Patient and Tumor Characteristics*.
BMI < 18.5 kg/m2 N = 21 |
BMI 18.5- 24.9 kg/m2 N = 120 |
BMI > 24.9 kg/m2 N = 263 |
p-value | |
---|---|---|---|---|
Age (years) | 53.5 | 50.5 | 52.3 | 0.41 |
Race | ||||
White | 16 (76.2%) | 81 (67.5%) | 152 (57.8%) | 0.07 |
Black | 5 (23.8%) | 37 (30.8%) | 102 (38.8%) | |
Other/Missing | 0 (%) | 2 (1.7%) | 9 (3.4%) | |
Smoker | ||||
Yes | 16 (76.2%) | 65 (54.2%) | 104(39.6%) | <0.01 |
No | 2 (9.5%) | 39 (32.5%) | 126 (47.9%) | |
Previous | 2 (9.5%) | 9(7.5%) | 18 (6.8%) | |
No data | 1 (4.8%) | 7(5.8%) | 15 (5.7%) | |
Co-morbidities | ||||
None | 11 (52.4%) | 65 (54.2%) | 95 (36.1%) | <0.01 |
≥ 1 | 10 (47.6%) | 54 (45%) | 162 (61.6%) | |
No data | 0 (0%) | 1 (0.8%) | 6 (2.3%) | |
Histology | ||||
Squamous | 17 (81%) | 101 (84.2%) | 226 (85.9%) | 0.68 |
Other | 4 (19%) | 17 (14.2%) | 33(12.5%) | |
No Data | 0 (0%) | 2 (1.7%) | 4 (1.6%) | |
Stage | ||||
I-II | 12 (57.1%) | 73(60.8%) | 195 (74.1%) | 0.01 |
III-IV | 9 (42.9%) | 46 (38.4%) | 65 (24.7%) | |
No data | 0 (0%) | 1 (0.8%) | 3 (1.2%) | |
Grade | ||||
1 | 0 (0%) | 4 (3.4%) | 10 (3.8%) | 0.75 |
2 | 12 (57.1%) | 66 (55%) | 129 (49%) | |
3 | 7 (33.4%) | 43 (35.8%) | 100 (38%) | |
No data | 2 (9.5%) | 7 (5.8%) | 24 (9.2%) | |
LN status | ||||
Positive | 8 (38.1%) | 33 (27.5%) | 75 (28.5%) | 0.61 |
Negative | 13 (61.9%) | 87 (72.5%) | 188 (71.5%) | |
IMRT | ||||
Yes | 3 (14.3%) | 34 (28.3%) | 83 (31.6%) | 0.24 |
No | 18 (85.7%) | 86 (71.7%) | 176 (66.9%) | |
Unknown | 0 (0%) | 0 (0%) | 4 (1.5%) |
Percentages represent proportion of column total. IMRT: intensity modulated radiation therapy.
The BMI for each patient was determined by using the National Institute of Health's (NIH) online calculator based on the formula of weight in kilograms divided by the square of the height in meters (http://www.nhlbisupport.com/bmi/). The World Health Organization (WHO) defines a normal body weight as a BMI of 18.5 to 24.9 kg/m2, underweight as a BMI ≤ 18.5kg/m2, overweight as a BMI of 25 to 29.9 kg/m2, and obesity as a BMI of ≥ 30 kg/m2. We divided our study subjects into the three categories of underweight (BMI ≤ 18.5 kg/m2), normal weight (BMI 18.5-25 kg/m2), and overweight (BMI >25kg/m2) for analysis.
Primary outcomes were overall survival (OS) and disease-free survival (DFS). Overall survival was defined as the time from diagnosis to death or last contact. Patients alive were censored at the date of last clinical contact. Disease-free survival was defined as the time from diagnosis to recurrence or death, whichever occurred first, while those alive and free of disease were censored at last contact. Site of recurrence was categorized as pelvic or distant. Secondary outcomes consisted of Grade 3-4 complications as defined by the Common Terminology Criteria for Adverse Events (CTCAE v4.0). Complications included radiation enteritis/proctitis/cystitis, bowel obstruction, vesico-vaginal fistula, recto-vaginal fistula, and lymphedema. Co-morbidities were categorized as cardiovascular disease, pulmonary disease, diabetes, rheumatic disease, liver disease, active nephropathy, or other. All biomarkers were analyzed as categorical variables (albumin: < or ≥ 3.5 g/dl, lymphocyte percentage: < or ≥ 20%, lymphocyte count: <1.8 or ≥ 1.8 K/cumm, creatinine: < or ≥ 1.0 mg/dl, hemoglobin: <8, 8-9.9, 10-12, or >12 g/dl).
Statistical Methods
The distributions of patient and clinical characteristics across different BMI groups were compared using the chi-square test or ANOVA as appropriate. Survival curves were estimated by the Kaplan-Meier product-limit method. Univariate Cox proportional hazard models were fitted to identify factors significantly related to OS or DFS. To assess whether BMI is an independent predictor of survival, multivariate Cox models were constructed via a backward selection procedure (using p-value of 0.10 for variable removal), while considering all factors showing p< 0.3 in the univariate analyses.
In addition, 3 sets of multivariate Cox models were explicitly construed and compared for DFS and OS. The first model (Model 0) only included conventional risk factors (co-morbidity, stage, lymph node status, and grade). The next multivariate models were constructed by entering BMI (Model 1) and then BMI plus the biomarkers albumin, lymphocyte percentage, and creatinine (Model 2). Overall predictive accuracy of the models was compared using concordance index (C-statistic) and calibration statistics.(9) The C-statistic reflects the ability of a model to discriminate participants who develop the events of interest (i.e., death or recurrence) from those who do not. Its value ranges from 0.5 to 1 with a value of 1 indicating a perfect predictive power. The calibration statistic is an analog of Hosmer-Lemeshow goodness of fit test in a censored data setting. It was calculated by dividing the patients into 10 levels of predicted risk and a value of 20 or less indicates good agreement between the predicted and the observed event rate.(10) All statistical tests were 2-sided and p-values < 0.05 were considered significant. All the analyses were performed with SAS version 9 (SAS Institute, Cary, NC).
Results
Of the 404 study subjects, 117 (29.3%) were stage I, 163 (40.8%) were stage II, and 120 (29.7) % had either stage III/IV disease. The mean age at diagnosis was 52 years (range 22-90 years). The mean follow-up time for survivors was 47.2 months (range 0.4- 131.4 months). There were 145 deaths (36%) noted during the follow-up period. In regards to pelvic radiation therapy, 128 (32%) received intensity-modulated radiation therapy (IMRT) and 274 (68%) received conventional radiotherapy. Almost the entire group also received cervical brachytherapy treatment (n=399, 99%) and concurrent chemotherapy (n=288, 92%). Patient and tumor characteristics are summarized in Table 1 by BMI category. Of note, there was a higher percentage of smokers and diagnoses of stage III/IV disease among subjects with a BMI < 18.5 kg/m2 compared to the other two groups. More subjects with a BMI > 25 kg/m2 had one or more co-morbidities. Among those who recurred, the percentage of patients who experienced a distant recurrence was similar between all BMI groups (p=0.4).
Of the 372 patients, 21 (5.2%) were classified as underweight, 120 (29.7%) were classified as normal weight, and 263 (65.1%) were considered overweight prior to treatment. The majority of patients lost weight during treatment with the underweight group increasing to 39 (4.5% increase) and the overweight group decreasing to 186 (19% reduction) by the end of treatment. Given the frequency of missing data for BMI at the finish of treatment, and for overall change (n= 63), this data was not included in further analyses, and only each patient's BMI at the start of treatment was used in the multivariate analysis.
The results for univariate analysis for overall survival and progression-free survival are shown in Table 2. Overall survival was 33%, 60%, and 68% for patients with a BMI of <18.5 kg/m2, 18.5-24.9 kg/m2, and >25 kg/m2, respectively. Body mass index was significant for both OS and DFS (Figures 1, 2). Additionally stage, tumor size, co-morbidity, lymph node status, and serum lymphocyte percentage and total lymphocyte count, and serum albumin level ≥ 3.5 g/dl were also significant for both OS and DFS. Grade was significant only for DFS.
Table 2. Univariate Analysis of Overall and Disease-Free Survival.
Overall Survival | Disease-Free Survival | |||||
---|---|---|---|---|---|---|
HR | 95% CI | p-value | HR | 95% CI | p-value | |
BMI* | < 0.001 | <0.01 | ||||
< 18.5 kg/m2 | 2.85 | 1.61, 5.05 | 2.05 | 1.17, 3.57 | ||
≥ 25 kg/m2 | 0.83 | 0.58, 1.19 | 0.84 | 0.61, 1.16 | ||
Age at diagnosis | 1.01 | 0.99, 1.02 | 0.19 | 1.00 | 0.99, 1.02 | 0.45 |
Race† | 0.84 | 0.8 | ||||
Black | 1.1 | 0.78, 1.54 | 1.11 | 0.82, 1.5 | ||
Other | 1.3 | 0.30, 4.99 | 1.12 | 0.28, 4.54 | ||
Comorbidity | < 0.01 | <0.01 | ||||
1 vs 0 | 1.88 | 1.28, 2.73 | 1.6 | 1.14, 2.25 | ||
2 vs 0 | 0.95 | 0.58, 1.59 | 0.87 | 0.56, 1.36 | ||
3 vs 0 | 2.08 | 1.17, 3.70 | 1.59 | 0.94, 2.72 | ||
Smoker# | 0.45 | 0.44 | ||||
Current | 1.25 | 0.88, 1.78 | 1.22 | 0.89, 1.68 | ||
Previous | 1.18 | 0.58, 2.39 | 1.22 | 0.66, 2.25 | ||
Albumin ≥ 3.5 g/dl | 0.56 | 0.38, 0.83 | < 0.01 | 0.64 | 0.44, 0.94 | 0.02 |
Creatinine > 1.0 mg/dl | 1.3 | 0.87, 1.95 | 0.19 | 1.31 | 0.91, 1.87 | 0.14 |
Hemoglobinº | 0.05 | 0.19 | ||||
8.0-9.9 mg/dl | 1.15 | 0.62, 2.12 | 0.97 | 0.56, 1.70 | ||
10-12 mg/dl | 1.11 | 0.66, 1.87 | 0.88 | 0.55, 1.40 | ||
> 12 mg/dl | 0.67 | 0.39, 1.13 | 0.67 | 0.42, 1.06 | ||
Total Lymphocyte | 0.7 | 0.50, 0.99 | 0.05 | 0.74 | 0.55, 1.01 | 0.05 |
Lymphocyte percentage ≥ 20% | 0.56 | 0.39, 0.78 | <0.01 | 0.63 | 0.47, 0.85 | <0.01 |
XRTα | 0.89 | 0.61, 1.32 | 0.58 | 0.85 | 0.61, 1.19 | 0.35 |
Reference category: Normal weight
Reference category: White
Reference category: Never
Reference category: < 8.0 mg/dl
XRT: conventional radiation; Reference category: IMRT (intensity modulated radiation therapy)
Figure 1.
Overall survival by BMI classification.
Figure 2.
Progression-free survival by BMI classification.
Cox proportional hazards modeling for disease recurrence showed that the addition of BMI into the model significantly improved its predictive ability (Table 3). Similarly, the addition of BMI into the model for overall survival also showed a statistically significant improvement in model predictive performance with a c-statistic of 0.719 (Table 3). A BMI of < 18.5 kg/m2 remained a significant predictor with in the model for overall survival (HR 2.38, 95% CI 1.58, 4.38, p-value <0.01). The biomarkers albumin, creatinine, lymphocyte percentage and total lymphocyte count did not contribute to model improvement for DFS or OS and were excluded from the final model. Comparison between the different models is shown in Table 4.
Table 3. Final Multivariate Proportional Hazards Model for Survival Outcome.
Disease-Free Survival | Overall Survival | |||||
---|---|---|---|---|---|---|
HR | 95% CI | p-value | HR | 95% CI | p-value | |
BMI < 18.5 kg/m2* | 1.64 | 0.90, 2.97 | 0.11 | 2.37 | 1.28, 4.38 | <0.01 |
BMI > 24.9 kg/m2 | 0.79 | 0.55, 1.12 | 0.19 | 0.76 | 0.51, 1.14 | 0.18 |
# of Comorbidities† | ||||||
1 | 1.70 | 1.19, 2.44 | <0.01 | 1.96 | 1.30, 2.93 | <0.01 |
2 | 1.25 | 0.79, 1.99 | 0.35 | 1.50 | 0.88, 2.57 | 0.14 |
3 | 2.16 | 1.22, 3.81 | <0.01 | 3.24 | 1.73, 6.07 | <0.001 |
Stageº | ||||||
II | 1.82 | 1.20, 2.77 | <0.01 | 2.19 | 1.35, 3.55 | <0.01 |
III/IV | 2.23 | 1.45, 3.42 | <0.001 | 2.46 | 1.49, 4.04 | <0.001 |
Positive LN | 2.30 | 1.67, 3.17 | <0.0001 | 2.46 | 1.72, 3.53 | <0.001 |
HR: Hazard ratio, 95% CI: 95% confidence interval
Reference category: BMI 18.5-24.9 kg/m2
Reference category: No comorbities
Reference category: Stage I
Table 4. Comparison of Cox Proportional Hazards Models for Survival Outcomes.
Disease-Free Survival | Overall Survival | ||||
---|---|---|---|---|---|
Model | Variables Included | C-statistic | Calibration Statistic | C-statistic | Calibration Statistic |
0 | Comorbidity | 0.6745 | 13.669 | 0.695 | 10.354 |
Stage | |||||
LN status | |||||
Grade | |||||
1 | Model 0 | 0.6883 | 11.497 | 0.719 | 10.965 |
+ BMI | |||||
2 | Model 1 | 0.6951 | 12.545 | 0.723 | 4.322 |
+ Albumin | |||||
+ Creatinine | |||||
+ Lymphocyte % |
During the study period 128 grade 3-4 complications were identified. There were 28 subjects with no information regarding complications [BMI < 18.5 kg/m2 (3), BMI 18.5-24.9 kg/m2 (11), BMI > 25 kg/m2 (14)]. Radiation enteritis/proctitis/cystitis was the most frequent complication noted (n=61). A lower body mass index was found to correlate with an increased frequency of total complications as well as with an increased frequency of bowel obstruction and lymphedema (Table 5). Overweight patients experienced fewer complications. Univariate analyses revealed BMI, smoking status and lymph node status to be correlated with increased risk for complication (Table 6). In multivariate analysis, smoking status was found to be an independent predictor of complications (Table 7).
Table 5. Grade 3 or 4 Complication Type and Rate*.
BMI < 18.5 kg/m2 N= 18 | BMI 18.5-24.9 kg/m2 N=109 | BMI > 24.9 kg/m2 N=249 | p-value | |
---|---|---|---|---|
All complications | 9 (50%) | 45 (41.3%) | 74 (29.7%) | 0.03 |
Radiation enteritis/proctitis | 3 (16.7%) | 24 (22%) | 34 (13.6%) | 0.03 |
Recto-vaginal or vesico-vaginal fistula | 2 (11.1%) | 12 (11%) | 22 (8.8%) | 0.05 |
Bowel obstruction | 6 (33.3%) | 8 (7.3%) | 11 (4.4%) | <0.001 |
Lymphedema | 1 (5.6%) | 3 (2.6%) | 3 (1.2%) | 0.02 |
Percentages represent proportion of column total (subjects within each BMI category with data available)
Table 6. Univariate Analysis for Grade 3-4 Complications.
HR | 95% CI | p-value | |
---|---|---|---|
Positive LN status | 0.58 | 0.35, 0.96 | 0.04 |
BMI† | 0.04 | ||
Normal (18.5 -24.9 kg/m2) | 0.70 | 0.26, 1.9 | |
Overweight (≥ 25 kg/m2) | 0.42 | 0.16, 1.1 | |
Smoking status° | 0.04 | ||
Current | 1.79 | 1.12, 2.88 | |
Previous | 1.97 | 0.87, 4.49 | |
Race* | 0.93 | ||
Black | 1.00 | 0.64, 1.58 | |
Other | 0.64 | 0.07, 6.27 | |
Co-morbidity# | 0.38 | ||
1 | 1.46 | 0.89, 2.45 | |
2 | 1.29 | 0.72, 2.34 | |
3 | 1.68 | 0.75, 3.79 | |
Histology% | 0.93 | ||
Squamous | 1.04 | 0.46, 2.41 | |
Other | 1.23 | 0.37, 4.04 | |
Grade 3 | 1.12 | 0.71, 1.75 | 0.63 |
Stage∧ | 0.15 | ||
II | 1.06 | 0.62, 1.81 | |
III/IV | 1.63 | 0.93, 2.84 |
Reference category: Underweight (<18.5 kg/m2)
Reference category: Never smoked
Reference category: Caucasian
Reference category: None
Reference category: Adenosquamous
Reference category: Stage I
Table 7. Multivariate Analysis for Grade 3-4 Complications.
HR | 95% CI | p-value | |
---|---|---|---|
BMI† | |||
Normal weight (18.5 kg/m2-24.9 kg/m2) | 0.80 | 0.28, 2.33 | 0.69 |
Overweight (≥ 25 kg/m2) | 0.52 | 0.18, 1.48 | 0.22 |
Smoking status* | |||
Current | 1.78 | 1.07, 2.94 | 0.03 |
Previous | 2.08 | 0.89, 4.88 | 0.09 |
Reference level: Underweight (<18.5 kg/m2)
Reference level: Never smoked
Discussion
Among women with locally advanced cervical cancer undergoing definitive chemoradiation therapy, our analysis demonstrated BMI to be independently associated with patient outcomes. A higher BMI conferred a significant advantage in terms of complication rates. In contrast, a lower BMI was an adverse prognostic factor associated with decreased overall survival and a significant increase in complication rate. These results differ from the general body of knowledge regarding obesity and cancer outcomes.
Previous literature specific to cervical cancer outcomes and BMI is inconsistent. Two large population studies have demonstrated an increased mortality in women with cervical cancer as body weight increases.(3, 11) This increased mortality in obese women could be due to lack of screening with pap smears as one large meta-analysis showed that women with class III (BMI > 40kg/m2) obesity are ∼40% less likely to undergo cervical cancer screening,(12) or it may simply be due to a detrimental impact of obesity on treatment outcomes. The potential biologic mechanisms responsible have yet to be fully elucidated and include theories involving increased levels of endogenous hormones and obesity-related chronic low-level inflammation.(13)
Two retrospective studies of approximately 372 and 175 subjects have not shown obesity to be a factor that negatively influences survival in women with cervical cancer.(14, 15) A more recent larger retrospective study of 738 cervical cancer patients demonstrated that a higher BMI was associated with a more favorable outcome.(16) The authors did not report any adverse influence of BMI on outcomes. These studies differ from our analysis in that we also found a lower BMI to be associated with adverse outcomes when compared to normal weight and obese cervical cancer patients.
The impact of BMI on survival and morbidity in cervical cancer patients undergoing radical hysterectomy has also been evaluated.(17) The retrospective study included 229 patients over a 17-year study period. It demonstrated that a low BMI (≤ 20 kg/m2) was a significant independent negative prognostic factor on survival. It also showed BMI to be a more important predictor of survival than the FIGO stages IB1 and IB2. Although the study did not include patients undergoing radiation therapy, their result supports our finding that a low BMI adversely affects treatment outcomes in cervical cancer patients.
Varying differences in outcomes between cervical cancer patients with higher versus lower BMI undergoing radiation therapy are due to heterogeneous studies that either did not evaluate clinically relevant outcomes or failed to find significant differences among obese patients. Eifel et al. reported an increased rate of bladder injury on multivariate analysis (RR, 1.51) when analyzing the complication rates in 867 women with BMI > 31 kg/m2 and cervical cancer undergoing low-dose rate brachytherapy and external beam radiation therapy over 34 years, but did not look at survival differences.(18) Kapp et al. performed a retrospective analysis of radiation treatments in patients with cervical cancer and failed to find a statistically significant difference in either acute or late radiation-induced complications in obese women compared to lean women. However, there were increased overall complications in the obese patients with cervical cancer (OR 1.96).(19) Another study of definitive radiation therapy in six morbidly obese women with class III BMI who had stage IB cervical cancer did not show any radiation therapy complications and no recurrences were detected.(20) None until our study showed a lower BMI to be a negative prognostic factor.
Current smoking status was also found to be an independent negative prognostic factor for risk of complications, confirming previous findings that smoking is correlated with late toxicity from radiation.(18) Our numbers were too small to evaluate associations between each individual complication and BMI.
Our results demonstrating BMI to be an independent prognostic factor is of considerable importance as the current standard of care for locally advanced disease is pelvic radiotherapy with concurrent cisplatin chemotherapy.(21) Using this treatment strategy, the two-year progression-free survival is approximately 67% with an estimated 5-year overall survival of 60%. Our study showed patients with a BMI of <18.5 kg/m2 to have a dramatically reduced overall 5-year survival of 33%. Such a reduction in the overall survival of these underweight patients suggests that although these patients receive the same standard chemotherapy and radiation regimen as their normal weight and obese counterparts they experience worse outcomes. Not only was overall and progression free survival reduced in underweight patients, but also complications were increased. Despite advances to improve patient tolerance of treatment and overall clinical outcomes, our data suggest that the patient specific factor of a BMI < 18.5 kg/m2 is still prognostic of a high complication rate.
Whether pre-treatment interventions in underweight cervical cancer patients with locally advanced disease will improve outcomes is unclear. The decreased complication rate among the overweight patient cohort suggests that achieving a higher weight prior to treatment may be of benefit in at risk patients. Biomarkers were included in the study in an effort to identify a potential new therapeutic target as well as to provide a clinical marker to help guide and monitor interventions, but none was found to be significant. There may be inherent biologic differences both within the patients themselves and their tumors that is reflected in their weight differences and thus their clinical response to treatment.
Increasing evidence exists for the cancer cachexia syndrome where complex interactions between host inflammatory responses and the tumors themselves may influence survival outcomes22-24. Various systemic inflammatory markers, singly or in combination, have been shown to be associated with patient clinical outcomes in numerous operable solid tumors including colorectal, small cell lung, renal, breast, and ovarian cancers25. C-reactive protein (alone and in combination with albumin), serum albumin, serum neutrophil:lymphocyte ratio, and serum platelet:lymphocyte ratio are the biomarkers with the most evidence for prognostic value within the literature25. Only albumin was included in our study.
We found no particular pattern for site of disease recurrence among underweight patients to explain their increased risk for recurrence. However, our analysis was limited due to the small number of events within this group (N=7).
Other limitations of our study include those associated with retrospective data collection and all results need to be validated prospectively. Our multivariate analysis showed a trend (p = 0.11) towards decreased DFS in underweight patients. We were likely unable to detect a statistically significant difference due to small sample size (n=21). With a larger cohort, the negative impact of a low BMI would likely be more pronounced and achieve significance. This study was performed at a single institution, and given the varying radiotherapy techniques between different institutions, our data may not be generalizable to populations treated elsewhere. BMI data were obtained from patient report at their initial office visit prior to treatment, and may not be accurate. However, other studies have shown patient report of height and weight to be highly reliable26, 27 and that, if anything, women tend to underestimate their weight28. Underestimation of weight would bias the results towards the null hypothesis and would make it less likely for our study to have shown a difference, strengthening our findings.
Overall, body mass index was found to be an independent predictor of clinical outcomes in patients with locally advanced cervical cancer undergoing definitive radiation therapy and chemotherapy. Underweight patients (BMI < 18.5 kg/m2) have diminished overall survival with an increased complication rate compared to normal weight and obese patients. This is of great clinical importance for underweight cervical cancer patients who may be cured by chemoradiation at the high cost of significant morbidity and increased mortality. Modifications to current counseling of these patients regarding nutrition and smoking as well as further research into the development of potential interventions should be considered.
Modifications to reduce radiation doses and subsequent toxicity may also benefit patients who are underweight. Assessment of such strategies is beyond the scope of this paper but should be considered in future research once our findings have been prospectively confirmed. These data, once confirmed, could also provide the springboard to promote research that specifically targets new treatment strategies for underweight cervical cancer patients.
References
- 1.http://www.cdc.gov/nccdphp/dnpa/obesity/trend/maps/index.htm.
- 2.Modesitt SC, van Nagell JR., Jr The impact of obesity on the incidence and treatment of gynecologic cancers: a review. Obstet Gynecol Surv. 2005;60:683–692. doi: 10.1097/01.ogx.0000180866.62409.01. [DOI] [PubMed] [Google Scholar]
- 3.Calle EE, Rodriguez C, Walker-Thurmond K, et al. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med. 2003;348:1625–1638. doi: 10.1056/NEJMoa021423. [DOI] [PubMed] [Google Scholar]
- 4.Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2008. CA Cancer J Clin. 2008;58:71–96. doi: 10.3322/CA.2007.0010. [DOI] [PubMed] [Google Scholar]
- 5.Albanes D. Caloric intake, body weight, and cancer: a review. Nutr Cancer. 1987;9:199–217. doi: 10.1080/01635588709513929. [DOI] [PubMed] [Google Scholar]
- 6.Brown P, Allen AR. Obesity linked to some forms of cancer. W V Med J. 2002;98:271–272. [PubMed] [Google Scholar]
- 7.Wolk A, Gridley G, Svensson M, et al. A prospective study of obesity and cancer risk (Sweden) Cancer Causes Control. 2001;12:13–21. doi: 10.1023/a:1008995217664. [DOI] [PubMed] [Google Scholar]
- 8.Tornberg SA, Carstensen JM. Relationship between Quetelet's index and cancer of breast and female genital tract in 47,000 women followed for 25 years. Br J Cancer. 1994;69:358–361. doi: 10.1038/bjc.1994.65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.D'Agostino RB, Sr, Grundy S, Sullivan LM, et al. Validation of the Framingham coronary heart disease prediction scores: results of a multiple ethnic groups investigation. JAMA. 2001;286:180–187. doi: 10.1001/jama.286.2.180. [DOI] [PubMed] [Google Scholar]
- 10.Harrell FE, Jr, Califf RM, Pryor DB, et al. Evaluating the yield of medical tests. JAMA. 1982;247:2543–2546. [PubMed] [Google Scholar]
- 11.Garfinkel L. Overweight and cancer. Ann Intern Med. 1985;103:1034–1036. doi: 10.7326/0003-4819-103-6-1034. [DOI] [PubMed] [Google Scholar]
- 12.Maruthur NM, Bolen SD, Brancati FL, et al. The association of obesity and cervical cancer screening: a systematic review and meta-analysis. Obesity (Silver Spring) 2009;17:375–381. doi: 10.1038/oby.2008.480. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Siegel EM, Ulrich CM, Poole EM, et al. The effects of obesity and obesity-related conditions on colorectal cancer prognosis. Cancer Control. 17:52–57. doi: 10.1177/107327481001700107. [DOI] [PubMed] [Google Scholar]
- 14.Chen RJ, Chang DY, Yen ML, et al. Prognostic factors of primary adenocarcinoma of the uterine cervix. Gynecol Oncol. 1998;69:157–164. doi: 10.1006/gyno.1998.4971. [DOI] [PubMed] [Google Scholar]
- 15.Hopkins MP, Morley GW. Prognostic factors in advanced stage squamous cell cancer of the cervix. Cancer. 1993;72:2389–2393. doi: 10.1002/1097-0142(19931015)72:8<2389::aid-cncr2820720816>3.0.co;2-r. [DOI] [PubMed] [Google Scholar]
- 16.Munstedt K, Wagner M, Kullmer U, et al. Influence of body mass index on prognosis in gynecological malignancies. Cancer Causes Control. 2008;19:909–916. doi: 10.1007/s10552-008-9152-7. [DOI] [PubMed] [Google Scholar]
- 17.Finan MA, Hoffman MS, Chambers R, et al. Body mass predicts the survival of patients with new International Federation of Gynecology and Obstetrics Stage IB1 and IB2 cervical carcinoma treated with radical hysterectomy. Cancer. 1998;83:98–102. doi: 10.1002/(sici)1097-0142(19980701)83:1<98::aid-cncr13>3.0.co;2-7. [DOI] [PubMed] [Google Scholar]
- 18.Eifel PJ, Jhingran A, Bodurka DC, et al. Correlation of smoking history and other patient characteristics with major complications of pelvic radiation therapy for cervical cancer. J Clin Oncol. 2002;20:3651–3657. doi: 10.1200/JCO.2002.10.128. [DOI] [PubMed] [Google Scholar]
- 19.Kapp KS, Stuecklschweiger GF, Kapp DS, et al. Carcinoma of the cervix: analysis of complications after primary external beam radiation and Ir-192 HDR brachytherapy. Radiother Oncol. 1997;42:143–153. doi: 10.1016/s0167-8140(96)01881-6. [DOI] [PubMed] [Google Scholar]
- 20.Ampil FL, Lojun SL. Radiotherapy for cancer of the cervix in morbidly obese women: preliminary results. Eur J Gynaecol Oncol. 2000;21:262–263. [PubMed] [Google Scholar]
- 21.Rose PG, Bundy BN, Watkins EB, et al. Concurrent cisplatin-based radiotherapy and chemotherapy for locally advanced cervical cancer. N Engl J Med. 1999;340:1144–1153. doi: 10.1056/NEJM199904153401502. [DOI] [PubMed] [Google Scholar]