effect of Body Mass Index on survival in Patients with epithelial ovarian cancer

Ovarian cancer is the sixth most commonly diagnosed gynecologic tumor and constitutes a leading cause of mortality within this group of tumors1. This is because of its typically insidious onset and late diagnosis. Subsequently, only about 40% of women with ovarian cancer can be still alive more than 5 years post-diagnosis2. This poor survival is largely attributable to the fact that most cases of ovarian cancer in developing countries are diagnosed with advanced stage3. In Egypt, ovarian cancer represents 4.7% of all female cancers. It is ranked as the third and comprises 40.7% of all gynecologic cancers4. Risk factors such as obesity, commonly measured by body mass index (BMI), have been found to associate with poorer survival in a number of cancers including breast, prostate, and colorectal cancer5. Relation between obesity and epithelial ovarian cancer is still unclear. It is hypothesized that elevated androgen level mediated by obesity in the presence of certain androgen receptor polymorphisms might promote aggressive epithelial ovarian cancer biology6. This study aimed to assess the association between BMI and survival in Egyptian patients with epithelial ovarian cancer. Methods


IntroductIon
Ovarian cancer is the sixth most commonly diagnosed gynecologic tumor and constitutes a leading cause of mortality within this group of tumors 1 .This is because of its typically insidious onset and late diagnosis.Subsequently, only about 40% of women with ovarian cancer can be still alive more than 5 years post-diagnosis 2 .
This poor survival is largely attributable to the fact that most cases of ovarian cancer in developing countries are diagnosed with advanced stage 3 .
In Egypt, ovarian cancer represents 4.7% of all female cancers.It is ranked as the third and comprises 40.7% of all gynecologic cancers 4 .
Risk factors such as obesity, commonly measured by body mass index (BMI), have been found to associate with poorer survival in a number of cancers including breast, prostate, and colorectal cancer 5 .
Relation between obesity and epithelial ovarian cancer is still unclear.It is hypothesized that elevated androgen level mediated by obesity in the presence of certain androgen receptor polymorphisms might promote aggressive epithelial ovarian cancer biology 6 .
This study aimed to assess the association between BMI and survival in Egyptian patients with epithelial ovarian cancer.

Methods
This retrospective study was carried out on patients with epithelial ovarian cancer at the Oncology Department, Faculty of Medicine, Menoufia University. .
In the study period, from January 2011 to April 2014, a total of 87 patients were enrolled, but the clinical data completed only for 66 (76%) patients.
The collected data include patients' personal and clinical characteristics (age, performance status, weight, height, co-morbidities, parity and surgical interference), in addition to the disease characteristics, baseline investigations, treatment course, time to progression (TTP) and overall survival (OS).
Standing height was determined without shoes using Harpenden stadiometer (to the nearest 0.1 Cm).. Weight was measured using a digital scale (to the nearest 0.1 Kg) wearing light clothing and without shoes.BMI was calculated using the formula weight (in kilograms) divided by height (in meters) squared.
The World Health Organization classification of BMI 7 was used to classify patients as shown in Table 1.
Results were collected, tabulated and statistically analyzed by an IBM compatible personal computer with SPSS statistical package version 20 (IBM SPSS statistics for Windows, version 20.0, Armnok, NY: IBM Corp.).
Descriptive statistics were expressed in number (No), percentage (%), mean (x̅ ) and standard deviation (SD).For quantitative variables, Student's t-test was used for comparison between two groups of normally distributed data, while Mann-Whitney's test was used in the case of abnormal distribution.ANOVA test was used for comparison of quantitative variables between more than two groups of normally distributed data with LSD test as post Hoc test; while Kruskal Wallis test was used for comparison of quantitative variables between more than two groups of abnormally distributed data with Tamhane's test as a post hoc test.The correlation between two continuous variables was tested using Pearson correlation for normally-distributed data and Spearman correlation for abnormally-distributed data.Chi-square test (χ 2 ) was used to study association between qualitative variables.Whenever any of the expected cells were less than five, Fischer's exact test was used.Kaplan-Meier method was used for survival analysis and Log rank was used to test the significance of difference in survival between groups.

results
The included 66 patients' age ranged from 23 to 73 years and their BMI ranged from 17.6 to 48.1.More than half of patients were obese (BMI ≥30).Patients' characteristics are shown in Table 2.
Disease characteristics are shown in Table 3. Serous adenocarcinoma was the commonest pathology among the studied group (80.3%).Fifty two patients (78.8%) had elevated tumor marker CA-125 at presentation.
Table 4 shows the relationship between some of the studied variables and BMI.The average BMI did not differ significantly according to any of these variables.
Thirty seven patients (56%) were followed up until death.The estimated median OS of the whole group of patients was 47 months (95%CI: 37.25 -56.74).
Table 5 shows the relation between OS and some of the studied variables.Overall survival was significantly better in patients with platinum sensitive tumors, those who underwent surgery (at presentation or interval) and those with normal CA-125 level at presentation.
On comparing OS and TTP of patients with BMI ≤25 and patients with BMI >25, no statistically significant difference was obtained between the two groups (p-value = 0.93 for both).Kaplan Meier curve for OS and PFS in these two groups are shown in figures 1 and 2.
There was no statistically significant difference between the different categories of BMI as regards to OS and TTP as shown in Table 6.Kaplan Meier curves showing OS and TTP in different BMI categories are shown in figures 3 and 4.   Obesity is an established risk factor for ovarian cancer, and is associated with poor outcome in several tumors like breast cancer and colorectal cancer.However, the relation between obesity and ovarian cancer survival is less clear.
In this retrospective study, we found no statistically significant effect of being obese (≤25BMI) or non obese (BMI > 25) on OS in ovarian cancer patients.Alike, Matthews et al 8 found no difference in survival between patients with normal or abnormal BMI provided that complete surgical resection was done.Subsequently, he concluded that BMI was not the main issue; it was the possibility of resection.Also Zhang et al 9 found no statistically significant relation between BMI at diagnosis of ovarian cancer and patients' OS.
Our results are consistent with previous studies, whereas no association noted between ovarian cancer survival and being pre-obese or obese 8, 10 -12 .
Till now, the adverse effect of obesity on ovarian cancer survival remains equivocal for BMI measured at the time of diagnosis, where Bae et al 13 reported that BMI does not influence the survival of advanced stage patients except in underweight patients, which may be related to cancer related cachexia.
Noteworthy, Tyler et al studied the impact of BMI on survival of younger patients with epithelial ovarian cancer and also found no association 15 .
On the contrary, a recent large study which included the original data of 12,390 patients with ovarian carcinoma from 21 studies revealed that high BMI associated with an adverse effect on survival 16 .Such discrepancy in results may be attributed to the difference in sample size.
In our current study, no significant correlation obtained between BMI and TTP, consistent with previous studies, whereas Skírnisdóttir and Sorbe 11 , Barrett et al 17 and Hew et al 18 found that TTP was identical in patients with elevated BMI and in those with normal BMI.
Regarding the pathological nature of the tumors and its relation to the BMI of the patients, our results showed an even distribution within the different BMI categories, which agrees with the results of Bae et al 13 and Hew et al 18 .
Similarly, we found no significant association between BMI and the tumor stage in epithelial ovarian cancer patients.This concurs with the results of Barrett et al 17 and Hew et al 18 .In addition, no significant differences detected in the BMI according to parity, medical history, surgical optimality and initial CA-125, as reported by Bae et al 13 .
The known prognostic factors in epithelial ovarian cancer (e.g., platinum sensitivity) did not correlate with BMI, as previously reported 18 .

conclusIons
The prevalence of obesity among Egyptian patients with ovarian cancer is relatively high.
Although there is causal relationship between obesity and the development of epithelial ovarian cancers, no significant correlation obtained between BMI and OS and TTP among Egyptian patients.Similarly, there was no significant association between BMI and known prognostic factors in epithelial ovarian cancer.
The prognostic value of obesity remains a controversial issue with no consensus.

Figure. 2 :Figure. 4 :
Figure.2: Kaplan Meier curves showing time to progression of patients with BMI ≤25 and those with BMI >25

table 1 :
7he International Classification of adult underweight, overweight and obesity according to BMI7.
ECOG: Eastern Cooperative Oncology Group

table 4 :
Comparison between different aspects of characteristics regarding BMI.

table 5 :
Overall survival (OS) according to different risk factors.

table 6 :
Estimated overall survival and time to progression according to BMI.