Are the body shape index, the body roundness index and waist‐to‐hip ratio better than BMI to predict recurrent pregnancy loss?

Abstract Purpose Etiology could not be determined in approximately 50% of recurrent pregnancy loss cases, and it was named unexpected recurrent pregnancy loss(URPL). A body shape index(ABSI), body roundness index(BRI), and waist‐to‐hip ratio(WtHR) are new indexes that are superior to BMI in showing body fat distribution. We aimed to investigate the potency of ABSI, BRI, and WtHR in URPL, their superiority to BMI, and their suitability for clinical use. Methods One hundred and thirty‐eight patients between the ages of 20‐40 who applied to our hospital for URPL between January 2016 and December 31, 2020 were included in our study. Weight, height, waist circumference, and hip circumference were measured, and indexes were calculated. Differences between the URPL and control groups were calculated using the IBM SPSS program. Results There was a significant difference between the two groups for BRI, ABSI, and WtHR values, while there was no significant difference in BMI. BRI(4.4 ± 1.7vs3.9 ± 1.5), ABSI(0.08 ± 0.005 vs 0.078 ± 0.004), and WtHR(0.84 ± 0.06vs0.82 ± 0.05) values were higher in the URPL group. ROC analysis showed us that BRI, ABSI, and WtHR have a diagnostic value for URPL(P < .05). When indexes were above the cutoff values, RPL risk increased 3.59 times in ABSI, 2.26 times in BRI, and 2.9 times in WtHR(P < .05). Conclusions The relationship between obesity and URPL can be explained more clearly by using effective indexes that show body fat distribution rather than BMI. Ethics committee approval was obtained from Erzincan Binali Yildirim University in 14.01.2021. Clinical Research Ethics Committee no: 01/01.


| INTRODUC TI ON
The spontaneous termination of pregnancy before the 20th week and when the fetal weight is below 500 g is called "abortion," and about 15%-25% of known pregnancies will end in a miscarriage. 1,2 Recurrent pregnancy loss (RPL) has been defined as the loss of three or more consecutive pregnancies before the 20th gestational week. 3 The rate of RPL ranges from 1-5% of women of reproductive age. 4 The etiology of RPL includes immunological, genetic, endocrine, anatomical, environmental factors, and infections. Despite all these investigations, etiology could not be determined in approximately 50% of RPL cases, and it was named unexpected recurrent pregnancy loss (URPL). 5 Obesity is a chronic disease characterized by an increase in body fat mass since the energy taken into the body is greater than the energy consumed. 6 Obesity is determined by the measurement of body mass index (BMI). The World Health Organization defined a BMI of 25 and above as overweight and 30 and above as obesity. 7 In the literature, studies have reported that infertility, abortion, recurrent pregnancy losses, and failure rates in assisted reproductive techniques increase in obese patients. 8,9 It has been found that many chronic diseases, especially diabetes and cardiac diseases, are related to body fat distribution and fat percentage rather than the weight-height ratio. 10,11 MRI, bio-impedance analysis, air displacement plethysmography, and dual-energy X-ray absorptiometry are direct methods used to show body fat distribution. However, researchers have been canalized to new indirect methods that can be applied more easily, since direct methods are not practical in the clinic, their cost is high, and experienced personnel are needed for shooting and interpretation. 12 It has been reported that waist circumference (WC) and waistto-hip ratio (WtHR) are superior to BMI in showing cardiometabolic diseases. 13 Based on this, an index named a body shape index (ABSI) was developed by Krakauer NY and Krakauer JC in 2012, and it was determined that it is superior to BMI and WC alone in showing premature mortality. 14 The Body Roundness Index (BRI), which was developed in 2013 by Thomas et al. has modeled the human body as an ellipse and considered it in two axes: the major axis consisting of height and the minor axis consisting of waist and hip. They defined the degree of body roundness as "eccentricity" between 1 and 16. Values approaching 1 indicate thin and narrow elliptical-shaped bodies, while values close to 16 indicate bodies with round and wide elliptical shapes. 15 It has been reported that BRI is more sensitive than BMI and WC in predicting metabolic syndrome and dyslipidemia. 16,17 Ovarian reserve expresses the reproductive potential of the woman in terms of number and quality. 18 Follicle stimulating hormone (FSH), estradiol (E 2 ), antimullerian hormone (AMH) are some tests used for determining ovarian reserve. 19 Obesity has been associated with a decreased ovarian reserve and impaired oocyte quality by affecting follicle functions and development. 20,21 Studies concluding that decreased ovarian reserve may cause RPL and infertility as a result of the effect on oocyte quality and number are available in the literature. 22,23 One of the important underlying factors of central obesity is leptin/adiponectin imbalance, which has been reported to be effective in determining the prognosis of diseases associated with abdominal obesity. 24 Besides its role in fetal growth and development, leptin also has a modulator role for syncytiotrophoblates. It takes part in autocrine / paracrine events in implantation and the continuation of pregnancy. 25 It was thought that the impairment of leptin balance and the resulting leptin resistance may be associated with poor reproductive performance and miscarriage. 26 In light of this information, in this study, we aimed to investigate the potency of ABSI, BRI, and WtHR in URPL cases, as well as their superiority to BMI, their suitability for clinical use, and their effects on ovarian reserve.

| MATERIAL S AND ME THODS
One hundred and thirty-eight patients between the ages of 20-40 who applied to our hospital for URPL between January 2016 and December 2020 were included in our study. Their files were re-

| Calculation of Anthropometric Indexes
The weight of the patients was measured in kg and height in cm.
While the patients wore thin clothes for weight measurement, shoes were removed for height measurement. The weight was measured at approximately 0.1 kg. The height was measured at approximately 0.1 cm. Waist circumference was measured over bare skin, midway between the lower rib margin and the iliac crest at the end of expiration. Hip circumference was measured as the maximum circumference over the buttocks to the nearest 0.1 cm using a soft tape measure.

| Statistical analyses
IBM SPSS version 21 (IBM Corp) was used for analyzing data.
Descriptive statistics of continuous variables were presented as mean ± standard deviation, median (minimum-maximum) value, and categorical variables as number (%). The compliance of the data to normal distribution was tested with the Shapiro-Wilk test. The Mann-Whitney U test was used when comparing continuous variables in groups. The Chi-square test was used in the analysis of categorical variables. While testing the diagnostic value of the indices, ROC analysis was used, and area under curve (AUC) was presented with 95% confidence intervals (CI). Youden's index was used while determining the optimum cutoff value, and diagnostic accuracy criteria for the cutoff were presented. RPL risk was given as odds ratio (OR) according to the index cutoff points determined. A P value of <0.05 was significant in all statistical tests. While determining post hoc powers for primary outcomes (BRI, ABSI, WtHR), effect sizes were taken as 0.311, 0.442, and 0.362, respectively. Type-I error was taken as 0.05 and post hoc powers found as 74.0%, 95.5%, and 85.1%, respectively. While determining the difference between groups, the sample size was adequate. For post hoc power calculation, G*power 3.1.9.2 was used.

| RE SULTS
The demographic and characteristics of both groups are shown in When AMH≤1 and AMH>1 were compared, the number of patients with AMH≤1 was more common in the URPL group; the difference was not statistically significant (P = .235). Similarly, among the poor ovarian reserve markers, serum FSH≥11 and serum E 2 ≥60 values did not show a significant difference between the URPL and control groups (P >.05) ( Table 1).
As BMI increased, the number of patients with AMH≤1 ng/mL increased for both the URPL and control groups (P < .01). While AMH≤1 was observed in 8.3% of the patients with normal BMI, ). this rate increased to 25% in patients with BMI≥25 (P < .05). There was no significant difference between the BMI groups in terms of FSH≥11 and E 2 ≥60 (P > .05).
As shown in Table 1 In regions above the cutoff value for BRI, ABSI, and WtHR, the number of patients with AMH≤1 µg/L was higher and statistically significant (P < .05). However, an FSH level ≥11 U/L, a serum E 2 level≥60 nmol/L did not have statistical significance in patients who were below or above the cutoff value (P > .05).
In normal BMI patients, the relations of the URPL and control groups with the BRI, ABSI, and WtHR cutoff values are shown in Table 3. The URPL rate had been significantly increased in the patients above the cutoff value for all three indexes (P < .05). RPL, which in 50% of cases is of unknown etiology, making it impossible for patients and physicians to determine future pregnancy outcomes, may cause serious emotional stress and depression for patients. 32 It is known that obesity increases the risk of first trimester and recurrent pregnancy loss. Therefore, studies have been conducted on the etiology of RPL, and its relationship with obesity has been examined using BMI. 33 In this study, the relationship between BRI, ABSI, and WtHR with URPL, their potential superiority to BMI, and its relationship with ovarian reserve have been investigated. The relationship with URPL and an index (other than BMI) that may be effective in showing the diagnosis of obesity and fat distribution has not been investigated before in the literature. Our work is a first in this respect.

| D ISCUSS I ON
Obesity is thought to have a negative effect on female fertility by affecting the hypothalamic-pituitary-gonadal-hormonal axis, oocyte quality, embryo development, and endometrial receptivity. 34 According to Cavalcante et al. when the patients were classified according to BMI in the meta-analysis, the relationship between the obese group and RPL was seen, but the risk was not determined in the overweight and underweight groups. 35  In metabolic syndrome, the potency of ABSI was found to be lower compared to BMI and WtHR, while BRI was reported to be the most effective index in predicting metabolic syndrome. 16 In another study, BRI was found to have the highest capacity to define diabetes mellitus, while BMI was found to be the least associated. 30 In this study, where we investigated the relationship between URPL and indexes that were easily calculated in clinical practice, it is striking that WtHR, BRI, and ABSI values were significantly higher and diagnostic power in the URPL group, while there was no difference between the groups in terms of BMI value and BMI classification. We found ABSI has the best predictive ability for URPL, and if ABSI was above the cutoff value, URPL risk increased 3.59 times. WtHR above the cutoff value increased the risk 2.9 times, and BRI above the cutoff value increased the URPL risk 2.26 times.
In the normal BMI class, 67.7% of the patients above the cutoff value calculated for ABSI and 72.2% of the patients above the cutoff value calculated for WtHR were in the URPL group. It was interesting that while nine patients were followed in patients with normal BMI above the BRI cutoff value, all of them were in the URPL group. In light of these results, the risk of URPL must increase with the accumulation of body fat distribution in the central region, even in patients who are calculated as a normal BMI group and who are thought to have excluded the effect of obesity on URPL, and we think that this should be considered.
It has been reported that diminished ovarian reserve (DOR) can be effective on RPL by affecting oocyte quality and oocyte number. 43 In a meta-analysis results, it was stated that there is a relationship between DOR and RPL, especially URPL. According to the results of the same study, it has been reported that low AMH levels (<1 ng/mL) and RLP are related, and the relationship between FSH and E2 values is not clear. 44 Atasever et al. found in their study that low AMH and high FSH values were associated with RPL. 23 In another study, a significant relationship was found between RPL and AMH, but no significant relationship was found with FSH. 22 The mechanism of action of obesity on ovarian reserve has not been clearly explained in the literature.  and massive obese groups as a result of BMI classification are the limitations of the study.

| CON CLUS ION
In conclusion, URPL is a devastating situation whose etiology has not fully resolved, and the subsequent pregnancy outcomes are difficult to predict by patients and clinicians. In this study, we demonstrated that BRI, ABSI, and WtHR have diagnostic value for URPL, as in obesity-related diseases. We think that the relationship between obesity and URPL can be explained more clearly by using more effective indexes in showing body fat distribution rather than BMI.
It should be kept in mind that central obesity is associated with an increased risk of pregnancy loss even in patients with normal BMI for URPL, whose etiology is questioned in every detail. Therefore, central region weight control and prevention of central obesity should be recommended to both patients with high BMI groups and patients with normal BMI groups.