Factors Associated with Overweight and Obesity among Women of Reproductive Age in Cambodia: Analysis of Cambodia Demographic and Health Survey

12 Introduction. Overweight and obesity are associated with increased rates of chronic 13 disease and death globally. In Cambodia, the prevalence of overweight and obesity 14 among women is high and may be growing. This study aimed to determine the 15 prevalence and factors associated with overweight and obesity among women of 16 reproductive age (WRA) in Cambodia. Methods. We analyzed data from the 2021-2022 17 Cambodia Demographic and Health Survey (CDHS) that used a two-stage stratified 18 cluster sampling design. Data analysis was restricted to non-pregnant women, resulting 19 in an analytic sample of 9,417 WRA. Multivariable logistic regressions were performed 20 using STATA V17 to examine factors associated with overweight and obesity. Results. 21 Prevalence of overweight and obesity among non-pregnant women of reproductive age 22 were 22.56% and 5.61% were overweight and obese respectively. Factors 23 independently associated with increased odds of overweight and/or obesity included 24 women aged 20-29 years [AOR=1.85; 95% CI: 1.22 - 2.80], 30-39 years [AOR=3.34; 25 95% CI: 2.21 - 5.04] and 40-49 years [AOR=5.57; 95% CI: 3.76 - 8.25], married women 26 [AOR=2.49; 95% CI: 1.71 - 3.62], women from middle wealth quintile [AOR=1.21; 95% 27 CI : 1.02 - 1.44], and rich wealth quintile [AOR=1.44 ; 95% CI: 1.19 - 1.73], having at 28 least three children or more [AOR=1.40; 95% CI: 1.00 - 1.95], ever drink alcohol 29 [AOR=1.24 ; 95% CI: 1.04 - 1.47], and current drink alcohol [AOR=1.2; 95% CI: 1.01 - 30 1.45]. On the contrary, following factors were independently associated with decreased 31 odds of having overweight and obese: women with at least secondary education 32 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice. 2 [AOR=0.73; 95% CI: 0.58-0.91], and working in in manual labor jobs [AOR=0.76; 95% 33 CI: (0.64 - 0.90]. Conclusion. Increased age, married women, having at least three 34 children and alcohol consumption were the main risk factors associated with overweight 35 and/or obesity. Conversely, higher education, and working in manual labor were 36 negatively associated with overweight and/or obesity. Cambodia’s non-communicable 37 disease (NCD) public health programs should consider these characteristic for targeting 38 interventions to further reduce overweight and/or obesity in the coming years. 39 40

Overweight, obesity, non-communicable diseases, NCDs, women at 41 reproductive age, CDHS, Cambodia 42 43

44 Overweight and obesity are major global public health challenges that have been rapidly 45 on the rise in the last four decades and are regarded as an epidemic [1]. In 2016, 46 approximately 39% (1.9 billion) of adults aged 18 years and older were overweight and 47 13% (650 million) were obese worldwide [1]. In general, being overweight and obesity 48 are major risk factors for several non-communicable diseases (NCDs) including 49 cardiovascular and kidney diseases, type 2 diabetes, some cancers, musculoskeletal 50 disorders, and other chronic diseases [2,3]. Moreover, among women of reproductive 51 age (WRA), overweight and obesity have been associated with increased risk of 52 pregnancy complications, caesarean section births, adverse birth outcomes, and infant 53 mortality [4]. Overweight and obesity are the fourth leading cause of risk-attributable 54 mortality [5], with a reduced life expectancy of 5-20 years, depending on the severity of 55 the condition and the presence of comorbidities [1]. According to the WHO, being 56 overweight and obese are the l eading risks for global deaths with at least 2.8 million 57 adults dying annually due to these conditions [1]. Overweight and obesity were higher 58 proportion among women as compared to men in both developed and developing 59 countries; for instance, overweight was 40% in women vs 39% in men and obesity 15% 60 in women vs. 11% in men [1]. The Global Nutrition Report 2019 showed that 26.1% of 61 women and 20.4% of men were overweight, while obesity was 6.3% in women 62 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 3 compared to 3.5% in men [6]. Along with the rapidly increasing population in Cambodia, 63 from 15.42 million in 2015 to nearly 16.21 million in 2019 [7], the prevalence of 64 overweight and obesity are also steadily on the rise. The 2021-22 Cambodia 65 Demographic and Health Survey (CDHS) report indicated that overweight and obesity 66 among non-pregnant women at reproductive age increased from 18% in 2014 to 39% in 67 2021-22 [8]. It was estimated that overweight and obesity contribute to rising healthcare 68 costs in Cambodia (approximately 1.7% of its annual gross domestic product (GDP) per 69 capita) and is a major contributor to mortality and decreased general health and 70 productivity [9]. Regarding predictors of overweight and obesity among women include 71 higher socioeconomic status, older age, marriage, living in an urban residence, and lack 72 of education [3,10-12]. Women with formal employment had higher odds of being 73 overweight or obese than informally employed women [3,13]. Also, women who used 74 hormonal contraceptives such as oral contraceptive pills, implants, patches, and rings 75 were at higher risk of being overweight or obese [14,15]. Globally, 30% of daily smokers 76 are overweight or obese [16], with women smokers at greater risk for overweight or 77 obesity than men smokers [17,18]. Women with frequent television watching [19], 78 alcohol drinking [20,21], frequent consumption of sweets foods and unhealthy foods [22] 79 were found to have higher odds of being overweight or obese. To our knowledge, 80 factors associated with overweight and obesity specifically among WRA in Cambodia 81 using an updated data have not been explored. A previous study on prevalence of 82 overweight and obesity among WRA and its associated factors utilized data since 2014 83 [3]. In the context of Cambodia where the prevalence of overweight and obesity in WRA 84 have increased rapidly, a thorough and comprehensive investigation of a wide range of 85 socio-demographic and behavioral factors is warranted to identify the factors 86 independently associated with having overweight and/or obesity. Identifying the key 87 modifiable socio-demographic and behavioral factors, as well as women who are at a 88 high-risk of having overweight and obesity, may help guide the timely development of 89 promising and feasible public health intervention strategies to address the growing 90 overweight and obesity pandemic. Therefore, we aimed to determine the prevalence 91 and examined sociodemographic and behavioral factors associated with overweight and 92 obesity among WRA in Cambodia. 93 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 4 94

95 Data sources 96 97 We followed the methods of Um et al., 2023 [3]. To analyze prevalence and factors 98 associated with overweight and obesity among WRA in Cambodia, we used existing 99 women’s data from the 2021-2022 CDHS. The CDHS is a nationally representative 100 population-based household survey implemented by the National Institute of Statistics 101 (NIS) in collaboration with the Ministry of Health (MoH) with financial support from the 102 Government of the Cambodia, and technical assistance from ICF International (USA) 103 thorough the Demographic and Health Survey program. from September 15, 2021 to 104 February 15, 2022 [8]. The sampling frame used for the 2021-2022 CDHS was taken 105 from the Cambodia General Population Census 2019 [7]. Thereafter, the participants in 106 CDHS 2021-2022 were selected using pr obability proportion based on two-stage 107 stratified cluster sampling from the chosen sampling frame. In the initial stage, 709 108 enumeration area (EAs) (241 urban areas and 468 rural areas) were selected. In the 109 second stage, an equal systematic sample of 30 households was selected from each 110 cluster, for a total sample size of 21,270 households. In total, 19,496 women aged 15-111 49 years were interviewed with a response rate of 98.2%. Survey data were obtained 112 through face-to-face survey interviews using a standardized survey instrument by 113 trained interviewers. Anthropometric of weight and height measurements for adult 114 women age 15-49 were taken by trained female field staff using standardized 115 instruments and procedures [7]. Weight measurements were taken using scales with a 116 digital display (UNICEF model S0141025). Height and length were measured using a 117 portable adult height measurement system (UNICEF model S0114540). The detailed 118 protocol and methods were published previously [8]. Eligible participants for this study 119 were women of reproductive age 15-49 years, with exclusion of pregnant women and 120 those who had a birth within 2 months before the survey, with available body mass 121 index (BMI) data. As a result, we excluded a total of 828 pregnant women and 9,249 122 women with missing BMI data. A final sample including in this analysis was 9,417 123 women of reproductive age 15-49 years. 124 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 5 125 Measurements 126 Outcome variable 127 128 The primary outcome variable of this study was overweight and obese for women of 129 reproductive age. BMI - defined by dividing a person’s weight in kilograms by the square 130 of their height in metres (kg/m 2) – was used to measure the outcome [8,23]. For adults 131 over 15 years old, the following BMI ranges were used: underweight (< /i1 18.0/i1 kg/m2), 132 normal weight (18.5-24.9 /i1 kg/m2), overweight (25.0-29.9 /i1 kg/m2) and obesity ( ≥ 133 30.0/i1 kg/m2) [8,23]. Overweight/Obese was defined as a binary outcome for which 134 women with a BMI ≥ 25.0 kg/m2 were classified as overweight and obese (coded = 1), 135 while women with a BMI < 25.0 kg/m2 were coded as other (coded = 0). 136 137 Covariate variables 138 139 Covariates included socio-demographic characteristics of women of reproductive age: 140 Women’s age in years (15-19, 20-29, 30-39, and 40-49), marital status (not married, 141 married or living together, and divorced or widowed or separated), educational level (no 142 formal education, primary, secondary or higher education), occupation (not working, 143 agriculture, manual labor or un skilled, professional or sealer or services), number of 144 children ever born (no children, one-two child, and three and more children). 145 Households’ wealth status was represented by a wealth index calculated via principal 146 component analysis (PCA) and using variabl es for household assets and dwelling 147 characteristics. Weighted scores divided into five wealth quintiles (poorest, poorer, 148 medium, richer, and richest) each comprising 20% of the population [8], place of 149 residence (rural vs urban). Cambodia’s provinces were regrouped for analytic purposes 150 into a categorical variable with 4 geographical regions: plains, Tonle Sap, costal/sea, 151 and mountains [7]. Behavioral factors including smoking (non-smoker vs smoker), 152 alcohol consumption in the past month (never drink, ever drink, and current drink) 153 current drink of alcohol corresponds to one can or bottle of beer, one glass of wine, or 154 one shot of spirits in past month, watching television at least once a week (yes vs no), 155 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 6 contraceptive usage (not used, hormonal methods (using a pill, emergency 156 contraceptive pill, Norplant, injection, and vaginal rings), non-hormonal methods 157 (condoms, the diaphragm, the IUD, spermicides, lactational amenorrhea, and 158 sterilization), and traditional methods (periodic abstinence and withdrawal) [8]. 159 160 Statistical analysis 161 162 All statistical analyses were performed using STATA version 17 (Stata Corp 2021, 163 College Station, TX) [24] and accounted for the CDHS sampling weights and complex 164 survey design. Socio-demographic characteristics and behavioral factors were 165 described in weighted frequency and percentage. The provincial variation in the 166 prevalence of overweight and/or obesity was done using ArcGIS software version 10.8 167 [24]. Shapefiles for administrative boundaries in Cambodia are publicly accessible 168 through the DHS website (URL: https://spatialdata.dhsprogram.com/covariates). 169 Descriptive analyses that coupled cross-tabular frequency distributions with chi-square 170 tests were used to assess associations between explanatory variables and the outcome 171 variable Overweight/Obese. Variables associated with the outcome variable with a 172 significance level of p-value ≤ 0.10 or that had theoretical justification (for example, 173 women’s age, wealth index, alcohol consumption, residence, and geographic of region) 174 were included in the unadjusted and adjusted logistic regression analyses [3]. 175 Unadjusted logistic regression was used to determine the magnitude effect of 176 associations between overweight and obesity with socio-demographic characteristics 177 and behavioral factors reported as odds ratios (OR) with 95% confidence intervals (CI). 178 Then, adjusted logistics regression was used to assess independent associations, 179 reported as adjusted odds ratios (AOR), with overweight and obesity after adjusting for 180 other independent variables included in the model. Multi-collinearity between 181 independent variables was checked before fitting the final regression model 182 (supplementary table S1). 183 184

185 186 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 7 Characteristics of the study samples 187 188 The mean age of women was 31 years old (SD = 9.6 years) in which age group 30-49 189 years old accounted for 57.11%. Almost 67.82% were married; 49.47% had completed 190 at least secondary education and 11.90% did not receive formal schooling. A total of 191 30.30% of women did professional work, 25.86% were unemployed. Of the total 192 sample, 35.97% women were from the poor households. Over half (57.27%) of women 193 resided in rural areas. About 27.79% had three and more children while 29.49% had no 194 children. Only 1.40% of women reported cigarette smoking, 16.60% reported currently 195 drink alcohols and 18.60 reported ever drink, 12.35% reported using hormonal 196 contraceptives. Mean women’s BMI was 22.9 kg/m 2 (SD = 3.9 kg/m 2) and 22,56% and 197 5.61% were overweight and obese respectively (Table 1). 198 199 Table 1. Socio-demographic and behaviors characteristics of the weighted samples of 200 women aged 15-49 years old in Cambodia, 2021-2022 (n/i1 =/i1 9,417, weighted count) 201 Variables Freq. Percent (95% CI) Women's mean age in years (SD) 30.9 (9.6) 15-19 1,488 15.80 20-29 2,551 27.09 30-39 3,220 34.19 40-49 2,158 22.92 Marital status Not married 2,429 25.79 Married or living together 6,387 67.82 Widowed/divorced/separated 601 6.38 Education No education 1,121 11.90 Primary 3,637 38.62 Secondary and higher 4,659 49.47 Current work status Not working 2,435 25.86 Agricultural 1,607 17.06 Professional 2,853 30.30 Manual labor and unskilled 2,331 24.75 Number of children born . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 8 No birth 2,777 29.49 1-2 child 4,023 42.72 Three and above 2,617 27.79 Household wealth index Poor 3,387 35.97 Middle 1,822 19.35 Rich 4,207 44.67 Smoking Non smoker 9,285 98.60 Smoker 132 1.40 Current drinking alcohol Never drink 6103 64.80 Ever drink 1751 18.60 Current drink 1562 16.60 Frequency of watching television Not at all 5,808 61.68 Less than once a week 1,493 15.85 At least once a week 2,116 22.47 Ever report of contraceptive use No method 5,152 54.71 Traditional method 2,398 25.46 Non-hormonal method 704 7.48 Hormonal method 1,163 12.35 Place of residence Urban 4,024 42.73 Rural 5,393 57.27 Region Plain 4,708 49.99 Tonle Sap 2,859 30.36 Coastal 595 6.32 Plateau/Mountain 1,255 13.33 BMI mean in kg/m2 (SD) 22.9(3.9) Underweight 1013 10.76 (9.9-11.7) Normal weight 5751 61.08 (59.7-62.5] Overweight 2124 22.56 (21.5-23.7) Obese 528.2 5.61 (5.0-6.3] Notes: Survey weights applied to obtain weighted percentages. Plains: Phnom Penh, Kampong 202 Cham, Tbong Khmum, Kandal, Prey Veng, Svay Rieng, and Takeo; Tonle Sap: Banteay Meanchey, 203 Kampong Chhnang, Kampong Thom, Pursat, Siem Reap, Battambang, Pailin, and Otdar Meanchey; 204 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 9 Coastal/sea: Kampot, Kep, Preah Sihanouk, and Koh Kong; Mountains: Kampong Speu, Kratie, Preah 205 Vihear, Stung Treng, Mondul Kiri, and Ratanak Kiri. 206 207 Prevalence of overweight and obesity among WRA by provinces 208 209 The prevalence of overweight and obesity is highest among WRA in Kampong Cham 210 (34.1%), Kandal (32.3%), Svay Rieng (32%), Preah Sihanouk (31.2%), Phnom Penh, 211 Kompong Thom, Pailin, and Tboung Khmum (30% each) and lowest among WRA in 212 Ratanak Kiri (9.4%) and Kampong Chhnang (14.6%) (Fig 1, supplementary table S2). 213 214 215 216 Fig 1. Prevalence of overweight and obesity in Women Reproductive Age by province. Map 217 created using ArcGIS software version 10.8 [24]. Shapefiles for administrative boundaries in 218 Cambodia are publicly accessible through DHS website (URL: 219 https://spatialdata.dhsprogram.com/covariates). 220 221 222 Factors associated with overweight and obesity in bivariate analysis 223 224 Socio-demographic characteristics and behavioral factors were significantly associated 225 with overweight and obesity ( Table 2 ). Women had higher prevalence of 226 overweight/obesity if they were aged 40-49 years (47.31%) compared to younger age 227 groups, (p value <0.001), married (35.87%) compared to other relationship status (p 228 value <0.001), had no education (37.91%) compared to higher education (p value 229 <0.001), were employed in a professional job (34.03%) compared to other jobs, or were 230 from the rich households (30.90%) compared to poor households (p value <0.001). In 231 addition, Overweight/obesity increased with parity. Women reported at least three 232 children had a significantly highest overweight and obesity (42.53%), (p value < 0.001). 233 Women who current drink alcohol had a significant higher proportion of overweight and 234 obesity (33.87%) than women did not (p value <0.001). Women used non-hormonal 235 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 10 contraceptives methods had a significant higher proportion of overweight and obesity 236 (37.92%) than women using other methods (p value <0.019). Geographic regions of 237 residence were likewise associated with a woman having overweight and obesity. 238 Women who lived in urban areas had a higher prevalence of overweight and obesity 239 than those living in rural areas (29.90% vs 26.87%, p value = 0.028). Plain regions were 240 positively associated with overweight and obesity (29.91%) compared other regions ( p 241 value = 0.006). 242 243 Table 2. Factors associated with overweight and obesity among women aged 15-49 244 years in Chi2 analysis (n/i1 =/i1 9,417, weighted count) 245 Characteristics Overweight/Obese (n=2,652) Normal/Underweight (n= 6,764) P value Freq. % Freq. % Women’s age group in years 15-19 80 5.38 1,408 94.62 <0.001 20-29 444 17.40 2,107 82.60 30-39 1,108 34.41 2,112 65.59 40-49 1,021 47.31 1,137 52.69 Marital status Not married 181 7.45 2,248 92.55 <0.001 Married or living together 2,291 35.87 4,096 64.13 Widowed/divorced/separated 181 30.12 420 69.88 Education No education 425 37.91 695 62.00 <0.001 Primary 1,254 34.48 2,383 65.52 Secondary and higher 973 20.88 3,686 79.12 Current work status Not working 586 24.07 1,849 75.93 <0.001 Agricultural 485 30.18 1,121 69.76 Professional 971 34.03 1,881 65.93 Manual labor and unskilled 563 24.15 1,768 75.85 Number of children born No birth 264 9.51 2,512 90.46 <0.001 1-2 child 1,275 31.69 2,748 68.31 Three and above 1,113 42.53 1,504 57.47 Household wealth index Poor 841 24.83 2,546 75.17 <0.001 Middle 512 28.10 1,311 71.95 Rich 1,300 30.90 2,907 69.10 Smoking Non smoker 2,621 28.23 6,664 71.77 0.325 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 11 Smoker 32 24.24 100 75.76 Current drinking alcohol Never drink 1,573 25.77 4,530 74.23 <0.001 Ever drink 550 31.41 1,201 68.59 Current drink 529 33.87 1,033 66.13 Frequency of watching television Not at all 1,623 27.94 4,186 72.07 0.901 Less than once a week 423 28.33 1,069 71.60 At least once a week 606 28.64 1,509 71.31 Ever report of contraceptive use No method 1,132 21.97 4,020 78.03 <0.001 Traditional method 821 34.24 1,577 65.76 Non-hormonal method 259 36.79 446 63.35 Hormonal method 441 37.92 722 62.08 Place of residence urban 1,203 29.90 2,821 70.10 0.028 rural 1,449 26.87 3,944 73.13 Region Plain 1,408 29.91 3,299 70.07 0.006 Tonle Sap 768 26.86 2,092 73.17 Coastal 171 28.74 425 71.43 Plateau/Mountain 306 24.38 949 75.62 246 247 248 Factors associated with overweight and obesity in adjusted logistic regression 249 250 As shown in Table 3 , several factors were independently associated with increased 251 odds of having overweight and obesity among women. These factors included age 252 group 20-29 years [AOR=1.85; 95% CI: 1.22 - 2.80], 30-39 years [AOR=3.34; 95% CI: 253 2.21 - 5.04] and 40-49 years [AOR=5.57; 95% CI: 3.76 - 8.25] married [AOR=2.49; 95% 254 CI: 1.71 - 3.62] and widowed/divorced/separated [AOR=1.14; 95% CI: 1.14 - 2.63], 255 middle wealth quintile [AOR=1.21; 95% CI : 1.02 - 1.44], and rich wealth quintile 256 [AOR=1.44 ; 95% CI: 1.19 - 1.73], having at least three children or more [AOR=1.40 ; 95% 257 CI: 1.00 - 1.95], ever drink alcohol [AOR=1.24 ; 95% CI: 1.04 - 1.47], and current drink 258 alcohol [AOR=1.21 ; 95% CI: 1.01 - 1.45]. On the contrary, following factors were 259 independently associated with decreased odds of having overweight and obese: women 260 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 12 with at least secondary education [AOR=0.73; 95% CI: 0.58-0.91], working in in manual 261 labor jobs [AOR=0.76; 95% CI: (0.64 - 0.90]. 262 263 Table 3. Risk factors associated with overweight and obesity in unadjusted and 264 adjusted logistic regression analysis 265 266 Characteristics Total (n=9,417) Total (n= 9,225) OR 95%CI AOR 95%CI Women’s age group in years 15-19 Ref. Ref. 20-29 3.71*** (2.61 - 5.27) 1.85*** (1.22 - 2.80) 30-39 9.25*** (6.70 - 12.77) 3.34*** (2.21 - 5.04) 40-49 15.84*** (11.68 - 21.47) 5.57*** (3.76 - 8.25) Marital status Not married Ref. Ref. Married or living together 6.96*** (5.67 - 8.54) 2.49*** (1.71 - 3.62) Widowed/divorced/separated 5.34*** (3.81 - 7.50) 1.73** (1.14 - 2.63) Education No education Ref. Ref. Primary 0.86 (0.72 - 1.03) 0.97 (0.80 - 1.17) Secondary and higher 0.43*** (0.35 - 0.53) 0.73*** (0.58 - 0.91) Current work status Not working Ref. Ref. Agricultural 1.37*** (1.14 - 1.64) 0.85 (0.69 - 1.03) Professional 1.63*** (1.40 - 1.89) 1.10 (0.93 - 1.31) Manual labor and unskilled 1.00 (0.85 - 1.19) 0.76*** (0.64 - 0.90) Number of children born No birth Ref. Ref. 1-2 child 4.41*** (3.61 - 5.38) 1.23 (0.89 - 1.68) Three and above 7.03*** (5.79 - 8.55) 1.40** (1.00 - 1.95) Household wealth index Poor Ref. Ref. Middle 1.18** (1.02 - 1.37) 1.21** (1.02 - 1.44) Rich 1.35*** (1.18 - 1.56) 1.44*** (1.19 - 1.73) Smoking . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 13 Non smoker Ref. - Smoker 0.81 (0.52 - 1.24) - Current drinking alcohol Never drink Ref. Ref. Ever drink 1.32*** (1.13 - 1.54) 1.24** (1.04 - 1.47) Current drink 1.47*** (1.23 - 1.76) 1.21** (1.01 - 1.45) Watching television Not at all Ref. Less than once a week 1.02 (0.85 - 1.23) - - At least once a week 1.04 (0.89 - 1.21) - - Ever report of contraceptive use No method Ref. Ref. Traditional method 1.85*** (1.61 - 2.12) 1.02 (0.86 - 1.21) Non-hormonal method 2.06*** (1.64 - 2.60) 0.84 (0.65 - 1.10) Hormonal method 2.17*** (1.75 - 2.69) 1.01 (0.81 - 1.26) Place of residence Rural Ref. Ref. Urban 1.16** (1.02 - 1.33) 1.11 (0.94 - 1.31) Region Plain Ref. Ref. Tonle Sap 1.33*** (1.11 - 1.58) 1.17 (0.95 - 1.43) Coastal 1.14 (0.95 - 1.36) 1.11 (0.91 - 1.35) Plateau/Mountain 1.25** (1.00 - 1.55) 1.10 (0.86 - 1.41) *** p<0.01, ** p<0.05, * p<0.10, OR= unadjusted odds ratio, AOR= adjusted odds ratio 267 268 269

270 271 The overall prevalence of overweight and obesity among WRA were 22,56% (95% CI: 272 21.5-23.7) and 5.61% (95% CI: 5.0-6.3) were overweight and obese respectively. 273 Factors associated with being overweight or obese for non-pregnant women include 274 older age, increased parity, marriage and ever married, middle and rich household 275 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 14 wealth index, and alcohol consumption. Women with higher education as well as those 276 working in manual labor were less likely to be overweight or obese. 277 278 Being older aged women were more likely overweight or obese as compared to younger 279 women aged 15-19 years. This is consistent with other studies that showed obesity was 280 more prevalent in older WRA [3,25]. The risk of women becoming overweight or obese 281 rises with age, possibly due to unhealthy food consumption and/or a lack of physical 282 activity [26]. Women with better socioeconomic status and who live in urban areas had 283 higher odds of being overweight and obese, which is consistent with prior research 284 [27,28]. Women with higher socioeconomic status tend to use more improved 285 technologies for a more comfortable lifestyle [19,29]. Women have at least three and 286 more children at risk of being overweight and obese, similar other studies [3,30]. During 287 pregnancy, factors such as stress, depression, and/or anxiety may play a role in 288 hypothalamic-pituitary-adrenal hyperactivity [31,32]. Women with several children may 289 also have gained weight as a result of their reduced physical activity and have less time 290 to focus on health behaviors including weight management. In addition, women ever 291 and current consuming alcohol were significantly associated with higher risks of 292 overweight and obesity. The association between alcohol consumption status and 293 overweight and obesity has been studied thoroughly in various populations such as in 294 Urban Cambodia and in Hawassa city, southern Ethiopia [20,21]. Commonly, alcohol 295 consumption is considered to increase appetite, resulting in excessive energy intake 296 and leading to overweight and obesity [24]. In contrast, women with higher education 297 are less likely to be overweight or obesity than women with no school due to increased 298 knowledge and awareness; for example, studies in Cambodia, Nigeria and South Korea 299 reported that educated women had a lower risk of being overweight or obese and that 300 this may be linked to education influencing healthy behavior [3,33-36]. Education is a 301 critical predictor of women’s healthy behaviors and health outcomes, including diet, 302 physical activity [37]. Although overweight and obesity was not significantly associated 303 with geographical regions of residence, those from Plain and Coastal region had a high 304 rate of overweight and obesity, 29.91% and 28.74%, respectively. Also, women who 305 lived in urban areas had a higher prevalence of overweight and obesity 29.90%. Prior 306 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 15 research on overweight and obesity among women reproductive age in Cambodia 307 found women resided in urban had significant associated with overweight and obesity 308 [3]; our findings suggest a need for more regionally representative studies. 309 310

311 312 The overall prevalence of overweight and obesity was very high among non-pregnant in 313 Cambodia. Women being older age, higher education levels, higher income and wealth, 314 and behavioral factors such as alcohol consumption, type of employment as risk factors 315 associated with overweight and obesity among women of reproductive age in Cambodia. 316 It is crucial to design intervention programs that target these socio-demographic factors 317 and to raise awareness on the importance of consuming healthy food as well as the 318 benefits of regular physical activity, especially among older women and those with 319 professional jobs. 320 321 322

and strengths of this study 323 324 Our study has several limitations. First, because CDHS 2021-2022 collects cross-325 sectional data, our analysis could not explore changes over time. Second, data on 326 several important variables, such as women’s food-consumption behaviors, were not 327 collected by CDHS, and CDHS did not objectively measure physical activity; therefore, 328 our ability to examine the association of these variables with overweight and obesity 329 was restricted. Third, numerous psychological factors (e.g., depressive and anxiety 330 disorders) and physiological factors may also be associated with obesity, but these 331 factors were not included in this study because they were not available in the data set. 332 Finally, because the data were available only for women aged 15 to 49, our results may 333 not be generalizable to girls younger than 15 and women older than 49. Despite these 334 limitations, our results contribute to the literature on the prevalence and association 335 between sociodemographic and behavioral variables and the status of overweight and 336 obesity among women reproductive age in Cambodia. The major strength of this study 337 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 16 is that it is the first to use nationally representative data with a high response rate of 97% 338 to examine the prevalence and predictors of overweight and obesity in this country. 339 Data were collected using validated survey methods including calibrated measurement 340 tools and highly trained data collectors, which contribute to improved data quality [38]. 341 Formally incorporating the complex survey design and sampling weights into the 342 analysis bolsters the rigor of the analysis and enables generalizing our findings to the 343 population of non-pregnant women in Cambodia. 344 345 346 Abbreviations 347 BMI, Body mass index; NCDs, Non-communicable diseases; WHO, World Health 348 Organization; CDHS, Cambodia Demographic Health Survey; WRA, Women 349 reproductive age; AOR, Adjusted odds ratio; PPS, Probability proportional to size. EA, 350 Enumeration areas 351 352 353 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 17 Data Availability 354 Our study used 2021-2022 Cambodia Demographic and Health Survey (CDHS) 355 datasets. The DHS data are publicly available from the website at 356 (URL:https://www.dhsprogram.com/data/available-datasets.cfm). 357 The Shapefiles for administrative boundaries in Cambodia are publicly accessible 358 through DHS website at (URL: https://spatialdata.dhsprogram.com/covariates). 359 360 Ethical Approval 361 The data used in this study were extracted from CDHS 2021-22 data [8], which are 362 publicly available with all personal identifiers of study participants removed. Permission 363 to analyze the data was granted through registering with the DHS program website and 364 submitting an application outlining the intended use of the datasets [8]. Informed 365 consent was obtained from the study participants before data collection. The data 366 collection tools and procedures for CHDS 2021-22 was approved by the Cambodia 367 National Ethics Committee for Health Research and the Institutional Review Board (IRB) 368 of ICF in Rockville, Maryland, USA. 369 370 Funding Statement 371 The authors received no specific funding for this work. 372 373 Consent 374 Consent for publication is not necessary because this manuscript did not contain any 375 personal details such as photos, images, videos, or quote data. 376 377 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 18 Competing interests 378 The authors have declared that no competing interests exist. 379 Authors’ Contributions 380 Samnang Um. contributed to conceptualization, methods, performed the data analysis, 381 interpreted the data, writing the original draft, and reviewing/editing the manuscript. 382 Yom An. contributed to technical methods evaluation, support data analysis, and 383 reviewing/editing the manuscript. Finally, all authors read and approved the final 384 manuscript. 385 Acknowledgments 386 The authors would like to thank DHS-ICF, who approved the data used for this paper. 387 We would like to acknowledge management team of National Institute of Public Health, 388 Phnom Penh, Cambodia namely Professor. Chhea Chhorvann . Professor. Heng 389 Sopheab for their continued support and encouragement. 390 391

392 1. World Health Organization, Fact sheets, Obesity and overweight, Available 393 from [ https://www.who.int/news-room/fact-sheets/detail/obesity-and-394 overweight], Accessed on April 10, 2023. 395 2. Trends in adult body-mass index in 200 countries from 1975 to 2014: a 396 pooled analysis of 1698 population-based measurement studies with 19·2 397 million participants. Lancet (London, England) 2016, 387(10026):1377-1396. 398 3. Um S, Yom A, J AM, Sopheab H: Overweight and obesity among women at 399 reproductive age in Cambodia: Data analysis of Cambodia Demographic 400 and Health Survey 2014. PLOS global public health 2023, 3(3):e0001768. 401 4. Yu Z, Han S, Zhu J, Sun X, Ji C, Guo X: Pre-pregnancy body mass index in 402 relation to infant birth weight and offspring overweight/obesity: a 403 systematic review and meta-analysis. PLoS One 2013, 8(4):e61627. 404 5. Stanaway JD, Afshin A, Gakidou E, Lim SS, Abate D, Abate KH, Abbafati C, 405 Abbasi N, Abbastabar H, Abd-Allah F: Global, regional, and national 406 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 19 comparative risk assessment of 84 be havioural, environmental and 407 occupational, and metabolic risks or clusters of risks for 195 countries and 408 territories, 1990–2017: a systematic anal ysis for the Global Burden of 409 Disease Study 2017. The Lancet 2018, 392(10159):1923-1994. 410 6. Report GN. Country Nutrition Profiles, Available from; 411 [https://globalnutritionreport.org/resources/nutrition-profiles/asia/south-412 eastern-asia/cambodia/]; Accessed on April 10, 2023. 413 7. National Institute of Statistics (NIS). (2020). General population census of 414 the Kingdom of Cambodia 2019; accessed on November 4th 2021, availbale 415 from: 416 [https://www.nis.gov.kh/nis/Census2019/Final%20General%20Population%2417 0Census%202019-English.pdf]. 418 8. National Institute of Statistics (NIS) [Cambodia], Ministry of Health (MoH) 419 [Cambodia], and ICF. 2023. Cambodia Demographic and Health Survey 420 2021–22 Final Report. Phnom Penh, Cambodia, and Rockville, Maryland, 421 USA: NIS, MoH, and ICF. 422 9. World Food Programme in Cambodia; 423 [https://www.wfp.org/countries/cambodia]: Accessed on April 10, 2023. 424 10. Balarajan Y , Villamor E: Nationally representative surveys show recent 425 increases in the prevalence of overweight and obesity among women of 426 reproductive age in Bangladesh, Nepal, and India . The journal of nutrition 427 2009, 139(11):2139-2144. 428 11. Abrha S, Shiferaw S, Ahmed KY: Overweight and obesity and its socio-429 demographic correlates among urban Ethiopian women: evidence from the 430 2011 EDHS. BMC Public Health 2016, 16:1-7. 431 12. Yeshaw Y, Kebede SA, Liyew AM, Tesema GA, Agegnehu CD, Teshale AB, Alem 432 AZ: Determinants of overweight/obesity among reproductive age group 433 women in Ethiopia: multilevel analys is of Ethiopian demographic and 434 health survey. BMJ open 2020, 10(3):e034963. 435 13. Oddo VM, Bleich SN, Pollack KM, Surkan PJ, Mueller NT, Jones-Smith JC: The 436 weight of work: the association between maternal employment and 437 overweight in low-and middle-income countries . International Journal of 438 Behavioral Nutrition and Physical Activity 2017, 14(1):1-10. 439 14. Lopez LM, Bernholc A, Chen M, Grey TW, Otterness C, Westhoff C, Edelman A, 440 Helmerhorst FM: Hormonal contraceptives for contraception in overweight 441 or obese women. Cochrane Database of Systematic Reviews 2016(8). 442 15. Aryeetey R, Lartey A, Marquis GS, Nti H, Colecraft E, Brown P: Prevalence and 443 predictors of overweight and obesity among school-aged children in urban 444 Ghana. BMC obesity 2017, 4:1-8. 445 16. de Munter Js, Tynelius P , Magnusson C, Rasmussen F: Longitudinal analysis 446 of lifestyle habits in relation to body mass index, onset of overweight and 447 obesity: results from a large population-based cohort in Sweden . 448 Scandinavian journal of public health 2015, 43(3):236-245. 449 17. Piirtola M, Jelenkovic A, Latvala A, Sund R, Honda C, Inui F, Watanabe M, 450 Tomizawa R, Iwatani Y, Ordonana JR: Association of current and former 451 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 20 smoking with body mass index: A study of smoking discordant twin pairs 452 from 21 twin cohorts. PloS one 2018, 13(7):e0200140. 453 18. Chiolero A, Jacot ‐ Sadowski I, Faeh D, Paccaud F, Cornuz J: Association of 454 cigarettes smoked daily with obesi ty in a general adult population . Obesity 455 2007, 15(5):1311-1318. 456 19. Ghose B: Frequency of TV viewing and prevalence of overweight and 457 obesity among adult women in Bangladesh: a cross-sectional study . BMJ 458 open 2017, 7(1):e014399. 459 20. Dagne S, Gelaw YA, Abebe Z, Wassie MM: Factors associated with 460 overweight and obesity among adults in northeast Ethiopia: a cross-461 sectional study. Diabetes, metabolic syndrome and obesity: targets and therapy 462 2019:391-399. 463 21. Tamaoki M, Honda I, Nakanishi K, Nakajima M, Cheam S, Okawada M, 464 Sakakibara H: Lifestyle Factors Associated with Metabolic Syndrome in 465 Urban Cambodia. International Journal of Environmental Research and Public 466 Health 2022, 19(17):10481. 467 22. Darebo T, Mesfin A, Gebremedhin S: Prevalence and factors associated with 468 overweight and obesity among adults in Hawassa city, southern Ethiopia: a 469 community based cross-sectional study. BMC obesity 2019, 6:1-10. 470 23. World Health Organization, Fact sheets, A healthy lifestyle - WHO 471 recommendations. Available from [ https://www.who.int/europe/news-472 room/fact-sheets/item/a-healthy-lifestyle---who-recommendations], 473 Accessed on April 10, 2023. 474 24. Yeomans MR: Alcohol, appetite and energy balance: is alcohol intake a risk 475 factor for obesity? Physiology & behavior 2010, 100(1):82-89. 476 25. Biswas T, Uddin MJ, Mamun AA, Pervin S, P Garnett S: Increasing prevalence 477 of overweight and obesity in Bangladeshi women of reproductive age: 478 Findings from 2004 to 2014. PloS one 2017, 12(7):e0181080. 479 26. Alemu E, Atnafu A, Yitayal M, Yimam K: Prevalence of overweight and/or 480 obesity and associated factors among high school adolescents in Arada 481 Sub city, Addis Ababa, Ethiopia . Journal of nutrition & food sciences 2014, 482 4(2):1. 483 27. Mangemba NT, San Sebastian M: Societal risk factors for overweight and 484 obesity in women in Zimbabwe: a cross-sectional study . BMC public health 485 2020, 20(1):1-8. 486 28. Kamal SM, Hassan CH, Alam GM: Dual burden of underweight and 487 overweight among women in Bangl adesh: patterns, prevalence, and 488 sociodemographic correlates. Journal of health, population, and nutrition 2015, 489 33(1):92. 490 29. Islam M, Sathi NJ, Abdullah HM, Naime J, Butt ZA: Factors Affecting the 491 Utilization of Antenatal Care Services During Pregnancy in Bangladesh and 492 28 Other Low-and Middle-income Countries: A Meta-analysis of 493 Demographic and Health Survey Data . Dr Sulaiman Al Habib Medical Journal 494 2022, 4(1):19-31. 495 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 21 30. Mbochi RW, Kuria E, Kimiywe J, Ochola S, Steyn NP: Predictors of overweight 496 and obesity in adult women in Nairobi Province, Kenya . BMC public health 497 2012, 12(1):1-9. 498 31. Li W, Wang Y, Shen L, Song L, Li H, Liu B, Yuan J, Wang Y: Association 499 between parity and obesity patterns in a middle-aged and older Chinese 500 population: a cross-sectional analysis in the Tongji-Dongfeng cohort study. 501 Nutrition & metabolism 2016, 13(1):1-8. 502 32. Pasquali R, Vicennati V, Cacciari M, Pagotto U: The 503 hypothalamic‐ pituitary‐ adrenal axis activity in obesity and the metabolic 504 syndrome. Annals of the New York Academy of Sciences 2006, 1083(1):111-128. 505 33. Hamad R, Elser H, Tran DC, Rehkopf DH, Goodman SN: How and why studies 506 disagree about the effects of education on health: A systematic review and 507 meta-analysis of studies of compulsory schooling laws . Social Science & 508 Medicine 2018, 212:168-178. 509 34. Chung A, Backholer K, Wong E, Palermo C, Keating C, Peeters A: Trends in 510 child and adolescent obesity preval ence in economically advanced 511 countries according to socioeconomic position: a systematic review . 512 Obesity reviews 2016, 17(3):276-295. 513 35. Barlow P: The effect of schooling on women's overweight and obesity: A 514 natural experiment in Nigeria. Demography 2021, 58(2):685-710. 515 36. McLaren L: Socioeconomic status and obesity . Epidemiologic reviews 2007, 516 29(1):29-48. 517 37. Liu H, Guo G: Lifetime socioeconomic status, historical context, and genetic 518 inheritance in shaping body mass in middle and late adulthood . American 519 sociological review 2015, 80(4):705-737. 520 38. BOWLER P: CUNNINGHAM, A. and JARDINE, N.(editors). Romanticism and 521 the sciences. Cambridge University Press, Cambridge: 1990. Pp xxii, 345; 522 illustrated. Price:£ 15, US $19.95. ISBN: 0-521-35685-7 (paperback). In.; 1992. 523 524 525 526 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint 22 Supplemental materials 527 Table S1. Results of checking multicollinearity using Variance Inflation Factor (VIF) 528 Table S2. Prevalence of overweight and/or obesity among women reproductive age, CDHS 529 2021-2022 (n=9,417) 530 531 532 533 534 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint Kratie Pursat Mondul Kiri Siemreap Koh Kong Preah Vihear Stung Treng Battambang Ratanak Kiri Kampong Thom Kampot Takeo Kandal Prey VengKampong Speu Oddar Meanchey Tboung Khmum Pailin Banteay Meanchey Kampong ChamKampong Chhnang Svay Rieng Preah Sihanouk Phnom Penh Kep Percent 9 - 15 16 - 25 26 - 30 31 - 35 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 28, 2023. ; https://doi.org/10.1101/2023.04.28.23289243doi: medRxiv preprint