**Background:** Extreme weight-loss behaviors are common among young Japanese women despite most being normal weight or underweight. Unhealthy weight-loss methods have been linked to depression in non-pregnant populations, but no prior studies examined whether pre-pregnancy weight-loss behaviors are associated with postpartum depression (PPD). The authors hypothesized that unhealthy and multiple weight-loss methods before pregnancy would increase PPD risk.
**Methods:** Data were drawn from the Japan Environment and Children's Study (JECS), a nationwide prospective birth cohort that recruited 97,413 pregnant women between January 2011 and March 2014 across 15 regional centers in Japan. After excluding women with miscarriage/stillbirth, multiple births, repeat participation, prior psychiatric/developmental disorders, and incomplete questionnaires, 62,446 women were analyzed. Weight-loss behaviors during the year before pregnancy were assessed via a dichotomous questionnaire administered in the second/third trimester, covering 7 methods: eating two-thirds as much or less, avoiding snacks, eating specific diet foods, taking diet pills, vomiting after eating, smoking cigarettes, and exercise. A weighted weight-loss score was calculated (1 point for healthy methods like exercise and avoiding snacks; 2 points for unhealthy methods like eating less and diet foods; 3 points for extremely unhealthy methods like diet pills, vomiting, and smoking). PPD was assessed at 1 month postpartum using the Japanese version of the Edinburgh Postnatal Depression Scale (EPDS), with a cut-off of ≥9 (75% sensitivity, 93% specificity). Binomial logistic regression was used to calculate crude and adjusted odds ratios (aOR), adjusting for age, BMI, marital status, education, income, emotional/physical abuse, psychological distress (K6), gestational weight gain, and other confounders. Sensitivity analysis used multiple imputation for missing data.
**Key Results:** At 1 month after delivery, 7,464 women (12.0%) had PPD (EPDS ≥ 9). Among women without antenatal psychological distress (K6 < 13), using at least one weight-loss method was significantly associated with PPD (aOR 1.318, 95% CI: 1.246–1.394, p < 0.001). Among women with psychological distress (K6 ≥ 13), the association showed a positive trend but was not statistically significant (aOR 1.250, 95% CI: 0.999–1.565, p = 0.052). In adjusted analyses for individual methods, all seven weight-loss behaviors were significantly associated with PPD. The strongest associations were for vomiting after eating (aOR 1.743, 95% CI: 1.465–2.065), smoking cigarettes (aOR 1.432, 95% CI: 1.287–1.591), and taking diet pills (aOR 1.308, 95% CI: 1.122–1.520). Even healthy methods like exercise (aOR 1.132, 95% CI: 1.072–1.194) and avoiding snacks (aOR 1.101, 95% CI: 1.045–1.160) were significantly associated with PPD. BMI-stratified analyses revealed important differences: in underweight women (BMI < 18.5, n = 9,726), significant associations were found for eating less, avoiding snacks, diet foods, and smoking, but not for vomiting, diet pills, or exercise. In normal-weight women (BMI 18.5–24.9, n = 45,945), all methods were significantly associated with PPD. In overweight/obese women (BMI ≥ 25, n = 6,389), only smoking cigarettes was significantly associated with PPD. The weighted weight-loss score showed a dose-response relationship: in normal-weight women, aORs increased progressively from score 1 (aOR 1.108, 95% CI: 1.039–1.182) to score 5 (aOR 1.832, 95% CI: 1.383–2.427). In underweight women, a significant increase was seen at score ≥ 3. In overweight/obese women, only score 4 was significantly associated (aOR 1.660, 95% CI: 1.058–2.604). Multiple imputation sensitivity analyses confirmed the main findings.
**Clinical Implications:** This is the first large-scale study to demonstrate that weight-loss behaviors before pregnancy are associated with increased PPD risk, with effects varying by pre-pregnancy BMI. The findings suggest that even "healthy" weight-loss methods like exercise and avoiding snacks may be problematic when used by normal-weight or underweight women who do not need to lose weight, possibly reflecting underlying body image concerns or disordered eating. The dose-response relationship between number of methods used and PPD risk underscores the importance of addressing excessive weight-loss behaviors. Clinicians should counsel women of reproductive age, particularly those planning pregnancy, about the potential mental health risks of unnecessary or extreme weight-loss behaviors. Limitations include recall bias (weight-loss behaviors reported retrospectively in second/third trimester), lack of data on frequency/intensity of behaviors, inability to fully exclude eating disorders, and the use of a non-validated weighted score. The study's strengths include its large nationwide sample, prospective design, and comprehensive adjustment for confounders.