Ramadan fasting and weight change trajectories: Time-varying association of weight during and after Ramadan in low-income and refugee populations

**Background** Obesity is a major driver of non-communicable diseases globally. Fasting, including intermittent fasting, has been shown to improve health outcomes, but the effects of Ramadan fasting—which involves abstention from both food and water from dawn to dusk—are less clear. Approximately 90% of the world’s 1.8 billion Muslims fast during Ramadan, making this a large-scale, recurring dietary intervention with enormous public health implications. Existing studies have not elucidated specific patterns of weight change during and after Ramadan over a long-term follow-up period. This study aimed to address that gap by evaluating weight trajectories over 52 weeks in a low-income and refugee population in Amman, Jordan.

**Methods** This was a secondary analysis of data from 913 participants recruited between 2012 and 2015 as part of a multicenter, 3-arm randomized controlled trial (NCT01596244) evaluating a Microclinic Program for diabetes management. Participants were recruited from three community health centers in Amman, Jordan. Eligibility included men and women aged 18 years or older with diabetes or at risk for diabetes. Weight was measured weekly using a standardized protocol (Health Scale SVR 160) starting at the beginning of Ramadan. The analysis combined all study arms into a single cohort. Two multi-level mixed effects modeling approaches were used: a discrete model (assuming stepwise weekly changes) and a continuous linear spline model (assuming piecewise linear relationships). Models were adjusted for age, sex, and trial arm. Sensitivity analyses tested knot location robustness. Subgroup analyses examined interactions by sex and by diabetes status (non-prediabetic/non-diabetic [HbA1c < 5.7%], pre-diabetic [HbA1c 5.7–6.5%], and diabetic [HbA1c > 6.5%]).

**Key Results** Baseline characteristics: mean age 54.2–56.6 years across arms, 65–67% women, mean BMI 33.4–33.6 kg/m², mean HbA1c 6.90–6.91%. The discrete model showed slight weight gain in the first two weeks of Ramadan (0.53 kg, 95% CI: 0.06, 1.01 by week 2), followed by gradual weight loss reaching 0.55 kg (95% CI: 0.05, 1.05) by week 8 (one month after Ramadan ended). Weight then gradually increased over the next 18 weeks, reaching an estimated gain of 2.54 kg (95% CI: 1.57, 3.53) by week 26. A sharp drop of approximately 2.66 kg occurred between week 26 and week 28 (−0.12 kg, 95% CI: −0.89, 0.56 at week 28), after which weight stabilized near baseline. The continuous spline model confirmed this pattern: weight decreased during the first eight weeks (estimated slope [ES]: −0.104, 95% CI: −0.158, −0.050), then increased until week 26 (ES: 0.080, 95% CI: 0.044, 0.116), with another drop around week 27 (ES: −0.136, 95% CI: −0.426, 0.154), and stabilization thereafter (ES: −0.011, 95% CI: −0.021, −0.001). Subgroup analyses showed similar overall trajectories by sex, with greater fluctuation in females than males. By diabetes status, non-prediabetic/non-diabetic individuals showed the largest weight fluctuations, while diabetic individuals peaked earlier (week 20) compared to pre-diabetic and non-diabetic groups (week 27).

**Clinical Implications** This is the first study to evaluate weight change patterns over a full 52-week period encompassing Ramadan and post-Ramadan months. The observed weight fluctuations—particularly the rebound after initial loss—may be clinically concerning, as bodyweight variability has been shown to increase the risk of incident type 2 diabetes. The study identifies critical turning points (week 8 for maximum loss, week 26 for peak regain, and week 28 for stabilization) that public health organizations can use to time interventions. Ramadan presents a valuable opportunity for targeted health campaigns, community awareness programs, and monitoring to sustain the modest weight loss benefits and prevent post-Ramadan weight regain. Limitations include the non-randomized design (randomization to fasting is not feasible), conduct in a single country (Jordan) during summer months (June–August), lack of data on physical activity and fasting adherence, and limited generalizability to non-fasting Muslims, high-income countries, or winter Ramadan seasons. Future studies should evaluate these temporal effects across different geographic, cultural, and seasonal contexts and among low- and high-risk subgroups.