Obesity is considered a 21st century epidemic and reliance on the measurement of body mass index (BMI) alone has been shown to be inadequate in assessing and managing obesity-related cardiometabolic risk (1, 2). Waist circumference—a marker of visceral adiposity—is strongly associated with all-cause (3, 4) and cardiovascular (5, 6) mortality. Visceral fat accumulates when adipose tissue has a limited ability to expand. This process is indicative of a hyperlipolytic state; marked by resistance to insulin and an enhanced adipokine output.

There are large differences in body composition between men and women. With an overall weight gain, waist circumference also increases in young adults; however, men have larger increases in waist circumference naturally accompanying their overall greater weight gain compared to women. Hence, a universally acceptable approach is to set a higher circumference cutoff value for abdominal obesity in men compared to women (7). For example, the International Diabetes Federation sets the threshold for abdominal obesity for men at 90 cm (or 94 cm for White men), and 80 cm for women across ethnic backgrounds. Metabolic syndrome—providing clinicians with a simple screening tool to identify individuals who are likely to have insulin resistance and related metabolic abnormalities—lists waist circumference (≥ 80 cm in women) as one of the five possible criteria for the syndrome, highlighting the importance of central adiposity on disease probability (8, 9).

The premenstruum is marked by an increase in food craving in a significant number of reproductive age women, who are particularly vulnerable to weight gain (10–13). A study by Hartlage et al. (14) administered a questionnaire listing 50 premenstrual symptoms to women from the community and showed that food craving ranked second in magnitude, immediately after bloating. Indeed, food craving, defined as an “intense desire to eat a specific food or food type” (15, 16) is endorsed by women in the premenstruum both retrospectively (17, 18) and prospectively (14, 19, 20). Functional magnetic resonance tomography studies show higher brain responses to food stimuli in the luteal vs. follicular phase, particularly in the corticolimbic areas involved in homeostasis and reward (21–24). Combined, the results of this research characterize functional and self-reported indices of increased food craving in the premenstruum.

No study to date has established an association between either the PMS/PMDD diagnosis, or a particular premenstrual symptomatology, and central adiposity. Data from the Nurses’ Health Study 2 (NHS2) was analyzed for central adiposity (waist circumference); however, only an association between PMS and BMI was found. PMS cases were identified by asking women whether they ever received a diagnosis for PMS—an approach which may have diminished the power to isolate an association. PMS diagnosis is based on various symptom types (see Section “Diagnostic Criteria”), some of which may be distal to central adiposity. As only one symptom must be met (and up to the total of four), a woman would meet the PMS diagnosis if she reported, for example, increased difficulty concentrating in the premenstruum. An association with the PMS diagnosis in this case may be more difficult to capture than if the diagnosis was based on food craving. Moreover, PMS/PMDD diagnosis is rarely established clinically using prospective recordings across two menstrual cycles, as required in DSM-5; hence, it is possible that there were diagnostic misclassifications of the ever/never diagnostic status in the Bertone-Johnson et al. (25) analysis because the diagnosed participants were likely not followed prospectively at the time of the diagnosis determination.

Our analysis is based on a stringent prospective symptom data collection of premenstrual food craving symptomatology. Importantly, it is adjusted for BMI, which is necessary as the full strength of the association between waist circumference with morbidity and mortality is realized only after adjustment for BMI (3, 26, 27). Based on the literature showing that women’s food craving increases premenstrually (14), and that there is an association between food craving and waist circumference in women even independent of the reproductive status (28), we hypothesized that there would be a positive association between premenstrual food craving and waist circumference. Our study identified the premenstruum as a window of opportunity for the development of targeted interventions, which is of critical importance for health promotion in reproductive age women, who are especially vulnerable to developing obesity.

In accordance with our hypothesis, the present study shows a positive association between premenstrual food craving and waist circumference. This association remained significant even after adjusting for BMI, suggesting a particularly dangerous central fat accumulation. Though preliminary in nature, this is the first study to link food craving in the premenstruum with central adiposity.

Changing sex hormones in the luteal phase are associated with decreased levels of amino acids and lipid species, presenting a physiological milieu of heightened energy requirement (36–41) and indicative of an anabolic state (41, 42). The amino acid decrease likely reflects the effect of rising progesterone levels in the luteal phase. Progesterone upregulates cell cycle and growth, resulting in protein biosynthesis for endometrial thickening (43). Therefore, women may feel an increase in food craving during this time to satisfy this underlying energy expenditure. In one study (41), as many as 39 amino acids were suppressed in the luteal phase. A potential relationship between premenstrual amino acid levels, food craving, eating patterns and central adiposity should be investigated in future studies. Understanding which amino acid suppression is potentially linked to heightened premenstrual food craving may provide a basis for a targeted intervention development.

Alternatively, the relationship observed in the present study is driven by increased food intake as a means of affect regulation in the premenstruum, though this hypothesis should be tested. Reward-driven behaviors are goal-oriented changes and are based on past experiences as well as positive associations we may make with food (44, 45). If eating provides a euphoric experience, the behavior to consume this food will be reinforced and repeated. Therefore, craving, operationalized as appetitive motivation, can be portrayed as the amount of work and effort an individual is willing to perform to obtain a type of food. Our finding possibly reflects deficiency in reward-driven behaviors in a segment of women with PMS/PMDD, who score high on reward dependence (46), show low positive affect (47, 48) as well as low frontal electroencephalography (49–51), which are associated with deficient reward processing (52, 53).

Temporal behavioral interventions should be developed in the event subsequent studies demonstrate that the association observed in the present study is mediated via an increase in food intake. Reducing central adiposity in reproductive age women would have a population-level impact, as interventions would improve not only their health and well-being, but also that of their offspring (54). However, unless carefully designed, such efforts may fail. Generic interventions are generally not successful in reproductive age women, as evidenced by lower recruitment and retention rates, inferior attendance and compliance, and worse weight loss outcomes in comparison to older adult participants (55).

There is some, though limited, evidence that individualized treatment in reproductive age women is effective. For example, Eiben et al. (56) targeted diet, exercise, and behavior change utilizing face-to-face contact followed by ongoing regular telephone and e-mail contact for 1 year (56), and showed reduced central adiposity. Levitskey and colleagues (57) showed that a minimal intervention, where young women weighed themselves daily for 10 weeks and were provided with feedback on energy needs for weight maintenance, resulted in weight stability compared with controls who gained weight. However, long-term evaluation of these or other behavioral interventions in reproductive age women is not known (58) and no studies to date incorporated premenstrual behavioral changes related to food cravings. Results of our study provide compelling evidence that an intervention designed to specifically target premenstrual cravings may result in an efficacious reduction of central adiposity given that food cravings may be conditioned by interventions (59–64).

By calculating effect size for premenstrual food craving, our goal was to emphasize its importance as a clinical problem due to its link to central adiposity. We show that nutrient craving is affected by changing sex hormones between menstrual phases, which may predispose a segment of women to high nutritional needs. The mean effect size for premenstrual food craving in our study was 0.69, while in the study by Hartlage et al. (14), it was 0.29 for the community sample and 0.96 for the PMDD clinical sample. The effect size difference in our study is likely due to our advertising strategy, which asked “Do you have PMS?,” drawing a sample of participants with PMS/PMDD symptomatology from the community pool.

The present study is the first to identify a segment of reproductive age women who have an adverse health marker (waist circumference) related to premenstrual food craving. The findings of our study are preliminary in nature, and future evaluations should be completed in larger participant samples. If replicated, the finding of the present study provides a platform for further novel nutrition interventions, with an emphasis on improved health and well-being in this at-risk population.

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

This study protocol was reviewed and approved by the Office for the Protection of Research Subjects (OPRS) at the University of Illinois at Chicago, approval number 2018-1533. The patients/participants provided their written informed consent to participate in this study.

ND analyzed the study data. DK, AN, and FS assisted with the manuscript preparation and submission. JS provided the statistical consulting. AH designed the study, analyzed the study data, and wrote the initial and final manuscript versions. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.