Re: “Impact of Stress on Menstrual Cyclicity During the COVID-19 Pandemic: A Survey Study” by Ozimek et al.

Dear Editors:

We were very interested to read the article “Impact of Stress on Menstrual Cyclicity During the COVID-19 Pandemic: A Survey Study” by Ozimek et al. ¹ In the context of widespread media attention (as cited by the authors) to the potential influence of coronavirus disease 2019 (COVID-19)–related stress on menstrual cycle changes, it is important to have data on which to base clinical and public health guidance. However, we are concerned that methodological limitations of the study were not fully presented.

First, the title of the article uses causal language (“impact”) that is not supported by the study's cross-sectional design. The authors ask women to report their current perceived stress and their self-reported experience of menstrual changes during the COVID-19 pandemic. Menstrual changes are reported qualitatively, rather than with any measure of change (e.g., cycle length and number of bleeding days) or indication of their clinical significance. The study does not define what might be considered a “change” such that even physiological fluctuations within the range of normal might be reported as a change. The prevalence and significance of reported menstrual changes might thus be overestimated.

Second, although the authors asked women to recall their perceived stress before and during the COVID-19 pandemic via the Perceived Stress Scale (PSS), the pre-COVID-19 data are not used in the analysis of menstrual changes. Any changes in stress due to the pandemic are not incorporated into the analysis of the association of stress with changes in menstrual cycles. Taken alone, the COVID-19 PSS score could reflect numerous other ongoing stresses during the study period (e.g., nationwide protests against racial violence).

Third, the article does not adequately acknowledge the selection bias inherent in convenience sampling recruitment via social media-based snowball sampling. The study recruitment statement, “Help us learn more about how the COVID-19 pandemic impacts women's reproductive health,” likely furthered selection bias by attracting individuals predisposed to having both stress and reproductive health problems, inclusive of menstrual abnormalities. Selection bias is demonstrated in the PSS scores reported in this study, which skewed toward moderate to high levels of stress. No individuals scored <16 such that no individuals could be categorized as having “low stress” at any point before or after the pandemic.

The conclusion that these data “suggest that the COVID-19 pandemic may have directly contributed to menstrual cycle irregularities in women experiencing both moderate and high degrees of stress” goes beyond the data presented. The authors rightly described the COVID-19 pandemic as creating “intense feelings of fear, uncertainty, or anxiety […] coupled with the disruption of daily routines and behaviors.” We are concerned that this cross-sectional study in a small convenience sample will exacerbate such anxiety.

In our observational study that analyzed data from 214,426 cycles from 18,076 individuals, we noted that although nearly half of individuals reported more pandemic-related stress, that changes in average prospectively collected cycle and menstrual length were not clinically significant. ² Populations of women who participate in online surveys and use mobile apps for menstrual tracking represent a select population, and as neither study directly assessed perceived stress before the pandemic, more prospective data are needed to answer this important question, which as the authors point out is of great interest to menstruating individuals and the media.