Occupational physical demands and menstrual cycle irregularities in flight attendants and teachers.

This was a secondary analysis of data from a study of occupational exposures and pregnancy outcomes among female flight attendants and teachers.[ 16 ] Flight attendants aged 18–45 years were selected from employee rosters from three airlines operating in three U.S. cities. Flight records were obtained from study airlines for those flight attendants who worked at least one month between August 1992 and July 1996. A comparison group of female teachers aged 18–45 years and teaching grades 5–12 was selected from local school districts in the same cities. The original study team selected teachers as the comparison group because, based on Census data, they were a female-dominated workforce with similar socioeconomic status to flight attendants but with few of the same occupational exposures related to flight, such as cosmic radiation exposure and circadian rhythm disruption. Because the primary objective of the original study was to investigate pregnancy-related outcomes, the study population in the original study included only women who were married and had no hysterectomy or tubal ligation at study entry. Between November 1999 and April 2001, the flight attendants and teachers were invited to complete a computer-assisted telephone interview that asked about their current work, menstrual function, and reproductive histories. This interview was the source of data for the present analysis. All participants gave informed consent for participation. The study was approved by the institutional review board of the National Institute for Occupational Safety and Health (protocol 96-DSHEFS-12). We analyzed six menstrual cycle characteristics experienced over the 12 months preceding the interview: (1) frequency of menstrual cramps (never, sometimes, often, always); (2) menstrual pain (none, mild, moderate, severe); (3) irregular cycles (yes [≥8 days between shortest and longest cycles], no [≤7 days between shortest and longest cycles]); (4) menstrual flow (light, moderate, heavy) (5) bleed length (short [1–3 days], medium [4–7 days], long [≥8 days]) and (6) cycle length (short [≤25 days], medium [26–31 days], long [≥ 32 days]). The full interview questions and original response options are available in the Supplementary Materials . We examined five occupational physical demands as reported in the questionnaire: (1) overall physical effort (self-reported as either light/moderate or hard), (2) standing (≤8, ≥9 hours per day), (3) frequency of lifting loads weighing ≥15 pounds (≥6.8 kg) (<1, 1–5, 6–10, ≥11 times per day), (4) frequency of pushing or pulling loads with ≥15 pounds (≥6.8 kg) of effort (<1, 1–9, 10–19, 20–30, ≥31 times per day), and (5) frequency of bending or twisting at the waist (<1, 1–25, 26–50, 51–75, ≥76 times per day). The full interview questions and original response options are available in the Supplementary Materials . In our first analysis, we compared the prevalence of the six menstrual characteristics between flight attendants and teachers. We used modified Poisson regression to estimate prevalence ratios (PRs) and 95% confidence intervals (CI) for associations between working as a flight attendant and each menstrual cycle characteristic. For multinomial outcomes, we created dichotomous outcomes for analysis by contrasting each non-reference group separately against the reference group. We included study city (A, B, C), age at interview (continuous), body mass index (BMI; continuous) and race/ethnicity (non-Hispanic White, other) in the multivariable models based on directed acyclic graphs (DAGs). Our second analysis—examining the relationship between occupational physical demands and menstrual cycle characteristics—was restricted to flight attendants because teachers reported a low prevalence of occupational physical demands and to avoid confounding by unmeasured differences between flight attendants and teachers. We used modified Poisson regression to estimate PRs and 95% CIs for associations between each measure of occupational physical demands and each menstrual cycle characteristic. The final models included study city (A, B, C), age at interview (continuous), body mass index (BMI; continuous), race/ethnicity (non-Hispanic White, other), and household income (<$75,000, ≥$75,000 U.S. dollars, not adjusted for inflation) as covariates based on DAGs. The covariates included in this model and the one described above differed because their DAGs differed. For example, income was on the causal pathway between exposure and outcome in the first model, but not the second. Age at menarche is presented in the participant characteristics but was not included in the models because it is correlated with BMI. We conducted all analyses in R version 4.0.4. We excluded participants for the following eight reasons: (1) no longer flight attendants or teachers at interview; (2) recruited as flight attendants but changed careers and became teachers by the time of the interview, and vice versa ; (3) aged ≥45 years at interview; (4) no menstrual periods in the past 12 months; (5) pregnant or breastfeeding within the past 6 months; (6) used hormonal contraception, hormonal medication, or intrauterine device (IUD) in the past 12 months; (7) self-reported endometriosis; and (8) missing data on any of the inclusion/exclusion criteria variables or variables included in the multivariable models. For each menstrual characteristic separately, we additionally excluded participants who had missing data on the outcome or who reported implausible or contradictory information about their menstrual cycles ( e.g. , bleed length longer than cycle length).

In the original study, 5,096 flight attendants were contacted to participate in the study, 2,595 were eligible, and 2,226 (86%) completed the interview; among teachers, 1,582 were contacted, 466 were eligible, and 376 (81%) completed the interview. Of the 2,602 participants completing the interview, we excluded 1,788 from all analyses. The top reasons for exclusion were: no longer working as a flight attendant or teacher at the time of the interview (n = 463); hormone or IUD use (n = 457); age ≥45 years (n = 345); and recent pregnancy or breastfeeding (n = 238). Participant flow charts in the Supplementary Materials describe these and additional exclusions in more detail. Following exclusions, 814 participants (694 flight attendants and 120 teachers) were included in one or more analyses. Flight attendants and teachers differed on most sociodemographic characteristics ( Table 1 ). Compared to teachers, flight attendants were younger, were more likely to be of minority race/ethnicity, had lower BMI, had higher household incomes, and had later age at menarche. Occupational physical demands also differed markedly between flight attendants and teachers ( Table 2 ). 61% of flight attendants described their job as involving hard physical effort compared to 12% of teachers. For each of standing, lifting, pushing or pulling, and bending or twisting, a larger proportion of flight attendants reported greater physical demands than did teachers. Menstrual cycle irregularities were common ( Table 3 ). Over half of participants reported menstrual cramps often or always, with 61% of flight attendants and 51% of teachers usually experienced moderate to severe pain. About one in five participants had short cycles, irregular cycles, or long bleed lengths. Nearly half reported heavy menstrual flow. Compared to teachers, flight attendants were more likely to report never having menstrual cramps (adjusted PR 2.4, 95% 1.0, 6.3, never vs. sometimes); however, among those who reported cramps, flight attendants had them more frequently than teachers (adjusted PR 1.6, 95% CI: 1.2, 2.1, always vs. sometimes). Flight attendants were also less likely to have mild menstrual pain than teachers (adjusted PR 0.6, 95% CI: 0.5, 0.8, vs. moderate pain) ( Table 3 ). Menstrual flow and irregular cycles did not differ between flight attendants and teachers. Our estimates for bleed length and cycle length were imprecise because of small sample size. We then restricted the analyses to flight attendants to investigate associations between specific occupational exposures and menstrual cycle irregularities. Flight attendants were more likely to always (vs. sometimes) have cramps if their job involved overall hard physical effort (PR 1.2, 95% CI: 1.0, 1.5 vs. light to moderate effort) or if they reported standing more than 8 hours per workday (PR 1.2, 95% CI: 1.0, 1.5 versus ≤8 hours) ( Table 4 ). All measures of occupational physical demands were associated with either greater likelihood of severe menstrual pain (vs. moderate pain) or lower likelihood of having no or mild menstrual pain (vs. moderate pain), with more physically demanding work associated with greater menstrual pain. Irregular cycles were most strongly associated with lifting, with PR 2.1 (95% CI: 1.3, 3.4) and PR 1.8 (95% CI: 1.1, 2.8) among participants who lifted heavy loads 6–10 times and >10 times per day ( Table 5 ). We also found suggestive associations between irregular cycles and frequent pushing or pulling and bending or twisting, but these estimates were imprecise. Menstrual flow was not associated with any measure of occupational physical demands. Short and long bleed lengths and menstrual cycles were fairly uncommon in our population, which made our estimates for these associations difficult to interpret because they were imprecise ( Supplementary Materials , Table A1 ).

An estimated 85% of U.S. women suffer from painful periods, with over one third of all women unable to perform daily activities, including work, because of their symptoms.[ 8 ] In this population of working flight attendants and teachers, menstrual pain was common, affecting 92% of flight attendants and 96% of teachers. Flight attendants with the greatest occupational physical demands were the most likely to have menstrual cramps, menstrual pain, and irregular cycles. Our results are consistent with reports dating to the 1960s in which flight attendants and flight nurses reported more frequent occurrence of menstrual cramps, menstrual pain, and irregular cycles when working on jet versus propeller aircraft, when on-duty versus off-duty, or when compared to non-flight attendants.[ 2 – 4 , 6 , 7 ] We found that flight attendants were more likely to report always having menstrual cramps than teachers, but due to the small number of teachers in the study the analysis was underpowered for most of our outcomes. Additional differences might exist between the two groups that we were unable to detect in this study. We used data from a study conducted between 1999 and 2001, and the working conditions of flight attendants could have changed in the intervening years. If work became less physically demanding over time ( e.g. , less pushing and pulling of service carts due to reductions in meal service), we would expect to see fewer flight attendants in the highest exposure category. Generally, our results were similar in the middle- and highest-exposure categories ( Tables 4 and 5 ), suggesting that the same associations might hold in lower exposed populations. Some exposures—such as prolonged standing—have likely not changed over time. If the prevalence of menstrual cycle irregularities changed over time, because our results are presented as prevalence ratios (whose maximum value is determined by the baseline prevalence), we might expect the associations to remain in the same direction although the magnitude of the association might be different. Studies of occupational physical demands and menstrual cycle characteristics have been conducted in other worker populations, but differences in exposure definition, outcome definition, and confounder control make it difficult to directly compare our numeric results to those from previous studies. For example, one of the largest studies of occupational physical exertion and menstrual cycle characteristics among nurses defined lifting as the frequency of lifting loads ≥25 pounds (≥11.3 kg) (0, 1–5, 6–15, or >15 times per day) and allowed study participants to self-define irregular cycles.[ 15 ] In contrast, we defined lifting as lifting a load of ≥15 pounds (≥6.8 kg) (0–5, 6–10, or >10 times per day) and irregularity as ≥8 days between the shortest and longest menstrual cycles within the past year. Despite these differences, our results are consistent with evidence from several studies indicating that frequent lifting is associated with greater menstrual pain and irregular menstrual cycles.[ 13 , 15 , 17 ] Few previous studies have examined occupational physical demands aside from lifting. Two studies found no association between standing and irregular cycles; we also found no association.[ 15 , 17 ] Strenuous occupational activity was associated with long cycles in one study, but the estimate was imprecise due to small sample size (risk ratio 1.6, 95% CI: 0.7, 4.2).[ 14 ] We have seen no prior studies of occupational pushing, pulling, or bending to which we can compare our results. We minimized confounding from healthy worker effects by restricting the analysis of physical factors to flight attendants only—who likely differ from teachers on unmeasured sociodemographic factors—in addition to controlling for measured potential confounders. Our analyses comparing flight attendants to teachers, however, might be subject to residual confounding. Although teachers were selected as a comparison group because they were thought to be sociodemographically similar to flight attendants, there are differences between these populations. For example, we did not have information on work schedules, work hours, and other occupational exposures that differ between these two occupational groups and that might be associated with menstrual cycle irregularities. This limitation also likely affected our analysis restricted to flight attendants, in that we were unable to determine if the associations we observed with occupational physical demands were in part explained by other correlated occupational exposures. Using self-reported measures of menstrual characteristics and occupational physical demands could have led to bias from misclassification, although for some of our measures—such as menstrual pain—self-reported outcomes are the best available. Previous studies have found that self-reported cycle length in particular is not always accurately reported, especially for those with the shortest or longest menstrual cycles.[ 18 , 19 ] The higher prevalence of physical demands among flight attendants than teachers in our study is consistent with data from the Occupational Information Network, which shows that compared to teachers, flight attendants expend a higher overall physical effort; spend more time standing, bending, or twisting; and perform more manual material handling (a measure of lifting).[ 20 ] Previous studies have found that participants generally accurately report sitting (i.e., not standing), overall physical activity, and lifting.[ 21 ] There is evidence that workers with musculoskeletal pain report occupational physical demands less accurately than those with no pain; we do not know if this would also be true for participants who have menstrual pain.[ 22 , 23 ] Like most studies of menstrual cycles, we had to exclude participants with hormone or IUD use, as these are commonly used to treat irregular cycles, heavy flow, and painful cramps. We were concerned that this exclusion could cause selection bias (collider stratification bias). For selection bias to occur, there had to be an association between occupational physical demands and hormone or IUD use. We found no association ( Supplementary Materials , Table A2 ), meaning that this exclusion likely did not cause selection bias in our study. We do not know the extent to which our study population is representative of all flight attendants or teachers at risk for menstrual cycle irregularities, with our exclusions related to hormone or IUD use meaning that we did not have information from workers who experienced occupational exposures prior to their initiation of hormone or IUD use. One variable that we did not consider in our analysis was circadian rhythm disruption, either through jet lag (crossing time zones) or working at night. In other occupational groups, circadian disruption through night shift work is associated with irregular cycles and short cycle length.[ 24 , 25 ] Circadian disruption was also hypothesized to be responsible for the more severe menstrual disruptions reported by flight attendants working on jet versus propeller aircraft because flights that cross multiple time zones were more likely to use jet engines.[ 2 ] An analysis of circadian disruption and menstrual cycles among flight attendants could provide additional explanation for the increased risk for menstrual cycle irregularities in this worker population. Similarly, we did not have work hours in the analysis, which has previously been associated with cycle length.[ 15 ] Non-occupational factors such as nutrition and other potential predictors of menstrual cycles were also not considered and could contribute to residual confounding. Occupational physical demands may contribute to the high burden of menstrual cramps, menstrual pain, and irregular cycles among flight attendants. Identifying ways to reduce occupational physical demands and menstrual irregularities among workers could benefit a large proportion of the workforce.

In the 1960s, the first reports of menstrual cycle irregularities in flight crew appeared in the scientific literature. In England, the former Soviet Union, Switzerland, and the United States, flight attendants and flight nurses reported heavy menstrual flow, long bleed lengths, irregular cycles, and severe menstrual cramps after flying.[ 1 – 4 ] In studies over the next 40 years, flight attendants in Mexico, Italy, and China continued to report menstrual cycle irregularities they believed were associated with their work.[ 5 – 7 ] Menstrual cycle irregularities and menstrual symptoms are responsible for considerable pain and disability, loss of productivity, and lower quality of life for the majority of menstruating people.[ 8 ] New research also demonstrates that menstrual cycle irregularities are associated with increased risks for chronic diseases later in life, including cancer and cardiovascular diseases.[ 9 – 11 ] Despite consistent reports of menstrual irregularities among flight attendants over half a century, the sum of the epidemiologic evidence is weak. The articles referenced above represent the studies of menstrual function among flight attendants that have been conducted to date. Among them, few presented formal statistical analyses and only one adjusted for potential confounders such as age.[ 6 ] Specific occupational exposures were rarely investigated, although two studies found that menstrual cycle irregularities were more common among flight attendants working on jet versus propeller aircraft.[ 2 , 3 ] Commercial flights on jet aircraft are typically longer, at higher altitudes, and have more passengers than propeller aircraft, but it is unclear from these studies if menstrual cycle irregularities are associated with properties of the aircraft environment or the increased occupational physical demands that accompany longer flights on jet aircraft. Contemporary employment as a flight attendant involves performing physically demanding job tasks that include lifting passenger baggage (typically over shoulder height), loading and pushing service carts, and spending most of the workday on their feet. Previous studies have found that strenuous occupational physical demands are associated with menstrual cycle irregularities across multiple different occupations, suggesting that the physical demands common to flight attendants are potential determinants of menstrual cycle irregularities. A study among slaughterhouse and cannery workers found that heavy lifting at work was associated with worse menstrual cramps.[ 12 ] Among nurses, heavy lifting was associated with worse menstrual cramps, irregular cycles, and short cycles; heavy physical exertion was associated with long cycles; and prolonged standing was associated with short cycles.[ 13 – 15 ] The goal of this study was to address methodologic limitations of the epidemiologic evidence regarding menstrual cycle irregularities among flight attendants, particularly related to confounder control and investigation of specific occupational exposures. We focused on two questions: if flight attendants have a higher prevalence of menstrual cycle irregularities than teachers, and if occupational physical demands are associated with menstrual cycle irregularities among flight attendants.