Abstract
BACKGROUND: Shift work disrupts the circadian rhythm and may cause menstruation disorders. This study assessed the impact of shift work on menstrual cycle in a population of Chinese nurses. METHODS: Questionnaires on menstrual characteristics and shift schedules were sent to female nurses of the First Affiliated Hospital of Sun Yat-sen University (FAHSYSU) and Guanghua Hospital of Stomatology (GHHS), affiliated to Sun Yat-sen University. Part I was a cross-sectional study and included 139 nurses in GHHS who had regular 8:00-17:30 working (non-shift group), and 334 nurses from FAHSYSU who worked shifts, a response rate of 67.5 % and 59.6 %, respectively (age ≤ 50 years). Menstrual patterns were compared and age-adjusted relative risks of shift work were analyzed. Part II was a nested case-control study. Cases were nurses in Part I who had regular cycle with mean cycle length (MCL) of 25-31 days and but at least 3 days variation in MCL after starting shift work (n = 45). Controls consisted of 67 nurses with matching shift patterns and age, but no MCL changes. A control non-shift age-matched group consisted of 30 GHHS nurses with no MCL changes. A follow-up second questionnaire was sent 2 years later. RESULTS: In Part I, the shift group had a significantly higher proportion of nurses with menstrual cycle irregularity. The proportion of nurses with a cycle of 25-31 days decreased from 81.7 to 67.8 % after changing to shift work. Logistic regression analysis showed that night shift frequency was the only risk factor associated with cycle shortening. After adjusting for age, MCL was shorter when night work was performed > 7 times per month. In Part II, the mean change in MCL in the case group, including prolongation or shortening, was 4.115 ± 2.084 days after shift working. In the 2 years' follow-up, the MCL of the study group did not recover to the original length. CONCLUSIONS: Rotating shift work can increase the prevalence of menstrual cycle irregularity. Night shift frequency was the only risk factor associated with cycle reduced. Changes in MCL did not show recovery over a follow-up period of 2 years.