A Simple Algorithm is Created for Identifying Intermenstrual Intervals Containing an Oscillatory LH Pattern That Associates With Vasomotor Symptoms Using Daily Urinary LH Excretion in the Study of Women’s Health Across the Nation (SWAN)

Background: A specific and unique pattern of luteinizing hormone (LH) excretion has been associated with vasomotor symptoms (VMS) in early menopausal women. Described as “oscillations” of LH excretion, this pattern is consistent with secretory “surges” of LH followed by pituitary “fatigue”. This pattern has not been observed in non-VMS intermenstrual intervals and supports the concept that a breakdown in the hypothalamic-pituitary ovarian axis feed-back loops leads to extreme and cyclic variations in gonadotropin hormone releasing hormone (GnRH) secretion that stimulates collateral nerves to alter core body temperature. Regardless of the precise mechanism, the pattern of LH secretion, as transduced in daily urine as oscillations, provides the basis for the development and validation of a VMS algorithm. Objective: The purpose of this study was to create a simple algorithm to identify intermenstrual intervals exhibiting oscillatory LH, to facilitate investigations into its associations with VMS and other symptoms during the menopausal transition (MT). Methods: As part of the Study of Women’s Health Across the Nation (SWAN), participants in the Daily Hormone Substudy (DHS) were asked to provide daily urine samples - from which LH, E1c, and PdG were measured - and complete a daily symptoms diary for one menstrual cycle (up to 50 days). Analyses included 144 participants whose first DHS collection did not meet the Kassam criterion for evidence of luteal activity; of these, 61 were assessed by an expert as having oscillatory LH and 83 as non-oscillatory LH. Proposed algorithm-based classifications regarding oscillatory LH included number of days with LH at least 50% of the collection maximum LH (number of large-LH days) and number of days with LH no more than twice the collection minimum LH (number of small-LH days). Agreement of these 2 criteria with rater-assigned oscillatory LH was assessed using nonparametric t-tests and binomial logistic regression. Associations of these with VMS frequency were assessed using Spearman correlations. Results: The number of large-LH days was strongly associated with oscillatory LH: median (interquartile range) = 13 (7,22) for oscillatory collections versus 4 (2, 11) for non-oscillatory collections (p<.0001) but number of small-LH days was unrelated (p=.98). Percentage of collection days with VMS was significantly correlated with number of large-LH days (Spearman r=.37, p<.0001) but not with number of small-LH days (Spearman r=.03, p>.05); adjustment for total collection length had negligible impact. Conclusion: A simple algorithm using urinary LH profiles can be used to identify intermenstrual collections that likely contain intervals of VMS.