MHC-correlated odour preferences in humans and the use of oral contraceptives

Previous studies in animals and humans show that genes in the major histocompatibility complex (MHC) influence individual odours and that females often prefer odour of MHC-dissimilar males, perhaps to increase offspring heterozygosity or reduce inbreeding. Women using oral hormonal contraceptives have been reported to have the opposite preference, raising the possibility that oral contraceptives alter female preference towards MHC similarity, with possible fertility costs. Here we test directly whether contraceptive pill use alters odour preferences using a longitudinal design in which women were tested before and after initiating pill use; a control group of non-users were tested with a comparable interval between test sessions. In contrast to some previous studies, there was no significant difference in ratings between odours of MHC-dissimilar and MHC-similar men among women during the follicular cycle phase. However, single women preferred odours of MHC-similar men, while women in relationships preferred odours of MHC-dissimilar men, a result consistent with studies in other species, suggesting that paired females may seek to improve offspring quality through extra-pair partnerships. Across tests, we found a significant preference shift towards MHC similarity associated with pill use, which was not evident in the control group. If odour plays a role in human mate choice, our results suggest that contraceptive pill use could disrupt disassortative mate preferences.

. Female participant sample size. The control group includes women who neither used, intended to use nor actually used oral contraceptives during the study. The Pill group includes women who intended to use oral contraceptives (for Session 1) and/or actually began using them between Sessions 1 and 2. Numbers in brackets denote women who were not from the UK or from a different ethnic origin, who were excluded from certain analyses.

Control group
Pill group Total Women tested in both sessions 1 60 (8) 37 (4)  97 Women tested in Session 1 only 2

(4) 1 (1) 13
Women tested in Session 2 only 2 0 3 (1) 3 Footnotes 1 A total of 193 women, aged 18-35, were registered as study participants, of whom 97 completed all study requirements (visited the lab for two testing sessions approximately 3 months apart; each session lasted approximately 1h and included some additional tasks such as rating faces). Of these, 60 did not use the pill and 37 began using it during the experiment.
2 In addition to the 97 who completed all aspects of the study, we were also able to include in selected analyses 13 who completed the Session 1 but not Session 2 (of whom 1 expressed intention to begin pill use) and 3 who began using the pill before Session 1 (they were tested only as pill users in the between-subjects comparison of users versus non-users). In summary, total sample sizes were therefore 110 for Session 1 (all non-users), 100 for Session 2 (60 non-users, 40 pill users) and 97 for the within-subjects comparisons (60 in control group, 37 in pill group). Twelve women were from outside the UK, or from a different ethnic origin, and were thus excluded in some analyses to control for potential differences in allele frequencies between populations (leaving 52 in the control group and 33 in the pill group). 5 of the 13 who completed only Session 1 were from outside the UK (including the one who expressed intention to begin pill use), as was 1 of the 3 women who were tested only as pill users.   Table S4. Mean scores given to individual male odour donors when odours were presented to MHC-similar or MHC-dissimilar women. All women were normally cycling (data are from Session 1). The upper set of scores relates to standardized scores to control for differential use of the rating scale by female raters, while the lower set shows raw scores. Within each set, the upper 3 rows relate to all ratings, the lower 3 is restricted to ratings from the core sample of women (from UK participants and odours that were not confounded by deodorants or tobacco smoke). N=79 and 52 (the numbers of men who were scored in both the similar and dissimilar conditions).   Figure S1. Repeatability of scoring for odour pleasantness Histograms are frequency distributions of Spearman rank correlation coefficients calculated for scores awarded to six shirts by individual raters, either within test sessions (inter-session interval: approximately 1h) or between test sessions (inter-session interval: approximately 3 months). Scores are repeatable for each comparison except for pill users between sessions.

Figure S2. Repeatability of scoring for odour intensity and partnership desirability.
Histograms are frequency distributions of Spearman rank correlation coefficients calculated for scores awarded to six shirts by individual raters, either within test sessions (inter-session interval: approximately 1h) or between sessions (inter-session interval: approximately 3 months). Left-hand column= non-users; right-hand column= pill-users. Statistics are onetailed t-tests against chance value of zero.