Background:  The effects of HIV and HAART on menstrual irregularity and bone mineral density (BMD) have not been defined.

Methods:  Women >35 years with or at risk for HIV had standardized interviews on menstrual status, demographic characteristics, drug use in past 5 years, and medical history. Weight, HIV antibody, CD4⁺ count, HIV viral load, thyroid stimulating hormone (TSH), prolactin, and follicle stimulating hormone (FSH) (on cycle day 1 to 6 if menstruating) were measured. BMD was assessed by GE Lunar Prodigy densitometer. We assessed the associations of HIV and HAART with:  menstrual irregularity by logistic regression, controlling for age, race, weight, and opiate use; and BMD of the spine using linear regression controlling on the above factors and steroids, smoking, alcohol use, and socioeconomic status.

Results:  Among 493 women, 53% were HIV⁺. Mean age was 45±5 years; mean weight 78±19 kg; 52% were black, 38% Hispanic, and 8% white. Hormone levels were available for 360 of the 493 (73%) women. Among 135 women with menstrual irregularity (oligo- or amenorrhea), 34% were postmenopausal (FSH >31 U/L), 26% had abnormal TSH (<0.63 or >4.19µU/mL), and 24% had an elevated prolactin (>14.6 ng/mL); 50% of women with amenorrhea >1 year were postmenopausal. The table shows that mean BMD decreased with menstrual irregularity. Overall, on multivariate analysis HIV and opiate use were associated with menstrual irregularity (p = 0.009 and p = 0.002, respectively) and with lower BMD (p = 0.007 and p = 0.04, respectively). HIV was independently associated with menstrual irregularities (p = 0.04) and lower BMD (p = 0.02) in premenopausal women; and lower BMD (p = 0.01) in postmenopausal women. Premenopausal menstrual irregularities were associated with lower BMD independent of HIV status (p = 0.02). In HIV⁺ women, current HAART use, not CD4⁺ count or viral load, was independently associated with higher BMD (p = 0.014)

*p <0.05

Conclusions:  These findings suggest that the detrimental effect of HIV on BMD is mediated at least in part by menstrual irregularity, and is attenuated by HAART. Longitudinal studies are needed to further explore these associations.