- T Cundy,
- J Cornish,
- M C Evans,
- H Roberts,
- I R Reid
- Department of Medicine, University of Auckland, Auckland 1
- New Zealand Family Planning Association of New Zealand, Auckland
- Correspondence to:Dr Cundy.
Depot medroxyprogesterone acetate works as a contraceptive by inhibiting the secretion of pituitary gonadotrophin and thus suppressing ovulation. Women who use it are partially oestrogen deficient, and we have previously shown that long term users have reduced bone density.1 In this study we investigated whether such bone loss is reversible.
Patients, methods, and results
We recruited 54 women (50 of European origin, four of Maori origin. Group 1 comprised 14 women who had used depot medroxyprogesterone acetate for at least three years; they were studied while taking the drug and after having stopped it. Group 2 comprised 22 long term users of the drug, who were studied while they were taking it. The control group, group 3, comprised 18 women who had never used the drug. The age range in the groups was similar, but a higher proportion of women who had ever used the drug were cigarette smokers. While using the drug all the women had amenorrhoea. Of the women in group 1, 12 resumed menstruation within 2-24 (median 8) months, and two remained amenorrhoeic. Their plasma concentrations of gonadotrophins and oestradiol indicated that oestrogen production had been restored; thus all had become oestrogen sufficient.
Bone density in the second to fourth lumbar vertebrae and the femoral neck was measured twice in each woman, at an interval of 9-20 (median 12) months, by dual energy x ray absorptiometry; body weight was also measured. Sequential measurements within groups were compared with the paired t test and comparisons between groups with the unpaired t test.
When first measured bone density in the lumbar spine was on average 9.0% lower in groups 1 and 2 than the controls (P< 0.02) and bone density in the femoral neck was on average 5.7% lower (NS). At the second measurement, bone density in the lumbar spine had not changed in groups 2 and 3 but had significantly increased in group 1 (P < 0.001) (table); bone density in the femoral neck had not changed in any group. The mean increase in bone density in the lumbar spine in group 1 was 3.4% a year (95% confidence interval 1.6% to 5.2%); this occurred despite a significant fall in body weight (mean 2.2 kg, P < 0.01). Additional measurements of bone density in eight women from group 1 were made two years after they had stopped taking depot medroxyprogesterone acetate; bone density in the lumbar spine increased significantly between months 12 and 24 (P < 0.002), the mean increase overall being 3.0% (0.4% to 5.6%) at 12 months and 6.4% (3.4% to 9.4%) at 24 months.
Spinal bone density increased when long term users of depot medroxyprogesterone acetate became oestrogen sufficient, suggesting that bone loss related to use of the drug results from oestrogen deficiency. The increase occurred despite a fall in body weight (which usually favours bone loss) and a delay in the resumption of regular menstruation. In the women who were followed up for two years spinal bone density increased further in the second year. In our earlier study women using depot medroxyprogesterone acetate had an average 7.5% deficit in bone density in the lumbar spine1. The mean gain in bone density of 6.4% two years after the women in this study stopped taking depot medroxyprogesterone acetate suggests that the bone loss may be almost completely reversible even after long term use of the drug.
The changes in bone density in the femoral neck were less striking. Although we had previously found the deficit in bone density in the women using depot medroxyprogesterone acetate to be similar in the femoral neck and spine,1 studies of young women with oestrogen deficiency have generally shown the spine to be more severely affected.*RF 2-4* In this study bone density in the femoral neck was at least maintained in women who stopped taking depot medroxyprogesterone acetate, whereas small losses occurred in the other groups.
This study was supported by the Health Research Council of New Zealand.