Do users of the intrauterine system (Mirena) have different genital symptoms and vaginal flora than users of the intrauterine contraceptive device?

INTRODUCTION

Bacterial vaginosis (BV) is the commonest cause of abnormal vaginal discharge in women, with a reported prevalence of 1.8–25%, depending on the population studied. ^1–3 Up to 50% of women with BV may be asymptomatic. ³ The copper intrauterine contraceptive device (IUCD) is strongly associated with abnormal vaginal discharge ¹ and BV, with a prevalence of up to 50% reported in users. ^1–3 Hormonal influences may play a role in the control of vaginal flora, but studies are conflicting. Abnormal vaginal flora has been shown to develop most frequently at the beginning of the menstrual cycle when oestrogen levels are low. ⁴ Similarly, a significant reduction in BV has been demonstrated in association with increased oestradiol levels in response to fertility treatment. ⁵ Conversely, Miller et al. ⁶ have shown that the hypo-oestrogenic state induced by use of intramuscular depomeproxyprogesterone (DMPA) appears to have no effect on the prevalence of BV, and Baeten et al. ⁷ demonstrated that DMPA users were at significantly decreased risk of developing BV. ⁷ However, it has been postulated that the mechanical effect of blood in the vagina may predispose to the development of BV. Therefore, the lack of menses associated with the use of DMPA may contribute to the apparent low prevalence of BV in its users. This is the first study to determine whether the local hormonal and mechanical effects of the progesterone-incorporated intrauterine system (IUS; Mirena) alter the prevalence of BV in its users as compared with users of the IUCD. We aimed to assess whether users of the IUS have different genital symptoms and vaginal flora from users of the copper IUCD.

METHODS

One hundred and seventy-two women requesting insertion of either an IUCD or an IUS were recruited from routine contraception clinics between August 2000 and April 2003. The women were assessed prior to insertion (visit 1) and again after 4–6 weeks (visit 2) and six months (visit 3). At each visit, the women were asked about symptoms and a high vaginal smear was obtained for Gram staining and microscopy. The vaginal flora was graded using the Ison-Hay criteria. ⁸ Microscopy of the sample from each visit was performed by one of three clinicians, who were blinded as to whether a particular smear had been obtained from a woman fitted with an IUCD or an IUS. Every tenth sample was graded by a separate clinician in order to validate the results, with a high concordance rate (40/45 [89%]). Where there was uncertainty or disagreement about a grading, the smear was reviewed by all three clinicians until a consensus was reached. The study received approval from the Local Ethics Committee. P values were calculated using either chi squared or Student's ‘t’-test as appropriate.

RESULTS

Denominators vary due to missing data.

Seventy-eight women had an IUCD and 94 had an IUS inserted. The mean time between visit 1 and visit 2 was 48 days. The mean time between visit 1 and visit 3 was 226 days. The mean age of women who had an IUCD inserted was 33 compared with 36 in those with an IUS (P = 0.02). The characteristics of women after insertion of the IUCD or the IUS at visit 2 and 3 are summarized in Table 1.

The mean interval between last menstrual period (LMP) and visit 2 was 18 days in women with an IUCD and 30 days in those with an IUS (P = 0.13). At visit 3, the mean interval from LMP was 23 days in women with an IUCD and 79 days in those with an IUS (P < 0.001).

At visit 2, 17/64 (27%) IUCD users reported that they had developed an abnormal discharge since visit 1, compared with 12/88 (14%) IUS users (P = 0.04, odds ratio [OR] 2.29 [95% confidence interval 1.01–5.22]). At visit 3, 11/57 (19%) IUCD users reported developing a discharge compared with 7/84 (8%) IUS users (P = 0.06, OR 2.63 [95% confidence interval 0.95–7.26]). Five women had reported an abnormal discharge at visit 1 and were excluded from the analysis.

At visit 1, 11/164 (7%) women were found to have BV and another 13/164 (8%) were found to have non-BV abnormal vaginal flora (i.e. grade 2, 4 or un-gradable). At visit 2, 2/51 (3.9%) IUCD users who had normal vaginal flora at visit 1 had developed BV compared with 1/71 (1.4%) IUS users (P = 0.38, OR 2.86 [95% confidence interval 0.25–32.39]). At visit 3, 3/41 (7.3%) IUCD users had developed BV compared with 2/71 (2.8%) IUS users (P = 0.30, OR 2.73 [95% confidence interval 0.44–17.02]). One woman with an IUCD developed intermediate vaginal flora (grade 2) at visit 2. Similar results to all the above analyses were produced if women with either an IUCD or IUS prior to visit 1 (n = 30) were excluded.

Twenty-two of 25 (88%) women who developed a symptomatic discharge at visit 2 and 13/17 (76%) at visit 3 were found to have normal vaginal flora.

DISCUSSION

Women were significantly more likely to have developed an abnormal vaginal discharge 4–6 weeks after insertion of an IUCD compared with an IUS, although this trend was not significant at the six-month follow-up visit. Most of the women who developed a symptomatic discharge were found to have normal vaginal flora. Previous research has demonstrated an association between use of the IUCD and cervical inflammation and ectropion, ¹ either of which could potentially lead to abnormal discharge. The appearance of the cervix was not documented, nor were vaginal samples for Trichomonas vaginalis and fungal culture obtained as part of this study.

There was a trend for more women who had an IUCD rather than an IUS inserted to develop BV at both follow-up visits. However, there were insufficient numbers of women with BV to elucidate whether the difference was statistically significant. The proportion of IUCD users who developed BV was low compared with the previously reported prevalence. ^1–3 Women with abnormal flora at visit 1 were excluded from this analysis in order to elucidate the effect of the device inserted. This may have excluded women with a predisposition to BV and lead to an apparent low prevalence in the IUCD group of this study.

There was an increased interval since LMP in women with an IUS compared with users of the IUCD at visit 2, which was statistically significant at visit 3. This oligomenorrhoea is likely to have been a consequence of the local hormonal effect of the IUS. It has been postulated that the presence of menses in the vagina predisposes to the development of BV, which has been shown to be more prevalent immediately following menstruation. ⁴ This may explain the previously described normal or low prevalence of BV in users of DMPA despite their hypo-oestrogenic state. ^6,7 The lower prevalence of BV found in the IUS group in this study may also have been due to the mechanical influence of oligomenorrhoea rather than direct hormonal affect.

Despite a substantial number of women being recruited in this study we were unable to demonstrate a significance difference in the development of abnormal vaginal flora in users of the IUS and IUCD. Further studies, involving larger numbers of women, are required to further investigate this area.