Abstract
Objective
Low serum vitamin D levels have been associated with increased prevalence of the reproductive tract condition bacterial vaginosis (BV). The objective of this trial was to evaluate the effect of high-dose vitamin D supplementation on BV recurrence.
Study design
This randomized, placebo-controlled, double-blinded trial enrolled 118 women with symptomatic BV from an urban STD clinic (clinicaltrials.gov registration NCT01450462). All participants received 500mg oral metronidazole twice daily for seven days. Intervention participants (n=59) also received nine doses of 50,000 international units of cholecalciferol (vitamin D3) over 24 weeks; control women (n=59) received matching placebo. Recurrent BV was assessed via Nugent scoring after 4, 12 and 24 weeks. We assessed the effect of the intervention using an intention-to-treat approach, fitting Cox proportional hazards models to evaluate recurrent BV over the follow-up period.
Results
Most participants (74%) were black, with a median age of 26 years. Median presupplementation serum 25-hydroxyvitamin D [25(OH)D] was similar across randomization arms: 16.6 ng/mL in the vitamin D arm and 15.8 ng/mL in the control arm. At trial completion, median 25(OH)D among women receiving vitamin D was 30.5 ng/mL, vs 17.8 ng/mL in control women; 16% of women receiving vitamin D and 57% receiving placebo remained vitamin D deficient (<20 ng/mL). BV prevalence among women randomized to vitamin D was very similar to those randomized to placebo at the 4- and 12-week visits, but by the 24-week visit, BV prevalence was 65% among women in the vitamin D arm and 48% among control women. BV recurrence was not reduced by vitamin D supplementation (intention-to-treat hazard ratio, 1.11; 95% confidence interval, 0.68-1.81). Among women experiencing recurrent BV, median time to recurrence was 13.7 weeks in the vitamin D arm and 14.3 weeks in the control arm.
Conclusions
Women receiving vitamin D experienced significant increases in serum 25(OH)D, but this increase was not associated with decreased BV recurrence in this high-risk STD clinic population.
Alphabetized key words: Bacterial vaginosis, randomized controlled trial, STD clinic, vitamin D
1. INTRODUCTION
Bacterial vaginosis (BV), the most common vaginal infection worldwide among reproductive-age women,1 is associated with increased acquisition and transmission of HIV2–3 and numerous obstetric complications.4–5 BV is highly persistent and recurrent after treatment.6 However, the etiology of BV is not clear. The most consistent risk factor for BV in US populations is non-white race, an association that persists after adjustment for all known confounding factors.7 According to the US National Health and Nutrition Examination Survey (NHANES), 52% of black women have BV compared with 23% of white women.8
Vitamin D is essential to immune function, both stimulating mechanisms associated with pathogen elimination and regulating immune response.9 Vitamin D levels, quantified clinically as serum 25-hydroxyvitamin D (25(OH)D, also vary by race: 90% of US blacks, Hispanics, and Asians have vitamin D levels below the threshold of sufficiency set by the US Institute of Medicine (IOM) (30 ng/mL), compared with 75% of non-Hispanic whites.10–12 Thus, we hypothesized that the link between race and BV could be a conflation of two other associations: a) race and vitamin D, and b) vitamin D and BV.
Several cross-sectional studies report that vitamin D insufficiency is associated with BV in pregnant women.13–16 In contrast, two vitamin D trials in pregnant women examined BV as a secondary endpoint;17–18 neither found an effect of daily vitamin D supplementation on BV.19 Two studies examined the correlation between vitamin D and BV in non-pregnant women. One reported no association.16 In the other, vitamin D deficiency was associated with BV in HIV-positive women but not HIV-negative women.20 A case-crossover study of non-pregnant women used season as a proxy for vitamin D level, and found no evidence of increased BV prevalence during seasons when vitamin D levels are expected to be lowest.21
We conducted a randomized controlled trial of high-dose vitamin D supplementation in non-pregnant, BV-positive women recruited from a sexually transmitted disease (STD) clinic. We aimed to quantify the effect of high-dose vitamin D supplementation on BV recurrence.
2. MATERIALS AND METHODS
2.1 Study design and setting
This randomized, placebo-controlled, double-blinded, single-site trial was conducted in a public STD clinic. Trial design and results are reported according to CONSORT guidelines.22–23 Participants attended four visits: enrollment and after 4, 12, and 24 weeks (Figure 1). The study was approved by the Ohio State University Institutional Review Board on May 25, 2011 (protocol 2011H0089) and was registered at clinicaltrials.gov (NCT01450462).
Figure 1.
Flow diagram of the screening, enrollment, follow-up, and analysis of study participants.
2.2 Recruitment, eligibility and informed consent
All female patients received study information at registration. Women who expressed interest were screened for self-reported eligibility criteria. Eligible women provided written consent.
Eligible women were 18–50 years old; pre-menopausal; had at least one ovary; spoke English; and were BV-positive on clinical exam by modified Amsel criteria (vaginal pH >4.5; thin, homogenous discharge; and positive “whiff” test).24 Excluded women were pregnant (by report or rapid urine test) or planning pregnancy within six months; breastfeeding; currently menstruating; or self-reporting kidney disease or kidney stones, hypercalcemia, hypercalciuria, sarcoidosis, histoplasmosis, thyroid disease, lymphoma, or tuberculosis. Concomitant infection with HIV or other STDs was not an exclusion criterion.
2.3 Treatment
All participants received standard BV therapy at no cost: 500 mg metronidazole orally twice daily for 7 days. Women also received no-cost treatment for other infections diagnosed at enrollment according to normal clinic procedures.
2.4 Randomization
Using a permuted block design developed using SAS (Version 9.2, Cary, NC), participants were randomized into four equal-sized treatment arms labeled by color. Two arms received vitamin D and two received placebo. The code linking which color corresponded to which product was not known to participants, staff, investigators or data analysts. Randomization assignments were placed in individual opaque envelopes, each labeled by participant identification number (PIN). Sealed envelopes were stored in PIN order in a locked cabinet. At enrollment, each participant, together with staff, opened the envelope corresponding to the next available PIN to reveal her color group assignment.
2.5 Intervention
Women received their assigned study product in pre-filled pill boxes. Women in the vitamin D groups received nine capsules, each containing 50,000 international units (IU) of vitamin D3 (cholecalciferol) (BioTech Pharmacal, Fayetteville, AR) and women in the control groups received nine placebo capsules with identical appearance to the vitamin D. Women were instructed to take the capsules 1, 2, 3, 4, 8, 12, 16, 20 and 24 weeks after enrollment. Participants received a pill schedule at enrollment, a duplicate schedule via post one week later, and text message reminders each day a capsule was to be taken. The target date for each capsule was also written in permanent marker directly on pill compartments. Women who reported at enrollment that they regularly took vitamin D supplements were requested to stop for the duration of the trial.
2.6 Specimen collection, processing and testing
At the enrollment and 24-week visits, women underwent comprehensive clinical examinations with collection of blood and vaginal and cervical swabs for STD and reproductive tract infection testing. Chlamydia and gonorrhea were diagnosed by nucleic acid amplification testing; syphilis through rapid plasma reagin testing and confirmed by Treponema pallidum particle agglutination assay; HIV by rapid testing on plasma; trichomoniasis by microscopy and culture; and yeast by microscopy. Total serum calcium was assessed within 7 days of enrollment, and women with levels above the normal range – for whom vitamin D supplementation could be unsafe – were unenrolled. At the 4- and 12-week visits, women self-collected vaginal swabs which were used to create slides of vaginal material. Slides from all four visits were stored and underwent Gram staining and Nugent scoring for BV detection after trial conclusion; technicians scoring the slides were blinded to treatment assignment. Stored serum from each visit was used to quantify 25(OH)D at the end of the trial, using the DiaSorin LIAISON 25 OH Vitamin D TOTAL Assay. Pregnancy was assessed via urine testing at each visit. Women found to be pregnant remained in the study but had their assigned study product discontinued immediately.
2.7 Questionnaire
At enrollment and each follow-up visit, women underwent face-to-face interviews to collect data on demographics, sexual behavior, adherence, and other information. Women were asked about 19 individual side effects previously associated with vitamin D supplementation or metronidazole treatment. Endorsement of any side effect was recorded as an adverse event.
2.8 Safety
Participants’ serum calcium levels were checked at enrollment, 4 weeks and 24 weeks. Safety was further monitored by an independent Safety Committee, which reviewed adverse events by blinded study arm after half the anticipated person-time for the trial had been accrued.
2.9 Statistical analysis
All analyses were conducted using SAS. We first examined changes in serum 25(OH)D levels by randomization group over time. Next, we examined BV prevalence at each visit, by randomization group. Nugent score of 7–10 on Gram-stained vaginal smear was interpreted as BV.25 We estimated BV-free survival time using the Kaplan-Meier method.26 We tested the homogeneity of survival functions across randomization groups using the log-rank test. Our primary intention-to-treat analysis used Cox proportional hazards models to estimate the hazard ratio (HR) and 95% CIs for the effect of randomization assignment on time to BV recurrence. Our secondary analysis used extended Cox proportional hazard models, and modeled the effect of time-varying 25(OH)D level (rather than randomization assignment) on time to BV recurrence; that analysis also controlled for age and race. In all Cox models, women who missed a visit but returned for later visits were assumed to have remained BV-negative during the missing visit.
We performed three exploratory sensitivity analyses. First, using a threshold of α=0.10 to classify imbalance in participant characteristics at enrollment, we ran an adjusted Cox model controlling for those factors which, due to chance, were not evenly balanced by randomization. Second, we limited the analysis to women who had 25(OH)D levels <20 ng/mL (the IOM threshold for vitamin D deficiency) at enrollment, to determine whether the effect of vitamin D supplementation on BV recurrence was different for this subgroup. Third, we assessed the impact of using Amsel criteria to diagnosis BV at enrollment, by restricting the analysis population to women also determined to be BV-positive at enrollment by Nugent scoring.
2.10 Power
We anticipated 20% loss-to-follow-up and BV recurrence in the placebo arm of 50% after 24 weeks.6,27 Under these assumptions, with α=0.05, we estimated that 120 women (60 per group) would yield 80% power to detect an absolute difference in BV recurrence of 24%, comparing the vitamin D and placebo arms.
3. RESULTS AND DISCUSSION
3.1 Screening
Between September 2011 and January 2013, 285 women underwent screening and 85 were ineligible for at least one reason (Figure 1). Of 200 women eligible by self-reported criteria, 128 were BV-positive by modified Amsel criteria and 126 enrolled. Following randomization, one woman began acting erratically and was immediately unenrolled. Seven women (6%) had abnormal serum calcium levels and were unenrolled. The final sample included 59 women randomized to vitamin D and 59 randomized to placebo. Ninety-seven participants (82%) returned after enrollment: 51 (86%) in the vitamin D arm and 46 (78%) in the placebo arm returned after 4 weeks, and 38 (64%) women in the vitamin D arm and 40 (68%) control women returned after 24 weeks (Figure 1). Among returning participants, median follow-up was 24.0 weeks in the vitamin D arm and 24.8 weeks in the control arm.
3.2 Participant characteristics at enrollment
Participants were primarily African-American (n=87, 74%), with a median age of 26 years (Table 1). Women’s median lifetime number of male sex partners was 10 (IQR: 5–20 partners). A large minority (n=48, 41%) reported lifetime experience with anal sex, and 32 (27%) reported sex with women. All women were BV-positive by modified Amsel criteria, and when vaginal smears from enrollment were Nugent scored at the end of the trial, 80% of the sample was found to be BV-positive at enrollment by Nugent score (Table 1).
Table 1.
Demographic, behavioral and clinical characteristics of participants at enrollment, overall and by randomization group.
| Vitamin D | Control | Total | p-value | ||||
|---|---|---|---|---|---|---|---|
|
| |||||||
| N=59 | (%) | N=59 | (%) | N=118 | (%) | ||
| Race/ethnicity a | |||||||
| Black | 44 | (75) | 43 | (73) | 87 | (74) | 0.84 |
| White | 15 | (25) | 16 | (27) | 31 | (26) | 0.84 |
| American-Indian/Alaskan Native | 3 | (5) | 3 | (5) | 6 | (5) | 1.00 |
| Asian/Pacific-Islander | 0 | (0) | 1 | (2) | 1 | (1) | 0.32 |
| Hispanic | 3 | (5) | 4 | (7) | 7 | (6) | 0.70 |
| Education | |||||||
| Finished high school | 52 | (88) | 49 | (83) | 101 | (86) | 0.43 |
| Finished college | 8 | (14) | 6 | (10) | 14 | (12) | 0.57 |
| Main partner is … b | |||||||
| Man | 34 | (58) | 31 | (53) | 65 | (55) | 0.85 |
| Woman | 2 | (3) | 2 | (3) | 4 | (3) | |
| No main partner | 23 | (39) | 26 | (44) | 49 | (42) | |
| Season of enrollment | |||||||
| Fall (September–November) | 19 | (32) | 21 | (36) | 40 | (34) | 0.85 |
| Winter (December–February) | 16 | (27) | 14 | (24) | 30 | (25) | |
| Spring (March–May) | 19 | (32) | 19 | (32) | 38 | (32) | |
| Summer (June–August) | 5 | (50) | 5 | (50) | 10 | (8) | |
| Cohabitate with main partner c | 10 | (28) | 11 | (33) | 21 | (30) | 0.75 |
| Employed full- or part-time | 40 | (68) | 31 | (53) | 71 | (60) | 0.09 |
| Food insecure in the last year d | 19 | (32) | 19 | (32) | 38 | (32) | 1.00 |
| Housing insecure in the last year e | 25 | (42) | 23 | (39) | 48 | (41) | 0.71 |
| Self-rated health | |||||||
| Excellent | 7 | (12) | 7 | (12) | 14 | (12) | 0.70 |
| Very good | 17 | (29) | 19 | (32) | 36 | (31) | |
| Good | 26 | (44) | 20 | (34) | 46 | (39) | |
| Fair | 9 | (15) | 11 | (19) | 20 | (17) | |
| Poor | 0 | (0) | 2 | (3) | 2 | (2) | |
| Depression screen positive f | 25 | (42) | 28 | (47) | 53 | (45) | 0.72 |
| Ever pregnant | 39 | (66) | 40 | (68) | 79 | (67) | 0.84 |
| Current user of vitamin D supplements | 11 | (19) | 6 | (10) | 18 | (15) | 0.20 |
| Currently using any method of contraception g | 23 | (39) | 26 | (44) | 49 | (42) | 0.58 |
| Hormonal methods (oral contraceptive pills, implants, injectable, patch, ring, hormonal IUD) | 6 | (26) | 7 | (27) | 13 | (27) | 0.95 |
| Male condoms | 14 | (61) | 13 | (50) | 27 | (55) | 0.44 |
| Female condoms | 1 | (4) | 1 | (4) | 2 | (4) | 1.00 |
| IUD (copper or unknown type) | 1 | (4) | 1 | (4) | 2 | (4) | 1.00 |
| Sterilization | 3 | (13) | 9 | (35) | 12 | (24) | 0.10 |
| Other (diaphragms, spermicides, douching, withdrawal, rhythm, abstinence, morning after pills) | 1 | (4) | 3 | (12) | 4 | (8) | 0.61 |
| Condom use, last 3 months | |||||||
| 0% of vaginal acts with men | 24 | (41) | 16 | (27) | 40 | (34) | 0.32 |
| >0% and <100% of vaginal acts with men | 24 | (41) | 32 | (54) | 56 | (47) | |
| 100% of vaginal acts with men | 7 | (12) | 6 | (10) | 13 | (11) | |
| Ever anal sex | 23 | (39) | 25 | (42) | 48 | (41) | 0.55 |
| Ever sex with women | 18 | (31) | 14 | (24) | 32 | (27) | 0.41 |
| Concurrent sexual partnerships in the last 3 months h | 11 | (19) | 20 | (34) | 31 | (26) | 0.03 |
| Feminine hygiene products before, during or after sex in last 3 months | 24 | (41) | 20 | (34) | 44 | (37) | 0.58 |
| Ever douched | 50 | (85) | 41 | (69) | 91 | (77) | 0.09 |
| Ever past BV diagnosis (self-reported, prior to enrollment) | 48 | (81) | 43 | (73) | 91 | (77) | 0.23 |
| Nugent score | |||||||
| Normal flora (Nugent 0–3) | 3 | (5) | 3 | (5) | 6 | (5) | 0.58 |
| Intermediate (Nugent 4–6) | 7 | (12) | 10 | (17) | 17 | (14) | |
| BV (Nugent 7–10) | 49 | (83) | 45 | (76) | 94 | (80) | |
| Insufficient specimen (unscorable) | 0 | (0) | 1 | (2) | 1 | (1) | |
| Infections | |||||||
| Chlamydia | 5 | (8) | 6 | (10) | 11 | (9) | 0.75 |
| Gonorrhea | 3 | (5) | 6 | (10) | 9 | (8) | 0.30 |
| Trichomoniasis | 6 | (10) | 14 | (24) | 20 | (17) | 0.05 |
| Syphilis | 0 | (0) | 0 | (0) | 0 | (0) | 1.00 |
| Yeast | 4 | (7) | 4 | (7) | 8 | (7) | 1.00 |
| Prevalent HIV (not new diagnoses) | 2 | (3) | 0 | (0) | 2 | (2) | |
| Median | (IQR) | Median | (IQR) | Median | (IQR) | p-value | |
| Age (years) | 28 | (22–36) | 25 | (22–32) | 26 | (22–33) | 0.45 |
| Age of main partner (years) c | 35 | (25.5–42) | 28 | (23–31) | 30 | (22–38) | 0.01 |
| Body mass index (BMI) | 28.0 | (24.6–32.1) | 25.8 | (22.9–32.0) | 27.3 | (23.3–32.1) | 0.21 |
| Gravidity | 1 | (0–3) | 1 | (0–3) | 1 | (0–3) | 0.72 |
| Age at first sex | 16 | (14–17) | 15 | (14–17) | 16 | (14–17) | 0.46 |
| Lifetime number male partners | 11 | (6–20) | 8 | (5–15) | 10 | (5–20) | 0.10 |
| Number male partners, last 3 months | 1 | (1–2) | 1 | (1–2) | 1 | (1–2) | 0.97 |
| Number vaginal sex acts with men, last 3 months | 15 | (5–36) | 10 | (3–24) | 11 | (4–30) | 0.11 |
| Lifetime number female partners i | 3 | (2–3) | 2 | (2–8) | 3 | (2–6) | 0.65 |
Some women reported more than one race, so totals sum to greater than 100%. Race and ethnicity were queried together: “What race and ethnicity do you consider yourself? If you consider yourself to be in more than one group, please tell me all the groups that you are part of.” Response options, which were read aloud to the participant, were: Black/African American; White; American Indian/Alaskan Native; Asian/Pacific Islander; Hispanic or Latino(a); Other race/ethnicity (with response recorded by interviewer).
Women were considered to have a main partner if they answered ‘yes’ to the question, “Do you have a main partner right now? By “main partner,” I mean a spouse or other committed partner whom you are romantically involved with.”
Among those reporting a main partner.
Women were considered food insecure if they answered ‘yes’ to the question, “In the last year, have you ever been concerned about having enough food for yourself or your family?”
Women were considered housing insecure if they answered ‘yes’ to the question, “In the last year, have you ever been concerned about having a place to live for yourself or your family?”
The number and proportion answering yes to either or both questions on the 2-question Patient Health Questionnaire (PHQ-2).
The denominator for the proportions of women using each contraceptive method type is the total number of women reporting use of any method. Because women could report multiple methods, proportions sum to greater than 100%.
Women were considered to have concurrent sexual partnerships if they answered ‘yes’ to the question, “In the past three months, did you have sex with one partner while involved in a sexual relationship with another partner during the same period of time? This may include times when you did not consider yourself to be in a committed relationship.”
Among women who report any lifetime sex with women.
Ninety-one women (77%) reported a past BV diagnosis before enrollment. Several participants tested positive for STDs at enrollment (Table 1), but only trichomoniasis varied significantly by randomization group (p=0.05).
Participant characteristics were generally balanced between randomization groups, but we also observed some important differences. Women randomized to the control arm were more likely to report concurrent partnerships in the last three months (p=0.03). These women were also somewhat less likely than women in the vitamin D arm to be employed (p=0.09) and to have ever douched (p=0.09). Control women had somewhat fewer lifetime male sexual partners than women randomized to vitamin D (p=0.10). Among women with main partners, control women reported significantly younger partners than women in the vitamin D arm (p=0.01) (Table 1).
We also compared the characteristics of women who completed the trial (n=78) with those who failed to complete the final visit (n=40). Very few differences were observed (Table 2). Women who completed the trial were somewhat more likely to have ever douched (p=0.07), and they also reported higher median numbers of lifetime male partners (p=0.03) and somewhat higher median numbers of sex acts with men in the last 3 months (p=0.09).
Table 2.
Selected demographic, behavioral and clinical characteristics of participants, comparing women completing trial (n=78) with those lost to follow-up (n=40).
| Completed trial | Lost to follow-up | p-value | |||
|---|---|---|---|---|---|
|
| |||||
| N=78 | (%) | N=40 | (%) | ||
| Randomization group | |||||
| Vitamin D | 38 | (64) | 21 | (36) | 0.70 |
| Placebo | 40 | (68) | 19 | (32) | |
| Race/ethnicity a | |||||
| Black | 56 | (72) | 31 | (78) | 0.51 |
| White | 22 | (28) | 9 | (23) | 0.51 |
| American-Indian/Alaskan Native | 3 | (4) | 3 | (8) | 0.40 |
| Asian/Pacific-Islander | 0 | (0) | 1 | (3) | 0.34 |
| Hispanic | 4 | (5) | 3 | (8) | 0.69 |
| Education | |||||
| Finished high school | 67 | (86) | 34 | (85) | 0.90 |
| Finished college | 11 | (14) | 3 | (8) | 0.29 |
| Employed full- or part-time | 48 | (62) | 23 | (58) | 0.67 |
| Current user of vitamin D supplements | 12 | (15) | 5 | (13) | 0.69 |
| Currently using any method of contraception b | 31 | (40) | 18 | (45) | 0.58 |
| Hormonal methods (oral contraceptive pills, implants, injectable, patch, ring, hormonal IUD) | 7 | (23) | 6 | (33) | 0.41 |
| Male condoms | 17 | (55) | 10 | (56) | 0.96 |
| Female condoms | 1 | (3) | 1 | (6) | 0.69 |
| IUD (copper or unknown type) | 2 | (6) | 0 | (0) | 0.52 |
| Sterilization | 8 | (26) | 4 | (22) | 0.78 |
| Other (diaphragms, spermicides, douching, withdrawal, rhythm, abstinence, morning after pills) | 4 | (13) | 0 | (0) | 0.28 |
| Condom use, last 3 months | |||||
| 0% of vaginal acts with men | 29 | (40) | 11 | (31) | 0.46 |
| >0% and <100% of vaginal acts with men | 37 | (51) | 19 | (53) | |
| 100% of vaginal acts with men | 7 | (10) | 6 | (17) | |
| Ever anal sex | 33 | (42) | 15 | (38) | 0.69 |
| Ever sex with women | 22 | (28) | 10 | (25) | 0.71 |
| Concurrent sexual partnerships in the last 3 months c | 23 | (30) | 8 | (22) | 0.37 |
| Ever douched | 65 | (83) | 26 | (68) | 0.07 |
| Ever past BV diagnosis (self-reported, prior to enrollment) | 63 | (82) | 28 | (76) | 0.44 |
| Nugent score at enrollment | |||||
| Normal flora (Nugent 0–3) | 4 | (5) | 2 | (5) | 0.81 |
| Intermediate (Nugent 4–6) | 12 | (15) | 5 | (13) | |
| BV (Nugent 7–10) | 62 | (79) | 32 | (82) | |
| Insufficient specimen (unscorable) | 0 | (0) | 1 | (3) | |
| Infections | |||||
| Chlamydia | 8 | (10) | 3 | (8) | 0.63 |
| Gonorrhea | 7 | (9) | 2 | (5) | 0.44 |
| Trichomoniasis | 14 | (18) | 6 | (15) | 0.68 |
| Syphilis | 0 | (0) | 0 | (0) | 1.00 |
| Yeast | 5 | (6) | 3 | (8) | 0.82 |
| Prevalent HIV (not new diagnoses) | 2 | (3) | 0 | (0) | |
| Median | (IQR) | Median | (IQR) | p-value | |
| Age (years) | 28 | (22–35) | 24.5 | (22–31) | 0.33 |
| Age of main partner (years) d | 31 | (24–38) | 28 | (23–34.5) | 0.36 |
| Gravidity | 1 | (0–3) | 1 | (0–3) | 0.49 |
| Lifetime number male partners | 10 | (7–20) | 8.5 | (4–13.5) | 0.03 |
| Number vaginal sex acts with men, last 3 months | 12 | (5–39) | 7.5 | (2.5–24) | 0.09 |
Some women reported more than one race, so totals sum to greater than 100%.
The denominator for the proportions of women using each contraceptive method type is the total number of women reporting use of any method. Because women could report multiple methods, proportions sum to greater than 100%.
Women were considered to have concurrent sexual partnerships if they answered ‘yes’ to the question, “In the past three months, did you have sex with one partner while involved in a sexual relationship with another partner during the same period of time? This may include times when you did not consider yourself to be in a committed relationship.”
Among those reporting a main partner.
3.3 Vitamin D levels
Pre-supplementation vitamin D levels were similar: women randomized to the vitamin D arm had median 25(OH)D of 16.6 ng/mL (IQR: 12.1–21.2 ng/mL), vs. 15.8 ng/mL (IQR: 11.8–23.3 ng/mL) for control women. Vitamin D deficiency at enrollment (25(OH)D <20 ng/mL) was present in 71% of women randomized to vitamin D and 68% of control women. Throughout follow-up, women in the vitamin D arm had significantly higher median vitamin D levels than control women (Figure 2). By the final visit, the median 25(OH)D level in the vitamin D arm was 30.5 ng/mL, vs. 17.8 ng/mL in the control arm.
Figure 2.
Median and interquartile range of serum vitamin D levels over the follow-up period, by randomization group.
3.4 Self-reported adherence to metronidazole therapy
At the 4-week visit, women self-reported their compliance with metronidazole therapy, including the number of doses taken and the timing of treatment. Among women randomized to vitamin D, 81% reported taking metronidazole as prescribed. Among control women, 80% reported taking metronidazole as prescribed.
3.5 BV prevalence over follow-up, by randomization group
BV prevalence increased over follow-up in both groups (Figure 3). At the 4-week and 12-week visits, BV prevalence was similar between the vitamin D and control arms. By the 24-week visit, 65% of women randomized to vitamin D had BV, compared to 48% of control women.
Figure 3.
Prevalence and 95% confidence intervals of bacterial vaginosis over the follow-up period, by randomization group.
3.6 Infections diagnosed during follow-up
We observed no meaningful differences between groups in STD and reproductive tract infections during follow-up: trichomoniasis (4 women in the vitamin D arm vs. 6 in the control arm); chlamydia (2 vs. 4 cases); gonorrhea (0 vs. 2 cases); syphilis (0 vs. 2 cases) and yeast infections (10 vs. 9 cases) were each similar by intervention arm.
3.7 The effect of vitamin D on BV recurrence
BV-free survival time was very similar by randomization group (Figure 4). The log-rank test was non-significant (p=0.86). Median time to BV recurrence was 13.7 weeks in the vitamin D arm and 14.3 weeks in the control arm.
Figure 4.
Kaplan-Meier survival curve of time to BV recurrence, by randomization group.
The intention-to-treat HR for the effect of vitamin D supplementation on BV recurrence was 1.11 (95% CI: 0.68, 1.81) (Table 3). In the secondary analysis examining the effect of time-varying 25(OH)D on BV recurrence, the adjusted HR for BV per one-unit increase in 25(OH)D also demonstrated no significant effect of vitamin D level on BV (HR: 1.02, 95% CI: 0.99, 1.04).
Table 3.
Multivariable modeling of the effect of high-dose vitamin D supplementation on recurrence of bacterial vaginosis.
| Analysis description | Na | Exposure | Adjustment variables | HR | 95% CI |
|---|---|---|---|---|---|
| Primary analysis: Intention-to-treat | 97 | Randomization group | None | 1.11 | 0.68, 1.81 |
| Secondary analysis: using women’s measured 25(OH)D levels as primary exposure, instead of randomization group | 97 | Time-varying serum 25(OH)D level | age, race | 1.02 b | 0.99, 1.04 |
| Sensitivity analysis 1: controlling for factors not balanced by randomization at enrollment | 92 | Randomization group | employment status, sterilization for contraception, concurrent partnerships in the last 3 months, ever douched, lifetime male partners, trichomoniasis status | 0.71 | 0.40, 1.25 |
| Sensitivity analysis 2: restricted to women with deficient 25(OH)D (<20 ng/mL) at enrollment | 64 | Randomization group | None | 1. 10 | 0.62, 1.97 |
| Sensitivity analysis 3: restricted to women with BV at enrollment by Nugent scoring | 76 | Randomization group | None | 0.96 | 0.55–1.66 |
The primary and secondary analyses included person-time from all women who returned for any follow-up visit (n=97; 51 randomized to vitamin D and 46 to placebo). Sensitivity analysis 1 excluded 5 women who had missing data for one or more variables in the adjustment set, leading to an analysis sample of n=92 (49 in the Vitamin D group and 43 in the placebo group). Sensitivity analysis 2 was restricted to the subgroup of women with 25(OD)D levels <20 ng/mL at enrollment and who returned for at least one follow-up visit (n=64 women: 36 in the vitamin D group and 28 in the placebo group). Sensitivity analysis 3 was restricted to the subgroup of women who had BV according to Nugent scoring (Nugent 7–10) at enrollment and who returned for at least one follow-up visit (n=76 women: 42 in the vitamin D group and 34 in the placebo group).
HR for this model is interpreted as the increased hazard for BV per 1-ng/mL increase in serum 25(OH)D.
3.8 Exploratory sensitivity analyses
Following adjustment for BV risk factors not balanced by randomization, the HR for the effect of vitamin D on BV recurrence was 0.71 (95% CI: 0.40, 1.25). When restricting the analysis to women with deficient serum 25(OH)D at enrollment (<20 ng/mL), the HR was 1.10 (95% CI: 0.62, 1.97). Exclusion of participants who had normal or intermediate vaginal flora at enrollment, as determined by Nugent scoring, led to an HR of 0.96 (95% CI: 0.55–1.66) (Table 3).
3.9 Adverse events (AEs)
A total of 975 adverse events occurred during follow-up; 948 (97%) were considered possibly, probably or definitely related to study products (Table 4). For nearly all categories, the number of AEs among women randomized to vitamin D was lower than, or similar to, control women. The only AEs seen substantially more often among women in the vitamin D arm were dry mouth and vaginal discharge or itching. Two ectopic pregnancies occurred, both in the vitamin D arm; one occurred after enrollment but before the participant had begun vitamin D supplementation. Two miscarriages occurred, one in the vitamin D arm and one in the placebo arm (Table 4).
Table 4.
Related adverse events by randomization group.
| Vitamin D (n=442) | Control (n=506) | Total (n=948) | |
|---|---|---|---|
| Number of AEs | |||
| 4 weeks | 124 | 126 | 250 |
| 12 weeks | 80 | 99 | 179 |
| 24 weeks | 110 | 121 | 231 |
| Via telephone | 1 | 3 | 4 |
| Interim visits | 127 | 157 | 284 |
| Expected events a | |||
| Headache | 61 | 65 | 126 |
| Increased thirst | 40 | 44 | 84 |
| Nausea | 35 | 46 | 81 |
| Fatigue | 27 | 29 | 56 |
| Dry mouth | 26 | 7 | 33 |
| Constipation | 24 | 21 | 45 |
| Dizziness/fainting/weakness | 17 | 38 | 55 |
| Increased urination | 19 | 34 | 53 |
| Itchy skin | 21 | 26 | 47 |
| Behavior change/increased irritability | 19 | 26 | 45 |
| Loss of appetite | 17 | 13 | 30 |
| Vomiting | 16 | 25 | 41 |
| Ringing in ears | 14 | 12 | 26 |
| Muscle or bone pain | 12 | 14 | 26 |
| Metallic taste | 12 | 10 | 22 |
| Weight loss | 11 | 8 | 19 |
| Confusion | 7 | 14 | 21 |
| Abdominal pain | 6 | 2 | 8 |
| Back pain | 3 | 7 | 10 |
| Muscle weakness | 2 | 3 | 5 |
| Rash | 2 | 2 | 4 |
| High serum calcium | 1 | 2 | 3 |
| Kidney stone | 0 | 1 | 1 |
| Unexpected and possibly related events b | |||
| Yeast infection | 10 | 9 | 19 |
| Vaginal discharge/itching | 10 | 1 | 11 |
| Chest pain | 6 | 10 | 16 |
| Trichomoniasis | 4 | 6 | 10 |
| Irregular heartbeat | 3 | 2 | 5 |
| Shortness of breath | 2 | 10 | 12 |
| Chlamydia | 2 | 4 | 6 |
| Urinary tract infection | 2 | 1 | 3 |
| Ectopic pregnancy | 2 | 0 | 2 |
| Miscarriage | 1 | 1 | 2 |
| Low calcium | 1 | 1 | 2 |
| Bladder infection | 1 | 0 | 1 |
| Bleeding following cervical swab | 1 | 0 | 1 |
| Cellulitis | 1 | 0 | 1 |
| Mucopurulent cervicitis | 1 | 0 | 1 |
| Pressure in arm | 1 | 0 | 1 |
| Genital warts | 1 | 1 | 2 |
| Pelvic inflammatory disease | 1 | 1 | 2 |
| Gonorrhea | 0 | 2 | 2 |
| Syphilis | 0 | 2 | 2 |
| Herpes | 0 | 1 | 1 |
| Pelvic pain | 0 | 3 | 3 |
| Throat swelled | 0 | 1 | 1 |
Events that were known side effects of vitamin D supplementation or metronidazole, as described in the informed consent form.
Events that were possibly, probably or definitely related to study products or procedures.
4. COMMENT
In this randomized trial of STD clinic patients with symptomatic BV, supplementation with high-dose vitamin D in addition to standard metronidazole therapy did not reduce BV recurrence. Adjustment for adherence, and sensitivity analyses to test the robustness of the primary finding, reinforced the lack of association between vitamin D supplementation and BV recurrence.
Several cross-sectional studies suggested an association between low vitamin D levels and BV prevalence. One study reported a significant 65% increase in BV prevalence in pregnant African-Americans with serum 25(OH)D <20 ng/mL, compared to those with levels above 32 ng/mL.13 Another found that BV predicted vitamin D deficiency (25(OH)D <20 ng/mL) in pregnant African-American adolescents (OR: 4.4, p=0.02).14 An NHANES analysis reported increased odds of BV in pregnant women with vitamin D levels less than 30 ng/mL (adjusted OR: 2.9, 95% CI: 1.1–7.3).16 A case-control study that examined 25(OH)D levels during pregnancy reported significantly lower vitamin D levels in BV-positive compared to BV-negative participants (18.0 vs. 24.3 ng/mL, p=0.04).15
In contrast, two randomized trials of vitamin D3 supplementation in pregnant women examined BV as a secondary endpoint.17–18 When the trials were combined and after adjusting for study and race, neither the 2000 IU daily dose nor the 4000 IU daily dose significantly affected BV (p=0.75 for 2000 IU vs. control; p=0.34 for 4000 IU vs. control).19 The existing literature is also mixed on the association between vitamin D and BV in non-pregnant women. In the NHANES analysis, 25(OH)D <30 ng/mL in non-pregnant women was associated with BV in unadjusted but not adjusted analyses.16 One cross-sectional analysis reported an adjusted OR of 3.1 (95% CI: 1.2–8.4) for the effect of 25(OH)D <20 ng/mL on BV prevalence in HIV-positive women, but an adjusted OR of 1.1 (95% CI: 0.2–5.1) in HIV-negative women.20
As a small trial conducted in a busy STD clinic, our study had several limitations. We assessed only three of the four Amsel criteria because of inadequate time for microscopy. Some studies demonstrate that BV diagnosis using only two Amsel criteria (vs. four) has similar sensitivity and specificity compared to Nugent scoring. 28–29 Other studies indicate that the sensitivity of Amsel criteria can be low compared with Nugent scoring.30–31 In this study, the variability in the two diagnostic methods was clearly demonstrated: all women had symptomatic BV at enrollment by modified Amsel criteria, but only 80% had BV according to Nugent scoring.
Despite the randomized design, some variables were unbalanced at enrollment. However, sensitivity analyses demonstrated that this imbalance is unlikely to explain our primary findings. In addition, we did not conduct tests of cure following completion of metronidazole treatment, so BV diagnosed at the 4-, 12- and 24-week visits was likely a mixture of persistent BV which never cleared and recurrent BV following successful metronidazole treatment. However, for a polymicrobial clinical syndrome such as BV, where the division between health and disease is a continuum rather than a sharp line, the distinction between prevalent and re-emergent disease may be less important. Finally, we had higher than expected loss-to-follow up at 24 weeks, leading to lower-than-anticipated statistical power. However, given the consistency of the observed effects in both intention-to-treat and sensitivity analyses, we do not believe that low statistical power is masking a significant effect of vitamin D on BV recurrence.
Our trial also has important strengths. Unlike many interventions that rely on self-reported protocol adherence, the demonstrated rise in 25(OH)D among women randomized to vitamin D compared to control women is strong evidence of compliance: median serum 25(OH)D levels in intervention women had cleared the 30 ng/mL threshold after just 4 weeks. Nevertheless, 30 ng/mL may not be a meaningful threshold for BV recurrence, and a higher target level may be more appropriate. Considerable debate exists about the appropriate dose, schedule, and target serum 25(OH)D level for optimal human health.10,32–33
Earlier studies primarily examined prevalent BV rather than recurrent BV, our trial endpoint. Whether sufficient vitamin D may prevent the initial development of BV remains an unanswered question, but our results suggest that short-term, high-dose vitamin D in women with existing BV is not effective in reducing BV recurrence. It is also possible that vitamin D’s effect on BV requires more than 24 weeks to be observed. Although serum 25(OH)D levels of women in the randomization arm rose rapidly, whether related immune parameters increased at the same pace is not known. Despite high-dose supplementation, nearly 50% of women in the vitamin D arm had 25(OH)D levels lower than the 30 ng/mL threshold for sufficiency at trial end. Finally, most existing data on the vitamin D-BV association focused on pregnant women. Perhaps an association between vitamin D and prevalent BV exists in pregnant women because of unique conditions of pregnancy (a high estrogen state, compromised immunity, etc.), and that low serum vitamin D does not have the same impact on BV in non-pregnant women.
A recent evaluation confirmed that African-Americans have lower serum 25(OH)D levels than whites, but that African-Americans also have reduced levels of vitamin D binding protein, resulting in similar concentrations by race of bioavailable 25(OH)D.34 Three-quarters (74%) of trial participants self-identified as African-American; their measured 25(OH)D levels may be systematically lower than their bioavailable 25(OH)D. However, this limitation has no impact on the primary intention-to-treat analysis, in which each woman’s randomized treatment assignment, not her measured 25(OH)D level, is the primary exposure.
In summary, previous studies suggested that low vitamin D levels may be associated with increased BV prevalence. Our trial explored a related question: does increasing vitamin D level, in addition to metronidazole therapy, lead to reduced BV recurrence? Our findings suggest that short-term, high-dose vitamin D supplementation does not reduce BV recurrence in nonpregnant women. Given the established associations between BV and negative health outcomes, effective interventions to reduce BV’s impact continue to be urgently needed.
Acknowledgments
Financial support:
This work was supported by KL2RR025754 through the Ohio State University Center for Clinical and Translational Science (OSU CCTS). The OSU CCTS is supported by the National Center for Advancing Translational Sciences (NCATS), 8UL1TR000090-05. ANT was also supported by R21AI095987 from the National Institute for Allergy and Infectious Diseases (NIAID). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health, NCATS or NIAID.
The authors thank Jose Bazan, Bryna Harrington, Steven Ing, Jesse Kwiek, Nicole Cartwright Kwiek, Carson Reider, Cara Rice, Mary Ellen Wewers, and the entire staff at the Sexual Health Clinic at Columbus Public Health for their support of this project. We also thank Charles Rivers and Jane Schwebke at the University of Alabama at Birmingham.
Footnotes
Disclosure statement of any potential conflicts of interest:
Bio-Tech Pharmacal (Fayetteville, AR) donated both the vitamin D and placebo products used in this trial. Bio-Tech Pharmacal had no involvement in any other aspect of the research, including protocol development, data collection, analysis, manuscript writing or any other component of trial implementation or interpretation. ANT, PCR, KSF, JA, ME, JAD, RNF, MWR, MAK and RDJ have otherwise no commercial or other association that might pose a conflict of interest.
Paper presentation information:
This work was presented at the joint meeting of the International Society for STD Research (ISSTDR) and the International Union Against Sexually Transmitted Infections (IUSTI), 13–17 July 2013, in Vienna, Austria (abstract O06.2).
Disclaimers:
No author was employed by the United States Federal Government or United States Armed Forces.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Hillier S, Marrazzo JM, Holmes KK. Bacterial vaginosis. In: Holmes KK, Sparling PF, Mårdh P, et al., editors. Sexually transmitted diseases. 4. New York: McGraw-Hill; 2008. [Google Scholar]
- 2.Cohen CR, Lingappa JR, Baeten JM, et al. Bacterial vaginosis associated with increased risk of female-to-male HIV-1 transmission: a prospective cohort analysis among African couples. PLoS Med. 2012;9(6):e1001251. doi: 10.1371/journal.pmed.1001251. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Cu-Uvin S, Hogan JW, Caliendo AM, et al. Association between bacterial vaginosis and expression of human immunodeficiency virus type 1 RNA in the female genital tract. Clin Infect Dis. 2001;33(6):894–6. doi: 10.1086/322613. [DOI] [PubMed] [Google Scholar]
- 4.Oleen-Burkey MA, Hillier SL. Pregnancy complications associated with bacterial vaginosis and their estimated costs. Infect Dis Obstet Gynecol. 1995;3(4):149–57. doi: 10.1155/S1064744995000500. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.van Oostrum N, De Sutter P, Meys J, et al. Risks associated with bacterial vaginosis in infertility patients: a systematic review and meta-analysis. Hum Reprod. 2013;28(7):1809–15. doi: 10.1093/humrep/det096. [DOI] [PubMed] [Google Scholar]
- 6.Boris J, Pahlson C, Larsson PG. Six years observation after successful treatment of bacterial vaginosis. Infect Dis Obstet Gynecol. 1997;5(4):297–302. doi: 10.1155/S1064744997000513. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Ness RB, Hillier S, Richter HE, et al. Can known risk factors explain racial differences in the occurrence of bacterial vaginosis? J Natl Med Assoc. 2003;95(3):201–12. [PMC free article] [PubMed] [Google Scholar]
- 8.Allsworth JE, Peipert JF. Prevalence of bacterial vaginosis: 2001–2004 National Health and Nutrition Examination Survey data. Obstet Gynecol. 2007;109(1):114–20. doi: 10.1097/01.AOG.0000247627.84791.91. [DOI] [PubMed] [Google Scholar]
- 9.Hewison M. An update on vitamin D and human immunity. Clin Endocrinol (Oxf) 2012;76(3):315–25. doi: 10.1111/j.1365-2265.2011.04261.x. [DOI] [PubMed] [Google Scholar]
- 10.Ross AC, Taylor CL, Yaktine AL, et al., editors. for the Committee to Review Dietary Reference Intakes for Vitamin D and Calcium. Dietary Reference Intakes for Calcium and Vitamin D. Institute of Medicine, National Academy of Sciences; 2010. p. 482. [Google Scholar]
- 11.Adams JS, Hewison M. Update in vitamin D. J Clin Endocrinol Metab. 2010;95(2):471–8. doi: 10.1210/jc.2009-1773. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ginde AA, Liu MC, Camargo CA., Jr Demographic differences and trends of vitamin D insufficiency in the US population, 1988–2004. Arch Intern Med. 2009;169(6):626–32. doi: 10.1001/archinternmed.2008.604. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Bodnar LM, Krohn MA, Simhan HN. Maternal vitamin D deficiency is associated with bacterial vaginosis in the first trimester of pregnancy. J Nutr. 2009;139(6):1157–61. doi: 10.3945/jn.108.103168. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Davis LM, Chang SC, Mancini J, et al. Vitamin D insufficiency is prevalent among pregnant African American adolescents. J Pediatr Adolesc Gynecol. 2010;23(1):45–52. doi: 10.1016/j.jpag.2009.05.005. [DOI] [PubMed] [Google Scholar]
- 15.Dunlop AL, Taylor RN, Tangpricha V, et al. Maternal vitamin D, folate, and polyunsaturated fatty acid status and bacterial vaginosis during pregnancy. Infect Dis Obstet Gynecol. 2011;2011:216217. doi: 10.1155/2011/216217. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Hensel KJ, Randis TM, Gelber SE, et al. Pregnancy-specific association of vitamin D deficiency and bacterial vaginosis. Am J Obstet Gynecol. 2011;204(1):41, e1–9. doi: 10.1016/j.ajog.2010.08.013. [DOI] [PubMed] [Google Scholar]
- 17.Hollis BW, Johnson D, Hulsey TC, et al. Vitamin D supplementation during pregnancy: double-blind, randomized clinical trial of safety and effectiveness. J Bone Miner Res. 2011;26(10):2341–57. doi: 10.1002/jbmr.463. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Wagner CL, McNeil R, Hamilton SA, et al. A randomized trial of vitamin D supplementation in 2 community health center networks in South Carolina. Am J Obstet Gynecol. 2013;208(2):137, e1–13. doi: 10.1016/j.ajog.2012.10.888. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Wagner CL, McNeil RB, Johnson DD, et al. Health characteristics and outcomes of two randomized vitamin D supplementation trials during pregnancy: A combined analysis. J Steroid Biochem Mol Biol. 2013;136:313–20. doi: 10.1016/j.jsbmb.2013.01.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.French AL, Adeyemi OM, Agniel DM, et al. The association of HIV status with bacterial vaginosis and vitamin D in the United States. J Womens Health (Larchmt) 2011;20(10):1497–503. doi: 10.1089/jwh.2010.2685. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Klebanoff MA, Turner AN. Bacterial vaginosis and season, a proxy for vitamin D status. Sex Transm Dis. 2014;41(5):295–9. doi: 10.1097/OLQ.0000000000000124. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Schulz KF, Altman DG, Moher D for the CONSORT Group. CONSORT 2010 Statement: updated guidelines for reporting parallel group randomised trials. BMJ. 2010;340:c332. doi: 10.1136/bmj.c332. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Moher D, Hopewell S, Schulz KF, et al. for the CONSORT Group. CONSORT 2010 Explanation and Elaboration: updated guidelines for reporting parallel group randomised trial. BMJ. 2010;340:c869. doi: 10.1136/bmj.c869. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Amsel R, Totten PA, Spiegel CA, et al. Nonspecific vaginitis. Diagnostic criteria and microbial and epidemiologic associations. Am J Med. 1983;74(1):14–22. doi: 10.1016/0002-9343(83)91112-9. [DOI] [PubMed] [Google Scholar]
- 25.Nugent RP, Krohn MA, Hillier SL. Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation. J Clin Microbiol. 1991;29(2):297–301. doi: 10.1128/jcm.29.2.297-301.1991. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53:457–81. [Google Scholar]
- 27.Ya W, Reifer C, Miller LE. Efficacy of vaginal probiotic capsules for recurrent bacterial vaginosis: a double-blind, randomized, placebo-controlled study. Am J Obstet Gynecol. 2010;203(2):120, e1–6. doi: 10.1016/j.ajog.2010.05.023. [DOI] [PubMed] [Google Scholar]
- 28.Gutman RE, Peipert JF, Weitzen S, et al. Evaluation of clinical methods for diagnosing bacterial vaginosis. Obstet Gynecol. 2005;105(3):551–6. doi: 10.1097/01.AOG.0000145752.97999.67. [DOI] [PubMed] [Google Scholar]
- 29.Mittal V, Jain A, Pradeep Y. Development of modified diagnostic criteria for bacterial vaginosis at peripheral health centres in developing countries. J Infect Dev Ctries. 2012;6(5):373–7. doi: 10.3855/jidc.1625. [DOI] [PubMed] [Google Scholar]
- 30.Schwebke JR, Hillier SL, Sobel JD, McGregor JA, Sweet RL. Validity of the vaginal gram stain for the diagnosis of bacterial vaginosis. Obstet Gynecol. 1996;88(4 Pt 1):573–6. doi: 10.1016/0029-7844(96)00233-5. [DOI] [PubMed] [Google Scholar]
- 31.Sha BE, Chen HY, Wang QJ, et al. Utility of Amsel criteria, Nugent score, and quantitative PCR for Gardnerella vaginalis, Mycoplasma hominis, and Lactobacillus spp. for diagnosis of bacterial vaginosis in human immunodeficiency virus-infected women. J Clin Microbiol. 2005;43(9):4607–12. doi: 10.1128/JCM.43.9.4607-4612.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Holick MF, Binkley NC, Bischoff-Ferrari HA, et al. Evaluation, treatment, and prevention of vitamin D deficiency: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab. 2011;96(7):1911–30. doi: 10.1210/jc.2011-0385. [DOI] [PubMed] [Google Scholar]
- 33.Rosen CJ, Abrams SA, Aloia JF, et al. IOM committee members respond to Endocrine Society vitamin D guideline. J Clin Endocrinol Metab. 2012;97(4):1146–52. doi: 10.1210/jc.2011-2218. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Powe CE, Evans MK, Wenger J, et al. Vitamin D-binding protein and vitamin D status of black Americans and white Americans. N Engl J Med. 2013;369(21):1991–2000. doi: 10.1056/NEJMoa1306357. [DOI] [PMC free article] [PubMed] [Google Scholar]