Effects of Vitamin D Supplementation on Semen Quality, Reproductive Hormones, and Live Birth Rate

A Randomized Clinical Trial

Martin Blomberg Jensen; Jacob Gerner Lawaetz; Jørgen Holm Petersen; Anders Juul; Niels Jørgensen


J Clin Endocrinol Metab. 2018;103(3):870-881. 

In This Article


The Copenhagen Bone-Gonadal Study is a single-center, triple-blinded, randomized clinical trial conducted at the Department of Growth and Reproduction, Rigshospitalet, Denmark (ClinicalTrials.gov no. NCT01304927). The study protocol, including prespecified subgroups and a detailed plan for biostatistical analysis, are available in the Supplemental Materials.

Participants and Trial Design

The Copenhagen Bone-Gonadal Study was approved by the Danish Health and Medicines authority, the local ethical committee, and the data protection agency (approval no. 2010-024588-42, H-4-2010-138, and 2010124801) and monitored according to good clinical practices (Good Clinical Practices Unit, Bisbebjerg Hospital, Denmark). All men eligible for inclusion were part of an infertile couple in whom where semen analysis had shown impaired semen quality (Supplemental Figure 1).

After referral to our andrological center, all men delivered two semen samples for routine analyses and blood samples for karyotyping, analyses of microdeletions on the Y chromosome, and for serum measurements of ionized calcium, 25OHD, testosterone, estradiol, sex-hormone binding globulin (SHBG), luteinizing hormone (LH), follicle-stimulating hormone (FSH), and inhibin B before their visit to the outpatient clinic. At the clinic, the patient's medical history was obtained and a physical examination including ultrasound of both testes was performed. Men were invited to participate in the trial if they had impaired semen quality (determined by WHO criteria), were vitamin D insufficient (25OHD level ≤50 nmol/L), and did not have serious comorbidities. Informed consent was obtained and a new date was assigned to start the intervention (Figure 1).

Figure 1.

Flow diagram of participants in the Copenhagen Bone-Gonadal study.

Included men were given a specific trial identity number determined by minimization using the computer program Minim.[21] Minimization was done using four groups based on serum 25OHD, sperm concentration, body mass index (BMI) and serum inhibin B. To avoid unblinding, the principal investigator gave the necessary clinical information to the sponsor, who had a list of numbers headed by X or Y. This ensured that both the principal investigator and the sponsor were unaware whether the patient was allocated to the vitamin D plus calcium (active) group or the placebo group (i.e., double blinding).

At the first visit, all men delivered a fasting blood sample and had a dual-energy X-ray absorptiometry scan performed. Afterward, men in the active group received a dose of 300,000 IU of vitamin D orally, and the nurse interviewed them and provided them tablets containing 1400 IU of vitamin D plus 500 mg calcium. Men in the placebo arm received oil once orally and placebo tablets to take for the following 90 days. All men received a telephone call to identify any incidents 28 days after start of the intervention. A follow-up visit after 90 days included fasting blood samples, an interview with the nurse, and receipt of active or placebo tablets for the remaining 60 days. Finally, two semen samples were delivered, separated by a median of 10 days, by days 140 and 150. A fasting blood sample, a dualenergy X-ray absorptiometry scan, and a final interview with the nurse were also completed on day 150.

Men who were part of an infertile couple who reported a clinical pregnancy at the final interview were contacted by telephone in 2016. The telephone interview was conducted to obtain the following specific information on the outcome of the pregnancy: spontaneous pregnancy or use of ART treatment (i.e., intrauterine insemination, IVF, ICSI), number of treatment attempts, abortions, time and type of abortion, live birth, gestational week at birth, and sex and health of the child. Eighty percent of the reported pregnancies were confirmed by telephone interview and the time of conception for live births was determined by back calculation from gestational week at birth. The remaining 20% of the reported clinical pregnancies could not be confirmed by telephone because there was no answer from the patients; however, these pregnancies were considered as live births in the analysis.

Sample Size Calculation

Power estimate was based on a cross-sectional study of 300 young healthy men from the general population and previous intervention studies.[16,22,23] Our study design was based on an expected increase in serum levels of 25OHD of 50 nmol/L and an expected participant dropout of 10%. A setup with 150 men in each group, a level of significance of 5%, and power of 80% enabled us to detect changes in sperm concentration, sperm motility, and inhibin B level of 27%, 15%, and 20%, respectively.

Medication and Compliance

Infertile men were randomly assigned 1:1 (in blocks of 10) to either placebo or an initial oral dose of 300,000 IU of cholecalciferol dissolved in oil, followed by receipt of tablets consisting of cholecalciferol 1400 IU and calcium 500 mg once daily for 150 days (Pfizer, Copenhagen, Denmark). All men were instructed to abstain from taking vitamin D supplementation >400 IU daily. Randomization and manufacture of the high initial dose of vitamin D and placebo were performed by Glostrup Apotek. None of the companies had further involvement in the trial. The trial remained blinded until all biochemical analyses, data handling, and statistical analyses by an independent statistician had been completed (i.e., triple blinding).

To determine efficacy and compliance, serum levels of 25OHD, 1,25(OH)2D3, calcium, and PTH were measured on day 1 and day 150 to allow evaluation of compliance, which was further supported by counting returned tablets on days 90 and 150. All patients were informed about their vitamin D status before inclusion and were offered vitamin D supplementation of 35 mg daily for 180 days after completion of the study independent of their group assignment during the study period.

Biochemical Analysis

Fasting blood samples were collected between 8:00 AM and 10:00 AM. Serum was analyzed immediately for PTH, total calcium, and ionized calcium levels. The remaining analyses were conducted on frozen serum samples from days 1 and 150 after study completion. Measurements of 25OHD and 1,25(OH)2D3 levels were conducted using isotope-dilution liquid chromatography–tandemmass spectrometry with interassay coefficients of variation (CVs) <10% and <18%, respectively. FSH, LH, and SHBG levels were determined using a timeresolved immuno-fluorometric assay (Delfia; Wallac, Turku, Finland) and inhibin B levels were determined by a specific twosided enzyme-linked immunoassay (inhibin B genII; Beckman Coulter, Brea, CA). The CVs were <4%, <4%, <6%, and <11% for FSH, LH, SHBG, and inhibin B, respectively). Testosterone and estradiol levels were measured using radioimmunoassay platforms from Siemens (Los Angeles, CA) and Pantex (Santa Monica, CA), respectively; the CV was <13%. Ionized calcium levels were measured using the Konelab 30i (Thermo Fisher Scientific, Waltham, MA) with a CV<2%, and total calcium and PTH levels were measured using the Cobas 8000 (Roche, Basel, Switzerland) with a CV <4%.

Semen Analysis

Semen samples were obtained by masturbation in a room adjacent to the laboratory and information on duration of ejaculation abstinence, fever, and spillage was obtained. The two semen samples were delivered 16 days apart before the start and 10 days apart at the end of the study. Semen analysis performed at our laboratory has been described in detail previously.[24] Briefly, semen volume was calculated by weighing, sperm concentration was determined using a Bürker–Türk hemocytometer, and total sperm count was calculated by multiplication. Sperm morphology according to strict criteria was evaluated on Papanicolaou-stained smears. Sperm motility classified as progressive motile (WHO class A+B), nonprogressive motile (class C), or immotile (class D) was determined in duplicate at two times and presented as AB or ABC motility. The mean of two semen samples was calculated for total sperm count, semen volume, sperm concentration, sperm motility, and morphology at baseline and after study completion. Most men delivered two samples at baseline (317 of 330 men) and study end (253 of 269 men), but data from men delivering only one sample were also included in the analysis.

In 82 men, spermatozoa were investigated for DNA fragmentation before and after intervention at SPZ Laboratory (Copenhagen, Denmark; Supplemental Table 1). Semen (0.5 mL) was diluted with TNE buffer, then mixed and frozen directly in liquid nitrogen until fluorescent staining according to the sperm chromatin structure assay protocol.[25] The samples were analyzed using a FACSCalibur (BD Biosciences, San Jose, CA) flow cytometer. Data were acquired using the CellQuest software (version 3.2.; BD Biosciences). Each analysis was run blinded in duplicate and recording was stopped after 5000 events. Variation was assessed by analyzing two semen samples from the same person (patient no. 28) before intervention. The analyses had a standard deviation of 3.3% and CV of 16.5%.


A full description of the statistical analysis can be found in Supplemental Material. Descriptive statistics are presented as means with standard deviation (95% confidence intervals) or medians with interquartile range (IQR) in Table 1 and Table 2 and Figure 2 and Figure 3. The primary analysis was performed according to the randomized group assignment, using first crude comparisons of means (t test) and distributions (Mann-Whitney test) for total sperm count and sperm concentration. Subsequently, all variables were analyzed using a mixed model that adjusted for baseline levels. Variables were logarithmically or cubic-root transformed when required to obtain approximate normality of the residuals.

Figure 2.

Vitamin D homeostasis and semen quality in men with vitamin D insufficiency. Serum concentrations (reference ranges are in parentheses) at the time of screening (baseline), at the start of the intervention (day 1), and at study completion (day 150) are shown for (A) 25OHD; (B) 1,25(OH)2D3 (50 to 160 pmol/L); (C) total calcium (2.15 to 2.51 mmol/L); and (D) PTH (1.6 to 6.9 pmol/L). (E) Total sperm count. (F) Sperm concentration. (G) Sperm motility. (H) Sperm morphology. (A–D, G, H) Data are presented as mean ± 95% confidence intervals. (E, F) Data given as median ± 25th and 75th quartiles. Ca, calcium.

Figure 3.

Vitamin D homeostasis and sperm production in a prespecified subgroup of men with severe vitamin D deficiency. Serum concentrations (reference ranges are in parentheses) at the time of screening (baseline), at the start of the intervention (day 1), and at study completion (day 150) are shown for (A) 25OHD; (B) 1,25(OH)2D3 (50 to 160 pmol/L); (C) PTH (1.6 to 6.9 pmol/L); and (D) inhibin B. (E) Total sperm count. (F) Sperm concentration. (A–D) Data are presented as mean ± 95% confidence intervals. (E, F) Data are presented as median ± 25th and 75th quartiles. Ca, calcium.

Live-birth rate and method used to achieve pregnancies were analyzed using two-sided x2 and Fisher exact tests. Comparison of live-birth rates was only performed at day 150, and couples with no available information about pregnancies or live births or pregnancy achieved before the start of the study were excluded from the analysis. The ART methodology used to achieve pregnancy was determined only in couples obtaining a live birth.

Further analysis was performed according to prespecified subgroup analyses in relation to primary and secondary end points. Several prespecified subgroups were listed, but only vitamin D deficiency defined by serum 25OHD level ≤25 nmol/L was used.

Live-birth rate was also investigated in men with oligozoospermia (sperm concentration <15 million/mL) and oligo/asthenozoospermia (sperm concentration <15 million/mL or ABC sperm motility <40%) because men in the placebo group had better semen quality compared with the vitamin D group from the start of the study. Investigations of associations between serum inhibin B and the difference in inhibin B levels between day 150 and day 1 (Dd150−1) with serum and Dd150−1 vitamin D metabolite levels were done using linear regression adjusted for potential confounders such as age, BMI, and smoking.