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


A total of 1427 men were referred with male infertility and screened from January 2011 to August 2014. Of these, 330 infertile men formed the study cohort: 1002 men did not meet all inclusion criteria or met an exclusion criterion, and 95 men were eligible but did not wish to participate. The main reasons for exclusion were high vitamin D status, comorbidities, or azoospermia. Table 1 summarizes the basic characteristics of the 330 included men; Supplemental Table 2 lists the differences in sperm production, reproductive hormones and vitamin D status between included and excluded men. Fifteen men withdrew consent and eight men unexpectedly met an exclusion criterion (comorbidities such as diabetes and previous tuberculosis) and, therefore, were not allowed to start the intervention at day 1. The remaining 307 men were assigned to treatment with either vitamin D and calcium (n = 151) or placebo (n = 156). The first patient was included in March 2011 and the final visit of the last patient was completed in February 2015. Twenty men in the placebo group and 18 in the vitamin D plus calcium group were lost to follow-up. In total, 269 of 307 men (87.6%) completed the study (Figure 1). By counting returned tablets, it was evident that one man in the vitamin D group and three in the placebo group were noncompliant; however, all data from these four men were included in all the analyses.

Efficacy of the Intervention

The mean serum 25OHD concentration was 35 nmol/L in both groups at baseline (i.e., day of screening; Supplemental Table 1; Figure 2). A similar but remarkable 10 nmol/L increase in both groups was found from screening to day 1, although no active treatment was initiated during this period. From the start of the intervention until day 150, serum 25OHD concentration increased by a mean of 43 nmol/L (96%) in men randomly assigned to the vitamin D plus calcium group, whereas 25OHD levels only increased, on average, 5 nmol/L (11%) in the placebo group (P < 0.0005; Figure 1A). Men treated with vitamin D plus calcium experienced a 1.5% higher serum concentration of calcium, 30% higher 1,25(OH)2D3, and 8% lower PTH at the end of the study (day 150) compared with the day they started the intervention (day 1; P,0.01 for all; Figure 1A, Figure 1B, Figure 1C and 1D). In contrast, in the men in the placebo group, who had no change in serum calcium, there was an 8% decrease in 1,25(OH)2D3 levels and a 6% increase in PTH levels. At day 150, serum concentrations of 25OHD, 1,25(OH)2D3, PTH, and calcium were all different (P < 0.05 for all) between men in the treatment group and those in the placebo group.

Safety and Types of Incidents

Incidents were reported in 206 of the 307 men starting the intervention (62%; Supplemental Table 2). The frequency of incidents was similar in the two groups, with a median relative risk of 1.34 (95% CI; 0.84 to 2.15) for men in the treatment group (P = 0.21). The most frequent incident was infections (66%), followed by musculoskeletal pain (11%) and gastrointestinal complaints (8%), but none of the incidents differed between groups (Supplemental Table 2). No cases of kidney stones or hypercalcemia were detected and the highest concentration of ionized calcium measured was 1.28 mmol/L (IQR, 1.18 to 1.32). An elevated total calcium concentration of 2.81 mmol/L was detected in serum once (IQR, 2.15 to 2.51). Hypercalcemia was probably caused by a technical error, because ionized calcium was 1.24 mmol/L at the same time. A follow-up sample taken 1 week later was normal. One serious adverse event was observed for one man, who was admitted to the intensive care unit with gallstone-induced pancreatitis. This man was in the placebo group.

Primary end Point

On average, the couples had tried to become pregnant for >26 months, and presence of male factor infertility was indicated by a median sperm concentration of 12 million/mL in the included men (Table 1). No differences in total sperm count or sperm concentration were found at day 150 between the treatment and the placebo groups (Figure 1E, Figure 1F; Table 2).

Men treated in the treatment group had a 4% lower percentage of motile spermatozoa (P = 0.108) and progressive motile spermatozoa (P = 0.091) compared with the placebo group at baseline. The observed differences between groups were also present at day 150 for motile and progressive motility (−4% for both, IQRs, −9% to 1% and −8% to 1%, respectively; Table 2; Figure 1G). Motility variables assessed 4 hours after ejaculation showed similar results (data not shown).

The percentage of spermatozoa with normal morphology was higher in the placebo group than in the treatment group from the start of the study (P = 0.007) and tended also to be higher at day 150 (P = 0.057; Figure 1H). Moreover, the treatment group had a higher semen volume at baseline [0.5 mL (IQR, 0.1 to 0.8); P = 0.02] compared with the placebo group, and this difference did not change at day 150 (P = 0.038; Table 2).

Secondary end Points

Twenty-nine of the 269 men completing the trial reported their partner was pregnant before start of the intervention, whereas five men lost their partner during the study period, leaving 235 with the possibility of effecting a pregnancy. Therefore, the remaining 235 men were included in the analysis. Eighty-nine pregnancies were registered in the case report form and 80% of the men from these couples were interviewed by telephone at least14 months after their last visit (telephone calls were made between May and November 2016). The telephone interviews revealed that 24 of the 89 registered pregnancies were not conceived during the study period (days 1 to 150), and three pregnancies were not correctly registered, leaving 62 pregnancies that occurred during the trial period. Three of the 62 pregnancies resulted in spontaneous abortions (two in the placebo group and one in the treatment group), and 59 of the pregnancies resulted in live births.

Six children were born before gestational week 36 (three in each study group), and one child in each group was under investigation for a congenital disease. The distribution of sex was similar in the two groups. The live-birth rate was not statistically different between groups (Table 2; Figure 4A). However, the methodology used to achieve pregnancy for the 59 live births (i.e., spontaneous, intrauterine insemination, IVF, ICSI) differed between the groups (P = 0.032; Figure 4C).

Figure 4.

Live-birth rate and use of ART to achieve pregnancy. (A) Cumulative live births conceived from start of intervention (day 1) until study end (day 150). (B) Cumulative live births from oligozoospermic fathers defined by a sperm concentration <15 million/mL. (C) Methodology used to achieve the pregnancy responsible for live birth. The Fischer exact test was used to compare the live-birth rate at day 150 between the two study groups. Ca, calcium.

The chance of achieving a spontaneous pregnancy tended to be higher in couples in which the men were treated with vitamin D plus calcium compared with couples in which the man was in the placebo group (7.3% vs 2.4%; 95% CI, −0.6% to 10.5%). Moreover, couples in which men had oligozoospermia and were in the treatment group had a higher chance for live birth compared with the placebo group (35.6% vs 18.3%; 95% CI, 1.6% to 32.9%; Figure 4B). However, couples in which the man had oligozoospermia and was in the treatment group did not have a statistically significant higher rate of spontaneous pregnancy compared with the placebo group. Couples in which the man had either oligozoospermia or asthenozoospermia and was allocated to the treatment group had higher frequency of live births compared with the placebo group (33.8% vs 18.3%; 95% CI, 1.3% to 29.7%). As expected, men who were oligozoospermic or oligo/asthenozoospermic who were in the placebo group were associated with couples with fewer live births than men in the placebo group who had normal sperm concentration >15 million/mL or sperm motility >40%.

Men in couples who had a live birth had a higher serum 25OHD level at baseline [3.1 nmol/L; (IQR,−0.06 to 6.27); P = 0.054] compared with men in couples having no children. Furthermore, men in the treatment group and part of a couple who had a live birth had a higher 25OHD level [8.9 nmol/L (IQR, 0.5 to 17.3); P = 0.037] compared with men in couples who did not achieve a pregnancy during the study.

Serum concentrations of inhibin B and FSH, and the inhibin B-to-FSH ratio were not different between the two arms and did not change when comparing Dd150−1 (Table 1 and Table 2). However, the Dd150−1 in activated vitamin D was positively associated with serum inhibin B levels (b = 0.205; P = 0.055) and the inhibin B-to-FSH ratio (b = 0.164; P = 0.083) at study completion, after adjustment for BMI, smoking, and age.

Prespecified Subgroup Analysis

A prespecified vitamin D deficiency subgroup defined by a threshold set at 10 ng/mL (25 nmol/L) was used to evaluate the reproductive effect of supplementation in men presumed to have a greater need of vitamin D. Of the 330 included men, 66 (20%) had vitamin D deficiency; the average serum 25OHD concentration was 18 nmol/L. Serum 25OHD concentration reached 87 nmol/L in the treatment group, whereas it remained low(36 nmol/L) in the placebo group at day 150 (P < 0.0005; Figure 2A and Figure 2C).

Serum 1,25(OH)2D3 concentration increased to 105 pmol/L in men in the treatment group at day 150, which clearly differed from the men in the placebo group, in whom serum 1,25(OH)2D3 concentration reached 66 pmol/L (P = 0.001). This difference between groups was also reflected in serum concentrations of calcium and PTH at day 150 (P < 0.01; Figure 4; Supplemental Figure 2). As expected, vitamin D treatment was associated with lower PTH levels (−17%) at day 150 compared with day 1, whereas serum PTH level increased 10% during the study period in the placebo group (P < 0.001).

Median total sperm count and sperm concentration were not statistically significantly different between the two study arms at day 150 (P = 0.07). However, men randomly assigned to the treatment group had an increase in total sperm count of 15 million and an increase in sperm concentration of 3 million/mL at day 150 compared with baseline levels (P > 0.05). This is in contrast to a decrease of 5 million for total sperm count and 4 million/mL for sperm concentration when comparing day 150 with baseline levels in men in the placebo group (P > 0.05; Figure 4D and Figure 4E; Supplemental Figure 2).

No differences in sperm motility variables or sperm morphology were found between the two study groups or when comparing baseline with day 150 within the groups. However, the median serum inhibin B level was 49 pg/mL (IQR, 8 to 91; P = 0.021) higher in the treatment group at day 150 compared with the placebo group [34% higher (IQR, 4%to 65%); Figure 4D]. The inhibin B-to-FSH ratio also was higher in the treatment group (63%; IQR, −20% to 138%), although this was not statistically significantly different from that of the placebo group (P = 0.138).

Moreover, changes in 25OHD concentration during the study period were positively associated with the observed changes in inhibin B-to-FSH ratio (b = 0.101; P = 0.038). Men in the treatment group had an 8.4% (IQR, −0.4% to 17.1%; P = 0.060) higher inhibin B-to-FSH ratio compared with those in the placebo group when data were analyzed after normalization of the inhibin B-to-FSH ratio to the ratio before start of the intervention. However, differences in serum inhibin B and FSH levels, and the inhibin B-to-FSH ratio between the treatment arms at the end of the study were not statistically significant when the mixed model was applied (P = 0.3850, 0.9683, and 0.6118, respectively).

DNA fragmentation did not change from day 1 to day 150, and there was no difference between the treatment and placebo groups in the subgroup of 82 men in whom the DNA fragmentation index was investigated (Supplemental Table 1).