Caffeine Intake and Semen Quality in a Population of 2,554 Young Danish Men

Tina Kold Jensen; Shanna H. Swan; Niels E. Skakkebæk; Sanne Rasmussen; Niels Jørgensen


Am J Epidemiol. 2010;171(8):883-891. 

In This Article


A total of 2,554 men participated (approximately 31% of those approached). Of these men, 141 (5.5%) had fathered a pregnancy and 15 (0.6%) had been examined for infertility. One hundred forty-nine men who did not provide information about caffeine intake did not differ from the others with respect to other lifestyle factors and semen quality (data not shown). More than 50% of the total caffeine intake was derived from coffee and 20% from cola (Table 1). Caffeine content in cola is quite low compared with coffee; one cup of coffee contains about 117 mg compared with 70 mg in 0.5 L of cola.

Although men who reported no caffeine intake (n = 72) had better semen quality (median sperm concentration, total sperm count, and morphological normal sperm: 62 mill/mL, 210 mill, and 7%, respectively), moderate consumption of caffeine was not associated with a reduction in semen quality. However, men with a high caffeine intake (>800 mg of caffeine per day) had a slight reduction in semen quality (not statistically significant) (Table 2). Men whose caffeine intake was >800 mg (about 7 cups of coffee) per day generally had a less healthy diet, eating more burgers and cheese; drank more alcohol; smoked more often; and had a high or low body mass index (Table 3). In addition, they were from a lower social class, more often had self-reported genital conditions in the reproductive organs or conditions found at the physical examination, and more often had been exposed to smoking in utero compared with men whose caffeine consumption was lower (Table 3).

After control for confounders, we found that low (101–200 mg) to moderate (201–800 mg) daily caffeine consumption was not associated with a reduction in semen quality (Table 4). Consumption of >800 mg of caffeine per day resulted in a nonsignificant reduction in semen quality. Caffeine consumption was also entered as a (ln-transformed) continuous variable. Over the entire range, only semen volume decreased significantly with increasing caffeine intake (Table 4).

Analyses were then performed among men consuming caffeine from different sources to determine the associations with each. After we controlled for confounders (including cola consumption), no association of coffee, tea, chocolate beverages or bars, or diet soft drinks with semen quality was observed. When all caffeine sources were included simultaneously in the model, only cola consumption was associated with a significant reduction in semen quality (data available on request).

A total of 2,114 men reported that they drank cola during the past week; of these, 93 (4.4%) drank more than 14 bottles per week (>1 L per day, 140 mg of caffeine). Men who drank cola had poorer semen quality than men who did not (Table 2). Men who drank >14 bottles (140 mg of caffeine) of cola per week generally also drank less milk and consumed less fruit, vegetables, and fish, but they more frequently consumed beef and burgers compared with men who drank ≤1 L of cola per day (Table 3). In addition, they reported more diseases in reproductive organs, drank more alcohol, had a high or low body mass index, and more often were smokers or had been exposed to smoking in utero than men who drank fewer than 14 bottles of cola per week.

After control for confounders, semen volume, sperm concentration, total sperm count, and percentage of spermatozoa with normal morphology decreased among cola-drinking men compared with nondrinkers (Table 4) and significantly decreased among men who drank more than 14 bottles (1 L) of cola per week. Men whose weekly cola consumption was 0, 1–7, 8–14, and >14 bottles had respective adjusted sperm concentrations (mill/mL) of 56 (95% confidence interval (CI): 50, 64), 47 (95% CI: 44, 51), 49 (95% CI: 43, 57), and 40 (95% CI: 32, 51) and respective total sperm counts (mill) of 181 (95% CI: 156, 210), 144 (95% CI: 132, 157), 153 (95% CI: 129, 182), and 121 (95% CI: 92, 160). When cola was entered as a continuous variable (ln transformed), a significant decline in semen volume, total sperm count, sperm concentration, and sperm morphology was found. No association with sperm motility was observed. The analyses were not adjusted for dietary factors because they did not have a significant impact on the associations. All analyses were repeated by excluding diet soft drinks from cola intake, which did not affect our findings.

The analyses were repeated for caffeine intake from sources other than cola (Table 4) to determine whether the association between caffeine and semen quality was attributable to an adverse effect of cola. The same magnitude of effect as for total caffeine intake was found for men whose caffeine intake was not derived from cola (Table 4).

We also examined associations of caffeine and cola consumption with serum reproductive hormones (testosterone, inhibin B, follicle-stimulating hormone, and luteinizing hormone). However, we found no statistically significant associations (data not shown).