- Email: [email protected]
Cigarette smoking and semen quality
FERTILITY AND STERILITY Copyright
©
Vol. 65, No.4, April 1996
Printed on acid-free paper in U. S. A.
1996 American Society for Reproductive Medicine
Cigarette smoking and semen quality
Marilyn F. Vine, Ph.D.t:!: Chiu-Kit J. Tse, M.8.P.H.t Ping-Chuan Hu, Ph.D.§ K. Young Truong, Ph.D.11 The University of North Carolina, Chapel Hill, North Carolina
Objective: To determine whether cotinine levels provide stronger evidence for an association between smoking and semen quality than the number of cigarettes smoked per day or years smoked controlling for potential confounders and effect modifiers. Design: Cross-sectional study. Setting: Male volunteers at the Reproductive Endocrinology-Fertility Laboratory. Participants: Eighty-eight men (ages 18 to 35 years) provided a semen, urine, and blood specimen and completed a self-administered questionnaire concerning smoking and demographic information as well as caffeine and alcohol consumption. Urine, blood, and semen cotinine levels were analyzed via RIA. Main Outcome Measure: Standard clinical semen analysis. Results: Number of cigarettes smoked per day, years smoked, and log-transformed cotinine levels were associated negatively with semen quality (density, total count, and motility). The association was evident among men age ~22 years. For example, the correlation coefficient for the overall association between logged urine cotinine and logged sperm density was -0.23; those stratified by age were 0.13 (age <22 years) and -0.39 (age ~22 years). Potential confounders included in regression models did not diminish the associations. Conclusions: Smoking is associated with lowered semen quality. Fertil Steril 1996;65:835-42 Key Words: Smoking, sperm, semen quality, male, human
Many studies have assessed the association between cigarette smoking and sperm density, motility, and morphology with mixed results (1-3). Inadequate consideration of smoking dose or the potential Received May 3, 1995; revised and accepted September 26, 1995. * Supported by Cooperative Agreement RR13485-01 between Health Effects Research Laboratory, United States Environmental Protection Agency and the University of North Carolina, Department of Epidemiology. Although the research described in this article has been supported by the United States Environmental Protection Agency, it has not been subjected to Agency review and, therefore, does not necessarily reflect the view of the Agency and no official endorsement should be inferred. t Department of Epidemiology, School of Public Health. :j: Reprint requests: Marilyn F. Vine, Ph.D., Department of Epidemiology, School of Public Health, CB No. 7400 McGavranGreenberg Hall, University of North Carolina, Chapel Hill, North Carolina 27599-7400 (FAX: 919-966-2089). § Department of Pediatrics, School of Medicine. II Department of Biostatistics, School of Public Health. Vol. 65, No.4, April 1996
confounding and modifying effects of such factors as age as well as lack of statistical power in some studies may account for the contradictory findings. The primary goals ofthis study are to determine whether cotinine levels in semen, urine, and blood provide stronger evidence for an association between cigarette smoking and semen quality than the number of cigarettes smoked per day or years smoked and to assess the role of age, consumption of caffeine and alcohol, and month of specimen donation as potential confounders or modifiers of the association between smoking and semen quality. It is thought that cotinine, a metabolite of nicotine, may provide a better measure of smoke exposure than the number of cigarettes smoked per day. For example, infant birthweight was correlated more highly with mothers' serum cotinine levels than with the number of cigarettes reportedly smoked per day (4). The one study that used cotinine levels as the measure of tobacco smoke exposure to assess the Vine et a!.
Cigarette smoking and semen quality
835
effect of smoking on semen quality noted an association between seminal plasma cotinine levels and decreased sperm motility (5). Age, consumption of caffeine and alcohol, and season have been related to semen quality in various studies, but few studies have assessed these factors with regard to the association between cigarette smoking and semen quality. Leydig cell and Sertoli cell function as well as plasma T levels, sperm production, and semen quality decrease with age in men (6). Simon et al. (7) indicate that such decreases begin to occur among men in their 20s. In vitro studies show that caffeine stimula>;es human sperm motility in a dose-dependent manner (8) and coffee consumption has been shown to increase the risk of dyspermia (9). Although there are conflicting reports of the effects of moderate alcohol consumption on sperm quality (1, 9, 10), chronic high level consumption has been shown to decrease sperm concentration and increase the percent of abnormal forms in men with alcohol dependence syndrome (11). Seasonal variation also affects semen quality. Total sperm count, sperm concentration, and motility decrease most during the summer months, whereas the percent of abnormal sperm increases (12). Assessment of these factors as potential confounders and modifiers may help to clarify the association between cigarette smoking and semen quality. MATERIALS AND METHODS Study Participants Inclusion and Exclusion Criteria
Participants eligible for inclusion in this cross-sectional study were healthy white male volunteers, aged 18 to 35 years. Participants were restricted to white males in this age range to reduce the possibility of confounding by race and age. In addition, this age range was chosen because it is of greatest interest with respect to reproductive activity. Participants were excluded if they reported that they did not have a high school diploma or graduate equivalent degree; had ever been treated with radiation or chemotherapy in the past; had used recreational drugs including marijuana, cocaine or narcotics in the previous 3 months; had been hospitalized or missed school or work for a week or more during the previous 3 months; had ever had a vasectomy or a testis removed; had hepatitis B or tested positive for human immunodeficiency virus; or currently used tobacco products other than cigarettes or had used tobacco products more than once per month during the previous year. 836
Vine et al. Cigarette smoking and semen quality
Participant Recruitment and Enrollment
Study participants were recruited via newspaper advertisements in the Chapel Hill, North Carolina area between March and October of 1989. Eightyeight male volunteers (40 nonsmokers, 23 light smokers, and 25 heavy smokers) provided a blood, urine, and semen specimen and completed a selfadministered written questionnaire. Nonsmokers were defined as men who had not smoked 100 cigarettes in their lifetime. Light smokers included men who smoked between 1 and 19 cigarettes per day on average and heavy smokers were defined as men who currently smoked :=::20 cigarettes per day. All smokers had smoked cigarettes for at least 1 year. Heavy smokers were enrolled in the study first. Light smokers and nonsmokers were selected based on the age distribution of the heavy smokers (18 to 21, 22 to 25,26 to 29, and :=::30 years). An attempt was made to have heavy, light, and nonsmokers of the same age group provide specimens within a 2-week period of time. Individuals who did not show up for their appointments account for the imbalance in the age distribution between groups. Participants were paid $50 for providing blood, urine, and semen specimens and completing the questionnaire. Specimen Collection
All specimens were provided at the Reproductive Endocrinology-Fertility Laboratory at the University of North Carolina at Chapel Hill School of Medicine. Participants had been instructed to abstain from ejaculation for 3 to 5 days before their appointment to provide specimens. Semen and urine specimens were collected into separate sterile polypropylene containers. Semen specimens were collected via masturbation. Blood (45 mL) was drawn via venipuncture into clean heparinized vacutainers by trained laboratory personnel. Semen Analysis
An experienced laboratory technician at the Reproductive Endocrinology-Fertility Laboratory performed a standard clinical semen analysis using World Health Organization (WHO) criteria (13). The semen analysis included a determination of the sperm density (the number of cells/mL of semen), total sperm count, semen volume, motile sperm concentration (the number of motile cells/mL of semen), total number of motile sperm, percent of sperm with progressive motility (those moving forward), the percent of dead sperm, and the percent of abnormal sperm. Categories of abnormal sperm included amorphous sperm, coiled tails, double heads, cytoFertility and Sterility
plasmic droplets, bent midpieces, pin heads, double tails, and immature forms. Seventy-five percent of semen analyses were performed within 20 to 30 minutes of specimen donation and all but one specimen was analyzed within 1 hour. The time from specimen donation to analysis was missing for four nonsmokers and two light smokers. Cotinine Analyses
One milliliter samples of serum, urine, and semen supernatant for each study participant were analyzed for cotinine via RIA (14). Sample preparation and analysis was carried out according to procedures in Vine et aL (15). Semen cotinine values were not obtained for two study participants: a light smoker, because he did not produce enough specimen, and a nonsmoker, because of a laboratory accident that resulted in spillage of the sample. A 1-mL sample of urine from each participant also was analyzed for creatinine concentration at the University of North Carolina Hospitals Clinical Laboratory. Laboratory and semen analyses were conducted by personnel who were unaware of the characteristics of the participants from whom the specimens were provided. Questionnaire
Smoking and demographic information as well as alcohol and caffeine consumption were obtained with the self-administered questionnaire. A trained research assistant reviewed the questionnaire with each participant at the time of his clinic appointment to ensure completeness. Alcohol consumption was defined as the sum of the number of 12 oz. (355 mL) bottles of beer, 4 oz. (118 mL) glasses of wine, 12 oz. wine coolers, and shots of hard liquor consumed per week during the previous 3 months. Caffeine consumption was expressed as equivalent 8 oz. (236 mL) cups of coffee. This variable was calculated as the sum of the number of8 oz. cups of coffee + (0.5 X the number of 8 oz. cups of tea) + (0.25 x the number of 12 oz. glasses of caffeinated soft drinks) consumed per day during the previous 3 months. Relative amounts of caffeine in coffee, tea, and soft drinks were derived from values published by Pennington and Church (16). 1\
I
Statistical Analyses
1.1
!II
Categorical, correlation, and regression analyses were conducted to assess the association between cigarette smoking and semen quality. Analysis of variance was used to assess the association between the categorical variables, smoking status (nonsmoking, light smoking, or heavy smoking), and number of years smoked (0, less than the median, or greater Vol. 65, No.4, April 1996
than or equal to the median) with semen quality. Pearson correlation coefficients were calculated for the associations between continuous smoking variables (number of cigarettes smoked per day; semen, blood, and urine cotinine levels; and number of years smoked) and semen quality parameters. Cotinine values and semen quality parameters (except semen volume, motile progression, and percent abnormal sperm) were log-transformed in the correlation and regression analyses to produce measures that were less skewed and had more homogeneous variances than the unlogged measures. Because the correlation coefficient for the association between logged urine cotinine and logged urine cotinine divided by creatinine concentration was so high (r = 0.98), it was felt that analyses involving urine cotinine concentration need not be corrected for creatinine concentration. The associations between smoking dose (as indicated by the number of cigarettes smoked per day, years smoked, and logged cotinine levels) and the various semen quality parameters were stratified by tertiles of age, caffeine and alcohol consumption, and also by month of specimen donation (April to May, June to July, or August to October) to assess the role of these variables as modifiers of the association between smoking and semen quality. Linear regression analyses were conducted to confirm associations between smoking variables and semen quality parameters controlling for potential confounders and effect modifiers. RESULTS
Descriptive information regarding the study population by smoking status (nonsmoking, light smoking, and heavy smoking) is presented in Table 1. The number of cigarettes smoked per day, cotinine levels, number of years smoked, and coffee and soft drink consumption as well as overall caffeine and alcohol consumption were significantly higher among smokers than nonsmokers. Heavy smokers were, on average, 2 to 3 years older than light and nonsmokers. Tea consumption, the time from specimen donation to semen analysis, the number of days since the last ejaculation, and the month of specimen donation did not differ significantly by smoking status. In Table 2, the association between smoking status and semen quality parameters is presented. Analysis of variance indicated that smoking status (nonsmoking, light smoking, or heavy smoking) was not statistically significantly associated with any of the semen quality parameters evaluated. There was, however, a significant association between the number of years smoked (categorized as 0, <8, and :=::8 years [the median number of years smoked among Vine et al. Cigarette smoking and semen quality
837
Table 1 Descriptive Characteristics of the Study Population by Smoking Status* Nonsmokers (n = 40) No. of cigarettes per day Urine cotinine (nglmL) (logged) Years smoked Age (y) Coffee (8 oz. cups/d) Tea (8 oz. cups/d) Soft drinks (12 oz. glasses/d) Caffeine:j: (Equivalent 8 oz. cups of coffee/d) Alcohol (drinks/week) Days since last ejaculation Time from specimen donation to semen analysis II (h) Month of specimen donation~ April to May June to July August to October
0 ::':0 3.1 ::': 1.4 0 ::':0 24.3::': 4.6 0.5::': 0.8 0.6::': 0.9 1.2 ::': 1.1 1.1 ::': 1.0 5.1 ::': 6.4 4.6::': 4.3 0.5::': 0.2 16 (48.5) 13 (43.3) 11 (44.0)
Light smokers (n = 23) 10.8::': 7.7 ::': 5.6::': 23.2::': 1.1 ::': 1.3 ::': 2.3 ::': 2.4 ::': 10.5::': 4.6::': 0.4 ::':
4.9 1.4 3.3 3.7 1.5 2.1 1.7 1.7 9.6 3.3 0.3
10 (30.3) 8 (26.7) 5 (20.0)
Heavy smokers (n = 25) 26.8 ::': 8.6 ::': 10.4 ::': 26.3 ::': 1.7 ::': 1.2 ::': 1.9 ::': 2.8 ::': 10.6::': 4.4 ::': 0.4 ::':
8.0 1.2 6.2 4.8 2.2 1.4 1.5 2.1 11.7 2.4 0.2
7 (21.2) 9 (30.0) 9 (36.0)
P valuet
0.0001 0.0001 0.0001 0.0530 0.0053 NS 0.0075 0.0001 0.0233 NS§ NS NS**
* Values are means::': SD. Light smokers, 1 to 19 cigarettes per day; heavy smokers, ;;;,:20 cigarettes per day. t P value calculated by analysis of variance. :j: Caffeine consumption was calculated as follows: caffeine = number of 8 oz. cups of coffee/day + 0.5 x (number of 8 oz. cups of tea/d) + 0.25 x (number of 12 oz. glasses of caffeinated soft drinks/d). § NS, not significant (P value > 0.05). II Six specimens were missing. ~ Values in parentheses are percentages. ** P value calculated by X2 •
smokers]) and semen quality: sperm density (P = 0.06), total sperm count (P = 0.01), and total number of motile sperm (P = 0.02). Table 3 presents Pearson correlation coefficients for the associations between the continuous measures of smoking (number of cigarettes smoked per day, number of years smoked, and logged cotinine levels in urine, blood, and semen) and the semen quality parameters. All three measures of smoking were associated with decreased semen quality. Logged cotinine levels in the three body fluids and number of years smoked were statistically significantly associated with a decrease in sperm density, total sperm count, total number of motile sperm, and motile sperm concentration. The magnitude of the correlation coefficients was approximately -0.25. Logged cotinine levels in the three fluids were highly
Table 2
Semen Quality Parameters by Smoking Status*
Sperm density (x10 6/mL) Total count (x 106 ) Volume (mL) Motile sperm concentration (x106/mL) Total motile sperm (x 106 ) Motile progression (% moving forward) Percent dead Percent abnormal
Nonsmokers (n = 40)
Light smokers (n = 23)
93.9::': 276.2 ::': 3.1::': 47.8::': 135.3 ::': 52.8::': 9.8::': 22.3::':
73.5::': 165.4 ::': 2.5::': 37.7::': 83.4::': 48.9::': 13.8::': 22.4::':
70.2 246.2 1.8 32.9 109.8 17.2 9.2 9.9
* Values are means::': SD. Light smokers, 1 to 19 cigarettes per day; heavy smokers, ;;;,:20 cigarettes per day. 838
correlated (r = 0.9; P < 0.0001). Urine cotinine levels were an order of magnitude or more higher than those of the other two fluids (15). The magnitude of the correlation coefficients for the associations between the number of cigarettes smoked per day and the semen quality parameters were, in many cases, only slightly lower than those for the associations between logged cotinine levels and semen quality parameters but were less likely to be statistically significant. There was a positive association between the number of years smoked and the percent of abnormal sperm. This was due to the association only among those ;:::26 years of age (r = 0.48; P < 0.01). An association in this age group also was found with the other measures of smoking dose. The number of cigarettes smoked per day and the number of years smoked were highly correlated (r = 0.83). Multiplica-
Vine et al. Cigarette smoking and semen quality
55.3 126.0 1.0 33.8 73.3 14.8 10.5 8.8
Heavy smokers (n = 25) 71.6::': 189.7 ::': 2.9::': 38.5::': 100.1::': 54.0::': 10.2::': 24.2::':
59.7 157.5 1.3 35.0 84.5 16.3 10.0 10.9
P valuet
0.30 0.07 0.28 0.41 0.09 0.53 0.27 0.72
t Analysis of variance.
Fertility and Sterility
Table 3 Pearson Correlation Coefficients for the Associations Between Smoking Measures and Log-Transformed Semen Quality Parameters* No. of cigarettes per day (n = 88) Sperm density (x10 6/mL) Total count (X 106 ) Volume (mL)t Motile sperm concentration (x10 6/mL) Total motile sperm (X10 6 ) Motile progressiont (% moving forward) Percent dead Percent abnormalt
-0.21 (0.04) -0.18 -0.05 -0.21 (0.05) -0.16 -0.04 0.04 0.17
Logged cigarettes per day (n = 88) -0.25 -0.24 -0.09 -0.25 -0.22 -0.05 0.11 0.09
(0.02) (0.02) (0.02) (0.04)
Logged blood cotinine (n = 88) -0.26 -0.24 -0.01 -0.26 -0.21 -0.04 0.15 0.10
(0.01) (0.02) (0.01) (0.05)
Logged urine cotinine (n = 88) -0.23 -0.28 -0.15 -0.24 -0.26 -0.06 0.13 0.11
(0.03) (0.01) (0.03) (0.02)
Logged semen cotinine (n = 86) -0.22 -0.24 -0.10 -0.24 -0.22 -0.08 0.19 0.11
(0.04) (0.03) (0.02) (0.04)
No. of years smoked (n = 88) -0.26 -0.31 -0.08 -0.27 -0.29 -0.06 0.03 0.24
(0.01) (0.01) (0.01) (0.01) (0.03)
* Values in parentheses are P values. Those not shown are >0.05.
t Unlogged values were used because original parameter distributions were relatively normal.
tion of smoking dose by years smoked did not improve correlations between smoking variables and semen quality parameters. Stratification of the association between smoking variables and semen quality parameters was performed separately by age, caffeine intake, alcohol consumption, and month of specimen donation. Only stratification by age appeared to have a strong modifying effect on the association between smoking and semen quality. Stratification of the association between smoking (as indicated by the number of cigarettes smoked per day, years smoked, and logged cotinine levels) and semen quality parameters by tertiles of age «22, 22 to 25, 20:26) revealed an association between smoking and semen quality among men 20:22 years of age but not among the younger men. The magnitude of the association was similar for the upper two tertiles of age. For example, among men 20:22 years of age, logged urine cotinine levels were statistically significantly associated with sperm density (r = -0.39, P = 0.002), total sperm count (r = -0.41, P = 0.001), the total number of motile sperm (r = -0.38, P = 0.003), and the concentration of motile sperm (r = -0.40, P = 0.002). The correlation coefficients for the men < 22 years were r = 0.13, r = 0.06, r = 0.12, and r = 0.20, respectively (none was statistically significant). Smoking was associated with an elevation in the percent of abnormal sperm among men 20: 26 years of age (e.g., r = 0.43; P = 0.01: logged urine cotinine). The magnitude of the above associations was often just slightly lower when the number of cigarettes smoked per day and years smoked were used as the measures of cigarette smoking as when logged cotinine values were used. Figures lA and B present the scatter plots of the log-transformed values of sperm density and total number of motile sperm by logged urine cotinine for men < 22 and 20:22 years of age. Linear regression modeling (including an interaction term between age
category <22, 20:22 and logged urine cotinine) confirmed the associations between logged urine cotinine levels and the logged semen quality parameters, total count, total number of motile sperm, and motile sperm concentration, among the men 20: 22 years. All associations persisted when controlling for caffeine intake, alcohol consumption, and the time between specimen donation and clinical semen analysis (for the motility parameters). Month of specimen donation was not considered a potential confounder because analyses had shown that it was neither associated with the smoking measures nor the semen quality parameters. It was not possible to determine if there was an association between duration of smoking and semen quality independent of the effect of smoking dose because the number of cigarettes smoked per day and the number of years smoked were correlated so highly. Smoking dose as determined by logged cotinine levels produced models with slightly better R 2 s than models including the number of cigarettes smoked per day, the number of years smoked, or the multiplication of smoking dose by duration (logged sperm density: R2 = 0.15 as opposed to 0.11, 0.13, or 0.12, respectively, among those 20: 22 years).
Vol. 65, No.4, April 1996
DISCUSSION
Cigarette smoking is associated with lowered semen quality (including sperm density, total sperm count, total number of motile sperm, and motile sperm concentration) among healthy volunteers 20: 22 years of age. This association persisted after controlling for caffeine and alcohol consumption and the time from specimen donation to semen analysis. For example, the overall Pearson correlation coefficient for the association between logged urine cotinine and logged sperm density was -0.23 (P = 0.03); the correlation coefficients stratified by age «22, Vine et al.
Cigarette smoking and semen quality
839
Age <22
A
l 7
l 7
~ 6
•
5
s ~ 4 r
m
3
d
e
.
2
n s i
0
. . ...-... .. ..
0
d
. . , .. . . . . .·
.
•
T
t 4
a I
3 M 0
i I
• p
0 2
3
4
5
6
7
8
9
..
••
5
0
I
10
Logged Urine Cotinine (ng/mL)
•r
2 1 0 1
2
3
m
7
. . . . .. ..--... • •.. .. . .. · •• . . .... .. . .. ... ~ . . .. . . .. .
e d 5
'I •
s ~ 4 r m
3
d e n
2
s I I y
. .. . . . "........ .. . · .
0
6
1
0 1
2
3
4
5
6
6
8
7
10
9
Age 22+ l 7
L
9 9
5
4
Logged Urine Cotinine (ng/mL)
Age 22+ 0
. .. . . . . .. . . · .
.... . . · . ·
d
s 1
.
9 9 6
1
I y
Age <22
B
7
8
9
10
Logged Urine Colinine (ng/mL)
9
9 6
•d
T 5 0
!
4
I M
3
0
I
I I
2
.
.. ..•. •.. . • .. ..
..
. ·· ..
•
.
. .
•
s 1
·
.
..
.
p
r m
0 1
2
3
4
5
6
7
8
9
10
Logged Urine Cotinine (ng/mL)
Figure 1 (A), Sperm density (x106 cells/mLl versus logged urine cotinine (ng/mLl by age group: age < 22 years, r = 0.13, P = 0.5; age"" 22 years, r = -0.39, P = 0.002. (Bl, Total motile sperm (x10 6 l versus logged urine cotinine (ng/mLl by age group: age < 22 years, r = 0.12, P = 0.6; age"" 22 years, r = -0.38, P = 0.003.
~22)
were r = 0.13 (P = 0.52) and r = -0.39 (P 0.002), respectively. There was also an increase in the percent of abnormal forms of sperm among men ~ 26 years of age (r = 0.43; P = 0.01). The magnitude ofthe effect of smoking on semen quality was similar whether the number of cigarettes smoked per day, the number of years smoked, or logged cotinine levels in body fluids were used as the measure of tobacco smoke exposure. Categorization of smoking status (nonsmoker, light smoker, or heavy smoker) as was done in many of the early studies did not reveal an association between smoking and semen quality. Categorical measures of smoking status have less statistical power than continuous measures of smoking dose. Categorical measures of years smoked were, however, statistically significantly associated with semen quality. Previous studies of the effect of smoking on sperm =
840
Vine et al.
Cigarette smoking and semen quality
density, motility and morphology have noted contradictory results (1, 2). Vine et al. (3) using meta-analysis techniques showed that overall, smokers sperm density is, on average, 13% to 17% lower than that of nonsmokers. A more recent study assessed the relationship between smoking and semen quality using cotinine levels as the measure of tobacco smoke exposure and found a significant effect of semen cotinine on forward motility (r = -0.31; P < 0.005) among men 18 to 45 years of age who were attending an infertility clinic (5). Chia et al. (2) reported that dose times duration of smoking was associated with a decrease in semen quality. Results of this study agree with findings from several other studies that moderate alcohol consumption does not significantly affect semen quality (11, 10, 17). Furthermore, alcohol consumption does not affect the association between smoking and semen Fertility and Sterility
quality. Month of specimen donation was also not a significant factor. Specimen collection took place over a 6-month period between April and the beginning of October in North Carolina. Other studies have found that the most significant variation in semen quality occurs between the winter and summer months (12). Drinking more than 4 cups of coffee and smoking more than 20 cigarettes per day has been found to increase the percent of dead sperm and decrease sperm motility (17). Because only four individuals both smoked 2:20 cigarettes per day and drank the caffeine equivalent of 4 cups of coffee per day, it was not possible to assess this association directly in the present study. However, the strength of the association between smoking and semen quality did not increase with increasing tertiles of caffeine consumption «0.95,0.95 to 2.13, >2.13 equivalent cups of coffee per day) either among those <22 or 2:22 years of age. A possible explanation for the lack of an association between cigarette smoking and semen quality measures among those <22 years of age is that few heavy smokers were included in that age group. Only 4 of 13 (30%) smokers among those < 22 years were heavy smokers, whereas 21 of 35 (60%) smokers among those 2: 22 years were heavy smokers. Furthermore, Kriebel et al. (18) noted that younger smokers tend to smoke cigarettes with a lower content of tar per cigarette and that tar intake is correlated with urine mutagenicity. However, in our study, the average tar content of the brands of cigarettes smoked by the younger men was just slightly lower than that of the men 2: 22 years of age (19). Also, because age was associated with duration of smoking (r = 0.51), the men in the older age group, on average, smoked for more years than younger smokers. On the other hand, one cannot rule out the possibility of a biologically relevant combined effect of increasing age and cigarette smoke exposure on the decline in semen quality. Although an association between smoking and measures of semen quality has been noted, the mechanism by which cigarette smoking might affect semen quality is not known. Evidence suggests that nicotine and/or other compounds in tobacco smoke may be involved. Nicotine exposure in rats leads to atrophy of the testis and impaired spermatogenesis (20). The fact that nicotine and cotinine, a metabolite of nicotine, are detectable in semen (5, 15) suggests that perhaps other components oftobacco smoke are able to enter the semen and affect sperm development. For example, polonium-210, an a-emitting radioelement capable of damaging DNA and found in tobacco smoke, has been detected at higher concentrations in the semen of smokers as compared with nonsmokers (21). Exposure of male mice in utero Vol. 65, No.4, April 1996
to benzo(a)pyrene, a constituent of tobacco smoke, results in decreased testicular weight, atrophy ofthe seminiferous tubules, and altered spermatogenesis (22). Studies in humans indicate that cigarette smoke alters levels of hormones that are involved in spermatogenesis (23). Although the effect of smoking on semen quality is modest, given that 30% to 50% of the world's males smoke (24), smoking could have a significant negative effect on the distribution of semen quality in the population. Smokers with marginal semen quality may be pushed into the infertile range. More research is needed to determine the mechanisms and the extent to which smoking and other factors acting together affect sperm development. Acknowledgments. We gratefully acknowledge Christine Cabot, M.S., formerly of the Reproductive Endocrinology-Fertility Laboratory at the University of North Carolina Medical School, Chapel Hill, North Carolina for her help in performing the clinical semen analyses for this study. We thank Barbara S. Hulka, M.D., M.P.H., of the Department of Epidemiology at the University of North Carolina, Chapel Hill, North Carolina for her thoughtful suggestions upon review of the manuscript. REFERENCES 1. Oldereid NB, Rui H, Purvis K. Lifestyles of men in barren couples and their relationships to sperm quality. Eur J Obstet Gynecol Reprod BioI 1992;43:51-7. 2. Chia SE, Ong CN, Tsakok FM. Effects of cigarette smoking on human semen quality. Arch Androl 1994;33:163-8. 3. Vine MF, Margolin BH, Morrison HI, Hulka BS. Cigarette smoking and sperm density: a meta-analysis. Fertil Steril 1994;61:35-43. 4. Haddow JE, Knight GJ, Paolmaki GE, Kloza EM, Wald NJ. Cigarette consumption and serum cotinine in relation to birthweight. Br J Obstet Gynaecol 1987;94:678-81. 5. Pacifici R, Altieri I, Gandini L, Lenzi A, Pichini, Rosa M, et al. Nicotine, cotinine, and trans-3-hydroxycotinine levels in seminal plasma of smokers: effects on sperm parameters. Ther Drug Monit 1993; 15:358-63. 6. Vermeulen A. Environment, human reproduction, menopause and andropause. Environ Health Perspect 1993; 101 SuppI2:91-100. 7. Simon D, Preziosi P, Barrett-Conner E, Roger M, Saint-Paul M, Nahoul K, et al. The influence of aging on plasma sex hormones in men: the Telecom Study. Am J Epidemiol 1992; 135:783-91. 8. Imoedemhe DA, Sigue AB, Pacpaco EL, Olazo AB. The effect of caffeine on the ability of spermatozoa to fertilize mature human oocytes. J Assist Reprod Genet 1992;9:155-60. 9. Parazzini F, Marchini M, Tozzi L, Mezzopane R, Fedele L. Risk factors for unexplained dyspermia in infertile men: a case-control study. Arch Androl 1993;31:105-13. 10. Dunphy BC, Barratt CL, Cooke ID. Male alcohol consumption and fecundity in couples attending an infertility clinic. Andrologia 1991;23:219-21. 11. Kucheria K, Saxena R, Mohan D. Semen analysis in alcohol dependence syndrome. Andrologia 1985;17:558-63. 12. Levine RJ. Male factors contributing to the seasonality of human reproduction. Ann NY Acad Sci 1994; 709:29-45. 13. World Health Organization. WHO laboratory manual for the
Vine et al.
Cigarette smoking and semen quality
841
14.
15.
16.
17.
18.
842
examination of human semen and sperm-cervical mucus interaction. 3rd ed. New York: Cambridge University Press, 1993. Van Vunakis H, Gjika HB, Langone JJ. Method 16-radioimmunoassay for nicotine and cotinine. rARC Sci Pub 1987; 81:317-30. Vine MF, Hulka BS, Margolin BH, Truong YK, Hu PC, Schramm MM. Cotinine concentrations in semen, urine, and blood of smokers and nonsmokers. Am J Public Health 1993; 83:1335-8. Pennington JAT, Church HN. Bowes and Church's food values of portions commonly used. N ew York: Harper and Row Publishers, 1985. Marshburn PB, Sloan CS, Hammond MG. Semen quality and association with coffee drinking, cigarette smoking, and ethanol consumption. Fertil Steril 1989;52:162-5. Kriebel D, Henry J, Gold JC, Bronsdon A, Commoner B. The
Vine et al. Cigarette smoking and semen quality
19.
20. 21. 22. 23.
24.
mutagenicity of cigarette smokers' urine. J Environ Pathol Toxicol Oncol 1985; 6: 157 -70. United States Federal Trade Commission. Report of "tar" and nicotine content of the smoke of 176 varieties of cigarettes. Washington, D.C.: Federal Trade Commission, 1979. Weisberg E. Smoking and reproductive health. Clin Reprod Fertil 1985;3:175-86. Hunt V. Polonium-210 measurements in human semen. Health Phys 1990;58:511-4. MacKenzie KM, Angevine DM. Infertility in mice exposed in utero to benzo(a)pyrene. BioI Reprod 1981;24:183-91. Field AE, Colditz GA, Willett WC, Longcope C, McKinlay JB. The relation of smoking, age, relative weight, and dietary intake to serum adrenal steroids, sex hormones and sex hormone-binding globulin in middle-aged men. J Clin Endocrinol Metab 1994;79:1310-6. Liihdetie J. Cigarette smoking-a possible genotoxic hazard to male germ cells? Int J Androl 1987; 10:431-4.
Fertility and Sterility
©
Vol. 65, No.4, April 1996
Printed on acid-free paper in U. S. A.
1996 American Society for Reproductive Medicine
Cigarette smoking and semen quality
Marilyn F. Vine, Ph.D.t:!: Chiu-Kit J. Tse, M.8.P.H.t Ping-Chuan Hu, Ph.D.§ K. Young Truong, Ph.D.11 The University of North Carolina, Chapel Hill, North Carolina
Objective: To determine whether cotinine levels provide stronger evidence for an association between smoking and semen quality than the number of cigarettes smoked per day or years smoked controlling for potential confounders and effect modifiers. Design: Cross-sectional study. Setting: Male volunteers at the Reproductive Endocrinology-Fertility Laboratory. Participants: Eighty-eight men (ages 18 to 35 years) provided a semen, urine, and blood specimen and completed a self-administered questionnaire concerning smoking and demographic information as well as caffeine and alcohol consumption. Urine, blood, and semen cotinine levels were analyzed via RIA. Main Outcome Measure: Standard clinical semen analysis. Results: Number of cigarettes smoked per day, years smoked, and log-transformed cotinine levels were associated negatively with semen quality (density, total count, and motility). The association was evident among men age ~22 years. For example, the correlation coefficient for the overall association between logged urine cotinine and logged sperm density was -0.23; those stratified by age were 0.13 (age <22 years) and -0.39 (age ~22 years). Potential confounders included in regression models did not diminish the associations. Conclusions: Smoking is associated with lowered semen quality. Fertil Steril 1996;65:835-42 Key Words: Smoking, sperm, semen quality, male, human
Many studies have assessed the association between cigarette smoking and sperm density, motility, and morphology with mixed results (1-3). Inadequate consideration of smoking dose or the potential Received May 3, 1995; revised and accepted September 26, 1995. * Supported by Cooperative Agreement RR13485-01 between Health Effects Research Laboratory, United States Environmental Protection Agency and the University of North Carolina, Department of Epidemiology. Although the research described in this article has been supported by the United States Environmental Protection Agency, it has not been subjected to Agency review and, therefore, does not necessarily reflect the view of the Agency and no official endorsement should be inferred. t Department of Epidemiology, School of Public Health. :j: Reprint requests: Marilyn F. Vine, Ph.D., Department of Epidemiology, School of Public Health, CB No. 7400 McGavranGreenberg Hall, University of North Carolina, Chapel Hill, North Carolina 27599-7400 (FAX: 919-966-2089). § Department of Pediatrics, School of Medicine. II Department of Biostatistics, School of Public Health. Vol. 65, No.4, April 1996
confounding and modifying effects of such factors as age as well as lack of statistical power in some studies may account for the contradictory findings. The primary goals ofthis study are to determine whether cotinine levels in semen, urine, and blood provide stronger evidence for an association between cigarette smoking and semen quality than the number of cigarettes smoked per day or years smoked and to assess the role of age, consumption of caffeine and alcohol, and month of specimen donation as potential confounders or modifiers of the association between smoking and semen quality. It is thought that cotinine, a metabolite of nicotine, may provide a better measure of smoke exposure than the number of cigarettes smoked per day. For example, infant birthweight was correlated more highly with mothers' serum cotinine levels than with the number of cigarettes reportedly smoked per day (4). The one study that used cotinine levels as the measure of tobacco smoke exposure to assess the Vine et a!.
Cigarette smoking and semen quality
835
effect of smoking on semen quality noted an association between seminal plasma cotinine levels and decreased sperm motility (5). Age, consumption of caffeine and alcohol, and season have been related to semen quality in various studies, but few studies have assessed these factors with regard to the association between cigarette smoking and semen quality. Leydig cell and Sertoli cell function as well as plasma T levels, sperm production, and semen quality decrease with age in men (6). Simon et al. (7) indicate that such decreases begin to occur among men in their 20s. In vitro studies show that caffeine stimula>;es human sperm motility in a dose-dependent manner (8) and coffee consumption has been shown to increase the risk of dyspermia (9). Although there are conflicting reports of the effects of moderate alcohol consumption on sperm quality (1, 9, 10), chronic high level consumption has been shown to decrease sperm concentration and increase the percent of abnormal forms in men with alcohol dependence syndrome (11). Seasonal variation also affects semen quality. Total sperm count, sperm concentration, and motility decrease most during the summer months, whereas the percent of abnormal sperm increases (12). Assessment of these factors as potential confounders and modifiers may help to clarify the association between cigarette smoking and semen quality. MATERIALS AND METHODS Study Participants Inclusion and Exclusion Criteria
Participants eligible for inclusion in this cross-sectional study were healthy white male volunteers, aged 18 to 35 years. Participants were restricted to white males in this age range to reduce the possibility of confounding by race and age. In addition, this age range was chosen because it is of greatest interest with respect to reproductive activity. Participants were excluded if they reported that they did not have a high school diploma or graduate equivalent degree; had ever been treated with radiation or chemotherapy in the past; had used recreational drugs including marijuana, cocaine or narcotics in the previous 3 months; had been hospitalized or missed school or work for a week or more during the previous 3 months; had ever had a vasectomy or a testis removed; had hepatitis B or tested positive for human immunodeficiency virus; or currently used tobacco products other than cigarettes or had used tobacco products more than once per month during the previous year. 836
Vine et al. Cigarette smoking and semen quality
Participant Recruitment and Enrollment
Study participants were recruited via newspaper advertisements in the Chapel Hill, North Carolina area between March and October of 1989. Eightyeight male volunteers (40 nonsmokers, 23 light smokers, and 25 heavy smokers) provided a blood, urine, and semen specimen and completed a selfadministered written questionnaire. Nonsmokers were defined as men who had not smoked 100 cigarettes in their lifetime. Light smokers included men who smoked between 1 and 19 cigarettes per day on average and heavy smokers were defined as men who currently smoked :=::20 cigarettes per day. All smokers had smoked cigarettes for at least 1 year. Heavy smokers were enrolled in the study first. Light smokers and nonsmokers were selected based on the age distribution of the heavy smokers (18 to 21, 22 to 25,26 to 29, and :=::30 years). An attempt was made to have heavy, light, and nonsmokers of the same age group provide specimens within a 2-week period of time. Individuals who did not show up for their appointments account for the imbalance in the age distribution between groups. Participants were paid $50 for providing blood, urine, and semen specimens and completing the questionnaire. Specimen Collection
All specimens were provided at the Reproductive Endocrinology-Fertility Laboratory at the University of North Carolina at Chapel Hill School of Medicine. Participants had been instructed to abstain from ejaculation for 3 to 5 days before their appointment to provide specimens. Semen and urine specimens were collected into separate sterile polypropylene containers. Semen specimens were collected via masturbation. Blood (45 mL) was drawn via venipuncture into clean heparinized vacutainers by trained laboratory personnel. Semen Analysis
An experienced laboratory technician at the Reproductive Endocrinology-Fertility Laboratory performed a standard clinical semen analysis using World Health Organization (WHO) criteria (13). The semen analysis included a determination of the sperm density (the number of cells/mL of semen), total sperm count, semen volume, motile sperm concentration (the number of motile cells/mL of semen), total number of motile sperm, percent of sperm with progressive motility (those moving forward), the percent of dead sperm, and the percent of abnormal sperm. Categories of abnormal sperm included amorphous sperm, coiled tails, double heads, cytoFertility and Sterility
plasmic droplets, bent midpieces, pin heads, double tails, and immature forms. Seventy-five percent of semen analyses were performed within 20 to 30 minutes of specimen donation and all but one specimen was analyzed within 1 hour. The time from specimen donation to analysis was missing for four nonsmokers and two light smokers. Cotinine Analyses
One milliliter samples of serum, urine, and semen supernatant for each study participant were analyzed for cotinine via RIA (14). Sample preparation and analysis was carried out according to procedures in Vine et aL (15). Semen cotinine values were not obtained for two study participants: a light smoker, because he did not produce enough specimen, and a nonsmoker, because of a laboratory accident that resulted in spillage of the sample. A 1-mL sample of urine from each participant also was analyzed for creatinine concentration at the University of North Carolina Hospitals Clinical Laboratory. Laboratory and semen analyses were conducted by personnel who were unaware of the characteristics of the participants from whom the specimens were provided. Questionnaire
Smoking and demographic information as well as alcohol and caffeine consumption were obtained with the self-administered questionnaire. A trained research assistant reviewed the questionnaire with each participant at the time of his clinic appointment to ensure completeness. Alcohol consumption was defined as the sum of the number of 12 oz. (355 mL) bottles of beer, 4 oz. (118 mL) glasses of wine, 12 oz. wine coolers, and shots of hard liquor consumed per week during the previous 3 months. Caffeine consumption was expressed as equivalent 8 oz. (236 mL) cups of coffee. This variable was calculated as the sum of the number of8 oz. cups of coffee + (0.5 X the number of 8 oz. cups of tea) + (0.25 x the number of 12 oz. glasses of caffeinated soft drinks) consumed per day during the previous 3 months. Relative amounts of caffeine in coffee, tea, and soft drinks were derived from values published by Pennington and Church (16). 1\
I
Statistical Analyses
1.1
!II
Categorical, correlation, and regression analyses were conducted to assess the association between cigarette smoking and semen quality. Analysis of variance was used to assess the association between the categorical variables, smoking status (nonsmoking, light smoking, or heavy smoking), and number of years smoked (0, less than the median, or greater Vol. 65, No.4, April 1996
than or equal to the median) with semen quality. Pearson correlation coefficients were calculated for the associations between continuous smoking variables (number of cigarettes smoked per day; semen, blood, and urine cotinine levels; and number of years smoked) and semen quality parameters. Cotinine values and semen quality parameters (except semen volume, motile progression, and percent abnormal sperm) were log-transformed in the correlation and regression analyses to produce measures that were less skewed and had more homogeneous variances than the unlogged measures. Because the correlation coefficient for the association between logged urine cotinine and logged urine cotinine divided by creatinine concentration was so high (r = 0.98), it was felt that analyses involving urine cotinine concentration need not be corrected for creatinine concentration. The associations between smoking dose (as indicated by the number of cigarettes smoked per day, years smoked, and logged cotinine levels) and the various semen quality parameters were stratified by tertiles of age, caffeine and alcohol consumption, and also by month of specimen donation (April to May, June to July, or August to October) to assess the role of these variables as modifiers of the association between smoking and semen quality. Linear regression analyses were conducted to confirm associations between smoking variables and semen quality parameters controlling for potential confounders and effect modifiers. RESULTS
Descriptive information regarding the study population by smoking status (nonsmoking, light smoking, and heavy smoking) is presented in Table 1. The number of cigarettes smoked per day, cotinine levels, number of years smoked, and coffee and soft drink consumption as well as overall caffeine and alcohol consumption were significantly higher among smokers than nonsmokers. Heavy smokers were, on average, 2 to 3 years older than light and nonsmokers. Tea consumption, the time from specimen donation to semen analysis, the number of days since the last ejaculation, and the month of specimen donation did not differ significantly by smoking status. In Table 2, the association between smoking status and semen quality parameters is presented. Analysis of variance indicated that smoking status (nonsmoking, light smoking, or heavy smoking) was not statistically significantly associated with any of the semen quality parameters evaluated. There was, however, a significant association between the number of years smoked (categorized as 0, <8, and :=::8 years [the median number of years smoked among Vine et al. Cigarette smoking and semen quality
837
Table 1 Descriptive Characteristics of the Study Population by Smoking Status* Nonsmokers (n = 40) No. of cigarettes per day Urine cotinine (nglmL) (logged) Years smoked Age (y) Coffee (8 oz. cups/d) Tea (8 oz. cups/d) Soft drinks (12 oz. glasses/d) Caffeine:j: (Equivalent 8 oz. cups of coffee/d) Alcohol (drinks/week) Days since last ejaculation Time from specimen donation to semen analysis II (h) Month of specimen donation~ April to May June to July August to October
0 ::':0 3.1 ::': 1.4 0 ::':0 24.3::': 4.6 0.5::': 0.8 0.6::': 0.9 1.2 ::': 1.1 1.1 ::': 1.0 5.1 ::': 6.4 4.6::': 4.3 0.5::': 0.2 16 (48.5) 13 (43.3) 11 (44.0)
Light smokers (n = 23) 10.8::': 7.7 ::': 5.6::': 23.2::': 1.1 ::': 1.3 ::': 2.3 ::': 2.4 ::': 10.5::': 4.6::': 0.4 ::':
4.9 1.4 3.3 3.7 1.5 2.1 1.7 1.7 9.6 3.3 0.3
10 (30.3) 8 (26.7) 5 (20.0)
Heavy smokers (n = 25) 26.8 ::': 8.6 ::': 10.4 ::': 26.3 ::': 1.7 ::': 1.2 ::': 1.9 ::': 2.8 ::': 10.6::': 4.4 ::': 0.4 ::':
8.0 1.2 6.2 4.8 2.2 1.4 1.5 2.1 11.7 2.4 0.2
7 (21.2) 9 (30.0) 9 (36.0)
P valuet
0.0001 0.0001 0.0001 0.0530 0.0053 NS 0.0075 0.0001 0.0233 NS§ NS NS**
* Values are means::': SD. Light smokers, 1 to 19 cigarettes per day; heavy smokers, ;;;,:20 cigarettes per day. t P value calculated by analysis of variance. :j: Caffeine consumption was calculated as follows: caffeine = number of 8 oz. cups of coffee/day + 0.5 x (number of 8 oz. cups of tea/d) + 0.25 x (number of 12 oz. glasses of caffeinated soft drinks/d). § NS, not significant (P value > 0.05). II Six specimens were missing. ~ Values in parentheses are percentages. ** P value calculated by X2 •
smokers]) and semen quality: sperm density (P = 0.06), total sperm count (P = 0.01), and total number of motile sperm (P = 0.02). Table 3 presents Pearson correlation coefficients for the associations between the continuous measures of smoking (number of cigarettes smoked per day, number of years smoked, and logged cotinine levels in urine, blood, and semen) and the semen quality parameters. All three measures of smoking were associated with decreased semen quality. Logged cotinine levels in the three body fluids and number of years smoked were statistically significantly associated with a decrease in sperm density, total sperm count, total number of motile sperm, and motile sperm concentration. The magnitude of the correlation coefficients was approximately -0.25. Logged cotinine levels in the three fluids were highly
Table 2
Semen Quality Parameters by Smoking Status*
Sperm density (x10 6/mL) Total count (x 106 ) Volume (mL) Motile sperm concentration (x106/mL) Total motile sperm (x 106 ) Motile progression (% moving forward) Percent dead Percent abnormal
Nonsmokers (n = 40)
Light smokers (n = 23)
93.9::': 276.2 ::': 3.1::': 47.8::': 135.3 ::': 52.8::': 9.8::': 22.3::':
73.5::': 165.4 ::': 2.5::': 37.7::': 83.4::': 48.9::': 13.8::': 22.4::':
70.2 246.2 1.8 32.9 109.8 17.2 9.2 9.9
* Values are means::': SD. Light smokers, 1 to 19 cigarettes per day; heavy smokers, ;;;,:20 cigarettes per day. 838
correlated (r = 0.9; P < 0.0001). Urine cotinine levels were an order of magnitude or more higher than those of the other two fluids (15). The magnitude of the correlation coefficients for the associations between the number of cigarettes smoked per day and the semen quality parameters were, in many cases, only slightly lower than those for the associations between logged cotinine levels and semen quality parameters but were less likely to be statistically significant. There was a positive association between the number of years smoked and the percent of abnormal sperm. This was due to the association only among those ;:::26 years of age (r = 0.48; P < 0.01). An association in this age group also was found with the other measures of smoking dose. The number of cigarettes smoked per day and the number of years smoked were highly correlated (r = 0.83). Multiplica-
Vine et al. Cigarette smoking and semen quality
55.3 126.0 1.0 33.8 73.3 14.8 10.5 8.8
Heavy smokers (n = 25) 71.6::': 189.7 ::': 2.9::': 38.5::': 100.1::': 54.0::': 10.2::': 24.2::':
59.7 157.5 1.3 35.0 84.5 16.3 10.0 10.9
P valuet
0.30 0.07 0.28 0.41 0.09 0.53 0.27 0.72
t Analysis of variance.
Fertility and Sterility
Table 3 Pearson Correlation Coefficients for the Associations Between Smoking Measures and Log-Transformed Semen Quality Parameters* No. of cigarettes per day (n = 88) Sperm density (x10 6/mL) Total count (X 106 ) Volume (mL)t Motile sperm concentration (x10 6/mL) Total motile sperm (X10 6 ) Motile progressiont (% moving forward) Percent dead Percent abnormalt
-0.21 (0.04) -0.18 -0.05 -0.21 (0.05) -0.16 -0.04 0.04 0.17
Logged cigarettes per day (n = 88) -0.25 -0.24 -0.09 -0.25 -0.22 -0.05 0.11 0.09
(0.02) (0.02) (0.02) (0.04)
Logged blood cotinine (n = 88) -0.26 -0.24 -0.01 -0.26 -0.21 -0.04 0.15 0.10
(0.01) (0.02) (0.01) (0.05)
Logged urine cotinine (n = 88) -0.23 -0.28 -0.15 -0.24 -0.26 -0.06 0.13 0.11
(0.03) (0.01) (0.03) (0.02)
Logged semen cotinine (n = 86) -0.22 -0.24 -0.10 -0.24 -0.22 -0.08 0.19 0.11
(0.04) (0.03) (0.02) (0.04)
No. of years smoked (n = 88) -0.26 -0.31 -0.08 -0.27 -0.29 -0.06 0.03 0.24
(0.01) (0.01) (0.01) (0.01) (0.03)
* Values in parentheses are P values. Those not shown are >0.05.
t Unlogged values were used because original parameter distributions were relatively normal.
tion of smoking dose by years smoked did not improve correlations between smoking variables and semen quality parameters. Stratification of the association between smoking variables and semen quality parameters was performed separately by age, caffeine intake, alcohol consumption, and month of specimen donation. Only stratification by age appeared to have a strong modifying effect on the association between smoking and semen quality. Stratification of the association between smoking (as indicated by the number of cigarettes smoked per day, years smoked, and logged cotinine levels) and semen quality parameters by tertiles of age «22, 22 to 25, 20:26) revealed an association between smoking and semen quality among men 20:22 years of age but not among the younger men. The magnitude of the association was similar for the upper two tertiles of age. For example, among men 20:22 years of age, logged urine cotinine levels were statistically significantly associated with sperm density (r = -0.39, P = 0.002), total sperm count (r = -0.41, P = 0.001), the total number of motile sperm (r = -0.38, P = 0.003), and the concentration of motile sperm (r = -0.40, P = 0.002). The correlation coefficients for the men < 22 years were r = 0.13, r = 0.06, r = 0.12, and r = 0.20, respectively (none was statistically significant). Smoking was associated with an elevation in the percent of abnormal sperm among men 20: 26 years of age (e.g., r = 0.43; P = 0.01: logged urine cotinine). The magnitude of the above associations was often just slightly lower when the number of cigarettes smoked per day and years smoked were used as the measures of cigarette smoking as when logged cotinine values were used. Figures lA and B present the scatter plots of the log-transformed values of sperm density and total number of motile sperm by logged urine cotinine for men < 22 and 20:22 years of age. Linear regression modeling (including an interaction term between age
category <22, 20:22 and logged urine cotinine) confirmed the associations between logged urine cotinine levels and the logged semen quality parameters, total count, total number of motile sperm, and motile sperm concentration, among the men 20: 22 years. All associations persisted when controlling for caffeine intake, alcohol consumption, and the time between specimen donation and clinical semen analysis (for the motility parameters). Month of specimen donation was not considered a potential confounder because analyses had shown that it was neither associated with the smoking measures nor the semen quality parameters. It was not possible to determine if there was an association between duration of smoking and semen quality independent of the effect of smoking dose because the number of cigarettes smoked per day and the number of years smoked were correlated so highly. Smoking dose as determined by logged cotinine levels produced models with slightly better R 2 s than models including the number of cigarettes smoked per day, the number of years smoked, or the multiplication of smoking dose by duration (logged sperm density: R2 = 0.15 as opposed to 0.11, 0.13, or 0.12, respectively, among those 20: 22 years).
Vol. 65, No.4, April 1996
DISCUSSION
Cigarette smoking is associated with lowered semen quality (including sperm density, total sperm count, total number of motile sperm, and motile sperm concentration) among healthy volunteers 20: 22 years of age. This association persisted after controlling for caffeine and alcohol consumption and the time from specimen donation to semen analysis. For example, the overall Pearson correlation coefficient for the association between logged urine cotinine and logged sperm density was -0.23 (P = 0.03); the correlation coefficients stratified by age «22, Vine et al.
Cigarette smoking and semen quality
839
Age <22
A
l 7
l 7
~ 6
•
5
s ~ 4 r
m
3
d
e
.
2
n s i
0
. . ...-... .. ..
0
d
. . , .. . . . . .·
.
•
T
t 4
a I
3 M 0
i I
• p
0 2
3
4
5
6
7
8
9
..
••
5
0
I
10
Logged Urine Cotinine (ng/mL)
•r
2 1 0 1
2
3
m
7
. . . . .. ..--... • •.. .. . .. · •• . . .... .. . .. ... ~ . . .. . . .. .
e d 5
'I •
s ~ 4 r m
3
d e n
2
s I I y
. .. . . . "........ .. . · .
0
6
1
0 1
2
3
4
5
6
6
8
7
10
9
Age 22+ l 7
L
9 9
5
4
Logged Urine Cotinine (ng/mL)
Age 22+ 0
. .. . . . . .. . . · .
.... . . · . ·
d
s 1
.
9 9 6
1
I y
Age <22
B
7
8
9
10
Logged Urine Colinine (ng/mL)
9
9 6
•d
T 5 0
!
4
I M
3
0
I
I I
2
.
.. ..•. •.. . • .. ..
..
. ·· ..
•
.
. .
•
s 1
·
.
..
.
p
r m
0 1
2
3
4
5
6
7
8
9
10
Logged Urine Cotinine (ng/mL)
Figure 1 (A), Sperm density (x106 cells/mLl versus logged urine cotinine (ng/mLl by age group: age < 22 years, r = 0.13, P = 0.5; age"" 22 years, r = -0.39, P = 0.002. (Bl, Total motile sperm (x10 6 l versus logged urine cotinine (ng/mLl by age group: age < 22 years, r = 0.12, P = 0.6; age"" 22 years, r = -0.38, P = 0.003.
~22)
were r = 0.13 (P = 0.52) and r = -0.39 (P 0.002), respectively. There was also an increase in the percent of abnormal forms of sperm among men ~ 26 years of age (r = 0.43; P = 0.01). The magnitude ofthe effect of smoking on semen quality was similar whether the number of cigarettes smoked per day, the number of years smoked, or logged cotinine levels in body fluids were used as the measure of tobacco smoke exposure. Categorization of smoking status (nonsmoker, light smoker, or heavy smoker) as was done in many of the early studies did not reveal an association between smoking and semen quality. Categorical measures of smoking status have less statistical power than continuous measures of smoking dose. Categorical measures of years smoked were, however, statistically significantly associated with semen quality. Previous studies of the effect of smoking on sperm =
840
Vine et al.
Cigarette smoking and semen quality
density, motility and morphology have noted contradictory results (1, 2). Vine et al. (3) using meta-analysis techniques showed that overall, smokers sperm density is, on average, 13% to 17% lower than that of nonsmokers. A more recent study assessed the relationship between smoking and semen quality using cotinine levels as the measure of tobacco smoke exposure and found a significant effect of semen cotinine on forward motility (r = -0.31; P < 0.005) among men 18 to 45 years of age who were attending an infertility clinic (5). Chia et al. (2) reported that dose times duration of smoking was associated with a decrease in semen quality. Results of this study agree with findings from several other studies that moderate alcohol consumption does not significantly affect semen quality (11, 10, 17). Furthermore, alcohol consumption does not affect the association between smoking and semen Fertility and Sterility
quality. Month of specimen donation was also not a significant factor. Specimen collection took place over a 6-month period between April and the beginning of October in North Carolina. Other studies have found that the most significant variation in semen quality occurs between the winter and summer months (12). Drinking more than 4 cups of coffee and smoking more than 20 cigarettes per day has been found to increase the percent of dead sperm and decrease sperm motility (17). Because only four individuals both smoked 2:20 cigarettes per day and drank the caffeine equivalent of 4 cups of coffee per day, it was not possible to assess this association directly in the present study. However, the strength of the association between smoking and semen quality did not increase with increasing tertiles of caffeine consumption «0.95,0.95 to 2.13, >2.13 equivalent cups of coffee per day) either among those <22 or 2:22 years of age. A possible explanation for the lack of an association between cigarette smoking and semen quality measures among those <22 years of age is that few heavy smokers were included in that age group. Only 4 of 13 (30%) smokers among those < 22 years were heavy smokers, whereas 21 of 35 (60%) smokers among those 2: 22 years were heavy smokers. Furthermore, Kriebel et al. (18) noted that younger smokers tend to smoke cigarettes with a lower content of tar per cigarette and that tar intake is correlated with urine mutagenicity. However, in our study, the average tar content of the brands of cigarettes smoked by the younger men was just slightly lower than that of the men 2: 22 years of age (19). Also, because age was associated with duration of smoking (r = 0.51), the men in the older age group, on average, smoked for more years than younger smokers. On the other hand, one cannot rule out the possibility of a biologically relevant combined effect of increasing age and cigarette smoke exposure on the decline in semen quality. Although an association between smoking and measures of semen quality has been noted, the mechanism by which cigarette smoking might affect semen quality is not known. Evidence suggests that nicotine and/or other compounds in tobacco smoke may be involved. Nicotine exposure in rats leads to atrophy of the testis and impaired spermatogenesis (20). The fact that nicotine and cotinine, a metabolite of nicotine, are detectable in semen (5, 15) suggests that perhaps other components oftobacco smoke are able to enter the semen and affect sperm development. For example, polonium-210, an a-emitting radioelement capable of damaging DNA and found in tobacco smoke, has been detected at higher concentrations in the semen of smokers as compared with nonsmokers (21). Exposure of male mice in utero Vol. 65, No.4, April 1996
to benzo(a)pyrene, a constituent of tobacco smoke, results in decreased testicular weight, atrophy ofthe seminiferous tubules, and altered spermatogenesis (22). Studies in humans indicate that cigarette smoke alters levels of hormones that are involved in spermatogenesis (23). Although the effect of smoking on semen quality is modest, given that 30% to 50% of the world's males smoke (24), smoking could have a significant negative effect on the distribution of semen quality in the population. Smokers with marginal semen quality may be pushed into the infertile range. More research is needed to determine the mechanisms and the extent to which smoking and other factors acting together affect sperm development. Acknowledgments. We gratefully acknowledge Christine Cabot, M.S., formerly of the Reproductive Endocrinology-Fertility Laboratory at the University of North Carolina Medical School, Chapel Hill, North Carolina for her help in performing the clinical semen analyses for this study. We thank Barbara S. Hulka, M.D., M.P.H., of the Department of Epidemiology at the University of North Carolina, Chapel Hill, North Carolina for her thoughtful suggestions upon review of the manuscript. REFERENCES 1. Oldereid NB, Rui H, Purvis K. Lifestyles of men in barren couples and their relationships to sperm quality. Eur J Obstet Gynecol Reprod BioI 1992;43:51-7. 2. Chia SE, Ong CN, Tsakok FM. Effects of cigarette smoking on human semen quality. Arch Androl 1994;33:163-8. 3. Vine MF, Margolin BH, Morrison HI, Hulka BS. Cigarette smoking and sperm density: a meta-analysis. Fertil Steril 1994;61:35-43. 4. Haddow JE, Knight GJ, Paolmaki GE, Kloza EM, Wald NJ. Cigarette consumption and serum cotinine in relation to birthweight. Br J Obstet Gynaecol 1987;94:678-81. 5. Pacifici R, Altieri I, Gandini L, Lenzi A, Pichini, Rosa M, et al. Nicotine, cotinine, and trans-3-hydroxycotinine levels in seminal plasma of smokers: effects on sperm parameters. Ther Drug Monit 1993; 15:358-63. 6. Vermeulen A. Environment, human reproduction, menopause and andropause. Environ Health Perspect 1993; 101 SuppI2:91-100. 7. Simon D, Preziosi P, Barrett-Conner E, Roger M, Saint-Paul M, Nahoul K, et al. The influence of aging on plasma sex hormones in men: the Telecom Study. Am J Epidemiol 1992; 135:783-91. 8. Imoedemhe DA, Sigue AB, Pacpaco EL, Olazo AB. The effect of caffeine on the ability of spermatozoa to fertilize mature human oocytes. J Assist Reprod Genet 1992;9:155-60. 9. Parazzini F, Marchini M, Tozzi L, Mezzopane R, Fedele L. Risk factors for unexplained dyspermia in infertile men: a case-control study. Arch Androl 1993;31:105-13. 10. Dunphy BC, Barratt CL, Cooke ID. Male alcohol consumption and fecundity in couples attending an infertility clinic. Andrologia 1991;23:219-21. 11. Kucheria K, Saxena R, Mohan D. Semen analysis in alcohol dependence syndrome. Andrologia 1985;17:558-63. 12. Levine RJ. Male factors contributing to the seasonality of human reproduction. Ann NY Acad Sci 1994; 709:29-45. 13. World Health Organization. WHO laboratory manual for the
Vine et al.
Cigarette smoking and semen quality
841
14.
15.
16.
17.
18.
842
examination of human semen and sperm-cervical mucus interaction. 3rd ed. New York: Cambridge University Press, 1993. Van Vunakis H, Gjika HB, Langone JJ. Method 16-radioimmunoassay for nicotine and cotinine. rARC Sci Pub 1987; 81:317-30. Vine MF, Hulka BS, Margolin BH, Truong YK, Hu PC, Schramm MM. Cotinine concentrations in semen, urine, and blood of smokers and nonsmokers. Am J Public Health 1993; 83:1335-8. Pennington JAT, Church HN. Bowes and Church's food values of portions commonly used. N ew York: Harper and Row Publishers, 1985. Marshburn PB, Sloan CS, Hammond MG. Semen quality and association with coffee drinking, cigarette smoking, and ethanol consumption. Fertil Steril 1989;52:162-5. Kriebel D, Henry J, Gold JC, Bronsdon A, Commoner B. The
Vine et al. Cigarette smoking and semen quality
19.
20. 21. 22. 23.
24.
mutagenicity of cigarette smokers' urine. J Environ Pathol Toxicol Oncol 1985; 6: 157 -70. United States Federal Trade Commission. Report of "tar" and nicotine content of the smoke of 176 varieties of cigarettes. Washington, D.C.: Federal Trade Commission, 1979. Weisberg E. Smoking and reproductive health. Clin Reprod Fertil 1985;3:175-86. Hunt V. Polonium-210 measurements in human semen. Health Phys 1990;58:511-4. MacKenzie KM, Angevine DM. Infertility in mice exposed in utero to benzo(a)pyrene. BioI Reprod 1981;24:183-91. Field AE, Colditz GA, Willett WC, Longcope C, McKinlay JB. The relation of smoking, age, relative weight, and dietary intake to serum adrenal steroids, sex hormones and sex hormone-binding globulin in middle-aged men. J Clin Endocrinol Metab 1994;79:1310-6. Liihdetie J. Cigarette smoking-a possible genotoxic hazard to male germ cells? Int J Androl 1987; 10:431-4.
Fertility and Sterility