Cigarette Smoking and Semen Quality

EUROPEAN UROLOGY 70 (2016) 635–645 available at www.sciencedirect.com journal homepage: www.europeanurology.com

Platinum Priority – Review – Andrology Editorial by Christian Leiber and Ulrich Wetterauer on pp. 646–647 of this issue Cigarette Smoking and Semen Quality: A New Meta-analysis Examining the Effect of the 2010 World Health Organization Laboratory Methods for the Examination of Human Semen

Reecha Sharma a, Avi Harlev b,c, Ashok Agarwal c,*, Sandro C. Esteves d a Department of Health Services, Saint Joseph’s University, Philadelphia, PA, USA; b Fertility and IVF Unit, Soroka University Medical Center, Ben Gurion University of the Negev, Beer Sheva, Israel; c American Center for Reproductive Medicine, Cleveland Clinic, Cleveland, OH, USA; d ANDROFERT, Andrology and Human Reproduction Clinic, Referral Center for Male Reproduction, Campinas, SP, Brazil

Article info Abstract

Article history: Objective: Approximately 37% of men of reproductive age smoke cigarettes, with Europe Accepted April 6, 2016 having the highest tobacco use among all the World Health Organization (WHO) regions. Toxins from tobacco smoking can potentially affect sperm development and function, Associate Editor: with a negative effect on semen parameters. Given the high prevalence of smoking and Christian Gratzke recent changes in the WHO laboratory methods for the examination of human semen, the role of this exposure in face of new WHO methods needs to be clarified. Evidence acquisition: We conducted a systematic review, followed by a meta-analysis, Keywords: to determine whether cigarette smoking affects human semen parameters. PubMed, Fertility Saint Joseph’s University Discover, and Google Scholar were used to identify relevant studies published after release of the latest WHO methods for laboratory evaluation of Male infertility human semen. Participants were from fertility/urologic clinics and andrology laborato- Semen analysis ries. The outcome measures were semen volume, sperm concentration, motility, and Sperm count morphology, the parameters usually used in clinical settings to assess fertility. Sperm motility Evidence synthesis: Twenty studies with 5865 participants were included in the meta- Sperm morphology analysis. Exposure to cigarette smoking was associated with reduced sperm count (mean difference [MD]: À9.72 Â 106/ml;[2_TD$IF] 95% conﬁdence interval [CI], À13.32 to Semen quality À6.12), motility (MD: À3.48%; 95% CI, À5.53 to À1.44), and morphology (MD: Smoking À1.37%; 95% CI, À2.63 to À0.11). Subgroup analyses indicated that effect size was Spermatozoa higher in infertile men than in the general population and in moderate/heavy smokers World Health Organization than in mild smokers. The overall effect size on semen volume, sperm count, and motility remained similar when 2010 and earlier WHO manuals were used for semen analysis but was lower with regard to sperm morphology. Conclusions: Our results suggest that cigarette smoking has an overall negative effect on semen parameters. The latest WHO laboratory methods for the examination of human semen had a minimal impact on the magnitude of effect size, thus conﬁrming the observed negative effect of smoking on conventional semen parameters. Patient summary: A new systematic review and meta-analysis comprising 5865 men shows that cigarette smoking is associated with reduced sperm count and motility. Deterioration of semen quality is more pronounced in moderate and heavy smokers. # 2016 European Association of Urology. Published by Elsevier B.V. All rights reserved.

* Corresponding author. Department of Urology,[23_TD$IF] [24_TD$IF]American Center for Reproductive Medicine, Cleve- land Clinic, Mail Code X-11, 10681 Carnegie Avenue, Cleveland, OH 44195, USA. Tel. +1 216 444 9485; Fax: +1 216 445 6049. E-mail address: [email protected] (A. Agarwal). http://dx.doi.org/10.1016/j.eururo.2016.04.010 0302-2838/# 2016 European Association of Urology. Published by Elsevier B.V. All rights reserved. 636 EUROPEAN UROLOGY 70 (2016) 635–645

1. Introduction 2015. For SJUD, which includes 189 databases (http://www. sju.edu/int/resources/libraries/drexel/), and PubMed, the Approximately 37% of male adults worldwide use tobacco, Medical Subject Headings search terms used were mainly cigarettes [1]. Smoking rates have gradually *smoking* OR cigarette AND semen or sperm* OR *fertil* in declined in the United States, but Europe still has the any language. The following selections were made for the highest tobacco use among all the World Health Organiza- advanced search: ‘‘Only studies in humans’’ and ‘‘date range tion (WHO) regions [1,2]. Tobacco combustion produces from 2010 onwards.’’ Article types selected were clinical approximately 4000 chemical compounds, and smokers study, comparative study, journal article, observational inhale a host of toxins including nicotine, carbon monoxide, study, randomized controlled trial, review, and systematic cadmium, and other mutagenic compounds, all with review. Trial registers searched were Current Controlled potential deleterious effects on male germ cells [3]. The Trials (http://www.controlled-trials.com), ClinicalTrials.- toxins originating from cigarette smoke can decrease sperm gov (http://clinicaltrials.gov), and the World Health Orga- mitochondrial activity and damage the chromatin structure nization International Trials Registry Platform search portal in human sperm, therefore impairing fertilization capacity (http://www.who.int/trialsearch). The gray literature was both in vivo and in vitro [4,5]. searched using Google Scholar. Smoking cigarettes has been associated with a deterio- The titles and abstracts retrieved were initially screened. ration of sperm quality including motility, concentration, Full texts of selected abstracts matching inclusion criteria and morphology, which are the parameters most frequently were obtained. Review articles and reference lists were used in clinical settings to assess fertility [6–8]. However, hand-searched. Studies were analyzed for inclusion inde- the evidence is not unequivocal, and some studies have pendently by two of the authors, and any discrepancies found no effect on semen quality [9–11]. were resolved by discussion. In 2010, the WHO established new criteria for the laboratory examination of human semen [12]. The changes 2.2. Selection of studies and validity assessment specifically included assessments of (1) volume by weight rather than graduated pipette; (2) motility by two catego- Articles were included only if full texts were available, ries, namely progressive and nonprogressive, in contrast enrolled human participants, did not use workplace passive with four categories in previous versions; and (3) morphol- cigarette smoking, and were not review articles. Smokers ogy by strict criteria (Tygerberg) as opposed to the WHO were defined as those smoking only cigarettes, whereas criteria in previous manuals. New reference values for nonsmokers were those men who did not smoke at all at the semen characteristics were also proposed that were time of each particular study. Studies focusing on the effect markedly lower than those reported in previous manuals. of smoking on semen parameters in men with existing In a recent study evaluating the impact of these changes, specific conditions were excluded (Supplementary Table 1). semen characteristics of approximately 15% of patients Authors of unpublished or incomplete data sets were not considered abnormal according to the 1999 WHO reference contacted to request their data for this meta-analysis. values were reclassified as normal based on the 2010 WHO Participants were males aged 13 yr regardless of reference values [13]. Therefore, varying results may be population size and origin. We specified the primary expected with regard to semen analysis when one manual is outcome measures a priori as semen volume, sperm count, followed versus another, with obvious implications for motility, and morphology because these parameters are the counseling, diagnosis, and management of men seeking most commonly used measures in the investigation of male fertility evaluation. These data reinforce the need for a fertility. Some studies provided data on all of these careful examination of the 2010 WHO criteria for the measures and others on just some of them. One exposure laboratory examination of human semen to help health care (smoking) was compared at once (no multivariate analysis). providers understand the clinical impact of such changes. The following characteristics were assessed for each study: We conducted a systematic review and aggregated the study population (infertile vs general population), cigarette available published data on the effect of cigarette smoking consumption per day, and data collection methods (eg, on semen parameters using a meta-analysis. The aim was to semen analysis according to the edition of the WHO summarize the evidence of the effect of cigarette smoking manual). on human semen characteristics in view of the new We adhered to the Preferred Reporting Items for 2010 WHO criteria for the laboratory examination of Systematic Reviews and Meta-analysis statement to report human semen. the results [14]. The initial search yielded 15 119 articles. Of those, 145 abstracts were selected, and 124 full-text articles 2. Evidence acquisition were obtained. From these, 104 studies were excluded with reasons (Supplement 1, item 5). Twenty studies fulfilled all 2.1. Search strategy criteria and were included in the meta-analysis. To determine validity, each included study was assessed We conducted a systematic search using PubMed, Saint according to the criteria for nonrandomized studies to Joseph’s University Discover (SJUD), and Google Scholar to assess the risk of bias [15] (Supplement 1 and Supplemen- identify all relevant studies published from 2010 to August tary Table 2). EUROPEAN UROLOGY 70 (2016) 635–645 637

2.3. Data abstraction Table 2 shows the MD on individual sperm parameters and results of heterogeneity tests between smokers and Each semen parameter was evaluated separately and nonsmokers for infertile and healthy men combined and by independently. First, we analyzed the studies for smoking subgroups based on study participant selection (general and sperm quality comparing smokers with nonsmokers for population and infertile men). Homogeneity was ob- infertile and healthy men combined. Then we compared served only for the association between smoking and sperm infertile smokers versus matched smokers from the general morphology among studies comparing moderate and heavy population. We also conducted a subgroup analysis based smokers with nonsmokers (I2[26_TD$IF]<50%). Studies on the on the WHO method for the examination of human semen remaining parameters varied significantly (p < 0.01, I2 (2010 edition vs previous editions). Lastly, we analyzed >50%). Despite including only studies published after the results based on the consumption of cigarettes per day. We release of the 2010 WHO manual, only seven studies classified the smokers as mild (1–10 cigarettes), moderate specifically utilized this edition for semen analyses. The (10–20 cigarettes), or heavy (>20 cigarettes) based on the remaining studies utilized previous versions, namely the report about smoking intensity in the United States and the 1999 (4th ed.) and 1992 (3rd ed.). level of cigarette consumption necessary for nicotine regulation [16,17]. 3.2.1. Semen volume Mean semen volume ranged from 2.2 to 3.78 ml in smokers 2.4. Analysis and from 2.2 to 3.7 ml in nonsmokers. Overall, semen volume was not significantly affected by smoking (REM Statistical analysis was undertaken using RevMan v.5.0 MD: À0.16; 95% CI, À0.33 to 0.01). software (Cochrane Collaboration, Oxford, UK) and Metafor We conducted two subgroup analyses to examine the package for R (R Foundation for Statistical Computing, Vienna, potential causes of the heterogeneity. Performing analyses Austria; http://www.R-project.org) [18]. Pooled mean differ- separately according to the WHO editions reduced the ences (MDs) between comparison groups were calculated to heterogeneity estimates, and the MDs became significant in determine the effect size. Both fixed effects models (FEMs) both subgroups, albeit not statistically different between the and random effects models (REMs) were fitted to assess the subgroups (Fig. 1). Likewise, heterogeneity estimates were model types that were most suited to the data. Heterogeneity reduced when participants were selected from the general was evaluated using the Q test and the I2 statistic population (I2 = 0%). There was also an increase in the overall [19]. Statistical significance was set at a p value <0.05. effect size in studies where participants had been selected We used the REM when heterogeneity was high (I2 from the general population (MD: À0.21 ml; 95% CI, À0.34 to >50%) [19]. Subgroup analyses were carried out to identify À0.09; p < 0.001) compared with those including only potential sources of the heterogeneity. Sensitivity analyses infertile men (MD: À0.15 ml; 95% CI, À0.36 to 0.07) and assessment of publication bias were conducted to (Supplementary Fig. 2 and Table 2). Sensitivity analyses assess the leverage of individual studies on the results [20] demonstrated that the observed pooled effect size was not (Supplement 1). materially affected by the removal of any of the studies (Supplementary Fig. 3). These results suggest that most of the 3. Evidence synthesis differences between studies is explained by the method of semen analysis and the type of study participants. 3.1. Description of studies 3.2.2. Sperm count 6 Supplementary Figure 1 provides a diagram of the review Mean sperm count ranged from 33.78 to 113.2 Â 10 /ml in 6 process. The characteristics of the included studies are smokers and 42.03 to 132.5 Â 10 /ml in nonsmokers. The provided in Table 1 [8,21–39].Participantswerefrom results indicated that the count was lower in smokers than in 6 infertility and urologic centers as well as andrology nonsmokers. The pooled MD by REM was À8.92 Â 10 /ml (95% laboratories. All semen analyses followed the WHO CI, À12.40 to À5.44; p < 0.001) (Table 2). Sensitivity analyses guidelines. indicate that removing the paper by Joo et al [23] slightly increased the MD to À10.0 Â 106/ml, whereas removing the 3.2. Quantitative analysis studybyMitraetal[21] decreased the MD to À6.9 Â 106/ml (Supplementary Fig. 4). When any of the other studies were Overall, 20 studies were identified, and these included data removed, the observed pooled effect size was not affected. on 5865 participants. The number of studies included in To analyze the potential causes of the heterogeneity, two each meta-analysis varied according to the sperm param- subgroup analyses were conducted. Heterogeneity eters reported: 13 provided data on semen volume, remained high (I2 >50%) for the 1999 and earlier WHO 20 provided data on sperm count, 15 provided data on editions but decreased for the 2010 WHO edition (I2 = 53%). motility, and 10 provided data on morphology. Because In both separate analyses, smoking had a negative effect on sperm parameters are continuous data, the MD and count (p < 0.001), but the effect size was not different associated confidence intervals (CIs) between comparison between WHO editions (Fig. 2). groups were calculated to determine the effect size of In contrast, heterogeneity estimates were not materially smoking on individual semen parameters. affected by performing subgroup analyses by type of 638 EUROPEAN UROLOGY 70 (2016) 635–645

Table 1 – Characteristics of included studies reporting an association between cigarette smoking exposure and sperm quality

Study Study design Study participants Comparison groups Semen parameters Main outcomes reported

Collodel et al [29] Retrospective cohort 271 infertile men and 118 smokers Count, motility, No difference in semen parameters in 25 controls with 153 nonsmokers morphology infertile patients subgroups. Lower proven fertility 25 fertility-proven count in heavy smokers compared controls with mild and nonsmoker Liu et al [33] Prospective cohort 147 infertile men 79 nonsmokers Count, motility, Lower count, motility, and morphology 68 smokers morphology in smokers than nonsmokers. Semen parameters abnormalities were correlated with zinc levels El-Melegy et al [22] Case control 130 fertile and 40 infertile smokers Volume, count, Infertile smokers had lower infertile subjects 30 infertile motility, morphology morphology and higher DNA nonsmokers fragmentation than other groups. 30 fertile smokers Heavy smokers had lower count, 30 fertile nonsmokers motility, and morphology than moderate and mild smokers Al-Matubsi et al [8] Cross-sectional 204 randomly selected 111 smokers Volume, count, Reduced count and motility in adults[7_TD$IF] 93 nonsmokers motility smokers, but no difference in volume. Higher testosterone and LH levels, but no difference in SHBG and FSH levels in smokers versus nonsmokers Aryanpur et al [25] Cross-sectional 180 infertile men 68 nonsmokers Volume, count, Lower sperm count in smokers. No 112 smokers motility, morphology effect of smoking on morphology and motility Aghamohammadi Cross-sectional 280 infertile men 144 nonsmokers Volume, count, No difference in semen parameters and Azfari [39] 40 smokers motility, morphology between smokers and nonsmokers Joo et al [23] Prospective cohort 62 healthy donors 13[8_TD$IF] nonsmokers[9_TD$IF] Count,[2_TD$IF] motility, Sperm count decreased in heavy 22 smokers[10_TD$IF] morphology smokers 27 heavy[1_TD$IF] smokers Mitra et al [21] Cross-sectional 304 infertile men 178 smokers Volume, count, Lower sperm motility, reduced 126 nonsmokers motility, morphology morphology, lower DNA integrity, higher FSH and LH levels, and decreased testosterone levels in smokers compared with nonsmokers Davar et al [30] Prospective cohort 151 infertile men 98 nonsmokers Count, motility, No difference in semen parameters 53 smokers morphology between groups. No correlation between semen parameters and cigarette consumption Gaisamudre et al [31] Prospective cohort 100 infertile men 50 nonsmokers Count Lower count and higher seminal MDA 50 smokers levels in infertile smokers than nonsmokers. Inverse correlation between cigarette[1_TD$IF] consumption[12_TD$IF] and sperm count Caserta et al [28] Cross-sectional 648 infertile men 200 smokers Count, motility, Lower count and motility in smokers 448 nonsmokers morphology than nonsmokers. No correlation between sperm parameters and smoking intensity Meri et al [26] Retrospective cohort 960 infertile men 564 nonsmokers Volume, count, Reduced motility and morphology in 266 heavy smokers motility, morphology smokers compared with nonsmokers. 130 non–heavy Lower count, motility, and morphology smokers in heavy smokers compared with non– heavy smokers Anifandis et al [24] Cross-sectional 207 infertile patients 98 nonsmokers Volume, count, Dose–response association between 76 moderate smokers motility smoking and semen volume, and 33 heavy smokers sperm DNA fragmentation. No signiﬁcant difference on count and motility among groups Zhang et al [34] Retrospective cohort 1512 infertile men 775 nonsmokers Volume, count, Lower volume, progressive motility, 180 mild smokers motility, morphology viability, and morphology, and higher 327 moderate smokers number of leukocytes in smokers than 230 heavy smokers nonsmokers. Dose–response correlation between morphology and amount of smoking Yu et al [27] Cross-sectional 322 infertile patients 147 heavy smokers Volume, count, Decreased morphology in heavy 175 nonsmokers motility, morphology smokers, but no difference in progressive motility and vitality between groups. Sperm histone replacement abnormalities were correlated with sperm motility, viability, count, and cotinine levels EUROPEAN UROLOGY 70 (2016) 635–645 639

Table 1 (Continued )

Study Study design Study participants Comparison groups Semen parameters Main outcomes reported

Hamad et al [32] Cross-sectional 54[13_TD$IF] infertile men 19[14_TD$IF] nonsmokers Volume, count, Lower volume, count, motility, and 35[15_TD$IF] smokers motility, morphology morphology in smokers than nonsmokers Jeng et al [35] Cross-sectional 192 healthy men 89 nonsmokers Volume, count, Lower morphology and testosterone 109 smokers motility, morphology levels in smokers compared with nonsmokers Kumar et al [38] Cross-sectional 240[16_TD$IF] male[17_TD$IF] partners[18_TD$IF] of[19_TD$IF] 102 nonsmokers Count, motility, No difference in semen parameters couples[20_TD$IF] consulting for 46[21_TD$IF] smokers morphology between smokers and nonsmokers infertility problem Moretti et al [36] Cross sectional 110 men attending an 71 nonsmokers Count, motility, No difference in semen parameters andrology laboratory 39 smokers morphology between smokers and nonsmokers Al-Turki [37] Retrospective cohort 258 infertile men 168 nonsmokers Volume, count, Lower volume, count, and motility in 90 smokers motility, smokers than nonsmokers

FSH = follicle-stimulating hormone; LH = luteinizing hormone; MDA = malondialdehyde; SHBG = sex hormone-binding globulin.

Table 2 – Results of the meta-analysis of the studies for smoking and sperm quality comparing smokers with nonsmokers for infertile and healthy men combined (overall) and by subgroups based on study participant selection (general population and infertile men)

Sperm parameter Study participants IV 95% CI Q(p)I2, % Test for overall effect

Sperm count Overall À8.92 À12.40 to À5.44 <0.001 90 <0.001 General population À0.22 À8.36 to 7.92 <0.001 86 0.96 Infertile men À11.29 À15.13 to À7.44 <0.001 84 <0.001 Sperm motility Overall À3.48 À5.53 to À1.44 <0.001 78 <0.001 General population À4.78 À9.29 to À0.27 <0.001 86 0.04 Infertile men À2.58 À4.10 to À1.05 0.01 38 0.0009 Semen volume Overall À0.16 À0.33 to 0.01 <0.001 85 0.07 General population À0.21 À0.34 to À0.09 0.37 0 <0.001 Infertile men À0.15 À0.36 to 0.07 <0.001 84 0.19 Sperm morphology Overall À1.37 À2.63 to À0.11 <0.001 87 0.03 General population À1.70 À5.26 to 1.86 <0.001 95 0.35 Infertile men À1.22 À2.62 to 0.18 <0.001 81 0.09

CI = conﬁdence interval; IV = inverse variance; Q(p) = Q test (probability). [(Fig._1)TD$IG]

Fig. 1 – Forest plot showing the effect of smoking on semen volume, including subgroup analyses according to different World Health Organization (WHO) criteria. CI = confidence interval; IV = inverse 1 640 EUROPEAN UROLOGY 70 (2016) 635–645 [(Fig._2)TD$IG]

Fig. 2 – Forest plot showing the effect of smoking on sperm count, including subgroup analyses according to different World Health Organization (WHO) criteria. CI = confidence interval; IV = inverse variance.

participants. However, the mean effect size was markedly edition (Fig. 3). Heterogeneity estimates were also reduced reduced when participants were selected from the general when participants were infertile men (I2 = 38%). For both population (REM MD: À0.22 Â 106/ml; 95% CI, À8.36 to subgroups, smoking was a significant risk factor for reduced À7.92) compared with infertile men (REM MD: motility. The pooled MD for participants selected from À11.29 Â 106/ml;[27_TD$IF] 95% CI, À15.13 to À7.44; p < 0.001) general population was À4.78% (95% CI, À9.29 to À0.27; (Supplementary Fig. 5). This indicates that some of the p = 0.04) compared with À2.58% (95% CI, À4.10 to À1.05; differences between studies might be explained by the p = 0.0009) in infertile men (Supplementary Fig. 7 and participant types (Table 2) and is consistent in suggesting a Table 2). These results are therefore consistent in suggesting negative association between smoking and sperm count. a negative association between smoking and sperm motility However, further studies are warranted to elucidate the size and indicate that the heterogeneity in the data can be of this association in the general population of smokers. explained by study participant type and the method of semen analysis. 3.2.3. Sperm motility Motility ranged from 16.6% to 72.3% in smokers and from 3.2.4. Sperm morphology 21.7% to 74.3% in nonsmokers. Overall, smoking was a risk Overall, smoking was a risk factor for impaired morphology. factor for reduced motility. The pooled MD by REM was The pooled MD by REM was À1.37% (95% CI, À2.63 to À0.11; À3.48% (95% CI, À5.53 to À1.44; p < 0.001) (Table 2). p = 0.03) (Table 2). Sensitivity analyses indicated that the Sensitivity analyses indicated that removal of the study by observed pooled effect size was not affected by removal of Al-Matubsi et al [8] slightly decreased the MD to À2.9% any of the studies (Supplementary Fig. 8). (Supplementary Fig. 6). Removal of any of the other studies Heterogeneity was dramatically reduced (I2 = 0%) for the had no effect on the observed pooled effect size. 2010 WHO subgroup, whereas it remained high for Subgroup analyses conducted according to WHO edi- 1999 and the earlier WHO subgroup (I2 = 89%). The overall tions and type of participants revealed that the overall effect effect of smoking on sperm morphology depended on the size remained similar in both subgroups. However, hetero- WHO edition that was used. The pooled MD by REM for the geneity decreased for the 2010 WHO subgroup (I2 = 37%). 2010 edition was 1.36% (95% CI, À0.63 to 3.35). In contrast, The pooled MDs were À3.63% (95% CI, À6.33 to À0.93; the pooled MD for 1999 and earlier editions was À1.88% p < 0.001) for the WHO 1999 and earlier editions and (95% CI, À3.26 to À0.51; p = 0.007). These results suggest À3.12% (95% CI, À5.73 to À0.52; p = 0.02) for the 2010 WHO that the method for semen analysis affected the magnitude EUROPEAN UROLOGY 70 (2016) 635–645 641 [(Fig._3)TD$IG]

Fig. 3 – Forest plot showing the effect of smoking on sperm motility, including subgroup analyses according to different World Health Organization (WHO) criteria. CI = confidence interval; IV = inverse variance.

of effect size (Fig. 4). In contrast, heterogeneity estimates that counts were significantly lower in moderate were not materially affected by performing analyses smokers (REM MD: –9.93; 95% CI, À18.04 to À1.82; separately according to study participant type. In both p = 0.02) and heavy smokers (REM MD: À28.06; 95% CI, subgroups, smoking was not significantly associated with a À42.27 to À8.86; p = 0.004) than in nonsmokers. The negative effect on morphology. The pooled MD by REM for effect size was more pronounced in heavy smokers the general population subgroup was À1.70% (95% CI, À5.26 than in moderate smokers (p = 0.0003)[4_TD$IF] (Supplementary to 1.86) compared with À1.22% (95% CI, À2.62 to 0.18; Fig. 11). p = 0.09) in infertile men (Supplementary Fig. 9 and Table 2). The overall effect size estimated by the analysis of all the 3.3.3. Sperm motility studies may therefore be conservative due to the influence Three studies [22–24] reported the effects of cigarette of confounding factors such as semen analysis method and consumption on sperm motility. The pooled results study participant type. indicated that motility was decreased by moderate smoking (REM MD: À3.98%; 95% CI, À6.84 to À1.11; p < 0.006) and 3.3. Subgroup analysis for the effect of cigarette consumption heavy smoking (REM MD: À4.62%; 95% CI, À11.08 to 1.84), on sperm quality albeit not significantly different in the latter. Sperm motility was higher in mild smokers than in moderate and heavy 3.3.1. Semen volume smokers (REM MD: 4.14%; 95% CI, 1.03–7.25; p < 0.01) Three studies [22–24] evaluated the effects of cigarette (Supplementary Fig. 12). consumption on semen volume. Pooled results indicated that volume was lower in moderate smokers (REM MD: 3.3.4. Sperm morphology À0.32 ml; 95% CI, À0.62 to À0.02; p = 0.04) and heavy Two studies [22,23] reported the effects of cigarette smokers (REM MD: À0.77 ml; 95% CI, À0.96 to À0.58; consumption on sperm morphology. The pooled results p < 0.001) than in nonsmokers. Mild smokers were shown showed that morphology was decreased by mild smoking to have higher volumes than moderate and heavy smokers (FEM MD: À0.9%; 95% CI, À1.68 to À0.12; p = 0.02), (p = 0.001) (Supplementary Fig. 10). moderate smoking (MD: À2.47%; 95% CI, À3.31 to À1.64; p < 0.001), and heavy smoking (MD: À4.24%; 95% CI, À5.02 3.3.2. Sperm count to À3.46; p < 0.001). The higher the cigarette consumption, Four studies [21–24] compared the effects of cigarette the higher the magnitude of the effect size (p < 0.0001) consumption on sperm count. The pooled results showed (Supplementary Fig. 13). 642 EUROPEAN UROLOGY 70 (2016) 635–645 [(Fig._4)TD$IG]

Fig. 4 – Forest plot showing the effect of smoking on sperm morphology, including subgroup analyses according to different World Health Organization (WHO) criteria. CI = confidence interval; IV = inverse variance.

3.4. Discussion changes in the methods for conducting such analyses compared with previous editions, as mentioned earlier With changes in the reference ranges and laboratory [12,41,42]. We therefore reassessed the effect of cigarette methods for the evaluation of human semen, the relation- smoking on conventional semen parameters in view of the ship between cigarette smoking and semen quality needs to changes introduced by the WHO manual. This is impor- be clarified [12,40]. To our knowledge, this meta-analysis is tant because WHO manuals are widely used by laborato- the first to summarize the evidence currently available on ries worldwide, and smokers with abnormal semen the association between cigarette smoking and sperm analysis results based on older WHO criteria might be quality after the publication of the latest WHO laboratory considered normal according to the new criteria. Hence manual for the examination of human semen. We included thenegativeassociationbetweensmokingandsperm studies evaluating the effect of cigarette smoking on semen quality, as suggested by earlier studies and meta-analyses volume, sperm count, motility, and morphology, the semen [6–11,43,44], might be perceived as nonsignificant or parameters most commonly used to assess male fertility nonexistent, which could influence counseling and potential. management given by health care providers and decision Our results suggest that cigarette smoking has an overall makers. negative effect on conventional semen parameters. Ciga- Our study aggregated the published evidence on the rette smoking was associated with reduced sperm count, effect of cigarette smoking on semen parameters in a sperm motility, and sperm morphology, but the effects on limited time period (2010–2015) to reflect best the changes semen volume were equivocal. These effects were overall introduced by the WHO. Given that 13 of the 20 studies more pronounced in infertile men than in the general published after 2010 had not followed the latest WHO population, and deterioration of semen quality was manual for the laboratory examination of human semen, we particularly associated with moderate and heavy smoking. conducted a subgroup analysis taking into account the The consistency in the direction of overall effects estimated method of semen analysis (new or old WHO criteria) to for all outcomes adds confidence to our findings. However, determine whether or not the method for conducting these utilization of the 2010 WHO method decreased the effect analyses would interfere with the magnitude of the effect size for sperm morphology, but the results were not size. Our results are nonetheless reassuring of an overall significantly different for volume, count, and motility negative effect of smoking on sperm parameters in the compared with previous WHO editions. studies that applied the 2010 WHO methods. The overall In 2010, the WHO introduced the first semen criteria effect size on semen volume, sperm count, and motility based on a population-based study of fertile men [40]. The remained similar when 2010 and earlier WHO manuals reference values were markedly lower than those thought were used, but it differed with regard to sperm morphology. to be compatible with normal male fertility [13]. Subse- Assessment of morphology by strict criteria represented a quently, the WHO issued its latest laboratory manual (5th major change in the 2010 WHO laboratory manual [12].A ed.) for the examination of human semen with important possible reason for the discrepancy in the WHO criteria used EUROPEAN UROLOGY 70 (2016) 635–645 643 in the included studies is that it might take several years for immature chromatin or fragmented DNA [55]. Of note, guidelines to be adopted into routine practice, and even the pregnancy as an outcome measure was not assessed in any broadly accepted guidelines are often not followed fully of the studies included in our review. [45]. This study has[29_TD$IF] some limitations. Heterogeneity was high The mechanisms through which smoking affects semen in most of the meta-analyses conducted, which may be parameters are not fully understood, but chemical com- explained by the relative small number of participants in pounds produced by cigarettes were shown to have each study and other confounders. For instance, not all risk deleterious effects on the development of male germ cells factors such as participant age, use of medication, and [3]. Nicotine was reported to have a negative influence on obesity, which might have affected semen quality, were sperm morphology and sperm count; seminal cotinine had consistently reported. The population of infertile men and a negative effect on sperm motility [46]. Nevertheless, smoking status also varied across studies. Although random- studies included in our meta-analysis were mostly obser- ization is not an issue with observational studies, selection of vational, which prevents us from making inferences on controls may lead to bias. Another limitation refers to the cause-and-effect mechanisms. quality of included studies that also varied. Smaller trials are Our results indicate that the negative effect of cigarette usually analyzed with less methodological rigor than studies smoking was more pronounced in infertile smokers than in involving larger cohorts, and an asymmetrical funnel plot their counterparts from the general population. Because (Supplement 1) suggests that selective reporting may have semen parameters are usually compromised in infertile led to an overestimation of effect sizes in small trials. Bias men, we speculate that the spermatozoa of infertile men are might have also been introduced because data not published more vulnerable to the negative effects of the[28_TD$IF] inhaled as full articles were excluded from our meta-analysis, and a toxins. Whereas levels of cytokines and radical oxygen few studies without data to calculate the mean and standard species are often elevated in infertile men, the protective deviation were also excluded. mechanisms against oxidative stress are less likely to be Nevertheless, our meta-analysis included >5000 parti- disrupted in the male general population [47]. Notwith- cipants, and the methods used were applied rigorously to standing, the association between oxidative stress markers, explain the source of heterogeneity. The pooled MDs on semen quality, and smoking status in infertile and healthy individual semen parameters indicated consistency of the men needs further investigation. results between the subgroup types. Sensitivity analyses Cigarette consumption also influenced the magnitude of also demonstrated minimal differences when individual effect size. We found that heavy and moderate smokers had studies were excluded, which suggests our results to be worse semen quality than mild smokers and nonsmokers. conservative. Others have corroborated our results by reporting an inverse correlation between cigarette consumption and 4. Conclusions semen parameters [48]. Although smoking status was essentially self-reported in the studies included in our Cigarette smoking was found to be a significant risk factor meta-analysis, accuracy between smoking status reported for decreased semen parameters in adult men. The WHO and biochemical elements in the blood is high [49]. On the laboratory manual used for the examination of human contrary, the effect of smoking duration and smoking semen had a negligible impact on the observed pooled effect cessation on semen parameters remains unclear even after size, except for sperm morphology. Given that most studies our meta-analysis. Earlier studies published before the published after 2010 utilized[30_TD$IF] previous manuals for semen release of the current WHO manual for the evaluation of analysis and considering that WHO methods will remain a human semen have suggested that smoking cessation was reliable source for the laboratory evaluation of human associated with an improvement in semen parameters semen, further research is needed to fully understand the [50,9]. However, none of the studies included in our review impact of modified methods on the association between evaluated the impact of smoking duration/cessation on the smoking and sperm quality. Studies focusing on subcellular semen characteristics. sperm damage such as sperm DNA integrity and oxidative The results of this study have implications for men of stress markers, which are not assessed during conventional reproductive age, health care providers, and decision semen analyses, are also required. This would improve the makers alike. Since smoking is a modifiable lifestyle factor precision of the estimated effect sizes and allow better that is particularly prevalent among such men [1–3], health judgment of the likely clinical importance of the findings. programs focusing on smoking cessation are expected to have a positive impact on semen quality and consequently Author contributions: Ashok Agarwal had full access to all the data in the male fertility. The clinical implication of smoking on study and takes responsibility for the integrity of the data and the fecundity rates remains inconclusive because conventional accuracy of the data analysis. semen analysis is limited in its ability to predict fertilization Study concept and design: Agarwal. capacity and fecundity according to its results [51]. Al- Acquisition of data: Sharma,[31_TD$IF] Harlev. though a few observational and epidemiological studies Analysis and interpretation of data: Sharma,[32_TD$IF] Esteves, Harlev. have suggested an association between conventional semen Drafting of the manuscript:[3_TD$IF]Esteves, Sharma,[32_TD$IF] Harlev, Agarwal. parameters and time to pregnancy [52–54], routine semen Critical revision of the manuscript for important intellectual content: Harlev,[34_TD$IF] analysis does not evaluate sperm dysfunctions such as Sharma, Esteves,[35_TD$IF] Agarwal. 644 EUROPEAN UROLOGY 70 (2016) 635–645

Statistical analysis: Sharma. [13] Murray KS, James A, McGeady JB, Reed ML, Kuang WW, Nangia Obtaining funding: None. AK. The effect of the new 2010 World Health Organization Administrative, technical, or material support: None. criteria for semen analyses on male infertility. Fertil Steril Supervision: Agarwal. 2012;98:1428–31. Other (specify): None. [14] Moher D, Shamseer L, Clarke M, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) Financial disclosures: Ashok Agarwal certifies that all conflicts of 2015 statement. Syst Rev 2015;4:1. interest, including specific financial interests and relationships and [15] SterneJAC, Higgins JPT, Reeves BC. ACROBAT-NRSI. A Cochrane Risk of affiliations relevant to the subject matter or materials discussed in the Bias Assessment Tool: for Non-Randomized Studies of Interventions manuscript (eg, employment/affiliation, grants or funding, consultan- (ACROBAT-NRSI), v. 1.0.0, September 24, 2014. Cochrane Collabora- cies, honoraria, stock ownership or options, expert testimony, royalties, tion Web site. http://bmg.cochrane.org/cochrane-risk-bias- or patents filed, received, or pending), are the following: None. assessment-tool-non-randomized-studies- interventions-acrobat- Funding/Support and role of the sponsor: None. nrsi. Accessed November 25, 2015. [16] Shiffman S. Tobacco ‘‘chippers’’–individual differences in tobacco Acknowledgments: Amy Moore, manager of the Medical Writing dependence. Psychopharmacology (Berl) 1989;97:539–47. Education Program at the Cleveland Clinic, helped edit the article. Ken [17] Pierce JP, Messer K, White MM, Cowling DW, Thomas DP. Preva- Celebucki, a graphic specialist at the Center for Medical Art and lence of heavy smoking in California and the United States, 1965– Photography at the Cleveland Clinic, created the figures. 2007. JAMA 2011;305:1106–12. [18] Viechtbauer W. Conducting meta-analyses in R with the metafor package. J Stat Software 2010;36:1–48. Appendix A. Supplementary data [19] Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring incon- sistency in meta-analyses. BMJ 2003;327:557–60. Supplementary data associated with this article can be [20] Duval S, Tweedie R. Trim and fill: a simple funnel-plot-based found, in the online version, at http://dx.doi.org/10.1016/j. method of testing and adjusting for publication bias in meta- eururo.2016.04.010. analysis. Biometrics 2000;56:455–63. [21] Mitra A, Chakraborty B, Mukhopadhay D, et al. Effect of smoking on semen quality, FSH, testosterone level, and CAG repeat length in References androgen receptor gene of infertile men in an Indian city. Syst Biol [1] Prevalence of tobacco use. World Health Organization Web site. Reprod Med 2012;58:255–62. http://www.who.int/gho/tobacco/use/en/. Accessed November 27, [22] El-Melegy NT, Ali ME. Apoptotic markers in semen of infertile 2015. men: association with cigarette smoking. Int Braz J Urol 2011;37: [2] Jamal A, Agaku IT, O’Connor E, King BA, Kenemer JB, Neff L. Current 495–506. cigarette smoking among adults–United States, 2005–2013. [23] Joo KJ, Kwon YW, Myung SC, Kim TH. The effects of smoking and MMWR Morb Mortal Wkly Rep 2014;63:1108–12. alcohol intake on sperm quality: light and transmission electron [3] Zenzes MT. Smoking and reproduction: gene damage to human microscopy findings. J Int Med Res 2012;40:2327–35. gametes and embryos. Hum Reprod Update 2000;6:122–31. [24] Anifandis G, Bounartzi T, Messini CI, Dafopoulos K, Sotiriou S, [4] Sharma R, Biedenharn KR, Fedor JM, Agarwal A. Lifestyle factors and Messinis IE. The impact of cigarette smoking and alcohol con- reproductive health: taking control of your fertility. Reprod Biol sumption on sperm parameters and sperm DNA fragmentation Endocrinol 2013;11:66. (SDF) measured by Halosperm. Arch Gynecol Obstet 2014;290: [5] Calogero A, Polosa R, Perdichizzi A, et al. Cigarette smoke extract 777–82. immobilizes human spermatozoa and induces sperm apoptosis. [25] Aryanpur M, Tarahomi M, Sharifi H, et al. Comparison of sperma- Reprod Biomed Online 2009;19:564–71. tozoa quality in male smokers and nonsmokers of Iranian infertile [6] Kunzle R, Mueller MD, Hanggi W, Birkhauser MH, Drescher H, couples. Int J Fertil Steril 2011;5:152–7. Bersinger NA. Semen quality of male smokers and nonsmokers in [26] Meri ZB, Irshid IB, Migdadi M, Irshid AB, Mhanna SA. Does cigarette infertile couples. Fertil Steril 2003;79:287–91. smoking affect seminal fluid parameters? A comparative study. [7] Ozgur K, Isikoglu M, Seleker M, Donmez L. Semen quality of Oman Med J 2013;28:12–5. smoking and non-smoking men in infertile couples in a Turkish [27] Yu B, Qi Y, Liu D, Gao X, Chen H, Bai C, Huang Z. Cigarette smoking is population. Arch Gynecol Obstet 2005;271:109–12. associated with abnormal histone-to-protamine transition in hu- [8] Al-Matubsi HY, Kanaan RA, Hamdan F, Salim M, Oriquat GA, Al man sperm. Fertil Steril 2014;101(1), 51–57.e1. http://dx.doi.org/ Hanbali OA. Smoking practices in Jordanian people and their impact 10.1016/j.fertnstert.2013.09.001. on semen quality and hormonal levels among adult men. Cent Eur J [28] CasertaD,BordiG,DiSegniN,D’AmbrosioA,MallozziM, Public Health 2011;19:54–9. Moscarini M. The influence of cigarette smoking on a population [9] Trummer H, Habermann H, Haas J, Pummer K. The impact of of infertile men and women. Arch Gynecol Obstet 2013;287: cigarette smoking on human semen parameters and hormones. 813–8. Hum Reprod 2002;17:1554–9. [29] Collodel G, Capitani S, Pammolli A, Giannerini V, Geminiani M, [10] Swan SH, Brazil C, Drobnis EZ, et al. Geographic differences in semen Moretti E. Semen quality of male idiopathic infertile smokers quality of fertile U.S. males. Environ Health Perspect 2003;111: and nonsmokers: an ultrastructural study. J Androl 2010;31: 414–20. 108–13. [11] Li Y, Lin H, Ma M, et al. Semen quality of 1346 healthy men, results [30] Davar R, Sekhavat L, Naserzadeh N. Semen parameters of non- from the Chongqing area of southwest China. Hum Reprod infertile smoker and non-smoker men. J Med Life 2012;5:465–8. 2009;24:459–69. [31] Gaisamudre KB, Waghmare AR, Naghate GR, Muneshwar J. Study on [12] World Health Organization. WHO laboratory manual for the exami- effect of cigarette smoking on sperm count and seminal malon- nation and processing of human semen. ed. 5. Geneva, Switzerland: dialdehyde levels of infertile men. Int J Med Res Health Sci WHO Press; 2010. 2013;2:451–7. EUROPEAN UROLOGY 70 (2016) 635–645 645

[32] Hamad MF, Shelko N, Kartarius S, Montenarh M, Hammadeh ME. [44] Marinelli D, Gaspari L, Pedotti P, Taioli E. Mini-review of studies on Impact of cigarette smoking on histone (H2B) to protamine ratio in the effect of smoking and drinking habits on semen parameters. Int human spermatozoa and its relation to sperm parameters. Androl- J Hyg Environ Health 2004;207:185–92. ogy 2014;2:666–77. [45] Grimshaw JM, Russell IT. Effect of clinical guidelines on medical [33] Liu RZ, Gao JC, Zhang HG, Wang RX, Zhang ZH, Liu XY. Seminal practice: a systematic review of rigorous evaluations. Lancet plasma zinc level may be associated with the effect of cigarette 1993;342:1317–22. smoking on sperm parameters. J Int Med Res 2010;38:923–8. [46] Paciﬁci R, Altieri I, Gandini L, et al. Nicotine, cotinine, and trans- [34] Zhang ZH, Zhu HB, Li LL, Yu Y, Zhang HG, Liu RZ. Decline of semen 3-hydroxycotinine levels in seminal plasma of smokers: effects on quality and increase of leukocytes with cigarette smoking in infer- sperm parameters. Ther Drug Monit 1993;15:358–63. tile men. Iran J Reprod Med 2013;11:589–96. [47] Yousefniapasha Y, Jorsaraei G, Gholinezhadchari M, Mahjoub S, [35] Jeng HA, Chen YL, Kantaria KN. Association of cigarette smoking Hajiahmadi M, Farsi M. Nitric oxide levels and total antioxidant with reproductive hormone levels and semen quality in healthy capacity in the seminal plasma of infertile smoking men. Cell J adult men in Taiwan. J Environ Sci Health A Tox Hazard Subst 2015;17:129–36. Environ Eng 2014;49:262–8. [48] Ramlau-Hansen CH, Thulstrup AM, Aggerholm AS, Jensen MS, Toft [36] Moretti E, Collodel G, Mazzi L, Campagna M, Iacoponi F, Figura N. G, Bonde JP. Is smoking a risk factor for decreased semen quality? A Resistin, interleukin-6, tumor necrosis factor-alpha, and human cross-sectional analysis. Hum Reprod 2007;22:188–96. semen parameters in the presence of leukocytospermia, smoking [49] Patrick DL, Cheadle A, Thompson DC, Diehr P, Koepsell T, Kinne S. habit, and varicocele. Fertil Steril 2014;102:354–60. The validity of self-reported smoking: a review and meta-analysis. [37] Al-Turki HA. Effect of smoking on reproductive hormones and semen Am J Public Health 1994;84:1086–93. parameters of infertile Saudi Arabians. Urol Ann 2015;7:63–6. [50] Soﬁkitis N, Miyagawa I, Dimitriadis D, Zavos P, Sikka S, Hellstrom [38] Kumar S, Murarka S, Mishra VV, Gautam AK. Environmental & W. Effects of smoking on testicular function, semen quality and lifestyle factors in deterioration of male reproductive health. Indian sperm fertilizing capacity. J Urol 1995;154:1030–4. J Med Res 2014;140(Suppl):S29–35. [51] Esteves SC. Clinical relevance of routine semen analysis and [39] Aghamohammadi A, Zafari M. The impact of cigarette smoking on controversies surroundingthe2010WorldHealthOrganization criteria sperm parameters: a cross-sectional study. International Confer- for semen examination. Int Braz J Urol 2014;40:443–53. ence on Environmental, Biomedical and Biotechnology Web site. [52] Zinaman MJ, Brown CC, Selevan SG, Clegg ED. Semen quality and http://www.ipcbee.com/vol16/17-E20004.pdf human fertility: a prospective study with healthy couples. J Androl [40] Cooper TG, Noonan E, von Eckardstein S, et al. World Health 2000;21:145–53. Organization reference values for human semen characteristics. [53] Slama R, Eustache F, Ducot B, et al. Time to pregnancy and semen Hum Reprod Update 2010;16:231–45. parameters: a cross-sectional study among fertile couples from [41] World Health Organization. WHO laboratory manual for the exam- four European cities. Hum Reprod 2002;17:503–15. ination of human semen and sperm-cervical mucus interaction. ed. [54] Buck Louis GM, Sundaram R, Schisterman EF, et al. Semen quality 3. Cambridge, UK: Cambridge University Press; 1992. and time to pregnancy: the Longitudinal Investigation of Fertility [42] World Health Organization. WHO laboratory manual for the exam- and the Environment Study. Fertil Steril 2014;101:453–62. ination of human semen and sperm-cervical mucus interaction. ed. [55] Esteves SC, Gosalvez J, Lopez-Fernandez C, et al. Diagnostic accura- 4. Cambridge, UK: Cambridge University Press; 1999. cy of sperm DNA degradation index (DDSi) as a potential noninva- [43] Vine MF, Margolin BH, Morrison HI, Hulka BS. Cigarette smoking sive biomarker to identify men with varicocele-associated and sperm density: a meta-analysis. Fertil Steril 1994;61: 35–43. infertility. Int Urol Nephrol 2015;47:1471–7.