Risk for reduced sperm quality among metal workers, with special reference to welders

MORTENSEN JT. Risk for reduced sperm quality among metal workers, with special reference to welders. Scand J Work Environ Health 14 (1988) 27-30. The purpose of this study was to investigate whether men employed in the metal industry have sperm of poorer quality than men in other types of work. A postal questionnaire was sent to men employed in the metal industry, certain other types of nonmetal industries, and other types of employment in which the factors suspected to influence sperm quality were not present. By means of this questionnaire survey, it was hoped to define the possible influences of the work environment on sperm quality. Out of the total of 3 119 men included in the investigation, 2 517 (81 070) filled out the questionnaire satisfactorily. Semen analysis was performed for a113 119men. There was a greater risk for poor sperm quality among welders than among men not employed in welding. The risk for poor sperm quality was increased for those welders who worked with stainless steel. Welding in general, and specifically with stainless steel, is connected with a risk of reduced sperm quality .

Earlier investigations have shown th at influ ences in the work env ironment can lead to a reduc tion in sperm qu ality. Examples of such influences a re exposur e to lead , expo sure to ionizing rad iation , and exposure to pesticides such as d ibromochloropropane (DBCP) (7,9).
Th e purpose of the present study was to investigate th e po ssibility of men employed in the metal industr y hav ing poorer sperm quality than men in oth er types of work.
Th e po ssible exposures in th e met al indu str y are metal dust, cutting oil, refrigerant greasing agents, and organ ic so lvents. In welding work, the possible exposu res are metal s such as alum inum, cadmium, iron , chromium, and nickel, as well as nitrous gases and ozone (4,10,11).
Attention has been directed towards the contents of smo ke generated by welding on stainless steel. This smo ke contains chromium, nickel, mang anese, copper, cadmium, and iron (4,9).
Other exposures that could po ssibly influence the quality of sperm are heat and ion izing radiation. This last type of exposure is present when welders check the welding seams (12).

Subjects and methods
The investigation was of the muiticentered, casereferent type. The men included in the study were tho se who, in co nnection with fertility pr oblems, had deliv- During December 1984, a po stal qu estionnaire was sent to the men in Aalborg, Oden se, and Soenderborg. Th e men in Aarhu s were not sent the qu estionnaire becau se th ey filled out a similar que stionna ire at the time of their co nsultat ion concerning their fertility problem .
The info rmation ob tained from the qu estionnaires from th e men in Aalborg, Odense, and Soenderborg has been defined as primary data. The result s of the semen a na lyses o f a ll the men in the investigatio n, as well as the informa tio n obtained from the que st ionnair es of the men in Aarhus, have been defined as secondary data . The semen analyses in all four hospitals were performed by trained laboratory technicians. The morphological diagnosis was performed by a physician.
The crit eria used to classify the men as cases were as follow s: (i) sperm concentration less than 20 million/rnl, (ii) less than 50 % of the sperm cells motile, and (iii) less than 50 % of the sperm cells with a normal morphological appearance. If anyone, two , or all three of these criteria were fulfilled, the qual ity of the man 's sperm was considered to be poor and the man was classified as a case. Men with sperm that did not full fill any of the criteria were con sidered to ha ve normal sperm and were classified as referents.
In cases in which more than one semen sample was submitted for ana lysis, the follow ing deci sion s were made: (i) if one of the specimens fulfilled the criteria for poor semen quality in the analysis and the other was rated normal , sperm quality was con sidered normal , and the man was classified as a referent; (ii) if both semen specimens were given a rating of poor, the Months Coup les' • Information on 241 couples was missing . Table 2. Dist ribut ion of the num ber of mon ths tr ied to conceive among all participants .

Results
There was an added risk for poor sperm quality among welders when th ey were compared with men not exposed to chemica l or physi cal agents suspected of influenc ing sperm quality (table 4). includ ed . There were also que stions concerning health (including health status at the time of the semen analysis), illness, and disease, as well as quest ions concern ing social and educational sta tus .
In accordance with the information obtained from the questionnaires, the men under investiga tion could be pla ced in one of the follow ing fo ur groups: (i) welders, (ii) metal wo rkers not exposed to welding, (iii) ot her indu strial workers, and (iv) unexposed workers, ie, men not exposed to chemical or physical agents suspected of influ encing sperm quality . This last group included , for instance, clerks, teachers , and hospita l staff.
T he risk odd s ratio (OR) was calculated according to a log-likelihood method as a measurement for an added risk of poor sperm quality. After data were stratified by hospital location , a Mantel-Haenszel odd s ratio was calculated (5). P ossible confou nde rs such as living quarters, age, smo king hab its, alcohol intake , medicine consumption, and earlier diagnoses of mumps with or without orchitis were investigated by means of a logit regre ssion an alysis to dete rmine whet her the se factors had an y influence on the odds ratios of welders working with stainless steel and welders working with nonstainless steel metals.
Out of the 3 11 9 men, 2 517 men filled out the questionnaires satisfactorily. The response rate was 81 070, and a distribution of the response rate is shown in table I.
The average age of the respondents was 29.9 yea rs , and that of the nonrespondents was 31.0 years . The welders had an average age of 29. 5 years .
Th e time the co up le in question had tr ied to conceive averaged 34.7 months . Among the 2 276 men an swering this qu estion, 142 or 6 0J0 of the couples in que stion had tried to concei ve in a period of less than one yea r, whereas 359 or 16 0J0 had tried to concei ve over a period of more than five years. The distribution of the time periods the couples tried to con ceive is shown in table 2.
Th e percentage of respondents employed at the time of the investigation was 89.7. The percentage of men tha t were child less was 77, and the corresponding percentage for their wives was 74. Table 3 shows the distribution of the num ber of semen specimens by location of hospital. Aa lborg required two semen specimens, rega rd less of the result of the first semen analys is. Aarhus, Odense, and Soenderborg asked for a second semen specimen on ly if the result s of the first semen analysis showed poor sperm qualit y.  man was classified as a case; (iii) if the results of both semen specime ns were normal, the ma n was classified as a referent. Information co ncern ing th e fert ility status (diagno sis) of the wife could be obtained from the questionnaires of the men in Aal borg , Ode nse, and Soenderborg. Thi s information was not avai lable for the men in Aarh us, no r was information on specific occupational exposure of these men .
It was po ssible to use the information obtained on the que stionn aires to gro up the men and their wives according to the Danish Ind ustry Code and the Occupational Code of Denmark published by the Stat istical Registry in Denmark.
The po sta l questionnaire included que stions on the specific exposu res of the men in their work en vironment six mont hs before they submitted their semen specimens. Quest ions on the nonoccupati onal exposure of the men , as well as on smoking ha bits, a lcoh ol intake , and their consumption of medicines, were also  a Men not exposed to chemical or physical agents su spected of influencing sperm qua li ty.

Discussion
Experiments with animals have demonstrated that man gan ese and cadmium cau se atrophy of the teste s (2). Furthermore, it has been demonst rat ed with rodents that cadmium and chro mium are stored in the testes (6). A Dan ish investi gation (8) of per sons examined in a fertilit y clinic showed an odds ra tio of 1.7 for poor sperm qu alit y among welders working with stai nless steel. Wh en the welders were defined by the Dani sh Industry Code , the relation was the same. This association, however, was not statistically significant. On the other hand , the association bet ween welding and pr olonged co nception tim e (mo re than one year) was statisticall y signific ant (8).
A German investigation sho wed that the frequ ency of welder s am ong an infertility clientele was greater th an th e frequ ency of welder s amo ng men con sultin g the same clinic for derm atological problems. There were some methodological problems in thi s study, however, such as selection bias and lack of confounder control (12).
In another Germ an investi gation of 61 welder s, semen an alysis showe d th at more th an 50 % of these men had a sperm count of less than 4 million sperm/ ml (3).
In the present study, even after geographic stratification , the welders had an increa sed risk for poor sperm qu ality, and thi s risk was statistica lly significant. The asso ciation was present even after control for the fertilit y status of the wife (by comparing cases whose wives were normal with refer ent s who se wives had a dia gno sis indicating abnormal fertility status).
Except for the men in Aarhus, information was availabl e co ncerni ng the extent to which the men had been exposed to welding on stainless steel or on nonstainless steel metals daily or weekly over the six-month peri od pri or to the submission of the semen sampl e for ana lysis. The odd s rati o was high for the men exposed to welding when they were compared to the nonwelders (all men who an swered "no" to the que stion concerning exposure to welding). How ever, the results were not statistically significant (OR 1.20,95 % confidence limits 0.92-1 .57). Thi s result co uld be due to the possibility of the group of welding-exposed men including pers on s who were only spo radically em-plo yed as welders or who could ha ve been onl y in the vicinity of welding work and thu s onl y been passi vely expo sed to welding.
When the welders were grouped according to work on nonstainless steel metals and work on stainless steel, the odd s ratio for the welders working on stainless steel showed a sharp rise (OR 2.34,95 % confidence limits 0.95-5.73) and that for the welders working on nonstainless steel metals dropped in comparison to the nonwelders.
There was an increased frequency of medication use among the welder s as compared to that of the men not expo sed to chemical or ph ysical agents suspected of influencing sperm qu ality. One welder took a betablocker, and this type of medicine is suspected to have a negat ive influence on sperm quality. The other types of medications consumed covered a wide spectrum, but they were not suspected of influencing sperm quality.
There was no association between the other possible confounders and occupation as a welder.
As earlier mentioned , ther e was a difference in the number of semen samples per man between the different hospitals . Men with one semen sample showing poor sperm quality were classified as having poor sperm quality. Thi s pr ocedu re introduced a po ssibilit y for misclassification since these men would have been classified as having normal sperm quality if they had had the opportunity to deliver a seco nd sample and it had been normal.
About 20 0"/0 (602 of 3 119)of the men did not return the questionnaire to th e investigators. If these men had had pr ior knowledge of the influence of the work environment of sperm qu ality, motivation to participate in the investigation could have been influenced by this prior kno wledge. Thus one can speculate that exposed men agreed to participate in the investigation , whereas unexposed men did not. Th is selection mechanism is presumably not very meaningful, however, in that there has been no public discussion concernin g the possibility of occupational expo sure s influencing sperm quality.
There was no difference in the sperm quality of the participant s as opposed to the nonparticipants. For Soenderborg, it was possible to evaluate the difference in type of jobs between the participants and nonparticipants. There was no difference.