Chromosome aberrations in lymphocytes of high-voltage laboratory cable splicers exposed to electromagnetic fields.

Chromosome in of to 19:29- 34. Thirteen high-voltage laboratory employees and 20 referents participated in a cross-sec tional, matched-pairs study of cytogenetic damage. During cable testing the workers were exposed to static, alternating, or pulsed electric and magneticfields. The alternating magnetic field levels of 50 Hz were 5-10 J1T, occasionally much higher. Chromosome aberrations, sister chromatid exchanges. and aneuploidy were studied in peripheral blood lymphocytes. In addition, chromosome aberrations were investigated in lymphocytecultures treated with hydroxyurea and caffeine. to inhibit deoxyribonucleic acid synthesis and repair. Among seven smoking laboratory employees the mean number of chromo some breaks/200 cells was 2.3, as compared with 0.7 for the job-matched referents. The comparable figures for inhibited cultures were 12.0 versus 6.0. No increasewas detected in nonsmokers with either method. The other genetic parameters showed no differences between the exposed workers and the referents. The results support. to some extent, the hypothesis of an increased risk of genotoxic effects among high-voltage laboratory workers.

During the last decade several st udies have focu sed on adverse health effects from ex pos ure to elect romagneti c fie lds ( I) . Some ep idemiologic st udies of wo rke rs exposed to electromagnetic fields have shown an increased risk of leukemi a, and some found an e levated ris k of brain tumors among "electrical workers" (2). In a mo rtality study of workers exposed to electrom agnetic fie lds , Milham (3) found elevated mortality from lymphatic and hematopoietic cancers. The high est risk was observed for power statio n operators (proportionate mortality rat io 195). It has also been sug gested that pulsed magnetic fields produce teratogeni c effects in chicken em bryo s (4,5) In a hig h-voltage laboratory in Oslo, the management and workers were concerned abo ut several cases of cance r (although at variou s sites) among the em ployees . How ever , the number of workers ever employe d in the laborato ry was not sufficient for a ca ncer incidence study. In a cytogenetic study of oilexposed cable workers (6) a subgroup consisting of fo ur ca ble tes t splicers fro m the lab orato ry showed sig ns of increased chromosome aberrations in blood lymphocytes. T he present study was performed to investigate this ob servation furthe r. The hypothesi s Exposure Expos ure to e lectromagne tic fie lds was est imated from inform ation on work practice, obtained through five interviews, and from direct measurements, using four different probes. Great ca re was take n in the constru ctio n of the equipment, in the cali bration procedures , and in the software integration procedures of the 25.4-MHz sampled signal (for high-frequency sig nals after pul se testing) to obtain reliable measurements.
Th e workers were exposed to electromagnetic fie lds du ring three different main types of cabl e testing, namely, direct current (DC) tests, alternating current (AC) tests, and pulse test s. The tests were typically performed on a piece of cab le 15-m long.
In the DC and A C testing, high volt age s (up to 1000 kV DC or 900 kV 50 Hz AC) were used across the insulation of the cab le. Periodically, the cab le was heated by the c ircu lation of a high current (DC or 50 Hz AC) up to 2000 A through the cab le. During high-vol tage DC or AC testin g, there was an est imated static or al ternating electric fiel d of 5-10 kV · m-I at a distance of I m above the floor of the test ha ll. Du ring heating, there was either a static or 50 -Hz alte rnat ing magnetic fie ld of about S-IS IlT in the con tro l room and in the res ting room for test splicers. Workers sometimes touched the cable duri ng the testi ng and co uld the n be exposed to a magnetic field of -500 /-IT (body) and -IO000 /-IT (hand).
During pulse testing, a voltage pulse of up to 2000 kV was suddenly applied to the cable. Typically, the rise time for the voltage across the insulation was 1-3 us, and the half-time for the following exponential decay was 50 us, The peak current during the pulse was about 10 000 A. During such testing an electric field of up to about 2 kV . m' was measured in the shielded control room. The field oscillated in a "ringing" pattern for a few microseconds with a broad range of frequencies in the 0-3 MHz range. Magnetic field pulses in the control room were measured to be about 20 /-IT with a maximum rate of change in the magnetic field (dB· dt" of about 8 T . S-I. During and after the testing, the workers often felt spark discharges. Especially DC testing often resulted in charged objects that gave sparks at touch.
It has been difficult to get a reliable picture of the exposure time for different fields for each category of employees. With few exceptions, there was only moderate testing activity during the last weeks prior to the blood sampling. AC testing, followed by DC testing, was probably the most dominant test in this period. During AC and DC testing, the various fields were present for several hours every day. During pulse testing, a worker might produce up to 100 pulses during a workday, but for most workers the mean exposure time to such pulses was less than 1 d a week. Test splicers were definitely more exposed to electromagnetic fields during AC and DC testing than the engineers were, and probably also during pulse testing. Spark discharges seemed to be more common for the test splicers than for the engineers.

SUbjects and methods
The study was cross-sectional with additional retrospective information about exposure. It compared all of the employees in the high-voltage laboratory, on I June 1990, at Alcatel STK in Oslo to matched referents in pairs or triplets. Four engineers, one foreman, and one engineering assistant (treated as one group, called "engineers") and seven test splicers participated. To match the engineers, other office employees ("job-matched" referents) were selected. For each test splicer, two referents were selected. One reference group ("job-matched") consisted of other production workers. In addition, a second reference group of office workers was selected. The referents were chosen according to the following matching criteria: same age (±3 years) and same smoking status at the time of the examination (present smoker/nonsmoker). The referents were not exposed to any possible genotoxic chemicals at work and had no more occupational exposure to electromagnetic fields than the general population. All of the exposed subjects and the referents received a questionnaire with which to record occupational and nonoccupational exposures with possible influence on chromosome damage. Table I shows the characteristics of the exposed and reference groups. The test splicers had a higher Test splicers Referents a , ' Office referents'  coffee consumption than the referents. As for alcohol con sumption, use of prescribed medicine, infections durin g the last three month s, off-work chemica l expo sure, chroni c diseases, and diagnostic radiography, there were only minor differences between the exposed subjects and the matched referents. The mean expo sure time in cable testing was I I (range [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20] years; the test splicers had a mean emplo yment time of 12 years, and the corresponding value for the engineers was 9 years. Blood samples were collected by venipuncture. For each matched pair, the sampling was performed within 90 min on the same day. If one memb er of a match ed pair suffered from an acute infection , the blood sampling was postponed until at least two weeks after recove ry. Sample pairs were sent by mail to the Department of Occupational Medicin e, Telemark Central Hospital, and anal yzed with the same proced ures that were used in a previous study (6). Cytogenetic scoring was performed by three persons, without knowledge of the exposure status of the subjects. Samples from the exposed subjects and the matched referents (duplets or triplets) were treated simultaneously and alway s scored by the same person.
The cytogenetic analyses included assays for sister chromatid exchanges, chromatid and chromosome aberrations, and aneuploidy. Sister chromatid exchanges were scored in 30 cell s, and the other parameters in 200 cell s, per person. In addition , chromosome aberrations were scored after deo xyribonucleic acid (DNA) synthesis and repair were inhibited in vitro with hydroxyurea and caffeine (7). Hydroxyurea and caffeine were both added at a concentration of 7.5 . I0~2 M, togeth er with colcemid (0.3 ug . mil culture medium ) 3 h prior to harv esting. Since a considerably larger amo unt of cytoge netic damage per cell was expected than in conventional cultures , 50 cell s per person were scored. For comparison purposes, the counts have been given as per 200 ce lls. Scand J Work Environ Health 1993, vol 19. no 1 Stati stical analysi s was performed with the use of SPSS PC+ (statistical package for the social science s, personal computer-s ), version 3.0 (8). Mean values and standard deviations were calculated. The level s of cytogenetic dama ge were compared by the Wilcoxon match ed-pairs signed rank s test. Most of the tests were two-sided . However, in the eva luation of chromoso me breaks , one-sided tests were found to be appropriate, considering the hypothesis of the study. Table 2 shows the cytogenetic parameters for all 13 employees in the high-voltage laborato ry and for the referent s. When all of the exposed employees were compared with the jo b-matched referents, no difference reached a level of statistical significance. The mean number of chromosome breaks among the test splicers was 2. l , as compared with 1.1 amon g the job-m atched referen ts (P =0.17). The numb ers of chromoso me gaps, sister chromatid exchanges, and aneupl oid cells were similar in all of the subgroups . Table 3 shows the mean number of chromosome breaks among the employees in the cable test laboratory in relation to the work cate gory, recent exposure to electromagnetic fields, and smoking. Five of the empl oyee s who had been on sick leave had been recentl y transfe rred to other departments or had admini strative work only were defined as having no recent exposure. In the subgroup of three smoking test splicers with recent exposure, the mean number of chromosome breaks was 3.7, as compared with 1.3 among the job -matched referent s. These three test splicers smoked less than the referent s did. The subgroups of the enginee rs were also compared with their matched referents. The number of subjects in each subgroup was small, and there were no diffe rences that approached the level of statistical signifi-  cance. In a compari son of all of the laboratory employees with the job-matched referents, the smokers showed a more than threefold increa se in chromosome breaks. The mean values were 2.3 and 0.7, respectively (P = 0.04). Table 4 shows the cytogenetic parameters after in vitro DNA synthesis and repair inhibition for all of the exposed employees and for the job-matched referents. All of the outcome parameters showed larger values than the conventional cultures did. The mean number of chromosome breaks amon g the exposed employees was twice that of the job-matched referents , 7.4 versus 3.7 (P = 0.08 ), while the other parameters showed only minor differences. The mean number of chromosome breaks was higher among the test splicers than among the job-matched referent s (9.1 versus 5.l) (P = 0.05). An effect of smoking was also present in this assay. [The mean number of chromosome breaks in the smoking splicers was 12.0, and that of the smoking job-matched referents was 6.0 (P =0.05)].

Discussion
Previously, only one cytogenetic study of high-voltage laboratory workers has been reported (9). In that study workers in a univer sity electricity laboratory had slightly elevated total chromosome change s and sister chromatid exchange rates than radiolog ists did. In the present study the hypothesis-generating observation from our previous study (6) was repeated. In both of these studies the frequen cy of chromosome breaks among test splicers divided by that of referents was approximately three. The elevated number of chromosome breaks seemed to be confined to smokers. In a Swedish study of electrical power switchyard workers (10) the number of chromosome breaks was, as in our study, the most elevated among exposed worker s who were current smok ers. The observed effect in the present study, using con ventional cultures, was further strengthened by the results from DNA synthesis and repair inhibited cultures. Again the difference was mainl y present among smoking test splicers as compared with their job-matched referents. Enhanced detection of cytogenetic dama ge with the use of this method has been shown for smokers (7). Previous in vitro studies have indicated that unrepaired DNA lesions exist in mutagen-treated cells until the cell s enter mitosis. The repair of these lesions can be effectively inhibited in the G 2 stage (the last 3 h in culture) (II , 12). Kihlman & Andersson (13) have suggested that chromosome aberration s as such cause delay or block in the G 2 stage. The block is removed by caffeine, the result being an increa sed number of damaged cells in mitosis.
In their work, test splicers are expo sed to electromagneti c fields, cable oils, organic solvents, and ozone. The size of this study does not permit any advanced stati stical analysi s of the importance of combined exposure. Our previous study (6) indicated th at ex posure to cab le oil s alone is not ge no toxic, as th e cytogen eti c par ameters in cable sp licers doing field installation did not di ffer from those of re feren ts . Exp osure to ca b le oil s and organic solvents is high er among ins ta llati on spl icers than among lab oratory test spli ce rs . As for th e sig ni ficance of ex posur e to ozone , one study (9 ) indicated that the ozone levels experien ced in high-voltage laboratory work are not related to any increased ri sk of chromosome damage. The wo rke rs in the high-voltage te st laboratory were exp osed to various electric and magn etic fie lds, as well as to spark discharges. Th e 50-Hz m agnetic fie lds du rin g both DC and AC test ing are comparable to fie lds experi en ced by pe opl e living cl ose to po wer tr an smission lin e s. At pr esen t, the ev idenc e for cytoge ne tic damage from such ex pos ure level s is lack ing (l). Ho wever, occasiona l ex posure is consid erably higher, whe n ca ble s are tou ched during long-time ca b le testing. In ad dition, th e pu lsed electric and magnetic fields experienced during pulse testing differ from the experience of most othe r workers in our society . In a S wedish study of 20 high-vo ltage substatio n m aintenan ce work ers, ind ications of work-rel ated reproductive hazards we re found (14 ). This result was followed up in tw o studies of po wer substat ion worke rs ex pos ed to electro magnetic fields with transie nt currents (10,15). Th e resul ts fro m bo th studies were pooled , and an incr eased number of chro moso m e aberration s wa s found. El ect romagnetic fie ld s and electric di scharges were cons idered to be the most probable ca uses of th e findings. In contrast to the Swedish study, another human study of 380-kV sw itchyard worker s showed no dam age to chromoso mes (16). It ha s been ar gu ed that the level of exposure w as not th e same in th e two studies.
Se veral in vitro studies of human lymphocytes expo sed to pul sed alt ernating electro m ag ne tic fields hav e show n increased numbers o f chromosome aberra tions (17)(18)(19), bu t some studies ha ve been negative (20, 2 1). Sister chromatid exchang es were usually not elevated (2 1-23), ex ce pt after lon ge r expo sure times (17 ).

Concluding remarks
In s ummary of th e literature and the present study, it seems th at electro m agne tic fields m ay be ca pa ble of producing ch ro mosome aberrations, but not eleva te d sister chr omatid excha nges in human lymph ocytes, dep ending on the type and do se of electromagne tic fie ld ex pos ure. Our st udy and th e Swedish studies are una ble to exclude s park di sch arges as a causal fa ct or. Until further stu di es of similar expo sures are published, cable te st employees should be advised to avoid unnecessary exposure to electromagnetic fields during pulse testing and not to tou ch the ca b le during long-time tests. They should also avoid spark discharges by sufficiently grounding the cables after testing.