British Journal of Industrial Medicine 1980;37 :299-306

48-hour ambulatory electrocardiography in dynamiteworkers and controlsC HOGSTEDT,' B SODERHOLM,2 AND L BODIN3

From the Department of Occupational Medicine, Regional Hospital, Orebro, University Hospital,Linkdping,l the Department of Clinical Physiology, Regional Hospital, Orebro,2 and the Department ofStatistics, University of Orebro,3 Sweden

ABSTRACT Sudden deaths and chronic cardiovascular diseases have been reported in excess fre-quency from the explosives industry. Forty-two active dynamite workers and 43 healthy, unexposedworkers have been studied by ambulatory electrocardiographic monitoring during two 24-hourperiods covering an exposed shift and the "abstinence phase," 40-64 hours after the last exposure todynamite. To achieve comparability the non-exposed individuals were screened for risk factors ofheart disease in the same way as those employed in exposed work. No statistically significantdifferences were found between the groups concerning the mean individual number of ventricular or

supraventricular ectopic beats per hour or per 10 000 heart beats during the whole recorded time.Among those who showed only ectopic beats in one of the two periods dynamite workers tended tohave more ectopic beats in the second period. "Complicated" ventricular ectopic beats (multifocal,bigemeny, or coupled) were seen in six dynamite workers and four controls. The only observedperiod of ventricular tachycardia was seen in a dynamite worker on Monday morning. The mean

individual corrected QT-time was similar between the two groups, and there was no tendencytowards longer QT-times among the dynamite workers during the abstinence period. Continuousmonitoring detected several cases with pronounced ventricular arrhythmias despite normal short-time ECG and may be used to investigate the heart rhythm in active workers.

Sudden deaths in excess frequency have beenreported from the explosives industry since the 1 950sand so-called "Monday deaths" have occurred in the"abstinence phase"-that is, 36-72 hours after thelast exposure to nitroglycol or nitroglycerin.'2Nitrate esters act as vasodilators immediately afterabsorption in the body and have been used as such inmedicine for more than a century. Coronary vaso-spasm, however, has been shown in the abstinencephase as a withdrawal phenomenon and is thoughtto be the cause of the precordial pains often reportedfrom dynamite workers during that time.3Among men employed a long time in the explosives

industries an increased risk of dying from chroniccardiocerebrovascular diseases might also be pres-ent.45 The conditions of hygiene in the explosivesindustries have, however, improved considerablyduring the past two decades,67 but longitudinal,prospective epidemiological studies for evaluat-

Received 12 July 1979

Accepted 8 August 1979

ing a possible remaining risk from today's lowerexposure levels would take a considerable time tocomplete. Therefore, supplementary methods forindicating cardiac disturbances have been sought,such as ambulatory 48-hour electrocardiographicmonitoring.

In animal studies indications have been obtainedthat the sensitivity of the heart to epinephrineincreases 24-48 hours after the last exposure tonitrate esters, resulting in an increased mortality ofthe exposed animals compared to controls and morefrequent ventricular ectopic beats.8 9 Increases insympathetic activity during myocardial ischaemiahave been found to cause ventricular arrhythmias,including fibrillation, and prolonged QT-time.Prolonged QT-time has been associated withincreased risk for sudden death in people withischaemic heart disease.10 Furthermore, men with alarge number of ventricular ectopic beats (>100/10000 complexes) in electrocardiographic recordingsrun a significantly increased risk of death fromcoronary heart disease and sudden death.1112Lange et a13 studied a nitroglycerin worker with

299

on January 1, 2021 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.37.3.299 on 1 A

ugust 1980. Dow

nloaded from

300

coronary cineangiogram during the abstinence phaseand showed coronary arteriospasm, which was

readily reversed by nitroglycerin. We may entertainthe hypothesis that the increased risk of suddendeaths among dynamite workers could be caused byventricular fibrillation from an increased sympatheticactivity during the abstinence phase combined with a

relatively ischaemic myocardium because ofcoronaryspasm in sensitive individuals. Such a hypothesiswould be supported if the dynamite workers hadmore cardiac disturbances and longer QT-timescompared with controls, especially in the abstinencephase. Continuous ECG recording could be used fordisplaying the cardiac rhythm during differentperiods, but only a few studies have been designed toexamine the type or frequency ofcardiac disturbanceson continuous monitoring among healthy, workingmen screened for coronary heart disease. It was

assumed that electrocardiographic abnormalitieswould reflect the effect of explosives on the incidenceof sudden death, and with this assumption we under-took this study to assess the effectiveness of currentenvironmental control by means of long-termelectrocardiographic monitoring.

Material and methods

SELECTION OF SUBJECTS AND HEALTH

SCREENINGForty-nine healthy male dynamite workers, nativeSwedes, were available for study at the explosivesfactory of this investigation. They had been em-

ployed in dynamite work between four months and27 years and had different degrees of exposureto dynamite. Their ages ranged from 18 to 62.The controls were chosen from native Swedes whowere employed at the same company but had never

been occupationally exposed to dynamite.Pre-employment examinations and annual health

check-ups had excluded subjects with risk factors forheart disease from dynamite exposure. Therefore a

similar screening procedure had to be appliedamong the controls to avoid selection bias. Exposedas well as unexposed subjects were examined andcompleted the (translated) self-administered versionof the London School of Hygiene CardiovascularQuestionnaire.13 Questions were added about a

family history of heart disease and high blood pres-sure among parents and siblings under the age of 61and concerning smoking and alcohol habits (in thesame categories as Doll and Hill14 and Myrhed15respectively).The blood pressure was measured in the right arm

with a standard mercury sphygmomanometer whilethe subject was seated. The diastolic pressure was

recorded at the disappearance of the arterial

Hogstedt, Sdderholm, and Bodin

(Korotkov's) sounds. Serum cholesterol concentra-tion was measured at the examination and checkedwith a fasting sample if the value was above thenormal values, as recommended from the laboratory.Height and weight were measured. ECGs at rest wererecorded for 30 seconds and read by the physiologistnormally evaluating the ECGs from the dynamiteworkers but without knowing which group the sub-ject belonged to.As a result of the screening procedure some sub-

jects had to be successively excluded from the controlgroup because of ECG-changes that were regardedas"pathological" or "suspected pathological," hyper-tension (>160/>90), hypercholesterolaemia (>8mmol/l for those older than 40 and > 7 mmol/l foryounger individuals), a history of or treatment forheart disease, diabetes mellitus, alcoholism, hyper-tension, or asthma-that is, the same criteria thatdisqualify for dynamite work (either employment orcontinuation of exposed work). Thus after screening65 control candidates 48 were accepted into thecontrol group in this study.

ECG MONITORINGFour AVIONICS electrocardiocorders 385 were usedto record two 24-hour periods from each of the 97subjects. Three electrodes were used, one applied onthe manubrium sterni and the two others on theright and left fifth ribs in the nipple lines, to give anapex position. After the ECG-signal had beenchecked the men returned to their normal activity.Recordings were made from 1300 on Friday toroughly the same time on Saturday, when theelectrodes were taken away and the batteries charged.The subject returned on Sunday at about 1300 forthe same procedure and was recorded until Mondaynoon. On Friday the dynamite workers went directlyto their late shift work (1330-2200) while the greatmajority of the controls were working daytime(0700-1600). Saturday and Sunday were free for bothgroups. On Monday the dynamite workers had theearly shift (0500-1330) while the controls worked thesame hours as on Friday. This procedure allowed forrecording during the last shift of the week for theexposed men, including the higher exposed cleaningoperations, and during a substantial part of theabstinence phase (that is, during the period about40-64 hours after the last exposed hour). Four sub-jects were recorded each week-end and attempts weremade to register two exposed and two unexposed atthe same time. The recordings started in September1977 and the final recordings were made in May1978. After each 24-hour registration the subjectswere interviewed about their experiences duringwork, especially the exposure, and sleeping hours,leisure time, and types of activity.

on January 1, 2021 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.37.3.299 on 1 A

ugust 1980. Dow

nloaded from

48-hour ambulatory electrocardiography in dynamite workers and controls

ANALYSES OF THE TAPESThe recordings were analysed on an AVIONICScomposite electrocardioscanner combined with aReynold's medical path-finder and a Mingograf 61for paper-prints. The electrocardioscanner as well asthe path-finder provide an oscilloscope display of theECG at normal or 60 times standard speed. Anarrhythmia event computer (Reynold's MedicalPA 4) for counting ventricular or supraventricularectopic beats was found to give too many false-positives, and therefore all hours with more than fiveectopic beats had to be printed on paper (running at4-2 mm/s) while hours with less than five ectopicbeats were evaluated visually from the oscilloscopes.As a preliminary analysis a slow speed paper print

for heart rate counts was recorded and at the sametime a skilled technician checked on the oscilloscopewhether any arrhythmias occurred. All tapes thatseemed to be without arrhythmias were recheckedvisually by another technician. The tapes witharrhythmias were later checked by the participatingphysiologist (BS). In cases of complicated arrhyth-mias or doubtful nature of the ectopic beats,printouts at 50 mm/s were used. All analyses weremade without knowledge of the activity or occupa-tion of the recorded subject.

PROTOCOLSThe average heart rate for each 30-minute period wasevaluated from the slow speed frequency paperprints. The number and quality of ectopic beats orarrhythmias were notified on individual protocolsfor each recorded hour without knowledge of theactivity and exposure status of the subject. After theanalyses of the recordings were finished, the meanheart rate, the number of ectopic beats, etc weresummarised for different periods of exposed hours,working hours without exposure, leisure hoursawake, and sleeping hours for the respective 24-hourperiod.

QT-TIMESThe QT-intervals were measured from five non-consecutive beats and corrected for the heart rate

according to Bazett's'6 formula: QTc =-

Those calculations were made from three differentperiods in each 24-hour recording.DATA ANALYSISAll the information on the subjects and the ECGrecordings was computerised, and a data set com-prising the 85 subjects and 64 basic variables wasformed. From this original data set new combi-nations of variables for the analyses of specifichypothesis could be created.

The main part of the statistical analysis has beendevoted to the analysis of dichotomised variablesdescribing presence or non-presence of factors suchas exposure to dynamite, appearance of ectopicbeats, smoking, age over 40 years, and knownheredity for cardiovascular diseases. Distribution-free as well as distribution-dependent tests were usedin the analysis of the original metric variables. Two-tailed p-values were calculated where so applicable.Classified as non-significant results are those forwhich p > 0 05.

Exposure

Dynamite consists of about 60% inorganic nitrates,25% blasting oil, 10% dinitrotoluene, and a few percent nitrocellulose and fillers-for instance, woodflour and chalk. The blasting oil consisted of 80%nitroglycol and 20% nitroglycerin at the studiedindustry in 1977-8. Exposure monitoring for nitrateesters was performed by ten-minute sampling in thebreathing zone ofexposed workers in every productionhouse each autumn and spring. The manufacturingprocesses and details about the exposure monitoringdevice as well as the analytic procedures at the studiedindustry have been described elsewhere.7The number of recorded subjects in different

operations and the calculated mean 8-hour time-weighted average (TWA) concentrations of nitro-glycol, which completely dominates the vapoursfrom the blasting oil, are displayed in table 1.

Table 1 Number of dynamite workers in differentoperations and the calculated mean 8-hour TWA-concentration of nitrate esters from short-timesamples in 1977-8

Type ofoperation No ofrecorded No ofshort-time Nitroglycolsubjects samples concentration

(mgiM3)

Nitration 3 - <0-1Washing I - <0-1Mixing 12 14 0 5Cartridge filling 19 51 0-9Packing 5 10 0-1Transporting 1 2 <0-1Repairing I - Unknown

Thirty-one subjects were employed in the moreheavily exposed operations of mixing and cartridgefilling while 10 were exposed to low concentrationsof nitrate esters (unknown for one). The eight-hourTWA-concentrations have been calculated by takingthe mean value of each operation from the exposuremeasurements during the autumn of 1977 and thespring of 1978 and adjusting for the mean number ofhours in exposed work-that is, the exposuremeasurements were not taken during the actual work

301

on January 1, 2021 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.37.3.299 on 1 A

ugust 1980. Dow

nloaded from

302

shifts when ECG were recorded. Nitration, washing,and most of the mixing are automated operations,and the operators are only exposed for shorter timesand then instructed to wear breathing protectiveequipment as well as gloves, which, however, doesnot exclude exposure as nitroglycol vapours areabsorbed in the gloves and secondarily through theskin.17

Results

The main purpose of the study was to compare thefrequency of cardiac disturbances during an exposedwork shift and a period 40-64 hours after the lastcontact with nitrate esters among dynamite workersand controls at the corresponding time. Therefore,five dynamite workers had to be excluded from theanalyses as they were not exposed during the Fridayshift, due to temporary replacements. Technical taperecorder failures made analyses impossible from therecordings of seven subjects-five controls and twodynamite workers. Thus the reported calculationshave been performed on recordings from 42 dyna-mite workers and 43 controls.The dynamite workers and the controls were

comparable in age distribution, smoking habits,height, weight, blood pressure, and cholesterolconcentration (table 2). The dynamite workersstated a somewhat higher average alcohol con-sumption than the controls (p = 0-11) but no morethan the Swedish average consumption, 4-5 kg/yramong men.15 Twenty and a half per cent of thedynamite workers and 42-9% of the controls hadclose relatives who had died from or had beentreated for heart disease or hypertension before theage of 61 (p = 0-06).

Hogstedt, Soderholm, and Bodin

The total number of observed ventricular ectopicbeats (VEB) and supraventricular ectopic beats(SVEB) were 6623 and 7766 respectively. Thesenumbers were heavily influenced by a few subjectswith large numbers of extrasystoles (table 3). Themean values of the individual mean number of VEBper hour or per 10000 heart beats were rather similarin the two groups. For the SVEB variables there wasa tendency towards larger mean values for theexposed group, but they were not significantlydifferent from those of the control group.The difference in the mean number of ectopic

beats per hour and per 10000 heart beats respectivelybetween the second and the first 24-hour period wasstudied according to the hypothesis that exposeddynamite workers might have an increased numberof cardiac disturbances in the abstinence phase butno statistically significant differences were found(table 4). In fact the tendency seemed to be thereverse, but the figures are heavily influenced by afew subjects with large numbers of ectopic beats.

Table 5 shows the frequency distribution for theindividual mean number of ectopic beats an hourduring the two periods of observation, separately andtogether. The class limits have been chosen to beanalogous with those used in some previousstudies.1819 The distribution is skew with very fewindividuals in the classes with more than five VEBan hour on average. There are no significant dif-ferences between the two groups, either separately ortogether.

Further analyses were made concerning thenumber of workers in the respective group displayingany VEB or SVEB. Thirty-five subjects showed atleast one VEB during the 48-hour recording, 43 atleast one SVEB, and 34 had no extrasystoles at all.

Table 2 Background information for 42 exposed dynamite workers and 43 unexposed controls

Age Heredity Height Weight Bloodpressure Cholesterol Alcohol Smoking habitsfor (cm) (kg) (mmol/l) (kg/year)

Mean SD Range CVD* Systolic Diastolic Smokers Ex- NeverMean SD Mean SD smokers smoked

Mean SD Mean SD

Exposed 36-4 13 8 18-62 8 177 77 128 12-7 76 10-6 5 8 1-0 3 0 4-4 24 6 12Controls 38-1 109 21-61 18 176 75 128 9 0 74 8-8 5 7 0-8 1-7 2-3 21 9 13

*See text for definition.

Table 3 Number of analysed hours, total amount, and range of observed ectopic beats for individuals, andindividual mean number of ectopic beats per hour respectively per 10 000 heart beats (104 hb)

Analysed VEB SVEB VEB SVEBhours

Total Rangefor Total Rangefor Per Per Per Perindividuals individuals hour 10' hb hour 10' hb

Exposed 1768 2769 0-1294 7181 0-2988 1-5 3-3 41 10-6Controls 1823 3854 0-3245 585 0- 269 2-2 5 5 0 3 0-8

on January 1, 2021 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.37.3.299 on 1 A

ugust 1980. Dow

nloaded from

48-hour ambulatory electrocardiography in dynamite workers and controls

Table 4 Difference between 2nd and Ist 24-hourperiod concerning mean number of ectopic beats perhour and per 10 000 heart beats

VEB SVEB

Per hour Per 10' hb Per hour Per 1O4hb

Exposed -1-35 -3 0 -0-76 -2-64Controls 1-36 3-17 0 10 0-12

The proportion of subjects with and without VEBsand SVEBs respectively was similar in the two groups.

For those subjects, however, who showed ectopicbeats in only one of the two 24-hour periods therewas a tendency in the dynamite group for these to bein the second period (table 6). For the controls theopposite was true for VEB, whereas a more uniformdistribution was obtained for SVEB. The differencebetween the two groups was statistically significantfor VEBs (p = 0O05; Fisher's exact test) but not so

for SVEBs.The subjects were further dichotomised for age

(younger or older than 40 years), exposure, smokingstatus (smoker/non-smoker), and appearance or non-

appearance of VEB/SVEB during the 48-hourrecording. Age was the only factor that contributed

significantly to the appearance of ectopic beats (chi-square tests). Multiple regression analysis was usedto obtain the direction of the influence of theindependent variables (table 7). Smoking habits andexposure contributed differently but non-significantlywith respect to VEB and SVEB. Furthermore, nostatistically significant differences were obtainedwith either three-way analysis of variance of VEBand SVEB per hour or per 104 heart beats (inlogarithmic scale) or in analysis of covariance (withthe exact data of age as covariate).

"Complicated" VEBs (multifocal, bigemeny, orcoupled) were seen in six dynamite workers and fourcontrols (table 8). The only observed period ofventricular tachycardia (seven beats) was seen in adynamite worker just after getting out of bed onMonday morning at 0330.

Table 8 Number of workers with "complicated" VEBs

Bigemeny Multifocal Coupled VT Total Nowith acomplicatedVEB

Exposed 3 3 2 1 6Controls 2 3 1 0 4

Table 5 Frequency distribution for number of ectopic beats an hour during two 24-hour periods

Ectopicbeats/h Dynamite workers Controls

I st period 2ndperiod Both periods 1stperiod 2ndperiod Both periods

VEB SVEB VEB SVEB VEB SVEB VEB SVEB VEB SVEB VEB SVEB

0 29 29 26 23 23 23 28 24 34 23 27 19001- 099 10 8 14 13 16 13 12 17 6 19 13 221-00- 5-00 0 1 0 2 0 2 2 1 1 0 1 15 01-30 00 2 2 2 2 3 2 0 1 1 1 1 13001- 1 2 0 2 0 2 1 0 1 0 1 0

Table 6 Number of workers dichotomised on having ectopic beats or not in two 24-hour periods

Both periods VEB exclusively in period SVEB exclusively inperiod Neitherperiod

VEB SVEB 1st 2nd 1st 2nd VEB SVEB

Exposed 10 13 3 6 0 6 23 23Controls 8 15 7 1 4 5 27 19

Table 7 Regression coefficients in multiple regression analyses of appearance of ventricular and supraventricularectopic beats on dichotomised variables for exposure to dynamite, age over 40, smoker, and interaction factors

Dependent variable Constant Smoker Age Exposure Interactionfactors R' SEofterm regression

I 1.S 1 1,3 1 2,l l 1,3 coefficient

VEB 0-244 0-038 0-420 0-104 -0 034 -0 150 0 107 -0-014 0 20 0-12SVEB 0 660 -0 144 0 334 -0-078 -0-183 -0-144 0-084 -0 100 0.19 0-12

303

on January 1, 2021 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.37.3.299 on 1 A

ugust 1980. Dow

nloaded from

Hogstedt, Soderholm, and Bodin

The mean of the corrected QT-times was similarfor the two groups in each of the six calculatedperiods, and there was no tendency towards longerQT-times during the abstinence phase for thedynamite workers.

Discussiori

The validity of a study of this kind and the con-clusions that may be drawn depend on the compar-ability in background variables for the exposed groupand the controls, the probability of detecting seriouscardiac disturbances during the recorded time, andthe relevance of continuous ECG monitoring for theobservation of such cardiac disturbances as well asthe analytical and statistical treatment.

COMPARABILITY OF STUDY GROUPSThe dynamite workers were comparable with thecontrol group in most background variables, buthereditary background and stated alcohol con-sumption differed. Neither heredity nor alcoholconsumption, however, correlated with the occur-rence of ectopic beats according to multiple regres-sion analyses.

Chronic alcoholism has been associated with heartdisease and cardiac arrhythmias, particularly atrialfibrillation,20 but the effect of moderate alcoholconsumption in normal men seems to be moreuncertain.15

Since the late 1950s health check-ups with increas-ing strictness for cardiac risk factors have beenperformed in the studied industry, and no one withsigns of suspected heart disease has been allowed tocontinue in exposed work. Of the dynamite workers,31 had had health check-ups at least once as well asthe examination preceding this study. Some of thesehad been examined more than 20 times, while thecontrols were screened only once for selection to thisstudy. Increasing time of ECG recording has beenshown to increase the detection of cardiac disturb-ances.21 Therefore, the dynamite workers have beenbetter screened for cardiac disturbances and mighthave had a more favourable "heart health state."

CONTINUOUS ECG MONITORING FORDETECTING SERIOUS CARDIACDISTURBANCESTwenty-four-hour ECG monitoring was found todetect more ventricular ectopic activity and repeti-tive forms, such as couplets and ventricular tachy-cardia, compared with exercise tests in cardiacpatients.22 No study of comparison between exercisetest and 24-hour ambulatory monitoring among

healthy subjects has been found, but continuousrecording during normal activity has some attractionas it more accurately reflects the real condition andcovers much longer times and more activities.Nevertheless, 24-hour recording for routine healthcheck-ups imposes considerable strains on themedical health service, the analysing physiologicallaboratory, and the monitored subjects.

Risk factors for mortality among people withevidence of coronary heart disease might differ fromthose operating in subjects without such signs-forinstance, Rose et a123 found no correlation betweenVEBs on short-time ECG or QTc among healthymen initially free of evidence of ischaemia andsubsequent deaths from coronary heart disease.Furthermore, smokers and non-smokers had similarfrequency and quality of cardiac disturbance in thisstudy, as well as in another,18 although smoking is awell-known risk factor for increased cardiac mor-tality.24 An excess frequency of myocardial infarctionwas reported among staff workers compared withproduction line-workers in a car assembly factory,but no differences were found with 24-hour ambula-tory ECG monitoring.25 26Sudden deaths among dynamite workers are not

very frequentper se, although they have been reportedin excess frequency compared with the averagepopulation.2 In the studied plant, where about 100people have been employed in the dynamite line atthe same time, only one accepted case of "Mondaydeath" has occurred during the past 25 years (andthree more cases have been discussed). Ventricularfibrillation is a rare phenomenon among cardiacpatients and cannot be expected to occur amonghealthy subjects.27 The frequencies of ectopic beatsnoticed both among the dynamite workers and thecontrols agree reasonably well with the frequenciesamong healthy men reported by others.18 26

Ventricular tachycardia has been observed onlyoccasionally among those screened for cardiac riskfactors.9 21 Therefore, the single period of ventriculartachycardia observed in this study in a dynamiteworker during the abstinence phase and just aftergetting out of bed on Monday morning, might be ofsome interest. This 51-year-old man had workedwith dynamite for 22 years, firstly as a heavilyexposed mix-house worker but during later years asTV-supervisor with exposure only during 30-60minutes a shift. He was a moderate smoker anddrinker and health check-up examinations over 20years had shown perfectly normal resting ECGs, buthis record included 981 VEBs altogether during the48-hour period. This episode of ventricular tachy-cardia might have resulted from cardiac affectionfrom mechanisms affected during the abstinencephase in a sensitive person.

304

on January 1, 2021 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.37.3.299 on 1 A

ugust 1980. Dow

nloaded from

48-hour ambulatory electrocardiography in dynamite workers and controls

ANALYTICAL AND STATISTICALCONSIDERATIONSThe possibility of erroneous interpretations of thetapes has been limited by the procedure with print-outs for all hours with more than five ectopic beatsand not trusting the automatic counter system usuallyused in this type of studies. Such systems might bequite sufficient for clinical and gross comparison butnot for the exact calculations considered necessary

in this study. Furthermore, all questionable beats or

periods have been printed out with the normal rateand evaluated by an experienced physiologist.

There are some difficulties in quantifying andanalysing the results from continuous ECG record-ings among healthy subjects. The distribution ofobserved arrhythmias was skew as many subjects didnot show any abnormalities, while a few subjects hadlarge numbers of ectopic beats. Most of theseanalyses have been based on binary variables anddistribution-free methods for statistical testing, notrelying on specific normality or homoscedasticityassumptions.The applied methods are considered valid since

they are robust, adapted to the actual distributionalproperties of the data under study, and due to theclose similarity between the conclusions reachedwith all the tests.Some of the background variables have been

introduced in the analyses, but a more detailed studyof their influence was limited by the relatively smallnumber of subjects available for recording.

SUMMARY AND CONCLUSIONNo significant differences were observed between thedynamite workers and the control group in respect ofthe mean number of ectopic beats per hour or per10000 heart beats, and the frequency of subjects withany ectopic beat was also similar in the two groups.Among those who showed only ectopic beats duringone of the two periods, more dynamite workers thancontrols showed ectopic beats during the secondperiod, but the tendency was not very strong and thenumbers were small. Furthermore, the mean QT-time was not longer during the abstinence phasecompared with periods around or directly afterexposure.The results should not be taken as a guarantee that

the present exposure levels are safe for either acuteor chronic cardiac effects. Risk factors for manifestheart disease and sudden deaths among people withsigns of coronary heart disease might not have thesame prognostic value in healthy subjects withoutsuch signs. On the other hand, the results do noturge for diminished exposure levels.

Several cases with complicated and large numbersof VEBs were detected by continuous recording but

not so on repeated short-time ECGs. The accom-plishment of this study shows that continuous ECGmonitoring could be used for ambulatory surveil-lance of cardiac disturbances in active workers.

We thank Lena Janzon and Margareta Gunnarssonat the Nitro Nobel Company for performing therecordings, Birgitta Falt and Siw Persson at theDepartment of Clinical Physiology for technicalassistance, and Professor Olav Axelson, Universityof Linkoping, for comments on the manuscript. Thestudy was sponsored by the Swedish Work Environ-ment Fund, contract number 77/131.

References1 Symanski H. Schwere Gesundheitsschadigungen durch

berufliche Nitroglykoleinwirkung. Arch Hyg Bakteriol1952 ;136 :139-58.

2Carmichael P, Lieben J. Sudden death in explosivesworkers. Arch Environ Health 1963;7:424-9.

3 Lange RL, Reid MS, Tresch DD, Keflan MH, BernhardVM. Nonatheromatous ischemic heart disease followingwithdrawal from chronic industrial nitroglycerinexposure. Circulation 1972;46:666-78.

4Hogstedt C, Axelson 0. Nitroglyerine-nitroglycol exposureand the mortality in cardio-cerebrovascular diseasesamong dynamite workers. J Occup Med 1977;19:675-8.

5 Hogstedt C, Andersson K. A cohort study on mortalityamong dynamite workers. J Occup Med 1979 ;21 :553-6.

6 National Institute for Occupational Safety and Health.Criteria for a recommended standard. Occupationalexposure to nitroglycerine/nitroglycol. Rockville: USDHEW, 1978.

7Hogstedt C, Davidsson B. Nitroglycol and nitroglycerinein a dynamite industry 1958-78. Am Ind Hyg Assoc J1980;41 :373-5.

' Yoshikawa H. Effect of adrenaline on nitroglycol poisonedmice. Ind Health 1964;2:127-8.

9 Clark DG. The supersensitivity of the rat cardiovascularsystem to epinephrine after repeated injections of ethy-lene glycol dinitrate. Toxicol Appl Pharmacol 1970;17:433-42.

10 Schwartz PJ, Wolf S. QT interval prolongation as predictorof sudden death in patients with myocardial infarction.Circulation 1978 ;57:1074-7.

1 Chiang BN, Perlman LV, Ostrander LD Jr, Epstein FH.Relationship of premature systoles to coronary heartdisease and sudden death in the Tecumseh epidemiologicstudy. Ann Intern Med 1969;70:1159-66.

12 Hinkle LE, Carver ST, Argyros DC. The prognosticsignificance of ventricular premature contractions inhealthy people and in people with coronary heart disease.Acta Cardiol 1974; suppl XVIII, 5-32.

13 Rose G, Blackburn H. Cardiovascular survey methods.Geneva: World Health Organisation Monograph SeriesNo 56, 1968.

14 Doll R, Hill AB. Lung cancer and other causes of death inrelation to smoking. Br Med J 1956;ii:1071-81.

11 Myrhed M. Alcohol consumption in relation to factorsassociated with ischemic heart disease. A co-twin controlstudy. Acta Med Scand 1974; suppl 567.

16 Bazett HC. An analysis of the time relations of electro-cardiograms. Heart 1920;7:353-70.

305

on January 1, 2021 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.37.3.299 on 1 A

ugust 1980. Dow

nloaded from

Hogstedt, Sdderholm, and Bodin17 Hogstedt C, Stahl R. Skin absorption and protective

gloves in dynamite work. Am Ind Hyg Assoc J 1980;41:367-72.

18 Clarke JM, Hamer J, Shelton JR, Taylor S, Venning GR.The rhythm of the normal human heart. Lancet 1976;ii:508-12.

19 Lown B, Calvert AF, Armington R, Ryan M. Monitoringfor serious arrhythmias and high risk of sudden death.Circulation 1975;51-2, suppl III:189-98.

20 Brigden W, Robinson J. Alcoholic heart disease. Br MedJ 1964;ii:1283-9.

21 Kennedy HL, Chandra V, Sayther KL, Caralis DG.Effectiveness of increasing hours of continuous ambula-tory electrocardiography in detecting maximal ventri-cular ectopy. Am J Cardiol 1978;42:925-30.

2' Ryan M, Lown B, Horn H. Comparison of ventricularectopic activity during 24-hour monitoring and exercise

testing in patients with coronary heart disease. N Engl JMed 1975;292:224-9.

23 Rose G, Baxter PJ, Reid DD, McCartney P. Prevalenceand prognosis of electrocardiographic findings inmiddle-aged men. Br Heart J 1978;40:636-43.

24 Astrup P, Kjeldsen K. Carbon monoxide, smoking, andatherosclerosis. Med Clin North Am 1973;58:323-50.

25 Baxter PJ, White WG, Sanderson CFB, Barnes GM,Baxter CS, Acheson RM. Myocardial infarction in carassembly workers. Br J Ind Med 1976;33:1-8.

26 Baxter PJ, White WG, Barnes GM, Cashman PMM.Ambulatory electrocardiography in car workers. Br J IndMed 1978 ;35:99-103.

27 Hiss RG, Averill KH, Lamb LE. Electrocardiographicfindings in 67375 asymptomatic subjects. III Ventricularrhythms. Am J Cardiol 1960;6:96-107.

306

on January 1, 2021 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.37.3.299 on 1 A

ugust 1980. Dow

nloaded from