Development of trapezius myalgia among female workers performing light manual work.

The aim of this prospective study was to detect a possible development of trapezius myalgia likely to be work-related in initially healthy female packers. Symptoms were recorded on a five-step intensity scale in a daily diary and in an interview at ten-week intervals. All of the subjects were interviewed concerning assumed risk factors during work and leisure time, and a clinical examination was performed. Within their first year of employment, 17 of 30 subjects developed work-related trapezius myalgia of sufficient intensity to be classified as patients. The median time before a clinically confirmed patient status was 23 (range 6-51) weeks. The symptoms showed a fluctuating pattern, decreasing on the weekends and during some holidays. In conclusion, likely work-related trapezius myalgia of a nonchronic character showed a high one-year cumulative incidence at a plant with repetitive light work.

about the temporal development of these complaint s has mainl y been based on retros pectiv e inform ati on and sick leave statistics. Maeda et al (6) rep orted two incidence peaks of occupational ce rvicobrachial disorder for keypunchers approximately 6 and 24 months after the start of employment. Studie s on musculo skel etal sick leave for sew ing machine opera tors and workers on an electro mecha nica l assembly line showed an increase during the fir st two to three years of employ ment (7)(8)(9) . Th ese studies all conc ern ed job exposur e with high stati c muscular activity. It would be of value to kno w how the se complaints deve loped, and which risk factors are important at a plant with exposur e to low static load.
A pro spe cti ve study of 30 femal e workers at a chocolate manufacturing pl ant wa s carried out to address the se que stions. Th e relation of trapeziu s myalgia to work was examined, especially with the aid of vocational muscle activity evaluated by elec tromyography (10 ) and subjective ly assessed risk factors of the job and leisure time (11 ). The purpose of this study was (i) to evaluate a possible development of trapezius myalgia at a plant w ith repetitive light work and , if it was found , (ii) to analyze its time

Subjects and methods
Subjects and patient status Fifty-five wome n hired consecutively by a chocolate manufacturing plant were included in the study from the start of their employme nt. Three subje cts decl ined to participate, and the oth er 52 sig ned an informed consent form . The study was approved by the regional ethi cal committee. To co mply with the inclusi on criteria , the subj ec ts were not allo wed to (i) ha ve been working in the same or a simil ar factory during the last five years, (ii) suffer from either a systemic illness (rheumatologic, vascular, or malignant disease) or a local neck or sho ulder disord er (malformation , neurological or traum atic), and (ii i) have had a medic al consultation due to neck or shoulder pain during the year preceding thi s employ ment.
Th e subjects wer e included dur ing a two-year period from August 1987 to Au gust 198 9, and data we re collec ted until late January 1990 . The distribution of ob servation time and the cause of termination are show n in figure I. Th e longest participation time was 120 weeks .
Information on complaint s of di scomfort and pain in the neck and sho ulde rs was coll ected throu gh a daily pain diary and interviews every 10th week. A five-poi nt pain scale (12) and the anatomic region s (13) show n in figur e 2 were used . The coh ort was divided into pati ents and nonp atients after termination of the study according to the degre e of symptom s and signs of trap ez ius myalgia. Th e following four criter ia had to be met for patient status: (i) a minimum observation time of six months (also for nonpatients), (ii) at least one IO-week period with a maximum complaint score of 3 or 4, includin g two weeks with a continuous score of 2 or higher, (iii) one or more tender or trigger points in the trapezius muscle, and (iv) symptoms with no known causation outside work. Figure 3 illustrates the complaint development of a single patient based on her diary recordings. She was given patient status in the 20th week of employment with a subsequent sick leave between week 21 and 24. Note the report ing of initial slight complaints for both shoulders and the neck in the figure .
The agreem ent between the dail y report s of complaints and the information given in the interview every l Oth week was determined. The maximal complaint level reported in the diary for the neck and the right and left shoulder was compared with the cor- responding anamnestic information from the interview. A data set was included if at least 70% of the time covered by the interview (last 10 weeks) was recorded in the diary report s. The first 10-week period was excluded due to possible postexercise muscle soreness. A Pearson correlation test between the two recordin g methods resulted in a coeffi cient of 0.87 , based on 190 data sets from 31 subjects (8 not from the 30-subject cohort). No systematic differences were found when the three body region s and the 10-week periods were considered separately. A McNem ar test showed symmetry of the results (P =0.3), 77% with identical assessments , 16% with one level, 6% with two levels, and 1% with three levels of difference. The test of "validity" gave a kappa value of 0.67, indicating good agreement [14, 15 (program 4F)]. Th irty subjects without symptoms related to known factors outsid e work and employed for more than six months constituted the cohort study base. The 22 subj ects not included in the cohort had the same mean level of compl aints during their employment and a similar pattern for vocational muscle activity as the cohort (l 0). The mean age was the same (25 .5 years).
If a subj ect terminated her job before the end of the study, an interview was performed that included questions about neck and shoulder pain and the cause of terminati on (figure 1). A group of seven subjects refused to participate ("quit but employed "), six subjects became pregnant and left the study, and a large group (23 subjects) terminated their employment before the end of the stud y mainly because they preferred anothe r job. Nineteen subj ects were obser ved from the beginning of employment to the end of the study. No significant difference in the electromyographic parameters, development of complaints, or Complaints during work, making it necessary to maintain a conscious effort in order to carry out the work tasks. Need for breaks .
Slight complaints when performing the work tasks, but not inter f e r ing with job performance.

:
1: Almost no complaints, only slight feeling of discomfor~at breaks whe n not concentrating on the wor k t a sk s.

NECK
It is difficult to carry out work because of the complaints. The feeling of discomfort is not fully relieved during a break .

Right Shoulder
Left Shoulder in mean values in uni variate analyses . The Man n-Whitney test, the chi-squared test, and the Wilcoxon signed-ranks test for paired samples were used , dep ending on the data format and the hypothesis being tested (17). Wald-type chi-squared statistics were used in the unbalanced repeated-measures mode ls for the analysis of variance in the development of complaints during the week [15 (program 5V)]. The statistical significance was give n by two-tailed P-values of les s than 0.05 .

TIME OF EMPLOYMENT (WEEKS)
Work load exposure The subjects were employed fu ll-t ime, 5 d a week, 6 1% during the day (8 h/shift) and 39% in the even ing (7.5 hlshift), all on a fixed wage system. The subjects had the following three vacation or holiday periods per year: summe r (4 weeks), Ch ristmas (I week) , and Easter (I week) . Work at two different categories of machines, for production and packing , were inclu ded in the study. In bot h categories the following three tasks were performed: feedi ng the mach ine , controlling and packing the products, and, finally, collecting finished products for further distribution (10 ). Work rotation, alternating between tasks every 15-30 min, was performed at all of the machines. The work was mainly machine-paced, with no individual influence on speed or on the organization of the work .

Statisti cal methods
Medians, percentiles, and ranges were used in the description of the data, if nothi ng else is noted. Nonparametric statistics were used to test the differences

Procedure
The new employees were asked to join the study within a month afte r the first day at work. Stru ctured interviews were performed every 10th week in an undi sturbed room during workhours. Complaints of discomfort or pai n in the neck and shoulders during the last 10 weeks were recorded. Neither the interviewer nor the subject was allowed to check diaries for this information. The clinical examination was per formed with the subject seated in a co mfortable chair with her arms in her lap. Symmetrical flat onefinger palpation of the trapezius muscles was done in search for 16 bilaterally situated tender or trigger po ints according to Travell & Simons (16) . The interview and examination lasted abo ut 30 min. The same physician (KBV) performed all of the interview s and clinical examinations. The examination was followed by the collection of the pain diary forms for the preceding 10 weeks, and the subject was given new forms for the next 10 weeks. No economic reward was given for the completed form s. The number of subjects returning pain diaries diminished consi dera bly after the first 30 week s of the observ ation period and was low after the first year of emp loyment. The diagnosis of trapezi us myalgia was made either by the first author (KBV) on the basis of the interview and physical examination or by a physic ian outside the plant. In the latter case the diagnosis was confirmed by a telephone call.  Table 1 gives the complaint level according to the time of the week or the year, and also the total complaint level, for the patients and nonpatients. Each subject contributed with a mean complaint level for weekdays, weekends, autumn (all days from the summer vacation to the Christmas holiday) and spring (all days from the Christmas holiday to the summer vacation). The complaint level during weekends was significantly lower than during the workdays for both groups. No difference was found between spring and autumn. The mean complaint level before and after the holiday period or vacation was calculated on the basis of four-week periods, and it showed a significant reduction after the Christmas holiday for the patient group. The complaint level was reduced by two-thirds for the patients and about one-half for the nonpatients during the summer vacation (not significant -due to the small number of respondents after the summer vacation). 2-6); this value was significantly higher than for the nonpatients' mean value for the first year of employment (1.1 points with a range of 0--4). The nonpatient group (N = 13) stayed at a mean level of "almost no complaints ... " throughout the study (figure 5). There was no difference in complaints the last year before employment, nor during the first 10 weeks of employment. Thereafter the neck and shoulder complaints of the patient group increased gradually, and maintained a significantly higher level than for the nonpatient group. Three nonpatients had one short episode with pain at level 3, but with no continuous complaints at level 2 or higher for a twoweek period.
Nine of the 17 patients recorded a sick leave prescribed by a physician. The median duration of sick leaves due to neck or shoulder disorders was 2 (range I-IS) weeks. The other patients (eight subjects) had a similar level and time course of musculoskeletal complaints, but no sick leave. There were no significant differences between the patients with and those without sick leave concerning the time from start of employment to patient status [26 (range 9-51) weeks for those with sick leave and 19 (range 6-44) weeks for those without sick leave]. Neither palpable tender or trigger points distinguished between those with and those without sick leave . Complaint development through the week for patients (N = 11) and nonpatients (N = 12). Each subject contributed with the mean of all her reports on each day of the week, for patients also before contracting patient status. The standard errors are indicated by bars.

Discussion
This prospective study showed a high one-year cumulative incidence of likely work-related trapezius myalgia at a plant with repetitive light work. Complaints of discomfort and pain in the neck and shoulders increased during the week and declined during weekends and some holidays.

Study validity and patient definition
The number of subjects included was small and therefore resulted in low power for the study. It may have been possible to augment the cohort study base by an extension of the two and a half years of study, but probably only to a limited extent due to the decreasing turnover rate at the plant.
Complaints from the neck and .shoulders were reported in two ways. If the diary is considered the best approximation, the validity of the 10-week interview was good. The diary itself showed poor compliance, with a low and rapidly decreasing number of respondents. A solution would possibly have been an economic reward for the delivery of finished forms. All of the patients reported considerably more symptoms and signs than necessary to fulfill the criteria for patient status, and thereby contrasted with the nonpatients. The patients had more tender and trigger points, but, in agreement with Sola & Williams (18), muscle tenderness at palpation seemed not to be a good criterion for patient status, as this complaint was also reported by many nonpatients. Fifty-seven percent of the cohort contracted possible work-related trapezius myalgia within the first year of employment. This figure was a cumulative incidence for one year in a cohort of initially healthy subjects (ie, with few complaints during the year preceding the employment). Comparable data were difficult to find, but cross-sectional (3,5,7,19), as well as follow-up (20,21), studies suggest less frequent occurrence of neck and shoulder complaints. Explanations for this discrepancy could be that our patient definition also included sufferers without sick leave and that other studies have been based on populations with more prolonged experience in their present job, thus potentially being a selected population. The new employees were frequently young and inexperienced and therefore possibly more vulnerable to musculoskeletal disorders than elderly persons (I), as has also been shown for back injuries (22).

Time course of complaints
Only a few studies have used a prospective design to describe the development of neck and shoulder complaints. Takala et al (21) studied the development of complaints in a workplace with light sedentary work. Female bank cashiers responded to a postal questionnaire four times at three-month intervals. These responses disclosed a fluctuating occurrence of nonsevere complaints, with a higher frequency during the autumn and winter. Our study also showed a fluctuating pattern, but could not identify any significant seasonal changes, except for reduced complaints after some holidays. A clear alleviation of the complaints during the weekends, followed by a weak increasing trend during the weekdays, indicated that the complaints were not chronic. Kilbom & Persson (23) reported an increase in severe musculoskeletal disorders among workers in an electronics manufacturing plant; from 8 to 18 and 21 % after one and two years of follow-up, respectively. In a study with a similar design, Hagg & Suurktila (24) found an increase in severe symptoms, especially during the first year of the observation time. These studies were all based on work tasks with high static muscular load. Our study indicates that many new cases of nonchronic trapezius myalgia occur during the first year of employment, especially during the first six months, in a job with low static load.

Work-related etiology
Studies of work-related musculoskeletal disorders for packing in the food processing industry are rare, and the few existing studies show contradicting results. Luopajarvi et al (3) found that the repetitive work load of assembly-line packers did not incur any significant increased risk of tension neck syndrome when compared with the more dynamic work of shop assistants. Nevertheless, some indications of a high prevalence of disease for packers have been reported (19). Our attempt to include only assumed work -related trapezius myalgia contracted at the current plant was based on three conditions. First, only subjects with no medical consultation due to neck or shoulder pain the year preceding employment were included, and no significant difference was found in complaint level between future patients and nonpatients during that year. Second, the duration of the exposure in the present job was plausible with regard to the initiation or facilitation of nonchronic symptoms (6,9). Third, the relation between symptoms and work was built on several pieces of information from the subject, mainly the absence of competitive risk factors (II). Nevertheless, the relative risk for trapezius myalgia is impossible to establish in a study such as ours, since no reference group was used. The static muscular load of the work in this plant was very low (10). Consequently, work-related risk factors should probably be searched for in the temporal pattern of muscular activity at work. A crucial factor could be long sustained periods of (even low) activity caused by workplace design or psychogenetic tension due to psychosocial factors.