Tobacco Documents Online

Inter.national Journal of Epidemiology ¢) International Epidernlological Association 1997 B14Z NZ4e3 371 KNtI INT J EPIDERIOL 97 (O]OXFOR9 UNIV PRESS EN Vol. 26, No. 2 Printed in Great Britain Risk Factors for Raynaud's Phenomenon among Workers in Poultry Slaughterhouses and Canning Factories MONIQUE KAMINSKI,* MADELEINE BOURGINE,* MARIE ZINS,*'** ANNIE TOURANCHETt AND CHRISTIAN VERGERS Kaminski M (INSERM Unit 149, 16 ave Paul Vaillant Couturier, 94807 Villejuif Cedex, Prance), Bourgine M, Zins M, Touranchet A and Verger C. Risk factors for Raynaud's phenomenon among workers in poultry slaughterhouses and canning factories. International Journal of Epidemiology 1997; 26: 371-380. Background. Apart from the use of vibrating tools, little is known about dsk factors for Raynaud's phenomenon. However, it has been hypothesized that this disorder may have a multifactodal aetiology, involving potential causal or triggering factors which can be found in the workplace. The objective of the study is to identify individual and occupational risk factors of Raynaud's phenomenon in a population of workers not exposed to vibration, but exposed to cold. Methods. The survey was carried out in 1987-1988 in 17 poultry slaughterhouses and six canning factories and included 1474 employees. Data were collected at the annual visit to the occupational health physician. Finger sensitivity to cold and Raynaud's phenomenon were identified from a list of symptoms occurring from exposure to cold. The role of potential dsk factors was assessed using multiple logistic regression. Results. A high prevalence of symptoms of finger sensitivity to cold was observed. Raynaud's phenomenon was more common in women than in men,. was related to family history of the disease but not to smoking or alcohol consumption. After controlling for non-occupational factors, the following working conditions appeared as risk factors for Raynaud's phenomenon: use of plastic gloves, less than four rest breaks, breaks in an unheated place, continual repetition of the same sedes of operations, exertion of the arm or hand and being able to think of something else while working. Conclusion. The study showed that a number of working conditions were associated with an increased risk of Raynaud's phenomenon and finger sensitivity to cold. Changes in working conditions might reduce the dsk of this disorder in the food processing industry. Keywords: Raynaud's phenomenon, local physical workload, repetitive work, exposure to cold, demographic factors, tobacco use The role of vibrating tools in the occurrence of Raynaud's phenomenon has been extensively investi- gated.1-4 However, it has been hypothesized that this disorder may have a multifactorial aetiology, involving other potential causal or triggering risk factors which can be found in the workplace.5-6 Among such factors are cold, noise, repeated local trauma or stress, which may cause vasoconstriction in the peripheral circulation. The prevalence of Raynaud's phenomenon has been estimated in various populations.7-ts Among women, the observed prevalence varies from 5% or less in Japan,15 * INSERM Unit 149, 16 ave Paul Vaillant Couturier, 94807 Villejuif Cedex, France. ** Present address: INSERM Unit 88, St Maurice, France. t Labour Medical Inspection, Pays de Loire, Nantes, France. * Labour Medical Inspection, Bretagne and Institute of Occupational Health, Rennes, France. 371 South Carolina,~° Spain~7 to 15-20% in Sweden,x2 Denmark,13 England,r-s and a mountainous region of France.~s The same ordering is observed for men, with a very low prevalence in Japan,~4-~5 South Carolina,~° Spainx7 and a prevalence of more than 10% in Eng- land7-s and France.~8 Comparisons of these estimations are difficult, because the various studies differ in the definition of Raynaud's phemoncnon and in the method of assessment. These differences can have a large impact on the estimation of the prevalence,x6 However, these data show that the prevalence is higher for women than for men and that it correlates well with differences in climate, higher values being observed in colder regions. Data from four regions of China confirm this relation with temperature: in a study of workers exposed to vibration with the same occupa- tions, the prevalence of 'vibration white finger' varied from 19% in the North to 7% in the South.~9 TH~'$ ARTI(~LE I$ FOR rNOTVII]UAL USE ONLY AND MAY NOT BE FURTHER REPRODUCED OR STORED ELEOTRONZDALLY NI"THOUT NRITTEN, PERMISSION FROM THE COPYRIGHT HOLDER. UNAUTHORZZE0 REPRODUCTION MAY RESULT ~_'.[ IN FINAND]'AL AND OTHER PENALTIES. " '

372 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY In France, the food and agriculture industry employs 15% of the female industrial workforce.2° Work in this industry involves many risk factors such as repetitive manipulations, work on assembly lines and other phys- ical and mental stressors.2~ Moreover, in many instances, for sanitary reasons, food processing has to be carried out at low temperatures, sometimes as low as 0-6°C in slaughterhouses and meat processing factories. Data collected in 1987-1988 on French workers in poultry slaughterhouses and canning factories offer the opportunity to identify characteristics which may be risk factors for Raynaud's phenomenon, with a special interest in working conditions, in an occupational popu- lation of women and men exposed to cold, but not to vibration. POPULATION AND METHODS Study Population The survey was carried out in 1987-1988 in 17 poultry slaughterhouses and six canning factories, in western France. These factories were selected according to re- gion and size (number of employees). The study popu- lation included all workers employed on 1 January 1987, administrative staff excluded. In factories employing fewer than 200 workers, all employees satisfying the above criteria were included in the sample. In the larger factories, a sample of 1/2 or 1/3 was randomly selected. The total sample included 1686 people, 212 of whom were not investigated for the following reasons: sick leave 35, maternity or parental leave 51, lack of time of the physician 42, resignation 31, redundancy 7, retire- ment 7, refusal 10, other or unknown 29. The remaining 1474 workers were interviewed and examined. Data Collection Data collection took place at the annual visit to the occupational health physician (mandatory in France); the employees completed a questionnaire on type of work, social and demographic characteristics, and liv- ing conditions; this questionnaire was checked by the physician or his/her assistant during the visit; details on working conditions were obtained by interview and included work organization, physical environment including temperature, physical effort, repetitiveness; a medical questionnaire covering symptoms, health status, and use of health services was completed by the occupational health physician. Symptoms Associated with Exposure to Cold The medical questionnaire included a list of questions on symptoms associated with exposure to cold: change of colour of the fingers, pain in fingers, numbness in fingers, symptoms occurring in attacks, length and fre- quency of these attacks. In the present study, Raynaud's phenomenon has been defined as attacks of white and numb fingers provoked by exposure to cold. This defini- tion was very restrictive in comparison to the preval- ence of symptoms considered one by one (Table I), and a similar analysis was carried out for 'finger sensitiv.ity to cold', here defined as change of colour associated with pain or numbness. Non-Occupational Characteristics Potential risk factors considered for this study included both non-occupational characteristics of the employees and characteristics of their working conditions. The first category included factors for which an association with Raynaud's phenomenon has been established or suspected:5 sex, history of finger sensitivity to cold among parents, grandparents, brothers or sisters, age, weight, height, treated hypertension, headaches, men- tion of current worries whether linked to occupation or not, present or previous tobacco use, present alcohol consumption. For women, menopausal status and oral contraception use were also considered. Occupational Characteristics Working conditions considered in the study cover four aspects: exposure to cold, level of noise, physical work- load, and stress linked to mental load. Exposure to cold was characterized by the type of factory, the temper- ature generally being lower in poultry slaughterhouses than in canning factories, the room temperature de- scribed by the workers, and two factors influencing finger temperature--the use of plastic gloves and the manipulation of meat. The number of breaks, and whether these were taken in a heated place, were also consid- ered. Local physical workload was assessed from the following factors: repetition of the same series of opera- tions, exertion of the hand or the arm, exertion due to the weight of the tool or the object to be lifted, use of a knife. Stress linked to mental load was characterized by: lagging behind in work, being always or often able to think of something else while working, repetitive work, assembly line work, bad relationship with col- leagues or superiors, dissatisfaction with work. Joint pain of the hand or the wrist was also con- sidered, as it may be influenced by the same working conditions. Analysis The first stage was to study the prevalence of Raynaud's phenomenon and finger sensitivity to cold in relation to each of the individual ~ind occupational characteristics, for women and men separately. To save

374 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY TABLE 3 Distribution of Raynaud's phenomenon according to the characteristics of the employees Women Men No. % P< No. % P< Family history of cold sensitivity: no 694 7 0.001 521 2 0.001 yes 159 20 57 8 Age: <29 years 278 8 245 1 30-39 376 10 ns 175 2 0.10 40-49 115 I0 94 2 350 104 10 77 7 Treated hypertension: no 822 9 ns 566 3 ns yes 55 13 26 0 Headaches: no 369 7 0.01 372 3 ns sometimes or often 507 I 1 218 1 Worries: no 506 8 0.05 415 2 ns yes 370 12 175 2 ns = no significant difference. TABLE 4 Distribution of Raynaud's phenomenon according to tobacco and alcohol use Women Men No. % P< No. % P< Current smoker: no 669 9 ns yes 207 10 If smoker: cigarettes/day 1-9 87 6 I0-19 81 16 0.05 320 39 5 If non smoker: ex-smoker no 507 10 ns yes 102 6 Alcohol consumption (drinks/week): 0 480 8 I-6 189 14 7-13 85 7 ns 14-20 75 8 21-34 27 11 335 10 0 295 2 ns 297 3 95 4 115 3 ns 87 1 141 0 0.05 134 3 O~ , 103 2 CO 47 0 12~ ' 71 0 ns CO 77 4 130 2 147 3 ns = no significant difference. Occupational Characteristics and Raynaud's Phenomenon Raynaud's phenomenon was more frequently observed in poultry slaughterhouses than in canning factories and when the individual had been employed for at least 5 years (Table 5). Among women, the crude analysis of work- ing conditions showed a higher prevalence of Raynaud's phenomenon when breaks were taken in an unheated place. The use of plastic gloves and having to manipu- late meat were also associated with a higher prevalence

RISK FACTORS FOR RAYNAUD'S PHENOMENON 373 TABLE I Description of symptoms provoked by exposure to cold TABLE 2 Distribution of symptoms according to colour changes Women Men Colour change 881 593 Colour changes of the fingers 44% 22% of which: whitening 19% 9% Fingers becoming painful 33% 16% Fingers becoming numb 31% 17% Fingers becoming white and numb 15% 6% Raynaud's phenomenon 9% 2% (attacks of white and numb fingers) Fingers sensitivity to cold 37% 18% (fingers changing colour and becoming numb or painful) None Fingers Fingers turning turning red or blue white Women 485 229 157 Fingers becoming painful 8% 61% 65% Fingers becoming numb 5% 50% 83% Symptoms occurcing in attacks 3% 21% 53% Men 457 79 51 Fingers becoming painful 5% 56% 45% Fingers becoming numb 6% 48% 65% Symptoms occurring in attacks 1% 5% 29% space, only results for Raynaud's syndrome are shown in the tables, but results for finger sensitivity to cold are summarized in the text. The second stage was multiple logistic regressions to estimate the role of each charac- teristic, controlling for the others. At this stage individ- ual factors known to be associated with Raynaud's phenomenon were considered: sex, family history of finger sensitivity to cold, age, treated hypertension, headaches. Occupational factors were included in the models if they were significantly related to Raynaud's phenomenon or to finger sensitivity to cold in the first stage. The tables show adjusted odds ratios and 95% confidence intervals. The analysis was carded out with SAS software, at the INSERM computer centre (SC5). Numbers vary slightly according to the studied char- acteristics, because of missing data. Logistic regres- sions were done for all subjects for whom data on all variables included in the model were available. RESULTS Description of Symptoms Table 1 shows that all symptoms were more frequent among women than among men: 44% of women described a change in colour of fingers when exposed to cold, com- pared with 22% of men. Finger whitening was described by 19% of women, and 9% of men. Pain and numbness were more frequent in cases Where the colour changed, es- pecially with whitening (Table 2). Sudden attacks of symp- toms also occurred more frequently in cases of whitening. In'total, 15% of women and 6% of men had white and numb fingers, and 9% and 2% had Raynaud's pheno- menon respectively, according to the definition used in this study, i.e. attacks of whitening and numbness of fingers (Table 1); 91% of women and 79% of men with Raynaud's phenomenon said that attacks occurred while they were working, 43% of women and men said that, during these attacks, they were obliged to interrupt their work. Fifty-five per cent of women and men with Raynaud's phenomenon also said that they had symp- toms away from the workplace. More than 60% of the subjects had attacks at least once a week. In at least 75% of cases, first symptoms appeared after they had started working in the studied factories, and for 70% of cases this was after the age of 20. Finger sensitivity to cold, i.e. change of colour and numbness or pain, was described by 37% of women and 18% of men (Table I), and for 37% of them, symptoms were serious enough to oblige them to interrupt their activity. Employee Characteristics and Raynaud's Phenomenon The analysis has been carried out separately for women and for men. However there are few cases among male workers so the power of the tests is limited. Nonetheless, it is possible to verify if the trends are in the same direction for both sexes. Raynaud's phenomenon was more frequent among employees (women or men) with a family history of cold sensitivity, but no significant relationship was ob- served with age, or treated hypertension (Table 3), nor with weight or height (results not shown). Raynaud's phenomenon was significantly more frequent among women with headaches, or having worries; these .rela- tions were not observed for men. No clear assocation was found with tobacco use, or alcohol consumption (Table 4). Among women, no relation was observed with menopausal status or use of oral contraception, even after controlling for age (results not shown). How- ever, among women, Raynaud's phenomenon was more frequent if there had been joint pain of the wrist or hand during the previous year (17% versus 7%). Similar results were obtained for finger sensitivity to cold.

RISK FACTORS FOR RAYNAUD'S PHENOMENON TABLE 5 Distribution of Raynaud's phenomenon according to job characteristics and working conditions Women No. % P< No. 375 Men % P< Type of factory: poultry slaughter and processing 613 I 1 0.001 456 3 as canning industry 264 5 136 1 Number of years in factory: ~5 years 118 4 0.05 142 I ns >5 years 748 I0 446 3 Temperature of the room: ~12°C 236 7 134 1 9-I 1 101 7 ns 61 3 ns 7-8 169 11 50 6 ~6 141 14 110 4 variable 209 9 220 I Use of plastic gloves: no 235 6 0.05 346 2 0. I0 yes 637 11 236 4 Manipulation of meat: no 292 7 0.10 229 1 ns yes 571 11 359 3 Total number of breaks:a ~3 685 10 ns 504 3 ns 94 185 6 78 1 Breaks in a heated room: no 467 I l 0.05 329 2 ns yes 410 7 263 3 High noise level: no 433 7 0.10 341 2 ns yes 442 11 248 2 Continual repetition of the same series of operations: no 223 5 0.01 243 I ns yes 649 11 295 3 Exertion of the hand or the arm: no 558 8 0.05 398 2 ns yes 286 12 178 3 Exertion caused by weight:b no 619 8 0.05 356 2 ns yes 239 13 220 3 Lagging behind in work: no 516 8 0.10 345 2 ns yes 343 12 236 3 Able to think of something else while working: always or often 462 12 0.05 248 3 ns seldom or never 413 7 344 2 Satisfaction with work: no 150 13 ns 80 3 ns yes 721 9 506 2 ns = no significant difference. Lunch and snack breaks included. Weight of the tool or the object to be lifted. of Raynaud's phenomenon. In relation to physical workload, a higher prevalence of Raynaud's phenom- enon was observed when there was continual repetition of the same series of operations, exertion of the hand or the arm, or exertion caused by the weight of the tool or the object to be lifted, but not for the use of a knife. Lagging behind in work and being always or often able to think of something else while working were also associ- ated with a higher risk of Raynaud's phenomenon. Among men, the trend was the same, although not significant.

376 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY No relation was found with repetitive work, work on an assembly line, satisfaction with work, or bad relations with colleagues or superiors. The analysis of finger sensitivity to cold mostly gave the same results: the main differences included a significantly higher frequency of symptoms among employees working at the lowest temperatures and among those having less than four breaks, and stronger relationships with characteristics of stress and satisfaction. Adjusted Analysis Among non-occupational characteristics, the important roles of sex and family history were confirmed, but • there was no longer any relation with headaches. There remained a higher risk with increasing age. Adding tobacco use in the model did not modify the results, and Raynaud's phenomenon was not associated with smoking, even after controlling for other factors. After controlling for non-occupational factors and type of factory, a significantly increased risk of Raynaud's phenomenon was found with the following working conditions (Table 6): less than four breaks (including lunch break and snack break), taking breaks in an unheated room, continual repetition of a series of operations, exertion due to the weight of the tool or object to be lifted, being able to think of something else while working, and at the limit of significance, the use of plastic gloves. Each of these work characteristics was associated with a twofold increase in risk. No signific- ant association was observed with room temperature. No interaction was found between the risk factors. Even after controlling for working conditions, the risk of Raynaud's phenomenon remained higher in poultry slaughterhouses than in canning factories. An analysis carried out for slaughterhouses alone found the same results as those on the total sample. Similar results were obtained for finger sensitivity to cold (Table 7); however, there was no difference be- tween the two types of factories, but finger sensitivity to cold was significantly more common among workers exposed to the lowest temperatures, under 7°C. DISCUSSION This study showed that, among workers in poultry slaughterhouses and canning factories, the prevalence of Raynaud's phenomenon was 9% for women and 2% for men. These figures relate to a restrictive defini- tion of the disorder: whitening and numbness of fingers occurring in attacks. However, finger sensitivity to cold, change in colour associated with pain or numb- ness, was much more common, described by 37% of TABLE 6 Raynaud's phenomenon: risk according to employees characteristics and working conditions (logistic regression, No. = 1321) ORa CI 95%b P< ¢ Sex: women 3.3 1.6--6.6 0.001 Family history of cold sensitivity 4.0 2.4-6.8 0.001 Age: ~29 years I 30-39 1.8 1.0-3.1 0.05 40-49 1.7 0.8-4.0 350 3.8 1.7-8.4 Treated hypertension 1.0 0.4-2.5 ns Headaches 1,4 0.8-2.3 ns Factory: poultry slaughter and processing 3.6 1.6-7.8 0.01 Temperature of the room: ~12°C 1 9-I 1 0,8 0.3-2.0 7-8 1,4 0.6-3.1 ns ~6 1.5 0.7-3,5 variable 0.8 0.4-1.8 Use of plastic gloves 1.7 0.9-3.2 0.10 Total number of breaksa <4 2.4 1.2-4.7 0.05 Breaks in an unheated room 2.0 1.2-3.3 0.01 Continual repetition of the same series of operations 2.1 1.0-4.2 0.05 Exertion caused by weighff 1.8 1.1-2.9 0.05 Able to think of something else while working 1.7 1.1-2,8 0.05 Odds ratio adjusted for all factors in the table. 95% confidence interval for the adjusted odds ratio. Significance of the adjusted test. Lunch and snack breaks included. Weight of the tool or the object to be lifted. women and 18% of men. After controlling for sex, age and family history, the disorder was found to be sig- nificantly associated with several work characteristics. As in other epidemiological studies, we have con- sidered all cases of Raynaud's phenomenon, including those which may be secondary to, or associated with, other diseases such as scleroderma, and which are very rare in population studies.6'1°'16 Because the study was cross-sectional, the observed prevalence of Raynaud's phenomenon probably under- estimates the actual situation. It is possible that some of the most affected workers have left this sector of pro- duction, where they have to work in the cold for much of the time, and where their symptoms prevent them from performing their tasks: 40% of the affected em- ployees said they were obliged to interrupt their activity during attacks.

RISK FACTORS FOR RAYNAUD'S PHENOMENON 377 TABLE 7 Finger sensitivity to colda according to employees characteristics and working conditions (logistic regression, No. = 1314) ORb CI 95%c p< d Sex: women 2.4 1.7-3.3 0.001 Family history of cold sensitivity 4.2 3.0-5.8 0.001 Age: ~29 years 1 30-39 0.9 0.7-1.2 ns 40-49 1.1 0.7-1.7 250 0.9 0.6-1.4 Treated hypertension 0.9 0.5-1.7 ns Headaches 1.2 0.9-1.6 ns Factory: poultry slaughter and processing 1.2 0.8-1.7 ns Temperature of the room ~ 12°C 1 9-11 1.7 1.0-2.7 7-8 1.5 0.9-2.4 0.05 ~6 2. I 1.4-3.3 variable 1.2 0.8-1.8 Use of plastic gloves 1.3 1.0-1.8 0.05 Total number of breaks~ <4 1.7 1.2-2.4 0.01 Breaks in an unheated room 1.4 1.0-1.8 0.05 Continual repetition of the same series of operations 1.3 1.0-1.8 0.I0 Exertion caused by weightr 1.4 I. 1-1.9 0.05 Able to think of something else while working 1.3 1.0-1.7 0.10 Fingers changing colour and becoming numb or painful. Odds ratios adjusted for all factors in the Table. 95% confidence interval for the adjusted odds ratio. Significance of the adjusted test. Lunch and snack breaks included. Weight of the tool or the object to be lifted. It would have been interesting to know if finger sensitivity to cold and Raynaud's phenomenon were more frequent in this population occupationally exposed to cold than in the general population. Unfortunately, because of the differences between studies in definition and method of assessment mentioned in the introduction, and their impact on the estimation of prevalence,6'16'22 it is not possible to determine this. Although the study was cross-sectional, the working conditions we have identified as risk factors in the study were not known as such when the data were collected, either by the employees, or by the physicians. Thus, investigation bias should be at worse very limited. How- ever, employees'answers in relation to certain working conditions may have been inaccurate, thus decreas- ing the power in the analysis of the role of these work- ing conditions. For some working conditions, we were able to compare the description given by the workers and a description made by the occupational health physician. A fairly good agreement was found for char- acteristics of the physical workload, such as continual repetition of a series of operations and work on an assembly line; agreement was lower for characteristics of the mental workload.23 In our study, the prevalence of Raynaud's pheno- menon and finger sensitivity to cold were much higher among women than men. This commonly accepted result was found in other studies.7-s,I°'13-18'24 Prevalence in- creased slightly with age, a result 'observed in some studies,s'll but not in others.7'12'25 In an occupational population, as age is an indicator of length of exposure, increasing prevalence with age is consistent with the dis- ease having an occupational origin. Moreover, the se- lection phenomenon mentioned above, linked to an earlier departure of the most affected workers, can underestimate the relation with age. Family history of finger sensitivity to cold appeared as an important risk factor in our study. However the association may have been overestimated because those who are affected by Raynaud's phenomenon may be more aware than others of the existence of a similar disorder among members of their family. A familial pre- disposition to Raynaud's phenomenon is usually accepted, but there have been few studies: it has been described by Riera17 and by Maricq et al.9 and De Trafford et al.24 only for women with initial symptoms, and by Leppert et aL12 for women most severely affected. In our population, a family history was described by 19% of women and 10% of men. Such a difference can only be explained by differences in reporting. For most health problems, women describe more diseases and symptoms than men,26 and have more frequent use of the health services.27 It would seem that this difference in attitude towards one's own health also exists towards the health of one's next-of- kin. However, family history did appear as a risk factor for men as well as women. In the pathophysiology of Raynaud's phenomenon, similar mechanisms have been suspected for Raynaud's phenomenon and migraine; five studies have found an association between the two diseases.12'Z7"24'28"29 How- ever, in one study, the relation was found only for women,17 and in another, migraine was more common among su.bjects affected by Raynaud's phenomenon than among controls, but non-migrainous headaches were more common in controls.29 In our study, the rela- tion between Raynaud's phenomenon and headaches was no longer significant when other factors were con- trolled for; we llad no way of distinguishing migrainous from non-migrainous headaches. Hypertension and its treatments have also been associated with Raynaud's

378 phenomenon: the two epidemiological studies which" looked for such an association did not find it.12'13 How- ever, in a study on vibration white fingers among lum- berjacks,3° affected workers had higher blood pressure than non-affected ones. One study found a relation between blood pressure and Raynaud's phenomenon among subjects on antihypertensive treatment.2~ In our study, we did not find any relation between treated hypertension and Raynaud's phenomenon. We did not find any relation either with weight or height, a result similar to that of Leppert et al)2 However, one study found a lower Quetelet index in affected subjects.2~ For women, a possible link with hormonal status has been discussed;24'3~ we did not find any relation with meno- pausal status or oral contraceptive use. Raynaud's phenomenon has been described more frequently by employees who had joint pain of the wrist or the hand; a similar result was observed for muscle and joint pain at any location, ~z and, among subjects exposed to vibra- tion, for those with a history of hand injuries.3° How- ever, other studies are needed to understand this result. In our study, we did not find any relation between tobacco use and Raynaud's phenomenon. The potential role of tobacco smoking has been discussed because of its effects on peripheral circulation and finger temper- ature. 5.32 Several studies have looked for such a rela- tion, without finding it.5'13'17'25'28'3°'32 Two studies have found an association between smoking~2 or the quantity smoked13 and finger sensitivity to cold, but not with Raynaud's phenomenon. Among subjects with sus- pected vibration white finger, a higher proportion of smokers was observed in the most affected cases, and a higher reactivity to a cold provocation test was found among smokers;33 however, confounding factors were not controlled for. In our study as in that of Riera et al. 17 there was no relation with alcohol consumption. Only one study found an increasing risk of Raynaud's pheno- menon with the quantity of alcohol consumed.2~ For working conditions studied in relation to Ray- naud's phenomenon, literature is limited to vibrations and, to a lesser extent, to chemical exposures.2'4 The increased risk we found associated with the continual repetition of the same series of operations, and witli exertion of the arm is compatible with the hypothesis of the role of heavy local physical workload,s If cold ex- posure is considered as a triggering factor, its potential aetiologic role is also discussed,s In our study, althougti symptoms were more frequent among employees work- ing at the lowest temperatures, we did not find a strong relation of Raynaud's phenomenon with the temperat- ure described by the employees: it is possible that their estimation was not accurate enough, thus decreasing the power of the analysis. It is also possible that the INTERNATIONAL JOURNAL OF EPIDEMIOLOGY contrast between the workplaces surveyed was not large enough to allow for a real difference in prevalence of Raynaud's phenomenon. For local temperature, the use of plastic gloves had a greater impact than the manipu- lation of cold objects. In relation to temperature, a larger number of breaks and breaks in a heated place seemed to reduce the risk, whereas a negative effect of alternate rewarming and cooling was expected.34 It is also possible that cold exposure has an indirect effect, for example on the local physical load, because of the decreased muscular strength and manual dexterity, and the compensating effort which will be required.35 In spite of the hypothesis related to noise (Jansen, quoted by GemneS), we did not find an excess of Ray- naud's phenomenon in employees who said they were exposed to a high level of noise. Among indicators of mental load and stress, also mentioned as possible fac- tors in the disease,5 being always or often able to think of something else appeared as the only risk factor. Raynaud's phenomenon was more frequent in poultry slaughterhouses than in canning factories. Some of the working conditions found to be risk factors for acro- syndromes were significantly more common in slaughter- houses than in canning factories. However controlling for these working conditions did not abolish this differ- ence. Such results are not easy to interpret; we may however assume that a characteristic such as the type of production is an indirect indicator of a whole set of working conditions not completely accounted for in our analysis. This study, as well as those published previously, is cross-sectional, which can limit the interpretation of some of the results, as discussed above. However, a follow-up of the workers has just been completed, and it will be possible to look at changes in symptoms: worsening or improvement, evolution from cold sens- itivity to Raynaud's phenomenon, to analyse the role of Raynaud's phenomenon in keeping or losing one's job and to check inconsistencies in the associations be- tween working conditions and Raynaud's phenomenon. CON.CLUSION Our study showed that besides individual charac- teristics several related to the workplace and the work itself were risk factors for Raynaud' s phenomenon, and more generally finger sensitivity to cold, especially among employees of poultry slaughterhouses. These characteristics relate mainly to the number and place of breaks, the local physical workload, continual repetition of the same series of operations, exertion of the hand or the arm, the repetitive aspect of the work. Although several of these factors involve a high proportion of

RISK FACTORS FOR RAYNAUD'S PHENOMENON 379 employees in this production sector, each was asso- ciated with a twofold increase in the risk of Raynaud's phenomenon, independently from other factors. Even if other studies are needed to evaluate better the role of temperature and noise levels, the results show on which working conditions preventive actions could be focused. ACKNOWLEDGEMENTS This study has been made possible by the collaboration between INSERM Unit 149, the Regional Labour Medical Inspections in Bretagne and Pays de Loire, the Institute of Occupational Health in Rennes (Professor Curtis) and a team of occupational health physicians from both regions: Drs Bonneau, Bourgault, Brigadeau, Caillon, Cannec, Danielou, Dano, Georgelin, Hameau, Jublin, Kervennic, Le Breton, Lemercier, Lemeur, Leroux, Maupeu, Mauvieux, Mouret, Pilet, Rom6, Saluden, Tissot, Touzard, Trehen, and their assistants. We thank the directors of the participating factories, and the employees for their active contribution to the study. We also thank B Stengel, K Messing and N Brusseaux for their contribution. The study was partly funded by the Department of Labour (SES) and by the Regional Health Authorities (DRASS) in Bretagne and Pays de Loire. REFERENCES l Taylor W, Pelmear P L. Raynaud's phenomenon of occupational origin. An epidemiological survey. Acta Chir Scand 1976; (Suppl, 465): 27-32. 2Gemne G (ed.) Symptomatology and diagnostic methods in the hand-arm vibration syndrome. Scand J Work Environ Health 1987; 13: 265-388. 3 Cherniak M G. Raynaud's phenomenon of occupational origin. Arch Intern Med 1990; 150: 519-22. '*Morse L M. Unusual rheumatologic and musculoskeletal disorder. Occup Med 1992; 7: 425-32. 5Gemne G. Pathophysiology and multifactorial etiology of acquired vasospasti'c disease (Raynaud syndrome) in vibration-exposed workers. Scand J Work Environ Health 1982; 8: 243-49. 6Wigley F M. Raynaud's phenomenon. Curt Opin Rheumatol 1993; $: 773-84. 7 Silman A, Holligan S, Brennan P, Maddison P. Prevalence of symptoms of Raynaud's phenomenon in general practice. Br Meal J 1990; 301: 590-92. SHeslop J, Coggon D, Acheson E D. The prevalence of inter- mittent digital ischaemia (Raynaud's phenomenon) in a general practice. J Roy Coll Gen Pract 1983; 33: 85-89. 9 Maricq H R, Weinrich M C, Keil J E, Le Roy E C. Prevalence of Raynaud phenomenon in the general population. A preliminary study by questionnaire. J Chron Dis 1986; 39: 423-27. 10 Weinrich M C, Maricq H R, Keil J E, McGregor A R, Diat F. Prevalence of Raynaud phenomenon in the adult popula- tion of South Carolina. J Clin Epidemiol t990; 43: 1343-49. t t Walker D D, Jones B, Ogston S, Tasker E G, Robinson A J. A study of white finger in the gas industry. Br J Ind Med 1985; 42: 672-77. 12Leppert J, Aberg H, Ringqvist I, S~rensson S. Raynaud pheno- menon in a female population: prevalence and association with other conditions. Angiology 1987; 38: 871-77. ta Olsen N, Nielsen S L. Prevalence of primary Raynaud's pheno- mena in young females. ScandJ Clin Lab Invest 1978; 37: 761-64. 14Iwata H, Makimo S, Miyashita K. Prevalence of Raynaud's phenomenon in individuals not using vibrating tools. Sangyo lgaku 1987; 20: 500-03. I~ Harada N, Ueda A, Takegata S. Prevalence of Raynaud's pheno- menon in Japanese males and females. J Clin Epidemiol 1991; 44: 649-55. t6Bartelink M L, Wollersheim M, Van de Lisdonk E, Spruijt R, Van Weel C. Prevalence of Raynaud's phenomenon. Neth J Med 1992; 41: 149-52. 17Riera G, Vilardell M, Vaqu6 J, Fonollosa V, Bermejo B. Prevalence of Raynaud's phenomenon in a healthy Spanish population. J Rheumatol 1993; 20: 66-69. Is Maricq H R, Carpentier P H, Weinrich M C et aL Geographic variation in the prevalence of Raynaud's phenomenon: Charleston, SC, USA, vs Tarentaise, Savoie, France. J Rheumatol 1993; 20: 70-76. 19 Yu Z S, Chan H, Qiao L, Qian D S, Ye ~/ H. Epidemiologic survey of vibration syndrome among riveters, chippers and grinders in the rail road system of the People's Republic of China. Scand J Work Environ Health 1986; 12: 289-92. 20 Institut National de la Statistique et des Etudes Economiques (INSEE). Enqu~te sur l'emploi de 1986. Les collections de I'INSEE 1986; D 114. 21 Bue J, Cdstofad M C. Contraintes et nuisances duns la vie de travail. 2~me s6rie de r6sultats de l'enqu~te nationale sur les conditions de travail, r6alis6e en Mars 1984. Dossiers Statistiques du Travail et de l'Emploi 1986; 20: 5-95. Z2Brennan P, Silman A, Black C et al. Validity and reliability of three methods used in the diagnosis of Raynaud's phenomenon. Br J Rheumatol 1993; 32:357-6 1. 23Hays M, Saurel-Cubizolles M J, Bourgine M, Touranchet A, Verger C, Kaminski M. Agreement in descriptions of work- ing conditions between two sources: workers and occu- pational health physicians, lnt J Occup Environ Health 1996; 2: 10-17. 24 De Trafford J C, Lafferty K, Potter C E, Roberts V C, Cotton L T. An epidemiological survey of Raynaud's phenomenon. Eur J Vasc Surg 1988; 2: 167-70. 2~ Keil J E, Maricq H R, Weinrich M C, McGregor A R, Diat F. Demographic, social and clinical correlates of Raynaud phenomenon. Int J Epidemiol 1991 ; 20: 22"1-24. 26Lecomte T. La morbidit6 d6clar~e, description et 6volution. France 1970-1980. Solidaritd-Santd, dtudes statistiques 1984; 1-2: 41-56. 27 Charraud A. La consommation m6dicale: disparit6s biologiques et soeiales, Evolutions significatives entre 1970 et 1980. Solidaritd-Sant~, dtudes statistiques 1984; 1-2: 13-26. 2SZahavi I, Chagnac A, Hering R, Davidovich S, Kuritzky A. Prevalence of Raynaud's phenomenon in patients with migraine. Arch Intern Med 1984; 144: 742-44.

380 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY 290'Keeffe S T, Tsapatsaris N P, Beetham W P. Association between Raynaud's phenomenon and migraine in a random population of hospital employees. J Rheumatol 1993; 20: 1187-88. 30 Patri B, Lienard M, Vayssairat M. Donn6es cliniques dans le syndrome de Raynaud induit par les vibrations chez le bfieheron. Arch Malf Prof 1987; 48:551-55. 31Terregino C A, Seibold J R. Influence of the menstrual cycle on Raynaud's phenomenon and on cold tolerance in normal women. Angiology 1985; 36-" 88-95. 32Cooke E D, Nicolaides A N. Raynaud's syndrome. Br Med J 1990; 300: 553-55. Ekenvall L, Lindblad LIE. Effect of tobacco use on vibration white finger disease. J Occup Med 1989; 31." 13-16. 34Enander A. Performance and sensory aspects of work in cold environments: a review. Ergonomics 1984; 27: 365-78. Enander A E. Effects of thermal stress on human perform- ance. Scand J Work Environ Health 1989; 15 (Suppl.): 27-33. (Revised version received September 1996)