Results

Questionnaires were completed by 12 907 (58%) of the sample originally selected for study. Among the non-responders were 221 (1% of those selected) whom general practitioners advised us not to contact, 1028 (4.6%) who had moved to an unknown address or died, and 56 (0.25%) whose questionnaire was completed by someone other than the intended recipient. Non-response tended to be more common in the younger age bands and in men. The 12 907 responders comprised 6913 men and 5994 women, and their occupational distribution was similar to that in the 1991 national census except that the armed forces were overrepresented. Further details of the pattern of response are reported elsewhere.10

Altogether, 1835 (14.2%) of those who responded gave a history of finger blanching at some time, including 1529 (11.8%) in whom the blanching was cold induced, and 597 (4.6%) in whom there was also a clear edge to the blanched part of the finger. Table 1 shows the lifetime prevalence of blanching by age and sex. Complaints tended to be more common in women than men.

The prevalence of blanching (standardised to the age and sex distribution of all responders) was similar when questioned in the summer and winter, and was little different in subjects who responded to our first letter compared with those who only replied after a reminder. Nor were there any large regional differences in the age standardised prevalence of symptoms (table 2). There was, however, a lower age standardised frequency of complaints among members of the armed forces.

Table 3 shows the distribution of subjects with cold induced finger blanching according to the severity of symptoms and whether they had led to medical consultation. Most subjects reported that they had fewer than 10 episodes a year, but more than a third had at some time been obliged to discontinue an activity because of an attack, and one fifth (2.0% of all men and 2.8% of all women who responded to the questionnaire) had consulted a doctor about the problem.

Cold induced finger blanching was reported more often by men and women who also said that they had had frequent headaches (PR 1.3, 95% CI 1.1 to 1.5) or often felt tired or stressed (PR 1.6, 95% CI 1.5 to 1.8). Table 4 shows the association of cold induced finger blanching with smoking and exposure to HTV after adjustment for the presence or absence of these other symptoms and also for age. In men smoking was associated with an increased risk of cold induced blanching overall (PR 1.4, 95% CI 1.2 to 1.6), but not with medical consultation for the symptom. No association with smoking was found in women.

Occupational and leisure time exposure to vibration were both associated with symptoms, including those leading to medical consultation. The relation was evident in both sexes, but stronger in men than women. For example, in comparison with men who had never been exposed to HTV, the PR for cold induced blanching in those exposed both at work and in leisure was 2.5 (95% CI 2.1 to 3.1), whereas the corresponding risk estimate in women was 2.1 (95% CI 1.3 to 3.2). Furthermore, the association between HTV and cold induced finger blanching was specific. Thus, no increased risk of the symptom was found with exposure to whole body vibration from industrial vehicles (PR in men 1.1, 95% CI 0.7 to 1.6). Nor did exposure to HTV carry any marked increase in the risk of low back pain (PR in men 1.1, 95% CI 1.0 to 1.1).

Tables 3 and 4 record findings in relation to cold induced finger blanching, but similar patterns were apparent for finger blanching overall and for cold induced finger blanching with a clear edge. Thus, in men who had been exposed to HTV both at work and in leisure the PR for blanching overall was 2.4 (95% CI 2.0 to 2.9), and that for cold induced finger blanching with a clearly demarcated edge was 2.9 (95% CI 2.0 to 4.0).

Table 5 gives estimates of the proportion of cases of Raynaud's phenomenon in Britain that are attributable to exposure to HTV. In men about a third of cases are estimated to arise from exposure to HTV, and this estimate was little affected by varying the case definition. In men who consulted a doctor about cold induced blanching the attributable proportion was slightly higher (37.6%). In men who reported extensive blanching (that is, blanching affecting at least eight of the digits or 15 phalanges), the estimated attributable number of cases nationally was 222 000. In women, by contrast, estimates of the attributable proportion were much lower: <4% for blanching overall and 5.3% among those who had consulted a doctor about the problem.

Discussion

Our findings indicate that Raynaud's phenomenon is common in the general population, that more than 2% of adults have consulted a doctor about the problem, and that among men who present to a doctor with the disorder more than a third are likely to have developed it through exposure to HTV.

The sample selected for study was chosen to provide information that within each occupation would be representative of Britain as a whole, and it included more than 21 000 men and women from general practices across the country. In Britain almost all people register with a National Health Service general practitioner (apart from members of the armed forces, who were sampled separately). Nevertheless, bias could have arisen if subjects from certain occupations or with finger blanching responded preferentially to the questionnaire. To test for this, we compared the distribution of occupations in the responders with that at the national census and also the prevalence of symptoms in people who responded with and without a reminder (table 2). In both cases only minor differences were found.

Another potential source of error was inaccuracy in the information about disease and exposure that was obtained from the postal questionnaire. Previous community surveys which ascertained finger blanching by the same question15 or a similar question16 to that used in our study have found a predictive value of about 80% for Raynaud's phenomenon as judged by follow up medical interview; and before using our questionnaire, we tested its validity in a sample of patients from rheumatology and general practice clinics, including 18 with a clinical diagnosis of Raynaud's phenomenon. Against this standard, the questions about finger blanching were found to have a high sensitivity (94%–100%) and good specificity (75%–78%).17 Some errors of recall may have arisen as the analysis was based on a lifetime rather than a recent history of finger blanching, but the disease tends to persist, and almost all of our cases indicated that blanching had occurred in the past 12 months.

We explored the accuracy with which the questionnaire ascertained exposures to HTV in subjects' current jobs, and found that reports were generally plausible.18 None the less, associations between HTV and blanching could have been exaggerated if symptomatic subjects tended to recall exposure better than the average, or if people exposed to vibration were more aware of symptoms than those who were not exposed. That finger blanching was reported more often by subjects who also complained of frequent headaches or tiredness and stress suggests that people do vary in their propensity to report symptoms. However, the association of HTV with blanching persisted after adjustment for these other complaints, whereas no such association was found with back pain. Moreover, the risk of finger blanching was not increased in subjects exposed to whole body vibration from industrial vehicles. This specificity of the relation between HTV and finger blanching argues against major inflation of risk estimates from reporting bias. Also, if risks were spuriously increased from biased ascertainment of symptoms, associations might be expected to be stronger for finger blanching overall than for the more specific complaint of cold induced blanching with a clear edge; but this was not found.

Two previous surveys have examined the prevalence of Raynaud's phenomenon in British communities. Among 450 adults aged 20–59 years who were registered with a Hampshire general practice, the prevalence of episodic finger blanching was 8.3% in men and 17.6% in women,15 based on a question identical to that used in this study. And a survey of patients from five general practices in London, Merseyside, and Cheshire found that the prevalence of cold induced blanching associated with numbness and parasthesiae was 11%–16% in men and 19%–21% in women.19 Our findings, based on a much larger and geographically more representative sample, are broadly similar, with again a higher prevalence in women than in men. Studies in other countries have suggested that Raynaud's phenomenon is more common in colder climates.20 21 We found little variation in the prevalence by region or by season, but age adjusted rates were markedly lower in military personnel than civilians. This may reflect a different attitude in the services to the reporting of what may be perceived as minor symptoms, and it is possible that a healthy worker selection effect has also contributed.

Around one fifth of cases in our study had consulted a doctor about their symptoms. Thus, it is important for physicians to be aware of risk factors for the disorder, particularly where they might offer scope for secondary prevention. Smoking has been reported to cause acute digital vasoconstriction,22 and has been associated with Raynaud's phenomenon in a few previous studies.23 24 25 We found an increased risk of finger blanching in male smokers, but the PR was only 1.4, and there was no corresponding association in women. Our analysis was aimed at determining whether smoking predisposes to the development of Raynaud's phenomenon, and for this purpose we used lifetime histories of smoking and finger blanching. We cannot exclude the possibility that smoking acutely exacerbates pre-existing disease, but our findings suggest that any impact it has on the incidence of new disease is small.

By comparison, the influence of exposure to HTV is much greater, especially in men among whom it carried a doubling of risk and accounted for about a third of all cases. The impact of HTV at a national level has not been systematically investigated previously. Estimates of the frequency of disease associated with vibration have come largely from compensation claims under the Department of Social Security's Industrial Injuries Scheme and from employers' statutory reports under the Reporting of Injuries, Diseases and Dangerous Occurrences (RIDDOR) Regulations,26 but these official schemes, which are heavily influenced by awareness and willingness to report, do not provide an adequate basis for determining the number of attributable cases. Information is also available from two labour force surveys carried out by The Health and Safety Executive in 1990 and 1995.27 28 These suggest a lower attributable proportion—no more than 5% in men assuming a prevalence of 5% for Raynaud's phenomenon in the general population. However, in these surveys no direct inquiry was made about finger blanching (only work related illnesses in general). Moreover, cases were only ascertained if they themselves attributed their disease to an occupational exposure, and the link with occupation often may not be apparent to the person concerned.

Elsewhere, a study in South Carolina found crude odds ratios of 1.9–2.0 in men and 1.4 in women for the association of HTV exposure with cold sensitivity and colour changes in the fingers or toes29; and a multicentre investigation in France and the United States found an odds ratio of 1.93 for Raynaud's phenomenon in people who had used vibratory tools.20 These risk estimates are compatible with ours.

The fact that many of our cases had consulted a doctor about their finger blanching underlines the need for general practitioners to be aware of the relation between Raynaud's phenomenon and HTV. Treatment options for established cases are limited, but there is some evidence that symptoms may improve (over several years) in workers who withdraw from or limit further exposure to HTV.30-33 In one study, a diminished prevalence of symptoms in former chain saw users was accompanied by an improvement in measured finger blood flow, but there was less improvement in older subjects and those with more advanced disease, suggesting that early withdrawal may be particularly beneficial. Continued exposure, by contrast, may cause symptoms to worsen.24 34-39 It follows that the scope for avoiding or reducing exposure is worth exploring. How far this is pursued, however, will depend on the severity of disease and the potential cost to the patient—for example, in reduced income—from a change in the work he does. It is also possible that the patient will be eligible for compensation if their illness has been caused by work.

Our findings are also important in emphasising the public health impact of HTV. They suggest that the numbers in Britain who may benefit from screening, counselling, and compensation are larger than previously supposed, and that a higher priority should be accorded to controlling the risks from this common occupational hazard.