Cvetkovski RS, Zachariae R, Jensen H, Olsen J, Johansen JD, Agner T. Prognosis of Occupational Hand EczemaA Follow-up Study. Arch Dermatol. 2006;142(3):305-311. doi:10.1001/archderm.142.3.305
Copyright 2006 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2006
To identify prognostic risk factors in patients with occupational hand eczema (OHE).
Cohort study with 1-year follow-up.
Danish National Board of Industrial Injuries Registry.
All patients with newly recognized OHE (758 cases) from October 1, 2001, through November 10, 2002.
Participants received a questionnaire covering self-rated severity, sick leave, loss of job, depression, and health-related quality of life. One year after the questionnaire was returned, all responders (N = 621) received a follow-up questionnaire, and 564 (91%) returned it.
Main Outcome Measures
Persistently severe or aggravated OHE, prolonged sick leave, and loss of job after 1-year follow-up.
During the follow-up period, 25% of all patients with OHE had persistently severe or aggravated disease, 41% improved, and 34% had unchanged minimal or mild to moderate disease. Patients with atopic dermatitis fared poorly compared with other patients. Patients younger than 25 years fared clearly better than older groups. Furthermore, severe OHE, age 40 years or greater, and severe impairment of quality of life at baseline appeared to be important predictors of prolonged sick leave and unemployment. Patients with lower socioeconomic status also had a high risk of prolonged sick leave, job change, and loss of job. Contact allergy was not found to be a risk factor for poor prognosis.
Atopic dermatitis, greater age, and low socioeconomic status may be reliable prognostic factors in early OHE. Quality of life and standardized severity assessment may also be valuable tools to identify patients at high risk of prolonged sick leave and unemployment.
Occupational hand eczema (OHE) has become a disease of increasing importance during recent decades because of its serious consequences, such as frequent eruptions and risk of prolonged sick leave.1- 7 Previous studies also indicate that OHE has an appreciable impact on quality of life (QOL), although the number of publications is limited.8- 10 The magnitude of the problem is supported by the fact that OHE is the most frequently recognized occupational disease in Denmark and many Western countries.11,12 Occupational hand eczema may occur at any age but is most common among young female workers.11- 14 Previous studies have not identified any strong predictors of a poor prognosis, although the presence of atopic dermatitis (AD) or allergic contact dermatitis (ACD) has been related to more severe outcome.10,15- 17 It has been suggested that patients with contact dermatitis may experience depressive symptoms.18- 20 So far, no data are available on the possible associations between depression and the prognosis of patients with OHE. Poor socioeconomic status has been shown to affect outcome adversely in many chronic disorders,1- 28 but no current data are available for patients with OHE.The aim of this study was to identify risk factors for a poor prognosis, defined as persistent severe or aggravated OHE, prolonged sick leave, and loss of job due to OHE in the year following recognition by the Danish National Board of Industrial Injuries (DNBII).
The cohort comprised all patients with newly recognized OHE who were 18 years old or older at the time of registration in the DNBII registry from October 1, 2001, through November 10, 2002 (58 weeks). Our clinical observations suggest that patients are especially vulnerable at an early stage of the disease. We therefore decided to evaluate the prognosis after 1 year of follow-up.
Data on diagnosis and subdiagnosis of OHE, severity assessment of OHE, and socioeconomic status for each person enrolled were registered. Socioeconomic status was measured by means of the socioeconomic classification system Socio97 (1997; Statistics Denmark, Copenhagen), which is a system based on educational requirements and managerial responsibilities in the job. In the analysis we used 4 categories of socioeconomic status: student/trainee, lowest level of employee, basic level, and medium to high level.
To supplement the information from the DNBII, we used 2 postal questionnaires. The baseline questionnaire was mailed within 1 to 2 weeks of case registration. The follow-up questionnaire was mailed 1 year after the baseline questionnaire had been returned.
The classification of diagnosis was based on the clinical examination of each patient by a dermatologist, including patch test and skin prick test. All patients were patch tested with the European Standard Series as a minimum. Patients in specific occupations or with specific exposures were given additional tests with relevant allergens, such as patch test series for hairdressers, bakers, dentists, those with exposure to rubber gloves, and those with plastic exposure.29 All patients were categorized into only 1 of 5 main diagnostic groups: (1) occupational irritant contact dermatitis (ICD); (2) occupational ACD; (3) occupational contact urticaria (CU); (4) ICD + ACD; and (5) ICD + CU. Patients with occupational hyperkeratotic or frictional hand eczema were classified as having ICD.
The patients were also categorized into diagnostic subgroups by the presence or absence of AD and by the presence or absence of nonoccupational ACD (NOACD). In this study, AD was defined as past flexural eczema or AD currently diagnosed by a dermatologist. Nonoccupational ACD was defined as a positive patch test result for a nonoccupational exposure.
Severity was assessed both by the DNBII and by the patients. The DNBII severity assessment was based on medical certificates from dermatologists. This severity assessment took into account the intensity of the skin response as well as the frequency of eruptions and extension on the hands.30 We categorized the DNBII severity assessment into 3 groups: (1) no or minimal OHE, (2) mild to moderate OHE, and (3) severe OHE.
All patients were asked to rate their disease severity on a 100-mm visual analog scale. The exact wording of the question was as follows: “How would you grade your occupational hand eczema on a scale from 0 to 10 during the past 12 months? 0 meaning no eczema and 10 meaning extremely severe eczema.” The participants were also asked whether they had taken any sick leave because of their OHE during the past 12 months. If the answer was yes, they were asked for how long, and then asked to summarize all episodes of sick leave into 1 of the following fixed answering categories: less than 1 week, 1 to 2 weeks, 3 to 5 weeks, and more than 5 weeks. The participants were also asked whether they had lost their job during the past 12 months because of their OHE.
All participants were asked to complete the Beck Depression Inventory31 and the Dermatology Life Quality Index (DLQI).32 The Beck Depression Inventory is a 21-item questionnaire measuring depressive symptoms experienced during the past 2 weeks. Each item is assigned a score of 0 to 3, with 3 indicating the most severe symptoms. A cumulative score is determined by adding the scores of the individual items. The responders were grouped on the basis of their Beck Depression Inventory total scores by means of the following classification: (1) no or minimal depression (score, 0-13), (2) mild depression (score, 14-19), (3) moderate depression (score, 20-28), and (4) severe depression (score, 29-63).
The DLQI is a 10-item questionnaire measuring QOL in skin disease.32,33 The 10 items cover 6 aspects of daily life experienced during the past week: (1) symptoms and feelings (items 1 and 2), (2) daily activities (items 3 and 4), (3) leisure (items 5 and 6), (4) work and school (item 7), (5) personal relationships (items 8 and 9), and (6) treatment (item 10). Each item is assigned a score of 0 (“not at all”) to 3 (“very much”). The DLQI total scores are calculated by summing the score of each question, resulting in a maximum of 30 and a minimum of 0. The higher the score, the greater the impairment of QOL. The scores can be expressed as absolute or percentages scores.
Sick leave due to OHE lasting more than 5 weeks in the past 12 months was considered prolonged sick leave. If a patient was categorized as having prolonged sick leave at baseline, it was classified as previous prolonged sick leave in the subsequent analyses. Changes in severity after 1 year of follow-up were categorized as aggravation, no change, or improvement of OHE. If a patient had severe OHE at baseline and did not improve in 1 year, we classified this patient as having a poor prognosis. Thus, a poor prognosis was defined as either aggravated or persistent severe OHE.
All analyses were performed in Stata, version 8.2 (Stata Corp, College Station, Tex). The visual analog scale scores were transformed from millimeters to percentages, eg, 64 mm = 64%. The scores were then divided into 3 equal-sized groups as follows: (1) no or minimal OHE (<33.3%), (2) mild to moderate OHE (33.3%-66.6%), and (3) severe OHE (66.7%-100.0%). The DLQI total scores at baseline were dichotomized into high QOL (DLQI total score, 0-10) and low QOL (DLQI total score, 11-30).34 We dichotomized depression status into minimal to mild depressive symptoms (Beck Depression Inventory score, 0-19) and moderate to severe depressive symptoms (score, 20-63).
We measured prevalence proportions among patients for basic characteristics and risk ratios for comparison of a poor prognosis across different variables, such as age, sex, diagnoses, socioeconomic status, disease duration, and occupation. We also measured risk ratios for those taking or not taking sick leave and for those who lost their job during follow-up. We estimated risk ratios from Poisson regression models with robust variance estimators.35 Unemployed, early retired, and other pensioned subjects were omitted from the analyses of prolonged sick leave and loss of job.
The DNBII registry had data on 758 eligible patients. The proportion lost to follow-up was 18.1% at recruitment and 13.0% at follow-up. We found no large differences in age, sex, socioeconomic status, diagnosis, or severity between responders and nonresponders at baseline.13 More women (n = 386; 94.6% of female responders at baseline) than men (n = 178; 83.6% of male responders at baseline) returned the follow-up questionnaire. We found no major difference between responders and nonresponders at follow-up with regard to age, severity, socioeconomic status, or diagnosis (ACD, ICD, CU, ICD + ACD, or ICD + CU). Severity was rated by 540 patients in both questionnaires, and the Poisson regression model was based on these 540 responders. For the items regarding sick leave and loss of job, there were small differences in the number of responders and nonresponders, eg, some responders answered the item regarding sick leave and not the item regarding loss of job and vice versa.
The overall improvement was 41%, but for butchers it was 11% and for physicians, dentists, and veterinarians it was 67%. The overall proportion of persistently severe or aggravated symptoms was 25%, also with broad differences between subgroups, eg, 37% for hairdressers and 0% for physicians, dentists, and veterinarians (Table 1). The most severely affected subgroups were butchers, kitchen workers and cooks, hairdressers, and patients aged 25 to 29 years. As shown in Table 1, the proportion of improved cases was similar among the group who reported job change during the 12-month follow-up and those who reported no job change. Patients sensitive to chromium also did not show any signs of a poor prognosis. Only 12% of chromium-sensitive patients had aggravated or persistently severe disease (data not shown).
As shown in Table 2, presence of AD and age 25 years or greater appeared to be associated with poor prognosis. We found no statistically significant association between poor prognosis and sex, ACD, ICD, CU, NOACD, disease duration, occupation, or socioeconomic status, although a slight tendency toward a protective effect of high to medium socioeconomic status was found. Almost 48% of the study population reported job change during the 12 months of follow-up. Job change was associated with younger age groups (52% of those who reported job change were younger than 25 years) and lower socioeconomic status (57% from the lowest level reported job change compared with 24% from the highest level).
Prolonged sick leave was significantly associated with having occupational CU, being 25 years old or more (with statistical significance for the age group 40-49 years), and being categorized as having severe OHE at baseline by the DNBII (Table 3). Self-rated severity produced similar associations to those based on DNBII-rated severity (data not shown). Severe impairment of QOL at baseline, as well as previous reports of prolonged sick leave, were strongly associated with the risk of prolonged sick leave during follow-up. We found a favorable prognosis for patients with high to medium socioeconomic status (no patients with prolonged sick leave in this group). We found no important associations between prolonged sick leave and age, ACD, ICD, AD, NOACD, disease duration, occupation, or the presence of depression.
Table 4 shows strong associations between having severe OHE (risk ratio, 14.00; 95% confidence interval, 1.9-102.9) at baseline and loss of job during the follow-up. Again, we found a favorable prognosis for patients with high to medium socioeconomic status (no patients with loss of job in this group).
We found an overall improvement in OHE after 1 year of follow-up, as expected owing to regression toward the mean.36,37 However, we found large variability among different subgroups that cannot entirely be explained by this phenomenon. Almost 41% of the study population reported improvement during 12 months of follow-up, which is lower than what was reported in a recent Swedish study with 15 years of follow-up in a cohort of patients with occupational and nonoccupational hand eczema.7 In that study, Meding and colleagues reported that 74% improved during the 15 years of follow-up.
Age younger than 25 years was found to be a significant protective factor against a poor prognosis, which supports earlier findings.8 As in previous studies,17,38 we found no effect of sex on the prognosis, although contradictory findings have been published.7,8,39 Poor socioeconomic status has been shown to affect outcome adversely in many chronic disorders.23- 28 We found that the prognosis of OHE in terms of aggravated or persistently severe disease after 1 year was clearly better for patients with high to medium socioeconomic status at baseline, but the association was not statistically significant.
After 1-year follow-up we found that patients with AD had a 1.5-times higher risk of aggravated or persistently severe OHE compared with patients without AD, which is consistent with previous findings.17,30,38,40 The prognosis has been reported to be worse for ACD than ICD in earlier studies,17,39,41- 44 but our results do not support this. We previously reported that patients with ICD at baseline had a higher proportion of severe OHE than did patients with ACD30; however, at 1-year follow-up we could no longer detect any significant differences between the 2 diagnostic groups. Chromium sensitivity has previously been associated with a poor prognosis,40,43,45,46 but our findings could not confirm that; however, the number of chromium-sensitive patients was relatively small in the present study. In a study from 1975, Fregert41 reported that the prognosis was especially poor in women with nickel allergy. Danish legislation since 1991 has led to a reduction in nickel exposure, which may be one of the reasons why we found no association between NOACD and a poor prognosis. The NOACD subgroup in our study consisted predominantly of women with nonoccupational nickel allergy.13
Our results support previous reports of age as a risk factor for prolonged sick leave.3 For patients with high or medium socioeconomic status, there were no reports of taking prolonged sick leave or loss of job at all. This tendency is in accordance with earlier studies among dentists in Sweden,47,48 and the results indicate that lower socioeconomic status is an important risk factor for prolonged sick leave and loss of job. Not surprisingly, we found that patients with severe OHE at baseline had a significantly higher risk of taking sick leave in the year following recognition than did patients with mild cases, which supports results from previous studies.30 Patients with severe OHE were, however, also at considerably higher risk of losing their job during the follow-up period, which, to the best of our knowledge, has not previously been reported. Independently of OHE severity, severe impairment of QOL at baseline was a strong predictor of prolonged sick leave, which underlines the importance of measuring the patients' perceived health-related QOL. We found no significant associations between prolonged sick leave and the presence of depression. We also could not corroborate earlier findings that patients with ACD have a higher risk of prolonged sick leave.3,30,42 Occupational CU appeared to be associated with a high risk of prolonged sick leave, which may be related to strict hygiene regulations in food-related occupations but was most likely due to chance. We found no association between sex and prolonged sick leave, which supports earlier findings.7 Almost 50% of the study population had changed their job during the 12 months of follow-up, which is high compared with findings in the study by Meding et al.7 They found that only 3% of the study population had changed their job during 15 years of follow-up, indicating that change of job occurs at an early stage of OHE. In our study, we found no significant improvement in the disease after the change of job, which is in line with the findings of Meding et al.7 They found that many patients change jobs from low-risk jobs to high-risk jobs. Patients with high to medium socioeconomic status in our study had a tendency to change jobs less often than did patients with lower socioeconomic status, which may indicate that job modification is easier for this group.
Potential sources of bias in this study arise from selective referral and losses to follow-up. The number of notified and thereby recognized cases does not necessarily reflect the true number of patients with OHE, because mild cases with short disease duration may be missed, although we had more than 30% mild cases in the cohort. There were no appreciable differences in age, sex, diagnosis, socioeconomic status, or severity at baseline between those who returned both questionnaires and those who were lost to follow-up, except that there was a high proportion of male nonresponders at follow-up. Our information on job changes and changes in disease status do not include information on the time of change. Conclusions about job change and prognosis should therefore be made with caution.
Predictive factors could be used by clinicians to guide treatment and to select early risk management strategies. To avoid prolonged sick leave that may lead to social and economic decline, physicians must try to identify subgroups of patients who are at greater risk of a poor outcome. We identified AD as a prognostic risk factor for either persistently severe or aggravated disease, which supports earlier findings. Furthermore, we found that patients younger than 25 years and those with high to medium socioeconomic status compared with older age groups and low socioeconomic status fared clearly better with regard to prolonged sick leave and loss of job. This indicates that socioeconomic status may have a place in the currently rather limited number of reliable prognostic factors in early OHE. Also not previously reported, we found that severe OHE and severe impairment of QOL at baseline were strong prognostic predictors of prolonged sick leave. This indicates that QOL and standardized severity assessment may be valuable tools to identify patients at high risk of prolonged sick leave and unemployment.
Correspondence: Rikke Skoet Cvetkovski, MD, PhD, Department of Dermatology, Gentofte Hospital, University of Copenhagen, Niels Andersensveg 65, DK-2900 Hellerup, Denmark (email@example.com).
Financial Disclosure: None.
Accepted for Publication: July 12, 2005.
Author Contributions:Study concept and design: Cvetkovski, Zachariae, Olsen, Johansen, and Agner. Acquisition of data: Cvetkovski. Analysis and interpretation of data: Cvetkovski, Zachariae, Jensen, and Agner. Drafting of the manuscript: Cvetkovski and Zachariae. Critical revision of the manuscript for important intellectual content: Cvetkovski, Zachariae, Jensen, Olsen, Johansen, and Agner. Statistical analysis: Cvetkovski, Zachariae, and Jensen. Obtained funding: Cvetkovski and Agner. Administrative, technical, and material support: Cvetkovski and Agner. Study supervision: Olsen, Johansen, and Agner.
Funding/Support: This study was supported by the Danish National Board of Industrial Injuries, the Danish Council for Research Policy, and the Danish National Institute of Occupational Health, Copenhagen.
Acknowledgment: We thank Bent Mathiesen, MD, Coordinating Chief Medical Consultant, Head of the Medical Unit, Danish National Board of Industrial Injuries, for help and advice during the study. Permission to use the DLQI in the present study was obtained from the Department of Dermatology, Cardiff University, Wales College of Medicine, Cardiff.