Occupational Dermatitis in Welding: Does Nickel Exposure Account for Higher Rates in Women? Analysis of a Canadian Cohort

Abstract

Objectives

Women are reported to have higher rates of nickel sensitization than men, but there have been few studies of sex-related differences in dermatitis associated with occupational nickel exposure. This analysis examines dermatitis in a large cohort of women and men in welding and electrical occupations and considers how far differences in rates of dermatitis may be accounted for by nickel exposure.

Methods

Women and men were recruited to cohorts of workers who had entered welding and electrical apprenticeships (the WHAT-ME and WHAT-MEN studies). Participants completed questionnaires at baseline and every 6 months for up to 5 years. At each contact, cohort members were asked about current dermatitis and whether it was made worse by work. From the first follow-up after recruitment, those working in their trade completed detailed subroutines about tasks in their trade including, for welders, the process, base metal, and consumables. Exposures were considered by trade and, within welding, by stainless or high alloy steel (SOHAS) as the base metal. Urinary nickel concentration was also examined. Using only report of dermatitis that began after entry to the trade, new-onset dermatitis, all episodes of dermatitis, and dermatitis made worse by work were examined against exposure by multilevel, multivariable logistic regression, allowing for potential confounding.

Results

Among 1885 participants (welders; 447 women, 554 men: electrical trades; 438 women, 446 men), 200 reported dermatitis that started before they entered the trade, leaving 1685 for analysis. Women, but not men, who had entered the welding trades were more at risk of new onset [odds ratio (OR) = 1.54; 95% confidence interval (CI) 1.02–2.32] or dermatitis episodes (OR = 1.75; 95% CI 1.10–2.77) than those entering the electrical trades. Within welding, women were more at risk than men of new-onset dermatitis (OR = 1.85; 95% CI 1.15–2.96) and dermatitis episodes (2.14; 95% CI 1.24–3.68) but were not more likely to report these were made worse by work. Use of SOHAS as the base metal was associated with reports that dermatitis was made worse by work (3.54; 95% CI 1.04–12.03), but having adjusted for SOHAS use, women still remained at greater risk. A final analysis considered the effect on risk estimates of removing those welding SOHAS from the analysis. The risk for women of episodes of dermatitis was essentially unchanged, whether the comparison group was male welders or women in the electrical trades.

Conclusions

Welding is associated with risk of dermatitis in women. Although welding of SOHAS was associated with dermatitis that was made worse by work, it does not importantly explain the higher female rates in welding and other causes must be examined to support preventive measures.

Introduction

Contact dermatitis in welders is reported in case series of occupational dermatitis from the UK (Meyer et al., 2000), North America (Warshaw et al., 2019), Finland (Kanerva et al., 2000), and Western Australia (Wall and Gebrauer, 1991). In other studies, from Italy (Rui et al., 2010) and from Sheffield, a metal working city in the UK (Shah et al., 1998), welders were grouped with other metal workers. The number of welders in each study is small, with the largest 167 from the UK (Meyer et al., 2000), and information on the causative agent is scarce. In 19 welders assessed in Western Australia, 15 were said to have irritant contact dermatitis (Wall and Gebrauer, 1991); in the 167 UK workers in welding trades, colophony, solvents, nickel, petrol, epoxies, and resins were among agents mentioned (Meyer et al., 2000). In the Finnish series, which looked specifically at metal allergies, there were 7 welders with chrome sensitization but less than 3 with nickel (Kanerva et al., 2000) while in the North American case series, 5 welders were included among 268 occupationally related nickel allergy patients (Warshaw et al., 2019). Welders working with certain types of stainless steel may be expected to be at risk of nickel allergy (Haudrechy et al., 1997). A recent review of nickel contact dermatitis discusses measurement of the skin dose needed for nickel penetration and resultant allergy (Ahlstrom et al., 2019).

We are not aware of any study of dermatitis that includes women welders although there is general acceptance that women have a higher rate of nickel allergy than men (Ahlstrom et al., 2019) and indeed higher rate of irritant dermatitis (Meding, 2000). Few occupational cohorts of metal workers include dermatitis as an outcome, but an early study of 853 hard metal workers, including 368 women, found higher rates of nickel allergy, combined nickel and cobalt allergy and contact dermatitis in the women in the cohort (Rystedt and Fischer, 1983). Few studies have identified factors associated with susceptibility to occupational dermatitis other than occupational exposures, gender, age, and atopy (Diepgen and Coenraads, 1999), but recent reviews have suggested that contact dermatitis is more common in smokers (Zimmer et al., 2018) and atopic dermatitis found more often in those who were overweight (Zhang and Silverberg, 2015). Dermatologist visits for skin conditions showed marked seasonality in the USA but with no important seasonal variation for contact dermatitis (Hancox et al., 2004).

The WHAT-ME cohort of women in the welding and electrical trades was established to investigate whether welding during pregnancy affected the outcome (Cherry et al., 2018) but also provided the opportunity to examine occupationally related symptoms, including dermatitis. A parallel cohort of men in the same trades allowed the examination of gender differences in the reporting of occupationally related conditions. The present report considers whether women in welding are at increased risk of dermatitis and evaluates whether exposure to nickel is a cause of any such excess.

Methods

Recruitment

Preliminary work comprised focus groups to assess whether a cohort study of birth outcomes would be acceptable to women in the trades, together with development and validation of a self-report questionnaire that accurately reflected welding tasks (Cherry et al., 2011). In discussion with Alberta Apprenticeship and Industry Training (AAIT) we defined welding trades as welding, boiler-making, steam fitting, and pipefitting: these and the electrical trades are all compulsory certification (registered) trades in Alberta. The AAIT agreed to send out letters on our behalf from January 2011 to women who had started an apprenticeship in one of these trades in 2005–2012. With these letters we included an information sheet about the study and a consent sheet which we asked them to sign and send back to the research team, if they were willing to take part in the study. Over the next 2 years we approached apprenticeship boards (or a parallel body) across Canada and asked them to help us identify women who had started a welding or electrical apprenticeship in their jurisdiction since 2005: every province and territory agreed. For apprentices starting in 2013–2014 in Alberta, direct mailing proved possible, following signing of a confidentiality agreement with the AAIT. The study became known as WHAT-ME (Women’s Health in Apprenticeship Trades—Metalworking and Electrical).

A parallel cohort for men in the same trades was set up in 2013 with the first tradesmen recruited in January 2014. This part of the study, which used an identical recruitment procedure but with direct mailing throughout, was restricted to Alberta. To keep the numbers manageable, and similar to those in the WHAT-ME study, we approached a random sample of men to take part in the study and restricted date of apprenticeship start to 2005–2012. The male study is referred to as WHAT-MEN (Workers’ Health in Alberta Trades—MEN).

This approach gave us four cohorts of skilled workers: women and men in the welding and electrical trades. For the women, to include the largest possible number of pregnancies conceived while working in their trade, follow-up continued for up to 5 years. For men, where the focus was on new-onset work-related ill-health, follow-up was for 3 years. Questionnaires could be completed in English or French, by telephone or online. No financial incentive to join the study was offered to any of the cohorts.

Data collected

Baseline information was collected as soon as possible after a consent form was received and subsequently at 6-month intervals. At baseline we collected demographic, substance use, reproductive, and health data together with an occupational history since leaving secondary school and the date of starting training in the trade. At each periodic (6-month) follow-up data were collected on current health problems, use of alcohol, tobacco and medication and all jobs since the last contact were recorded. Detailed trade-specific questionnaires were completed about tasks and activities on the last day at work (in the previous 6 months) on which they carried out their trade. These included the main tasks for the day (yes/no to a list of prespecified tasks), ergonomic demands, the physical environment and, for welders, subroutines for each main type of welding, including types of metal welded, filler rods or wires, type of flux and shielding gas, and coating with paints or primers. For each welding type reported, the participant also answered detailed questions about ventilation and personal protective equipment. The periodic questionnaire for the electrical trades followed the same design but without the specifics of welding-related task exposures. Women who became pregnant while in the trade also completed the detailed trade-specific questionnaires in early pregnancy and these data were included with the periodic data for all other participants.

In addition, for those in Alberta who gave consent, information on physician contacts was extracted from the Alberta administrative health database: in Canada healthcare is free at the point of service and essentially all residents have a personal health number (PHN). For a physician to receive a fee for service they must record both the diagnosis and PHN. The diagnosis may be made available for research, if the patient consents.

Spot urine samples were mailed-in by women during the early months of the WHAT-ME study (Arrandale et al., 2015) together with a further sample during the first trimester of any pregnancy. An additional round of urine collection included all men and women still working in their trade.

Outcome

At baseline and on each follow-up questionnaire the participant was asked

• Do you have any of the following health problems now?

• Do you have dermatitis or itchy/inflamed skin when you do not have insect bites or sunburn? Yes/No

• If yes: Was this made worse by work? Yes/No.

Two sources were used to determine whether dermatitis was present before entry to the trade.

• (1) At baseline (only) the participant was asked (if yes to dermatitis now) How old were you when this started?

• (2) For those who had been recruited in Alberta and who gave consent to data linkage, information about contact with a physician for a skin condition was obtained from the Alberta administrative health database for the period 1 April 2002–31 March 2018.

Three definitions of dermatitis were used.

• (1) New-onset dermatitis. This was the first occasion on which a participant reported dermatitis since entry to the trade. Those with dermatitis before entry to the trade were excluded.

• (2) Episodes of dermatitis. Each episode of dermatitis ‘now’ reported on a questionnaire, at baseline or periodic follow-up, with the date of the episode taken as the date of the questionnaire. Again, those with onset known to be before the start of trade were excluded.

• (3) Dermatitis made worse by work. This was restricted to those who, on that questionnaire, reported dermatitis. They reported (yes/no) whether this dermatitis was made worse by work. If there were reports of dermatitis on multiple questionnaires, there would be multiple responses to this question. Those with known onset before entry to the trade were excluded.

Exposure

Three exposures were considered

• (1) Having undertaken an apprenticeship in the electrical or welding trade.

• (2) Working in the trade and carrying out trade tasks (welding or electrical tasks) within 31 days before the report (yes/no) of dermatitis. At baseline, when days since last trade tasks was not recorded, all who reported they were currently working in their trade were assumed to have carried out trade tasks within the last month.

• (3) Welding of stainless or high alloy steel (SOHAS) on the last day welding within the 31 days before the report of dermatitis. Such steel is a source of exposure to nickel and also to other potential sensitizers, such as chrome. This was limited to those welders who completed periodic (follow-up) questionnaires 6 months or more after the baseline questionnaire.

In addition, nickel concentrations were measured in urine samples collected within 31 days of the report of dermatitis. As reported previously (Arrandale et al., 2015) samples were analyzed at the University of Alberta Hospital by ICP-MS (PerkinElmer NexION 300) for 17 metals, including nickel. Creatinine was estimated by the Beckman Jaffe method.

Potential confounders and other exposure variables

Information on cigarette smoking, height, and weight were collected at baseline. Smoking and weight were updated on each follow-up questionnaire. Age, current cigarette smoking (yes/no), and body mass index were included as potential confounders. In addition, the month in which the questionnaire was completed was coded as winter (November–April) or summer (May–October) reflecting the 6 months of cold weather in Alberta.

Statistical methods

The characteristics of those entering the welding and electrical trades were compared, using analysis of variance for means and chi-square for proportions. Comparisons were made of men and women within trade and women between trades. Further analysis was restricted to those working in their trade within 31 days before responding to the periodic question on dermatitis. The odds ratios (ORs) from these analyses represented the risk of reporting dermatitis on any one questionnaire, not a lifetime risk of dermatitis. Using a multilevel, multivariable logistic regression with robust variance (melogit within Stata 14.2) which allowed for multiple time periods for the same participant, the relation of new-onset dermatitis, episodes of dermatitis, and dermatitis made worse by work was examined first by trade, stratified by sex. The difference in dermatitis between women and men in welders was then tested, again with melogit and a Cox regression, with robust variance and testing of the proportional hazards assumption carried out, comparing time to new-onset dermatitis in male and female welders. The effects, in welders, of working with SOHAS was examined in further multilevel, multivariable logistic regression models. The final models compared dermatitis in women and men in welding and women in the welding and electrical trades with the same models when those working with SOHAS were excluded. For those in Alberta a more complete record of dermatitis was available and a sensitivity analysis was carried out of the effect of incomplete exclusion of prior dermatitis on the three outcome variables for men in Alberta.

Results

Data were obtained for 1885 participants. These comprised 447 female and 554 male welders and 438 female and 446 male workers in electrical trades. At the time of recruitment to the study (the date of completing the baseline questionnaire) women were younger than men in both trades (Table 1) and, in welding, had started in the trade more recently and had fewer cumulative months in the trade. Women in welding were younger than women in the electrical trades, had started in the trade more recently, and had worked in the trade for fewer months. Women in welding were more likely than women in electrical trades to be a smoker at baseline. A somewhat greater proportion of women than men in both trades reported dermatitis that had started before the entry to the trade but this was significant only for those in the electrical trades. Overall, using the information from the Alberta administrative health records in addition to self-report, a much larger proportion had an episode of dermatitis between April 2002 (when records were available) and start of trade: if we compare just those from Alberta the difference between women and men is greater, with 27.7% of women and 16.4% of men (P < 0.001) having had dermatitis before they entered the trade. This difference is seen both in those entering the welding trade (26% women and 16% men) and the electrical trades (20.0% women and 16.4% men). The number that, by their own report, developed new-onset dermatitis while in the study is shown in the final row of Table 1. Women who had entered the welding trade were more likely than men in that trade to report dermatitis but were no more likely to do so than women (or men) in the electrical trades.

Table 1.

Characteristic of those in the study by sex and trade at baseline.

	Welders				Electrical trades
	Women		Men		Women		Men
	Mean	SD	Mean	SD	Mean	SD	Mean	SD
Age (years)	29.4	8.6	34.9*	9.4	32.3**	9.2	34.8*	9.7
Year started in trade	2008.0	5.3	2005.0*	7.7	2006.4**	6.5	2006.6	5.3
Months in trade	44.2	52.0	91.9*	80.2	59.6**	62.6	79.3*	60.9
BMI (kg m−2)	25.2	4.5	27.4*	4.8	25.5	4.5	26.9*	4.4
	%	N	%	N	%	N	%	N
Current smoker	29.1**	130	26.9	149	17.6	77	9.0*	40
Dermatitis before started in trade	12.5	56	9.2	51	13.5	59	7.6*	34
New-onset dermatitis while in the study	30.0*	134	24.4	135	29.0	127	28.5	127
N	447		554		438		446

BMI, body mass index.

*Difference (P < 0.05) between men and women within trade.

**Difference (P < 0.05) between women in welding and electrical trades.

Table 2 shows the number of responses to the dermatitis question on the baseline and all follow-up questionnaires, limited to those on which the participant reported carrying out trade tasks on the last day at work within 31 days of answering the dermatitis question. This allowed for the inclusion of multiple reports for each participant, but limited to periods in which they were active in their trade. For example, for women with first episode of new-onset dermatitis, 628 women contributed 1931 observations: that is each woman had, on average, 3.25 periods when she was at work in her trade in the 31 days before completing the dermatitis question. It should be noted that, in both trades, there are more observations for men than women and that welders have fewer observations than electricians, making comparisons of ‘lifetime’ comparisons between the groups (as in the last line of Table 1) more difficult to interpret. Further scanning of Table 2 suggests that smoking but not season of completing the questionnaire was related to each of the dermatitis outcomes and that the welding trade was a more important factor for women than men. This was examined more formally in Table 3. More women in welding than in the electrical trades were found to have new-onset dermatitis (OR = 1.54) and to have any episode of dermatitis (OR = 1.75) but were not significantly more likely to report that their dermatitis was made worse by work. In contrast, men in welding were no more likely than men in electrical trades to have new onset or any episode of dermatitis but, if dermatitis had started since he was in his trade, those in welding were much more likely to say this had been exacerbated by work. As can be seen from Table 2, this increased risk reflected a lower report among men in the electrical trades rather than a remarkably high rate among male welders. A sensitivity analysis, including only men with incident dermatitis using also information from the Alberta administrative health database, resulted in the same conclusions for men: no difference between the trades for new-onset dermatitis or dermatitis episodes but a much increased risk of reporting dermatitis made worse by work, here an OR of 9.33, 95% confidence interval (CI) 2.33–42.29 based on 182 participants and 324 observations. Smoking was a more important factor for men in these analyses and body mass index for women. Age was related to episodes of dermatitis in both women and men. Season of reporting was not related to any of the dermatitis outcomes.

Table 2.

Numbers of observations of new-onset dermatitis, episodes of dermatitis, and dermatitis made worse by work in women and men apprenticed to welding and electrical trades.

	New-onset dermatitisa			Dermatitis episodesa			Dermatitis made worse by worka
	N	N with new onset	%	N	N with episode	%	N	N made worse by work	%
Women
Trade
Electrical	1116	82	7.4	1407	174	12.4	170	104	61.2
Welding	815	79	9.7	975	143	14.7	138	96	69.6
Current smoker
No	1544	119	7.7	1904	238	12.5	231	141	61.0
Yes	387	42	10.9	478	79	16.5	77	59	76.6
Season
Summer	776	67	8.6	951	132	13.9	130	81	62.3
Winter	1155	94	8.1	1431	185	12.9	178	119	66.9
N (observations)	1931	161	8.3	2382	317	13.3	308	200	64.9
Men
Trade
Electrical	1370	105	7.7	1684	238	14.1	229	103	45.0
Welding	1163	85	7.3	1318	146	11.1	144	106	73.6
Current smoker
No	2150	149	6.9	2557	314	12.3	304	159	52.3
Yes	383	41	10.7	445	70	15.7	69	50	72.5
Season
Summer	1741	128	7.4	2084	275	13.2	266	152	57.1
Winter	792	62	7.8	918	109	11.9	107	57	53.3
N (observations)	2533	190	7.5	3002	384	12.8	373	209	56.0

^a Participants reporting dermatitis before started in trade are excluded. Only those working in their trade in the 31 days before the questionnaire are included.

Table 3.

New-onset dermatitis, episodes of dermatitis, and dermatitis made worse by work in women and men apprenticed to welding and electrical trades. Logistic regression analysis.

	New-onset dermatitisa			Dermatitis episodesa			Dermatitis made worse by worka
	OR	95% CI	P	OR	95% CI	P	OR	95% CI	P
Women
Trade
Electrical	1	—		1	—		1	—
Welding	1.54	1.02–2.32	0.039	1.75	1.10–2.77	0.018	1.13	0.40–3.14	0.820
Current smoker
No	1	—		1	—		1	—
Yes	1.47	0.92–2.34	0.105	1.43	0.87–2.34	0.156	3.61	1.10–11.89	0.035
Season
Summer	1			1			1
Winter	0.92	0.62–1.37	0.684	0.89	0.56–1.41	0.626	1.01	0.37–2.77	0.979
Age	1.03	1.00–1.05	0.032	1.04	1.01–1.07	0.003	0.99	0.93–1.04	0.627
BMI (kg m−2)	1.06	1.02–1.11	0.009	1.06	1.01–1.12	0.022	1.03	0.93–1.14	0.592
N (participants; observations)	628; 1931			643; 2382			171; 308
Men
Trade
Electrical	1	—		1	—		1	—
Welding	0.73	0.47–1.15	0.176	0.63	0.39–1.02	0.059	9.00	2.21–36.67	0.002
Current smoker
No	1	—		1	—		1	—
Yes	1.96	1.15–3.31	0.013	2.08	1.19–3.65	0.011	2.77	0.73–10.44	0.133
Season
Summer	1			1			1
Winter	1.15	0.74–1.78	0.540	1.04	0.63–1.72	0.864	1.00	0.34–2.98	0.995
Age	1.02	1.00–1.05	0.060	1.04	1.02–1.07	0.002	0.96	0.90–1.01	0.134
BMI (kg m−2)	1.00	0.96–1.04	0.944	1.00	0.95–1.05	0.964	0.97	0.87–1.09	0.627
N (participants; observations)	809; 2533			820; 3002			206; 373

BMI, body mass index.

^a Participants reporting dermatitis before started in trade excluded.

The difference between women and men in reporting dermatitis that developed while welding is shown in Table 4. Women were more at risk of both new-onset dermatitis and of dermatitis episodes, with smoking and age as additional factors, but not of dermatitis made worse by work. A Cox regression of time to first dermatitis in welders is shown in Fig. 1. Women were seen to have more, and earlier, dermatitis.

Table 4.

Dermatitis by sex within welders: logistic regression analysis.

	New-onset dermatitisa			Dermatitis episodesa			Dermatitis made worse by worka
	OR	95% CI	P	OR	95% CI	P	OR	95% CI	P
Sex
Male	1	—		1	—		1	—
Female	1.85	1.15–2.96	0.011	2.14	1.24–3.68	0.006	0.75	0.29–1.95	0.556
Current smoker
No	1	—		1	—		1	—
Yes	1.71	1.10–2.67	0.018	1.88	1.16–3.05	0.010	2.65	0.97–7.29	0.059
Season
Summer	1			1			1
Winter	1.17	0.77–1.80	0.465	1.23	0.74–2.05	0.416	0.99	0.39–2.50	0.982
Age	1.04	1.02–1.07	0.001	1.06	1.03–1.09	<0.001	0.98	0.93–1.03	0.347
BMI (kg m−2)	1.02	0.97–1.07	0.506	1.02	0.96–1.08	0.535	1.01	0.91–1.11	0.879
N (participants; observations)	735; 1978			747; 2293			175; 282

BMI, body mass index.

^a Participants reporting dermatitis before started in trade excluded.

Figure 1.

Survival curve of first onset dermatitis occurring after start of trade in welders working in their trade in last 31 days (n = 735). BMI, body mass index.

The role of nickel in causing new-onset dermatitis, episodes of dermatitis, or dermatitis made worse by work is considered in Table 5 which shows the risks associated with the handling of SOHAS as the base metal for welding tasks. This analysis includes only those with a work period meeting criterion for inclusion on or after the first (6-month) follow-up as information on metals worked was not collected at baseline. All the ORs for the risk of handling SOHAS were greater than unity, but only that for reporting dermatitis made worse by work was significant. In 215 welders meeting the inclusion criteria for Table 5 and who gave a urine sample within 31 days of completing the questionnaire, 25 reported new-onset dermatitis. The mean urinary concentration of nickel, corrected for creatinine, was 1.71 [standard deviation (SD) = 1.69] µg g⁻¹ creatinine in those with new-onset dermatitis compared with 1.96 (SD = 1.56) µg g⁻¹ creatinine in those not reporting new-onset dermatitis (P = 0.441). Nickel concentrations in electricians meeting the same inclusion criteria were, for those reporting new-onset dermatitis, 1.60 (N = 19, SD = 1.62) µg g⁻¹ creatinine, for those without 1.44 (N = 174, SD = 1.03) µg g⁻¹ creatinine (P = 0.54). Overall the mean urinary nickel concentration was higher in the 215 welders (1.93; SD = 1.57 µg g⁻¹ creatinine) than the 193 electricians (1.45; SD = 1.10) P < 0.01.

Table 5.

Welders: current exposure to SOHAS by dermatitis.

	New-onset dermatitisa			Dermatitis episodesa			Dermatitis made worse by worka
	OR	95% CI	P	OR	95% CI	P	OR	95% CI	P
Current exposure to SOHAS
No	1	—		1	—		1	—
Yes	1.82	0.74–4.46	0.192	1.29	0.66–2.50	0.455	3.54	1.04–12.03	0.043
Sex
Male	1	—		1	—		1	—
Female	2.67	0.83–8.64	0.100	2.08	1.09–3.97	0.027	0.63	0.21–1.93	0.418
Current smoker
No	1	—		1	—		1	—
Yes	1.49	0.73–3.08	0.277	1.83	1.02–3.28	0.042	1.94	0.67–5.63	0.222
Season
Summer	1	—		1	—		1	—
Winter	1.24	−0.63–2.48	0.532	1.45	0.78–2.67	0.238	1.71	0.59–4.95	0.324
Age	1.06	0.99–1.14	0.091	1.06	1.02–1.10	0.002	0.99	0.94–1.06	0.830
BMI (kg m−2)	1.01	0.93–1.10	0.730	0.99	0.92–1.05	0.704	0.95	0.843–1.07	0.387
N (participants; observations)	490; 1257			535; 1544			140; 217

BMI, body mass index.

^a Participants reporting dermatitis before started in trade excluded.

Table 6 seeks to address directly the question of whether exposure to nickel (or more specifically SOHAS) accounted for the higher rate of dermatitis in women welders. The analysis summarized in the table considered, in the upper part of the table, how much the risk of dermatitis associated with being female changes, among welders, when those working with SOHAS were removed. The table shows first the risk among all in this subpopulation of those working in their trade at the 6-month or later follow-up, and then the results when those handling SOHAS were removed. The ORs associated with being a woman welder reduced very little, if at all. The second comparison, in the lower half of the table, was of the risk associated with being in the welding trades, in comparison with the electrical trades. Again the ORs are shown for all in the subpopulation and when those welding SOHAS were removed. In this subpopulation the estimation of risk associated with each dermatitis outcome was reduced, but with a remaining significant risk of episodes of dermatitis in welding.

Table 6.

Change in risk of dermatitis when participants welding SOHAS are removed from the logistic regression analyses.

	New-onset dermatitisa			Dermatitis episodesa			Dermatitis made worse by worka
	ORb	95% CI	P	ORb	95% CI	P	ORb	95% CI	P
Welders
All in trade at follow-up
Men	1	—		1	—		1	—
Women	2.91	1.07–7.90	0.036	2.11	1.10–4.05	0.024	0.69	0.21–2.22	0.533
N (participants; observations)	490; 1257			535; 1544			140; 217
All in trade not including SOHAS
Men	1	—		1	—		1	—
Women	3.00	1.21–7.39	0.017	1.90	0.97–3.71	0.060	0.56	0.18–1.75	0.315
N (participants; observations)	455; 1101			501; 1348			133; 196
Women in welding or electrical
All in trade at follow-up
Electrical	1	—		1	—		1	—
Welding	1.67	0.92–3.04	0.094	1.86	1.09–3.15	0.022	1.15	0.40–3.30	0.791
N (participants; observations)	456; 1323			509; 1752			151; 257
Women in trade not including SOHAS
Electrical	1	—		1	—		1	—
Welding	1.45	0.78–2.68	0.238	1.63	0.96–2.78	0.032	0.85	0.30–2.36	0.749
N (participants; observations)	440; 1249			493; 1656			141; 237

^a Participants reporting dermatitis before started in trade excluded.

^b Adjusted for smoking, age, season, and body mass index.

Discussion

The analyses reported here aimed to assess whether any excess of dermatitis in women welders could be attributed to nickel exposure. It has demonstrated that women welders, without a history of dermatitis when they joined the trade, do indeed have more dermatitis than either women in the electrical trade or men in welding. This is true both of the risk of an initial episode (new onset) of dermatitis and of any dermatitis episode, but not of dermatitis made worse by work. In welding, nickel exposure is almost entirely through the welding of SOHASs, although the concentration of airborne (inhalable) nickel will depend also on the process and nickel content in consumables (rods and filler metals). In this analysis we showed that welding SOHAS rather than absorbed nickel (as seen in urinary nickel concentration), was associated with dermatitis made worse by work, supporting the expectation that skin contact was responsible for the reaction and consistent with type IV sensitization (Ahlstrom et al., 2019). The significant residual effect attributable to gender, having allowed for SOHAS welding, suggested that exposure to nickel (and perhaps other sensitizers) through SOHAS did not fully account for the excess of dermatitis in women welders.

This analysis has strength in that it was able to include a large number of women in welding together with the two comparison groups. It also collected data on changing exposures and episodes of dermatitis in a prospective cohort and so was able to include multiple episodes for the same participant, limiting these to periods during which they were exposed to welding. It was able to confirm a role for smoking and increasing age in dermatitis in both men and women in these cohorts, and the increased risk with higher body mass index in women. The season in which the questionnaire was completed was not related to reporting of dermatitis. Although urinary nickel was not related to dermatitis, suggesting skin contact rather than absorption was more relevant to dermatitis in this population, urinary concentration was found to be elevated in welders, both in relation to urinary concentrations in electricians and to mean nickel concentration in a population sampled from Canadians aged 20–39 years, for whom a mean concentration of 1.45 was reported with the 95% CI (1.30–1.60) excluding the arithmetic mean (1.93) for welders in this study (Health Canada, 2010).

The study also had weaknesses. The number of in-trade observations/participant differed between men and women, with more in-trade observations for men. There were fewer observations for welders than electricians for both men and women, reflecting the employment of many welders on short-term contracts in oil sands developments in the north of Alberta, resulting in more periods without trade employment and less frequent questionnaire completion. These differences did not undermine the analysis of dermatitis by observation (as in Tables 2–6) but did make interpretation of ‘lifetime’ dermatitis more difficult. There was no clinical validation of the responses to the question about ‘dermatitis or itchy/inflamed skin’ which may have been interpreted in different ways. It is likely that under or over-reporting would be non-differential with respect to trade, but it is possible that there were interpretation differences between men and women. Further, we did not know anything about the type of stainless steel welded which has been shown to be a determinant of nickel sensitization (Haudrechy et al., 1997) and we had no measure of skin deposition. It was not possible to test participants for sensitivity to nickel or other metals, nor to say which, if any, participants had been sensitized since joining the trade: it is likely that much of the dermatitis was irritant (Wall and Gebrauer, 1991; Meyer et al., 2000), although the worsening of dermatitis caused by work with SOHAS in those with no dermatitis before joining the trade would be consistent with sensitization. On average welders had been in their trade for more than 3.5 years at the time of recruitment. If sensitivity develops during the early months at work (Rysdedt and Fischer, 1983; Meding 2000), there may have been differential drop-out from the trade, with those leaving welding because of dermatitis not volunteering for the cohort. The data were also lacking in that we do not have history of atopy. Further, although we attempted to restrict the analysis to those who had developed dermatitis since joining the trade, comparison of self-report with episodes included in the Alberta administrative health database showed that participants under-reported a history of dermatitis. For men, all from Alberta, an assessment of the effect of excluding both self-reported and health record dermatitis prior to the trade suggested no major change in estimates of effect. In addition, detailed subroutines on welding tasks were available only for the follow-up questionnaires: by later follow-ups many women, in the trade at recruitment, were no longer in paid employment or were working outside welding, reducing the size of the population of greatest interest for this study. Finally, we have no information on the part of the body on which the dermatitis usually appeared. Most occupational contact dermatitis appears on the hands, wrists, and forearms (Sasseville, 2008). If that is the case here, more thought needs to be given to the type of protection that might best be employed: almost all reported wearing gloves while welding, although we know nothing about the type, efficiency, and frequency of changing them.

More than one in four welders reported dermatitis at some point during this study and the results after adjustment for SOHAS suggest that, if the rate of dermatitis is to be reduced substantially, causes other than simply the base metal must be considered. Possibly relevant exposures reported by these welders included solvents and other solutions used in surface cleaning and also paints or primers on metals being ground or welded. Irritant dusts and friction may also have contributed (Meyer et al., 2000) and photodermatitis from ultraviolet radiation may be an important contributor. Even amongst those whose dermatitis is made worst be handling SOHAS, there may be other responsible agents including chrome in addition to nickel: these possibilities have not been explored sufficiently in this population to comment here. The excess of dermatitis in women in welding, compared with the electrical trades, points strongly to occupational exposures in welding that could, perhaps, best be identified within an industrial cohort with more complete evaluation of the type of dermatitis (allergic or irritant) and the circumstances of exposure. The increasing numbers of women entering nontraditional trades provide both the incentive and opportunity to identify and mitigate risks that may be present for all workers but emerge earlier or more evidently in women.

Funding

Funding for this project was provided by WorkSafeBC, the Government of Alberta’s OHS Futures program, the Canadian Institutes of Health Research, and the Canadian Standards Association.

Ethical approval

The research described here was approved by the Health Ethics Review Board of the University of Alberta.

Conflict of interest

The authors declare no conflict of interest relating to the material presented in this article. Its contents, including any opinions and/or conclusions expressed, are solely those of the authors.