The main purpose of analytic studies, including cohort and case-control studies, is to identify factors that may influence the disease frequency and to precisely estimate their contribution to disease occurrence. The results of analytic studies are best expressed in terms of relative risk. The relative risk is the ratio of the disease incidence among those exposed to a purported causal factor (risk factor), to the incidence among the unexposed. When derived from case-control studies, odds ratios provide an estimate of the relative risks.
Limited information is available on risk factors for HS. A familial history has been repeatedly documented in subsets of patients [2, 9, 10]. In a study, 14 out of 26 consecutive probands (53.8%) had a positive family history. Data from 11 families (42.3%) suggested a single gene transmission with an autosomal dominant inheritance. However, the disease frequency in the analyzed families felt short of the 50% expected of an autosomal dominant condition . Among the reasons for the discrepancy, a possible hormonal influence on gene expression, polygenic rather than single gene transmission and genetic-environmental interaction may be considered.
Among non-genetic factors, smoking habits, body weight and obesity, hormonal factors, use of antiperspirants and deodorants, hair removal by a safety razor, and infections have all been considered (Table 8.2) [4, 8, 11-16]. The role of smoking was first assessed in a matched-pair case-control study in Germany . Patients who had received surgical treatment for hidrad-enitis suppurativa in two dermatological centers were cases. As controls, patients admitted for various other skin diseases were selected and matched for sex and age to cases. Out of 84 patients treated for HS, 63 subjects (27 men, 36 women) completed the questionnaire. The rate of active cigarette smokers was 88.9% among cases and 46% in the matched-pair control group with an odds ratio of 9.4 (95% confidence interval 3.7-23.7). A proportion as high as 27% of cases reported at least one affected first-degree relative. A drawback of the study was the inclusion of prevalent rather than incident cas-
Table 8.2. Selected studies concerning potential non-genetic risk factors for HS
Country Study design Factor(s)
Konig, 1999  Germany
TNS Sofres, France
Jemec, 1988  Denmark
Morgan, 1982  UK
Jemec, 1996  Denmark
Alcohol consumption BMI and metabolic factors
Case-control Signs of andro-in women genization
Case-control Endocrine factors in women
Case-control Shaving and the use of chemical depilatories, deodorants, and talcum powder
Case-control Education, menstrual and reproductive history, oral contraceptives, BMI, earlobe piercing, use of cosmetics
Jemec, 1996  Denmark Case-control Lapins, 2001  Sweden
STD history, current STD, STD risk factors
Record-linkage Association with cancer of hospital discharge diagnoses with the Swedish Cancer Registry
Association with smoking and BMI
No difference except for a shorter menstrual cycle and longer duration of menstrual flow in HS cases
No association after matching for BMI
No significant difference except for a more frequent application of talcum powder by controls as compared with HS cases
HS cases better educated, younger and fewer pregnant. Pierced earlobes more frequent among HS cases
SIRa for all cancers 1.5 (95% CI 1.1-1.8) SIR for NMSC 4.6 (95% CI 1.5-10.7) SIR for buccal cancer 5.5 (95% CI 1.8-12.9) SIR for liver cancer 10 (95% CI 2.1-29.2)
a SIR = standardized incidence ratio.
es, the lack of assessment of an exposure-effect gradient, and the lack of control for potential confounders on the exposure-effect relationship. However, the odds ratio was so high that confounding by unmeasured factors was unlikely. These results have been recently confirmed by data from the TNS Sofres population survey in France, where HS was strongly correlated with cigarette smoking . At this time, cigarette smoking appears to be a major modifiable risk factor for HS. The effect of quitting smoking has not been assessed in HS patients who are smokers and it is worthy of consideration. In addition, the role of smoking as a prognostic factor should be evaluated. Interestingly, several case reports and a few case series point to an association between HS and Crohn's disease [17, 18]. It is well established that Crohn's disease is associated with smoking and smoking has detrimental effects on the clinical course of the disease .
The relationship between body weight and HS has been repeatedly considered in the lack of convincing evidence [12, 14]. Recently, the TNS Sofres population survey has rejuvenated such a proposed association by documenting a higher proportion of people who are overweight and obese among individuals reporting a history of HS with respect to the general French population (53.5% versus 44.2%) . Some attention
has been repeatedly drawn toward the possible role of hormonal factors [4, 12, 14]. No convincing data are available. In a study of 66 women with HS, 23 had acne and 23 (34.8%) were significantly obese (body mass index, BMI, >30 kg/m2). Plasma androgens were compared with controls matched for BMI and hirsuteness and no difference was documented . The prevalence of HS has been found to be higher among patients attending a sexually transmitted disease (STD) clinic than in an unselected general population . In principle, a selection bias may explain such a difference. However, an association with chlamydial infections has been also suggested. A case-control study based on patients in a STD clinic was unable to confirm the association with chlamydial infection  but, quite unexpectedly, it found a higher prevalence of genital human papilloma virus (HPV) infection among HS cases as compared with controls. The significance of such a finding is unclear and a chance effect is quite plausible. Another unexpected finding from a case-control study was the association of HS with pierced earlobes . Since only univariate analysis was conducted and many confounding effects may operate (e.g., social class, age, etc.) this association should be taken cautiously.
An insight into potential risk factors for HS may come from analyses of disease associations since associated diseases may share common risk factors. In principle, associations may also result from exposures which follow the development of one of the diseases of interest (e.g., iat-rogenic factors) or even represent an artifactual effect if the presence of a concomitant disease influences the diagnosis of another disease or its referral (Berkson's bias). Interestingly, in a large-scale analysis of Swedish hospital discharge diagnoses linked with data from the Swedish National Cancer Registry, a strong association was documented between HS and non-melanoma skin cancer (NMSC), buccal cancer and liver cancer . These associations may at least partly reflect chance findings derived from multiple testing. However, the primary hypothesis of the study was centered on an association between HS and NMSC. Even if NMSCs complicating perineal or buttock HS localizations is a possible explanation, alternative hypotheses may involve a role for skin pheno-type and/or sun exposure on the development of HS. Such a hypothesis is worthy of consideration in future studies.
Was this article helpful?