Men and women are combined in this chart, stratified by birth cohort and sex. Vertical lines indicate 95% CIs.
eTable 1. Associations Between Childhood BMI at Each Age From 7 to 13 Years and Risks of HS (per z Score)
eTable 2. Associations Between Childhood Height at Each Age From 7 to 13 Years and Risks of HS (per z Score)
eFigure 1. Flowchart Describing the Eligibility in the Analytical Study Population
eFigure 2. Association Between Birth Weight (kg) and HS From a Restricted Cubic Spline Model With 3 Knots
Customize your JAMA Network experience by selecting one or more topics from the list below.
Jørgensen AR, Aarestrup J, Baker JL, Thomsen SF. Association of Birth Weight, Childhood Body Mass Index, and Height With Risk of Hidradenitis Suppurativa. JAMA Dermatol. 2020;156(7):746–753. doi:10.1001/jamadermatol.2020.1047
Is there an association of birth weight, childhood body mass index, change in body mass index during childhood, and childhood height with subsequent risk of hidradenitis suppurativa in adulthood?
This cohort study of 347 200 Danish children found that both the lightest and heaviest infants had increased risks of hidradenitis suppurativa. Childhood body mass index was positively associated with the risk of hidradenitis suppurativa development in adulthood; returning to normal weight before puberty reduced the risks of hidradenitis suppurativa to levels observed in children who were never overweight.
Early body weight monitoring provides the opportunity to implement preventive measures aimed at reducing body mass index and development of hidradenitis suppurativa.
There is a lack of evidence on the association of birth weight, childhood body mass index (BMI), change in BMI during childhood, and childhood height with subsequent risks of hidradenitis suppurativa (HS) in adulthood.
To investigate the association of birth weight, childhood BMI, change in BMI during childhood, and childhood height with subsequent risks of HS in adulthood in a large Danish population-based cohort.
Design, Setting, and Participants
This cohort study included 347 200 schoolchildren from the Copenhagen School Health Records Register born from 1930 to 1996 who were linked to the Danish National Patient Register of hospital discharge diagnoses to identify cases of HS. Birth weight was reported by parents or guardians, whereas childhood weight and height were measured by school physicians or nurses at ages 7 through 13 years. Cox proportional hazards regressions were used to estimate hazard ratios (HRs) and 95% CIs. Statistical analysis was performed from February 20, 2019, to May 15, 2019.
Main Outcomes and Measures
A diagnosis of HS as recorded in the Danish National Patient Register.
Among the 347 200 children included in the study (175 750 boys) during the follow-up period from 1977 to 2017, 1037 individuals (677 females; median age at diagnosis, 39 years [range, 15-73 years]) received a diagnosis of HS. A nonlinear (U-shaped) association was found between birth weight and HS, such that both the lightest (2.00-2.75 kg; HR, 1.36 [95% CI, 1.10-1.68]) and the heaviest babies (4.26-5.50 kg; HR, 1.39 [95% CI, 1.01-1.93]) had increased risks of HS compared with normal-weight babies (3.26-3.75 kg; P = .04 for deviation from linearity). The risk of HS increased significantly with increasing BMI z score at each age from 7 to 13 years, from an HR of 1.32 (95% CI, 1.24-1.40) per BMI z score at 7 years of age to an HR of 1.50 (95% CI, 1.40-1.61) per BMI z score at 13 years of age. Compared with children with a normal weight at 7 and 13 years of age, those with a normal weight at 7 years of age and overweight at 13 years of age had a significantly increased risk of HS (HR, 2.11 [95% CI, 1.63-2.74]) and children with persistent overweight at both ages also had an increased risk of HS (HR, 2.61 [95% CI, 2.02-3.38]). Children with overweight at 7 years of age but with normal weight at 13 years of age did not have a significantly increased risk of HS (HR, 1.05 [95% CI, 0.67-1.67]). Childhood height at all ages was not associated with risk of HS (children at 7 years had an HR of 1.00 [95% CI, 0.94-1.07], and those 13 years had an HR of 1.06 [95% CI, 0.99-1.13], per z score).
Conclusions and Relevance
This cohort study found that both the lightest and heaviest babies had increased risks of HS. Childhood BMI was positively and significantly associated with risk of HS development in adulthood. These findings suggest that returning to normal weight before puberty reduces risks of HS to levels observed in children who were never overweight. Childhood height was not associated with risk of HS.
Hidradenitis suppurativa (HS) is a painful, chronic inflammatory skin disease associated with substantial individual morbidity and costs to the health care system.1-7 Hidradenitis suppurativa has an estimated global prevalence of 1% to 4%,6 but the true prevalence is likely higher, because HS represents an underdiagnosed condition.
The cause of HS is not well understood. Risk factors include a family history of the disease, smoking, adult obesity, and female sex. In addition, the severity of HS is greater in patients with obesity and lower long-term remission rates are reported in patients with obesity than in patients without obesity.6,8-11
Studies on low and high birth weights have found an association with a range of cardiometabolic diseases and an increased risk of type 2 diabetes.12 Low birth weight has been reported as a risk factor for premature adrenarche.13-17 Furthermore, studies have proposed a possible association in girls between premature adrenarche and obesity, hypertension, dyslipidemia, and insulin resistance,18 which are risk factors also linked with HS.11,19 A recent study suggests that there may be clinically different HS subtypes in patients with overweight (body mass index [BMI] ≥25 [calculated as weight in kilograms divided by height in meters squared]) compared with patients with BMI values below this threshold, highlighting the importance of body weight in the cause of HS.8 The timing of the development of obesity in association with the onset and severity of HS is, however, not established, to our knowledge. As HS often emerges after puberty and during the childbearing years in women, it is plausible that hormones are involved in the cause of HS, but this link is not fully established.20
The aim of this study was to investigate the association between birth weight, childhood BMI, change in childhood BMI, and childhood height from 7 through 13 years of age with risks of developing HS in adulthood in a large Danish population-based cohort.
The Copenhagen School Health Records Register is a unique Danish database containing computerized information from health examinations for almost all children born from 1930 through 1996 and who attended public or private schools in Copenhagen.21 Weight and height were measured by school doctors or nurses at 7 through 13 years of age. Parents or guardians reported their child’s birth weight (either from recall or birth records) at the first school health examination from birth year 1936 onward. The anthropometric information was systematically recorded on individual health cards along with the birth name, sex, and date of birth; this information has been computerized. Unique government-issued identification numbers were identified for more than 89% of children in the register.22 This study was approved by the Danish Data Protection Agency. According to Danish law, ethical approval is not required for register-based studies.
Follow-up was enabled via participants’ personal identification numbers and linkages with the Register of Vital Statistics, as well as with the Danish National Patient Register (NPR)23,24 of hospital discharge diagnoses. Information on HS diagnoses was available from the nationwide NPR, which contains information on an individual’s first registered diagnosis for all inpatient hospital contacts since 1977 plus all outpatient contacts since 1995. Individuals with HS were identified according to codes from the International Classification of Diseases, Eighth Revision (ICD-8; used from 1977 to 1994 [code 705.91]) and the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) thereafter (code L732). During follow-up, 1037 individuals (677 women and 360 men) received a diagnosis of HS (291 using the ICD-8 and 746 using the ICD-10).
Within the total population of 406 350 boys and girls in the Copenhagen School Health Records Register born from 1930 through 1996, 43 261 children were excluded because they did not have a personal identification number (eFigure 1 in the Supplement). Furthermore, 7694 children had emigrated, died, or were lost to follow-up prior to January 1, 1977, when the NPR was established, or before they turned 15 years. Six children received a diagnosis of HS before they turned 15 years, 8181 children had missing information about BMI and/or height at all ages, and 8 children had extreme values of BMI and/or height at all ages. Hence, the final study population consisted of 347 200 children.
In this study, follow-up started on January 1, 1977, or at the age of 15 years, whichever came later, and ended on the date of a diagnosis of HS, death, emigration, loss to follow-up, or June 30, 2017 (the latest register update available for these analyses), whichever came first.
Statistical analysis was performed from February 20, 2019, to May 15, 2019. The association between early-life anthropometric values (birth weight, BMI, changes in BMI, and height) and HS risk was investigated using Cox proportional hazards regression, with age as the underlying time metric and stratified by birth cohort and sex. Hazard ratios (HRs) and 95% CIs were generated by these regressions and used to estimate the risks. We assessed whether there was a linear association between birth weight and HS using a restricted cubic spline model with 3 knots using the likelihood ratio test. As the linear assumptions for the model were not satisfied, birth weight was evaluated with 5 a priori chosen categories (2.00-2.75, 2.76-3.25, 3.26-3.75, 3.76-4.25, and 4.26-5.50 kg), with birth weight in the normal range (3.26-3.75 kg) serving as the reference. We selected these categories to reduce the effect of digit preference in the recording of birth weight.25
Childhood BMI values were calculated and BMI and height values were transformed into age-specific z scores using an internal reference. Body mass index z scores were based on a period when the prevalence of obesity was low and stable (1955-1960) and height z scores were birth cohort–specific to account for secular increases in height. A restricted cubic spline model with 3 knots was used to examine the linearity of the associations between childhood BMI and height and HS risks using the likelihood ratio test. We did not detect nonlinearity in any of these associations; thus, associations are presented per z score. Furthermore, childhood BMI at each age was classified into normal or overweight using age-specific and sex-specific values based on the International Obesity Task Force classifications.26 In analyses on changes in BMI during childhood, BMI was modeled as overweight or nonoverweight, where children with normal weight at both ages 7 and 13 years were the reference. All associations were tested for potential sex interactions using the likelihood ratio test. As we did not find any sex interactions, the results are presented with the sexes combined. Furthermore, we tested for potential birth cohort interactions, similarly using the likelihood ratio test. Moreover, a time-varying effect was included to test the proportional hazards assumptions underlying the Cox proportional hazards regression models. We did not find strong indications of any birth cohort effects or violations of the proportional hazards assumption.
The follow-up period lasted up to 41 years, from 1977 to 2017. During this period, 1037 individuals (677 females; median age at diagnosis, 39 years [range, 15-73 years]) received a diagnosis of HS.
Birth weight was available for 266 641 individuals (136 068 males), among whom 823 cases of HS (545 females) were diagnosed. For girls, median birth weight within the 3 birth cohorts were similar: 1930-1949, 3.30 kg (95% CI, 2.50-4.30 kg); 1950-1969, 3.30 kg (95% CI, 2.45-4.10 kg); and 1970-1996, 3.30 kg (95% CI, 2.50-4.15 kg). For boys, median birth weight was 3.50 kg (95% CI, 2.50-4.50 kg) for the 1930-1949 birth cohort, 3.45 kg (95% CI, 2.50-4.30 kg) for the 1950-1969 birth cohort, and 3.48 kg (95% CI, 2.50-4.30 kg) for the 1970-1996 birth cohort. As expected, the median BMI and height increased with age and over time (Table 1). The prevalence of overweight in girls and boys also increased over time, reflecting the emergence of the obesity epidemic.
We found a nonlinear (U-shaped) association of birth weight with HS. Both the lightest (2.00-2.75 kg; HR, 1.36 [95% CI, 1.10-1.68]) and the heaviest babies (4.26-5.50 kg; HR, 1.39 [95% CI, 1.01-1.93]) had increased risks of HS compared with normal-weight babies (3.26-3.75 kg; P = .04 for deviation from linearity) (Table 2; eFigure 2 in the Supplement).
We investigated the association between childhood BMI and the risk of HS using z scores at ages 7 through 13 years as continuous variables. The risk of HS significantly increased with higher BMI z scores at each age from 7 to 13 years (Figure; eTable 1 in the Supplement). For the sexes combined, HRs increased from 1.32 (95% CI, 1.24-1.40) per BMI z score at 7 years of age to 1.50 (95% CI, 1.40-1.61) per BMI z score at 13 years of age. Compared with children with normal weight, children with overweight had increased risk of HS, with an HR of 1.99 (95% CI, 1.63-2.42) at 7 years of age and an HR of 2.38 (95% CI, 1.98-2.86) at 13 years of age (Table 3).
Longitudinal growth analyses were conducted in a subsample of individuals with information on body size measurements at 7 and 13 years of age. In these analyses, 269 196 individuals (133 983 females and 135 213 males) and 826 individuals (526 women and 300 men) with incident cases of HS were included. Children were categorized as having normal weight or overweight at each age and different overweight patterns and their associations with HS risks were evaluated (Table 4). Compared with children with normal weight at both ages, children who had a normal weight at 7 years of age but who were subsequently overweight at 13 years of age had a significantly increased risk of HS (HR, 2.11 [95% CI, 1.63-2.74]). Children who had persistent overweight at both childhood ages also had an increased risk of HS compared with children with normal weight at both ages (HR, 2.61 [95% CI, 2.02-3.38]). Children with overweight at 7 years of age but not at 13 years of age were not at a statistically significant increased risk of HS in adulthood (HR, 1.05 [95% CI, 0.67-1.67]).
At no ages was childhood height statistically significantly associated with risk of HS. Children at 7 years had an HR of 1.00 (95% CI, 0.94-1.07), and those 13 years had an HR of 1.06 (95% CI, 0.99-1.13), per z score (eTable 2 in the Supplement).
In this large, population-based cohort study with long-term follow-up of 347 200 children, we found a U-shaped association between birth weight and risk of HS in adulthood such that both the lightest and the heaviest babies had increased risks of HS. Furthermore, we found that higher childhood BMIs significantly increased the risk of HS in adulthood. The risks of HS in adulthood increased across all BMI values and were not limited only to children with overweight or obesity. We found increased risks of HS at BMI values below those defined as overweight or obesity in children. We also found that returning to normal weight before age 13 years corresponded to a risk of HS similar to that observed in children who were never overweight. Last, we found that childhood height at any age was not associated with risk of HS.
We found that both low and high birth weights were associated with increased risks of HS. As no other studies have examined these associations, to our knowledge, a direct comparison with the literature is not possible. Nonetheless, low and high birth weights are associated with a range of cardiometabolic diseases. In an earlier study on this same data resource using a very similar categorization of birth weight, compared with birth weights from 3.25 to 3.75 kg, low (2.00-2.75 kg) and high (4.75-5.50 kg) birth weights were associated with increased risks of type 2 diabetes in women.12 In men, associations were only found for the low-birth-weight category. Given the prevalence of insulin resistance among patients with HS, these findings suggest that there may be some common underlying mechanistic pathways. Studies have also reported low birth weight as a risk factor for premature adrenarche.13-17 Adrenarche refers to the physiological process defined by increased adrenal androgen production shortly before the onset of gonadarche.18,27 Studies have also proposed a possible association in girls between premature adrenarche and obesity, hypertension, dyslipidemia, and insulin resistance,18 which are also linked with HS.11,19
We found that, as BMI increased, so did the risk of HS. To put this in perspective, we can compare two 7-year-old boys of the same height: the one who weighs 2.10 kg (equivalent to 1 z score of BMI) more has a 32% increased risk of HS. When comparing two 7-year-old girls of the same height, the one who weighs 2.30 kg more has a 32% increased risk of HS. Similar comparisons show that by 13 years, when comparing 2 boys of the same height, a 5.90-kg greater weight (equivalent to 1 z score of BMI) equates to a 50% increased risk of HS, as does a 6.80-kg greater weight when comparing 2 girls of the same height. Early-life BMI and height are often relevant risk factors for diseases emerging in adulthood. Large population-based cohort studies using the same cohort as the present study have established that early-life BMI represents a risk factor for coronary heart disease, ischemic stroke, endometrial cancer, psoriasis, and type 2 diabetes as well as several other chronic noncommunicable diseases.28-32 Studies have shown that the timing of the onset of overweight has implications for disease risk. One study found that returning to normal weight before puberty in boys reduces the risk of type 2 diabetes to levels observed in boys who were never overweight.33 This finding illustrates that early life is an important period during which risks of later diseases can be identified and potentially modified.
The mechanisms whereby childhood BMI might increase risk of HS in adulthood are not fully elucidated. It is well established that smoking, adult obesity, dyslipidemia, and diabetes are strongly associated with HS.11,19 Children with overweight often remain overweight or obese in adulthood, but the likelihood of persistence depends on several factors. The risk of persistence is higher with more severe overweight and obesity in childhood, with increasing childhood age, and if assessed at younger adult ages.34,35 Thus, the long-term association of childhood obesity with the risk of HS may at least partly be associated with adult obesity. The positive associations observed between childhood BMI and risk of HS in adulthood in this study may indicate shared underlying biological processes such as obesity-induced inflammation.19 The immunologic effects of obesity and diabetes are associated with production of interleukin (IL)-1β, IL-6, and tumor necrosis factor (TNF) through activated macrophages in adipose tissue,36,37 which may promote development of HS. Also, higher body weight in childhood is associated with the risk of dyslipidemia in children38; this association suggests a plausible mechanism linking higher childhood BMI with an increased risk of adult HS. Furthermore, studies have shown significantly elevated levels of interleukins (IL-1β, IL-6, IL-10, IL-17A, and IL-12) in lesional HS skin39-41 and higher levels of TNF in serum, lesional, and perilesional HS skin.39,42 High levels of IL-6 and TNF are also characteristic of overweight patients and in those with metabolic syndrome.43 In addition, the satiety hormone leptin has been shown to stimulate immunoregulators including TNF, supporting the hypothesis that being overweight might provoke or worsen HS.44 Furthermore, several clinical observations suggest that sex hormones may play a role in the pathogenesis of HS. Sex hormones are well recognized to be associated with inflammation.45-48 The prevalence of HS is reported to be higher in women and many women have reported flares of HS activity during premenstrual periods and decreased activity during pregnancy.49-53
This study has some strengths. To our knowledge, this is the first study to evaluate the association between birth weight, childhood BMI and changes in BMI during childhood, and childhood height with HS development. Previous studies have evaluated associations between static adult BMI and risk of HS rather than childhood BMI trajectories. Most studies on adult BMI and HS are based on BMI values at the time of diagnosis or by recall.50,54 A particular strength of the present study is the use of measured (rather than recalled) weight and height that precedes the diagnosis of HS and thus avoids reverse causality. Thus, this study’s findings suggest that a high BMI in childhood can precede a clinical diagnosis of HS. Other major strengths of the study are the large and unselected study population recruited during childhood (before the normal onset of HS) and followed up into adulthood (after the normal onset of HS). All socioeconomic groups are represented in the Copenhagen School Health Records Register, a database that contains records of virtually every school child attending public or private schools in the Copenhagen municipality from 1936 onward.21
This study also has some limitations. A potential limitation is the lack of adjustment for potential confounders such as family history of HS, lifestyle factors, smoking, and socioeconomic status. In general, smoking is considered a confounder for the association between adult body weight and HS. However, as this study examines a child population assumed to have limited tobacco use, smoking is an unlikely confounder in our data. A further potential limitation is the lack of information on adult body size, as studies have shown that BMI tracks strongly from childhood to young adulthood,35 a period of great relevance for HS. The results are limited to urban populations, because all study participants per definition lived in greater Copenhagen as children. Moreover, almost all study participants were of Danish ethnicity21 and the generalizability of the study is therefore likely applicable to white populations.
In Denmark, patients with HS are diagnosed in inpatient or outpatient hospital clinics. The NPR contains discharge diagnosis from all Danish hospitals, keeping selection bias to a minimum. Because Denmark has universal health care, access bias is reduced. However, before 1995, patients diagnosed and treated outside the hospital setting were not included in the NPR. For this reason, patients with recent onset of disease and less severe HS might be underrepresented in the present study. Therefore, our associations may possibly apply to patients with more severe HS.
This study found that high-birth-weight and low-birth-weight babies are at increased risk of adult HS. Furthermore, increasing BMI in childhood is associated with a greater risk of the development of HS in adulthood. We also found that returning to normal weight before puberty corresponds to a risk of HS similar to that observed in children who were never overweight. Childhood height at any age was not associated with risk of HS. Early body weight monitoring provides the opportunity to implement preventive measures aimed at reducing BMI and thus HS development. For treating clinicians and other health care workers, this monitoring may be particularly relevant in children with a hereditary predisposition to HS. As childhood obesity rates are increasing worldwide, our findings suggest that more children are at risk of developing HS in adulthood.
Accepted for Publication: March 6, 2020.
Corresponding Author: Astrid-Helene Ravn Jørgensen, MD, Department of Dermato-Venereology & Wound Healing Centre, Bispebjerg Hospital, Bispebjerg Bakke 23, 2400 Copenhagen NV, Denmark (firstname.lastname@example.org).
Published Online: April 29, 2020. doi:10.1001/jamadermatol.2020.1047
Author Contributions: Dr Aarestrup had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: All authors.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Jørgensen, Aarestrup, Baker.
Critical revision of the manuscript for important intellectual content: Aarestrup, Baker, Thomsen.
Statistical analysis: Aarestrup, Baker.
Obtained funding: Thomsen.
Administrative, technical, or material support: Jørgensen, Baker, Thomsen.
Supervision: Jørgensen, Baker, Thomsen.
Conflict of Interest Disclosures: None reported.
Funding/Support: The work presented in this article is supported in part by grant NNF17OC0028338 from the Novo Nordisk Foundation.
Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.