Key points

Introduction

AD is a chronic inflammatory skin disease affecting both children and adults. AD is associated with a substantial public health burden secondary to high prevalence in many regions and increased health care utilization and costs. The epidemiology of AD has evolved over the past few decades, with emerging trends and novel insights into the burden of disease. Studying the epidemiology of AD is complex. This review addresses recent developments in the epidemiology and public health burden of AD.

Introduction

AD is a chronic inflammatory skin disease affecting both children and adults. AD is associated with a substantial public health burden secondary to high prevalence in many regions and increased health care utilization and costs. The epidemiology of AD has evolved over the past few decades, with emerging trends and novel insights into the burden of disease. Studying the epidemiology of AD is complex. This review addresses recent developments in the epidemiology and public health burden of AD.

Challenges of studying the epidemiology of atopic dermatitis

There are several challenges of studying the epidemiology of AD. First, there are no widely accepted biomarkers or objective diagnostic tests for AD. Moreover, the lack of standardized nomenclature for AD internationally, for example, atopic neurodermatitis, eczema, atopic eczema, and childhood eczema, makes it difficult to develop consistent and valid questionnaires for epidemiology research. In particular, the term, eczema , has several different uses, including as the most commonly used lay synonym for AD, as a descriptive morphologic and/or histologic term encompassing multiple etiologies, and as a diagnostic term for AD. Furthermore, there is considerable heterogeneity of AD with respect to the distribution (eg, flexural, extensor, head and neck areas, and generalized), morphology (eg, oozing, scaling, lichenification, prurigo nodules, ill-demarcated, and psoriasiform), intensity and time course (intermittent, chronic persistent disease, seasonal variation, and episodic flares), and associated comorbidities. As such, numerous approaches have been used to assess the epidemiology of AD, but there is no universally valid approach.

Epidemiology of childhood atopic dermatitis

United States

Recent prevalence estimates of childhood AD in the United States range from 6% to 12.98%, depending on the study design and approach used to assess for AD. A household survey of 42,249 children and adults in 1998 found that 10.7% had empirically defined eczema, and 6% had empirically defined AD. This study did not distinguish, however, between AD in children and adults. Household surveys of 102,353 and 91,642 children ages 0 to 17 from the 2003–2004 National Survey of Children’s Health (NSCH) and 2007–2008 NSCH found the 1-year prevalence of caregiver-reported health care–diagnosed eczema to be 10.7% and 12.98%, respectively, with significant variation between states and districts (8.7%–18.1%) ( Fig. 1 A ). The question used to assess AD was subsequently validated and found to have good sensitivity and excellent specificity and positive predictive value. Comparison between 2003–2004 and 2007–2008 suggests that the prevalence of childhood AD is increasing over time. Data from the National Health Interview Survey (NHIS), a US population-based household survey, indicate that the prevalence of childhood AD steadily increased from approximately 8% in 1997 to more than 12% in 2010 and 2011 but may have plateaued in 2012 and 2013 ( Fig. 2 ).

The distribution of childhood AD and severe AD in the United States from the 2007–2008 NSCH. ( A ) AD prevalence (%); ( B ) severe AD prevalence (%). Data were divided into tertiles and color coded: tertile 1 = blue, tertile 2 = green, and tertile 3 = red.

The increasing prevalence of childhood AD (%) in the United States between 1997 and 2013 from the NHIS.

Several sociodemographic groups seem at higher risk for childhood AD in the United States. Several studies found higher prevalence of AD in African Americans/blacks, even after controlling for several potential confounding factors, such as household income, health insurance coverage, and parental education level. Multiple US population-based studies found no association between gender and AD. Children from the NSCH study also had a higher prevalence of caregiver-reported AD with increased household incomes, higher family education levels, smaller family sizes, and urban and metropolitan living.

Epidemiology of adult atopic dermatitis

The conventional dogma has been that AD is a disorder of childhood, with few adults having active disease. Several recent studies suggest, however, that AD may be more common in adults than previously recognized. International studies of AD in adults found prevalence ranging from 2.0% to 6.9% prior to 2000 depending on regional and methodological differences. Recent studies, however, of 27,157 and 34,613 adults (ages 18–85 years) from the 2010 NHIS and 2012 NHIS found 1-year prevalence of AD in 10.2% and 7.2%, respectively. Both of these studies randomly sampled adults from nationally representative cohorts from all 50 states and assessed for self-reported disease using an in-person survey. The 2010 NHIS used a less specific question about “dermatitis, eczema, or any other red, inflamed skin rash,” which overestimates the disease prevalence. In contrast, the 2012 NHIS used a more specific question about “eczema or skin allergy,” which was similar to the previously validated question used in the NSCH but not a health care diagnosis. In the 2012 NHIS, the prevalence of AD peaked in early childhood (14%), remained high throughout childhood (13%–14%), decreased somewhat during adolescence (8%), and then remained stable throughout adulthood (6%–8%) ( Fig. 3 ).

The US prevalence of AD (%) varies by age. Data are from the 2012 NHIS.

There are several possible explanations for the higher-than-expected prevalence of AD in adults. First, AD may not burn out, or dissipate, as much as previously thought. A study of the 7157 children from the Pediatric Eczema Elective Registry, a phase 4 registry of topical pimecrolimus users, found that at every age, more than 80% of subjects had symptoms of AD and/or were using medication to treat their AD. It was not until age 20 years that 50% of patients had at least 1 lifetime 6-month symptom-free and treatment-free period. This cohort comprised, however, children from mostly urban locations, some with already longstanding disease, who received treatment of their AD by a health care provider with a nonsteroidal prescription topical agent, all of which may not be generalizable to the entire population. A recent systematic review of 46 studies (mostly birth cohorts) found a higher proportion of children with AD achieving an episode of clear skin (80% in 8 years). Moreover, children whose AD started later in childhood or adolescence, was more severe, or already persistent for many years were less likely to achieve an episode of disease clearance. It is likely that some of the children with an episode of clear skin had recurrent disease afterward. Thus, the overall persistence of childhood AD may be even higher.

Second, adult-onset AD may occur more commonly than previously thought. A US population-based study of 60,000 households found that 54% of respondents with empirically defined AD reported disease onset in adulthood. Moreover, a retrospective Turkish study of 376 patients with AD found that 16.8% reported adult onset. The issue of adult-onset AD is a little controversial, because some investigators might suggest that such patients may have had childhood disease but were unaware of it. Regardless of whether such cases are adult onset per se or adult recurrence, there seems a larger subset of patients with activation of their skin disease in adulthood than previously recognized.

Finally, it may be that the prevalence of AD has increased in adults over the years, similar to increases observed in children. This point has not been well studied to date and warrants further investigation.

Atopic dermatitis severity

AD severity assessments reflect a combination of symptoms (eg, itch and sleep disturbance), lesional severity (redness, thickness, lichenification, scaling, and so forth), and/or the extent of disease. Recent international consensus was achieved for the objective assessment of AD severity, with the Eczema Area and Severity Index and Scoring Atopic Dermatitis emerging as preferred assessments. Although these assessments are routinely used in clinical trials or smaller-scale studies, they are rarely if ever used in a population-based setting. Thus, most studies of AD severity relied on self-reported or caregiver-reported disease severity. An analysis of the 2007–2008 NSCH demonstrated that 67.0% of US children had reportedly mild disease, 26.0% moderate disease, and 7.0% severe disease, amounting to 2.98 million children with moderate-severe disease, with significant statewide variation (see Fig. 1 B). Although the prevalence of AD decreased with age, the relative proportions of moderate (ages 0–11 years: 24.7%; ages 12–17 years: 29.1%) and severe (ages 0–11 years: 6.7%; ages 12–17 years: 8.4%) disease increased with age. AD severity was also found worse in African Americans/blacks and Hispanics. Other studies also suggested that African Americans/blacks may be prone to more severe disease and increased health care utilization for AD compared with whites. A population-based study of 1760 children living around Nottingham, England, found that 84% had mild, 14% moderate, and 2% severe disease as judged by a dermatologist.

A household survey of 42,249 children and adults in 1998 found that 30% reported mild, 53% moderate, and 18% severe AD. The high percentage of moderate-severe AD in this study may be due to selection bias. That is, patients with moderate-severe disease may have participated at higher rates than those with mild AD. Unfortunately, the study did not distinguish between pediatric versus adult patients.

Few population-based studies have directly assessed the prevalence of and risk factors for moderate-severe AD in adults. The 2013 National Health and Wellness Survey found that 50% of respondents with AD reported having moderate-severe disease. The question used to assess for AD was not previously validated, however, and likely not sensitive enough, thereby missing many cases of AD. There was also a 15% nonresponse rate to the question about AD severity, raising concern for nonresponse bias. More population-based studies are needed to determine the prevalence and risk factors of moderate-severe AD in adults from the United States and internationally.