The atopic march refers to the natural history of allergic diseases as they develop during infancy and childhood. Introduced by the American allergists A. F. Coca and R. A. Cooke in 1923, the term atopy became closely associated with the IgE molecule after its identification as the carrier of hypersensitivity.1 However, in the context of the atopic march, IgE is a pathophysiologic mediator of some, but not all, atopic diseases. As such, it is better to consider the atopic march as a progression of allergic conditions that have common genetic and environmental predisposing factors, share the immunologic feature of one or more allergen-specific TH2 responses, and are characterized by a type 2 effector phase that can include generation of specific IgE, activation of granulocytes, and other innate features such as mucous production and edema. Importantly, the presence of one allergic condition increases the risk for development of others, resulting in the additive feature of the atopic march. Classically, the atopic march begins with atopic dermatitis (AD) and progresses to IgE-mediated food allergy (FA), asthma, and allergic rhinitis (AR) (Fig 1).2 Each of these conditions carry a complex 132 D.A. Hill and J.M. Spergel / Ann Allergy Asthma Immunol 120 (2018) 131–137 pathophysiologic makeup that involves multiple facets of the immune system. For example, AD was once considered a manifestation of atopy itself, but is now thought to result from a combination of primary skin defect(s) and underlying genetic or environmental propensity to develop type 2 inflammation. Although non– type 2 inflammation likely contributes to the pathophysiologic makeup of AD, for the purposes of this review, we focus on the role of type 2 inflammation because it is the central tenet of the atopic march. Because type 2 inflammation is associated with AD lesions, AD also represents an important route of allergen exposure by which systemic TH2 responses are initiated. Once an individual has commenced on the atopic march, it is difficult to halt the progression. In this article, we present a clinical vignette and then review the epidemiologic and translational evidence that supports the concept of the atopic march. We briefly review the immunologic mechanisms that are thought to underlie the atopic march, and discuss some of the clinical interventions aimed at preventing or intervening on this process. Clinical Vignette A 1-year-old white male with a history of moderate AD was referred to an allergist because of concern for FA. The patient was exclusively breastfed until 1 year of age, when he developed acute hives, vomiting, and respiratory distress after his first consumption of cow’s milk. His medical history was otherwise notable for infantile-onset AD that was poorly controlled with intermittent topical corticosteroid use. His family history included AD and asthma in his father, AD in his mother, and AD, FA, and asthma in his older sister. Skin testing confirmed the diagnosis of FA to milk, and the patient was instructed to avoid this food. An alternative AD therapeutic regimen was initiated, which included more frequent and higher-dose corticosteroid use, and his skin improved during the following weeks. The child was followed in the allergy clinic and his AD significantly improved by the age of 2 years. That winter, he was admitted to the hospital for respiratory distress in the setting of a viral upper respiratory tract infection. The following spring, he developed AR and had positive skin test results to trees, grasses, and dust mite. By 3 years of age, he carried the diagnoses of FA, AR, and moderate persistent asthma that required combination inhaled corticosteroid and long-acting β-agonist therapy. At 4 years of age, skin testing to cow’s milk revealed a wheal and flare of 1 and 2 mm, respectively, and the serum IgE level to milk of 3.45 kU/L. The child subsequently passed a food challenge, and cow’s milk was introduced into his diet. AD and IgE-Mediated FA: Early Members of the Atopic March One of the most common pediatric conditions, AD is thought to affect between 7% and 12% of US children. AD is most commonly diagnosed in the first 6 months of life—before the development of FA, AR, and asthma.2,3 AD likely results from disruption of the skin barrier because of intrinsic defects of epithelial cells in an individual with a genetic and/or environmental predisposition for type 2 inflammation. Defects in the epidermal barrier protein filaggrin, which are associated with AD and allergic sensitization, are often cited as a quintessential example of such a barrier defect. However, other examples have now been identified, including lossof-function mutations in SPINK5 and the gene encoding corneodesmosin.4 Importantly, loss-of-function mutations in the filaggrin gene do not increase the risk of food or aeroallergen sensitivity independently of AD status.5,6 Therefore, although intrinsic epithelial barrier dysfunction is central to the pathophysiologic makeup of AD, there are other important genetic and/or environmental modifiers that are required for the development of allergic skin inflammation and progression of the atopic march. For example, polymorphisms in the gene that encodes thymic stromal lymphopoietin (TSLP) and its receptor influence AD risk, FA, and asthma,7–9 whereas polymorphisms in the genes that encode interleukin (IL) 33 and its receptor are associated with increased risk of AD and asthma.10,11 Together, these studies and others support the existence of a shared set of polymorphisms that predispose patients to AD and promote the subsequent development of other allergic conditions.8,12,13 Most proximally on the atopic march, the presence and severity of AD positively correlate with the risk of developing FA. One recent review estimated that children with AD are as much as 6 times more likely to develop an FA compared with their healthy peers.14 Food specific IgE responses can be detected in the first months of life and peak at approximately 10% prevalence at 1 year of age.15 The fact that sensitization occurs before food ingestion in most cases suggests that sensitization to foods occurs via exposure through inflamed skin, as opposed to the gastrointestinal tract. In addition, positive correlations have been found between the use of wheat- or peanut-containing skin products and the development of wheat or peanut allergy, respectively,16,17 and exposure to peanut dust in a child’s home positively correlates with the likelihood of developing peanut allergy.18 Finally, allergen-specific T cells isolated from peanut allergic patients express skin-related homing molecules,19,20 providing additional evidence of the skin being the site of allergen sensitization in FA. Together, these findings support the theory that there is transcutaneous sensitization to food allergens in susceptible individuals. Asthma and AR: Late Members of the Atopic March AD is also strongly associated with the development of asthma and AR.21 In contrast to food specific IgE responses, IgE responses to inhalant allergens develop later in childhood, providing a possible explanation for the delayed age of onset for these conditions.15 The association between AD and respiratory allergy is influenced by AD severity—whereas approximately 20% of children with mild AD develop asthma, more than 60% with severe AD develop asthma.21,22 The presence of AD is also associated with increased asthma severity, and greater asthma persistence into adulthood.23,24 Similar to FA, loss-of-function mutations in the filaggrin gene Figure 1. Age at diagnosis of common allergic conditions. D.A. Hill and J.M. Spergel / Ann Allergy Asthma Immunol 120 (2018) 131–137 133 correlate with asthma susceptibility and severity in patients with AD, but not in those without AD, indicating that skin inflammation is required for allergic sensitization.5 Despite these data, not every patient with AD develops asthma, and not every patient with asthma has preceding AD. A recent retrospective analysis of 2 birth cohorts found 8 separate patterns of atopic disease progression.25 Although limited by cohort size and parental survey as the reporting method, this study found that 10.5% of respondents followed the traditional pattern of the atopic march, whereas 15.5% had persistent AD, 5.7% had wheeze without AD, and 9.6% had rhinitis without AD. Together, these findings indicate that the atopic march is not present in all atopic individuals, and in particular those with adult-onset disease. The presence of FA is also an independent risk factor for the development of AR and asthma. In a retrospective birth cohort study of almost 30,000 children, we found that the presence of FA was associated with development of asthma (odds ratio [OR], 2.16; 95% confidence interval [CI], 1.94–2.40), and rhinitis (OR, 2.72; 95% CI, 2.45–3.03).2 Of the major food allergens, FA to peanut, milk, and egg were significantly associated with the subsequent development of asthma (ORs, 1.74, 1.38, and 1.60, respectively) and rhinitis (ORs, 2.59, 1.46, and 1.80, respectively). In addition, there was an additive effect as patients with multiple food allergies were at increased risk of developing respiratory allergy compared with patients with a single FA. A large meta-analysis of birth cohort studies confirmed that early-life food sensitization increases the risk of wheeze or asthma, eczema, and AR.26 It is difficult to control for confounding effects in these analyses, such as personal history of AD or family history of atopy. In a subsequent analysis of our study, we found that FA predisposed patients to AR and asthma independently of AD status, suggesting an additive risk of developing new allergic conditions as one progresses on the atopic march. In addition, the ability to more accurately predict a patient’s risk of developing respiratory allergy by considering the presence of FA is of considerable clinical utility to the practicing clinician. Finally, the clinical association between asthma and AR is well established, with up to three-quarters of individuals with asthma reporting rhinitis symptoms.27 This association holds after controlling for total IgE, parental history of asthma, and allergen sensitization, suggesting that the coexistence of AR and asthma is not solely attributable to atopic predisposition. AR is also positively correlated with asthma severity, and AR treatment improves asthma control.28 Together, these observations indicate that the upper and lower airways behave as a physiologic and pathophysiologic unit and have led to comprehensive recommendations from the Allergic Rhinitis and its Impact on Asthma guideline panel, which emphasize the importance of appropriate treatment of AR in individuals with asthma.29 Immunologic Mechanisms Underlying the Atopic March As described previously, allergen exposure through inflamed skin is thought to be the primary route by which individuals initiate the atopic march (Fig 2). This hypothesis is supported by data from animal models that indicate that transcutaneous allergen exposure promotes the development of specific T- and B-cell responses and subsequent allergic disease. These models can take the form of genetic disruptions (as is the case with mice deficient in filaggrin) or skin irritation via mechanical or chemical means. Protease allergens (such as those contained in peanut, papaya, mites, insects, fungi, and some pollens) are unique in that they can induce sensitization when exposed to healthy skin and act as an adjuvant for other allergens. For example, peanut exposure on the skin of mice increases the development of specific IgE responses to milk.30 Once an allergen has entered the skin, it has the opportunity to interact with the immune system. Skin is divided into 2 immune compartments: the epidermis, which contains predominantly Langerhans cells and CD8+ cytotoxic T lymphocytes, and the dermis, which contains dermal and plasmacytoid dendritic cells (DCs), macrophages, mast cells, and innate and adaptive lymphocyte subsets.31 The inflammation observed in AD is associated with increased production of IL-4, IL-25, IL-33, and TSLP, which recruit IL-5– and IL-13–producing type 2 innate lymphoid cells and contribute to the development of type 2 inflammation.32 DCs and other immune cells migrate from the skin to draining lymph nodes, where they stimulate naive T cells to differentiate into allergen-specific TH2 cells.33 Once allergen-specific TH2 responses are present, they can exert effects systemically. For example, in mice it is established that epicutaneous sensitization can induce allergen specific IgE, anaphylaxis,34 and allergic inflammation of the lung,35 esophagus,36 and gastrointestinal tract.37 In the case of esophageal inflammation, the mechanism is IgE independent, implying that pathogenic TH2 cells migrate from the skin to other tissue sites. An additional mechanistic question related to the atopic march is whether the presence of one allergen-specific TH2 response potentiates to the development of additional TH2 responses. One mechanism by which this may occur is via a bystander effect, in which existing inflammation acts as an adjuvant for the development of other TH2 responses. Basophils are one potential contributor to the bystander effect because they are potent sources of IL-4, are recruited to draining lymph nodes early in the response to infectious or allergic stimuli, and can cooperate with DCs to promote TH2 cell responses.38 Critically, blood basophil numbers in the steady state and basophil recruitment to lymph nodes after allergen exposure are directly related to serum IgE levels in an antigenindependent manner.39,40 Thus, elevated IgE levels may potentiate basophil-facilitated TH2 responses both in the skin and at distant tissue sites.41–43 Together, these observations offer one immunologic mechanism by which the presence of a TH2 response to one allergen could potentiate the development of additional TH2 responses. Clinical Strategies Aimed at Prevention and Intervention on the Atopic March The epidemic increases in prevalence and severity of allergic conditions observed in recent decades are generally thought to have occurred too quickly to be attributed to genetic drift alone. As a result, considerable effort has been devoted to understanding the role of diet, hygiene, infections, allergens, air pollution, and other environmental factors in susceptibility to allergic disease.44 Particular attention has been given to the influence of pathologic or commensal microbial stimuli (and in particular commensal bacteria) on the mammalian immune system.45 Epidemiologic studies were the first to establish an increased risk of asthma and other allergic conditions with childhood antibiotic exposure.46 These associations held after adjusting for infections and other potential confounders, suggesting detrimental effects of antibiotic exposure on nonpathologic microbes. Subsequent studies in animal models indicated that antibiotic exposure markedly restructures the intestinal microbiome and skews the immune system toward type 2 inflammation.39,47 Clinical associations have also been made with childhood exposure to household pets, livestock, unpasteurized milk, and endotoxins, all of which are generally protective against allergic manifestations. Some of the most compelling recent data supporting a role for microbial exposure in atopic risk come from studies that compared geographically distinct but genetically related pediatric populations. In one study of children in Finland and Russia, a significant, dose-dependent reduction in the risk of atopy was associated with microbial exposure and prevalence of enteroviruses.48 In another study of Amish and Hutterite farm children, preva134 D.A. Hill and J.M. Spergel / Ann Allergy Asthma Immunol 120 (2018) 131–137 lence of asthma and allergic sensitization was 4 and 6 times lower in the Amish (who follow more traditional farming practices). This finding correlated with dust in Amish homes that had 6.8 times more endotoxin, which induced MyD88 and Trif-dependent inhibition of airway hyperreactivity and eosinophilia in a mouse asthma model.49 Given our improved understanding of the influence of microbial factors on allergic sensitization, it is not surprising that multiple studies have now attempted to prevent progression along the atopic march through various diet or probiotic-based interventions aimed at modifying the microbiome of various mucosal sites. However, the data to date are mixed, with some positive and some negative studies.50,51 Studies have also examined the effects of optimal AD therapy and immunotherapy on the atopic march. One recent, 3-year, doubleblind study examined the benefits of pimecrolimus therapy in 3- to 18-month-old patients with recent-onset AD.52 No significant differences in the percentage of patients who developed FA, AR, or asthma were detected between the treatment and control groups, although baseline AD severity was positively correlated with development atopic comorbidities. Other small trials have found that routine use of emollients reduces the incidence of AD by approximately half during the therapy period, although it remains to be seen whether prophylactic emollients influence the risk of allergic sensitization.53 There is also some evidence to suggest that subcutaneous and sublingual immunotherapy may prevent the progression to asthma in high-risk atopic patients.54 For example, one study examined the effect of receiving oral house dust mite extract on asthma outcomes in 111 infants at high risk of atopy.55 Oral house dust mite therapy was well tolerated in the treatment group, but there was no significant preventive effect observed on HDM sensitization or the development of AD, FA, or wheeze. More indepth investigation of the role of immunotherapy in the prevention of atopic disease is warranted. Other efforts have attempted to modify the atopic march using prophylactic antihistamines. A large trial of 817 infants with AD aged 1 to 2 years found no significant difference between treatment with high-dose cetirizine or placebo with regard to development of asthma overall.56 However, a subgroup analysis indicated that infants sensitized to dust mite, grass, or both who were treated with cetirizine were significantly less likely to be diagnosed with asthma during the treatment period. However, a follow-up trial undertaken in children with AD and sensitivity to grass pollen and/or house dust mite allergens found no benefit to levocetirizine therapy in delaying asthma diagnosis.57 Finally, a very promising avenue for disrupting the atopic march stems from the decades of work investigating the immunologic mechanisms and clinical relevance of oral tolerance.58 This work culminated in a randomized clinical trial of peanut consumption in Figure 2. Model of the atopic march. Deficiencies in epithelial proteins, such as filaggrin, in conjunction with atopic predisposition result in skin inflammation. Epithelialderived cytokines, such as interleukin (IL) 25, IL-33, and thymic stromal lymphopoietin (TSLP) recruit and activate innate cell types, including innate lymphoid cells (ILCs) and basophils, which produce cytokines that promote the activation of dendritic cells (DC). Activated DCs process allergen, upregulate major histocompatibility complex (MHC), and circulate to draining lymph nodes, where they can interact with naive T and B cells to promote the development of allergen-specific T- and B-cell responses. Allergen-specific TH2 cells home back to the skin through expression of C-C chemokine receptor type 4 (CCR4), cutaneous lymphocyte antigen (CLA), and other molecules. TH2 cells also enter the systemic circulation, where they can exert effector response at distant tissue sites. Memory B cells (BM) recirculate in the blood and lymph, whereas plasmablasts (PB) home to the bone marrow, where they differentiate into plasma cells and produce allergen specific IgE. D.A. Hill and J.M. Spergel / Ann Allergy Asthma Immunol 120 (2018) 131–137 135 infants at risk for peanut allergy in 2015 that found significant benefit to early peanut introduction.59 These findings led to new recommendations from the National Institute on Allergy and Infectious Diseases on how to prevent FA that emphasize facilitating oral peanut exposure before transcutaneous sensitization in individuals that are at risk for developing FA.60 However, it remains to be seen whether early introduction of peanut or other food allergens influence the development of distal atopic conditions, such as asthma or AR. Conclusion The pronounced global increase in the prevalence and severity of atopic diseases during recent decades is of critical relevance to our population health. The concept of the atopic march has greatly improved our understanding of the pathophysiology of allergic conditions and led to the development of exciting new therapeutic strategies for the prevention of atopy. Future research should be directed at better understanding environmental and genetic factors that predispose children to atopic conditions (including the role of commensal and pathologic microbes), as well as the fundamental immunologic mechanisms that lead to development of a TH2 response on initial allergen exposure. Finally, it is important to consider the addition of new diseases to the atopic march, as supported by rigorous epidemiologic and mechanistic evidence. In sum, the atopic march remains a fundamental and well-tested concept in the field of allergy that has practical relevance to the practicing allergist and has led to exciting new research questions and therapeutic opportunities.