does atopic dermatitis cause food allergy? a systematic review · 4/15/2016 · foods,5 ad is...

Atopic dermatitis and skin disease

Does atopic dermatitis cause food allergy?A systematic review

Teresa Tsakok, MRCP, MA,a* Tom Marrs, MRCPCH, MSc,b,c* Mahrose Mohsin, MD,d Susannah Baron, MRCP,d

George du Toit, FRCPCH,b,c Stephen Till, FRCP, PhD,c and Carsten Flohr, FRCP, FRCPCH, PhDa London and Canterbury,

United Kingdom

Background: The association between atopic dermatitis (AD)and food allergy (FA) is not fully understood, although a causalrelationship has been suggested. This has importantimplications for prevention and treatment.Objective: We aimed to review the association between AD andFA, the effect of FA on AD severity, chronicity, and age of onset,and the temporal relationship between the two.Methods: Medline and Embase were systematically searchedfrom inception to November 2014 for studies investigating bothAD and FA.Results: Sixty-six studies were identified. Eighteen werepopulation-based, 8 used high-risk cohorts, and the restcomprised patients with either established AD or FA. Inpopulation-based studies, the likelihood of food sensitization wasup to 6 times higher in patients with AD versus healthy controlsubjects at 3 months of age (odds ratio, 6.18; 95% CI, 2.94-12.98;P<.001). Other population-based studies reported that up to 53%of subjects with AD were food sensitized, and up to 15%demonstrated signs of FA on challenge. Meanwhile, studiesincluding only patients with established AD have reported foodsensitization prevalences up to 66%, with challenge-proven FAprevalences reaching up to 81%. Sixteen studies suggested thatFA is associated with a more severe AD phenotype. Six studies

From aSt John’s Institute of Dermatology, Guy’s and St Thomas’ NHS Foundation Trust

and King’s College London; bthe Department of Children’s Allergies, Guy’s and

St Thomas’ NHS Foundation Trust, London; cthe Division of Asthma, Allergy

and Lung Biology, King’s College London; and dthe Department of Dermatology,

East Kent Hospitals University NHS Foundation Trust, Canterbury.

*These authors contributed equally to this work.

Disclosure of potential conflict of interest: S. Baron has served as a consultant for Abbvie,

receives payment for lectures regarding psychodermatology training and also received

travel support from Abbvie. G. du Toit receives research funding from National

Institute of Allergy and Infectious Diseases (NIAID)/National Institutes of Health

(NIH), Food Allergy & Research Education (FARE), the MRC & Asthma UK Centre,

the UK Department of Health through the NIHR, the National Peanut Board (NPB),

and Osem and received grants from UK Food Standards Agency (FSA) during the

conduct of the study. S. Till serves as a consultant for ALK-Abell�o, receives payment

for lectures for Thermo Fisher, and receives unrestricted research funding from ALK-

Abell�o. C. Flohr serves as a consultant for Roche/Regeneron and has received

payments for lecture from Nutricia. The rest of the authors declare that they have no

relevant conflicts of interest.

Received for publication January 21, 2015; revised September 28, 2015; accepted for

publication October 27, 2015.

Available online February 18, 2016.

Corresponding author: Carsten Flohr, FRCP, FRCPCH, PhD, St John’s Institute of

Dermatology, St Thomas’ Hospital, Westminster Bridge Road, London SE1 7EH,

United Kingdom. E-mail: [email protected].

The CrossMark symbol notifies online readers when updates have been made to the

article such as errata or minor corrections

0091-6749/$36.00

� 2016 American Academy of Allergy, Asthma & Immunology

http://dx.doi.org/10.1016/j.jaci.2015.10.049

Downloaded from ClinicalKey.com.au at UFor personal use only. No other uses without permission.

indicated that AD of earlier onset or increased persistence isparticularly associated with FA. Finally, one study found that ADpreceded the development of FA.Conclusions: This systematic review confirms a strong anddose-dependent association between AD, food sensitization, andFA. AD of increased severity and chronicity is particularlyassociated with FA. There is also evidence that AD precedes thedevelopment of food sensitization and allergy, in keeping with acausal relationship. (J Allergy Clin Immunol 2016;137:1071-8.)

Key words: Atopic dermatitis, eczema, food allergy, foodsensitization

Atopic dermatitis (AD; synonyms: atopic eczema and eczema)is the most common chronic inflammatory disorder of the skin,affecting more than 20% of children in industrialized countries1-3

and up to 3% of adults.4 Although up to two thirds of patients withAD do not show sensitization to environmental allergens orfoods,5 AD is often associated with other atopic diseases, suchas IgE-mediated food allergy (FA),6 and around one third of allchildren with early-onset AD progress through the so-calledatopic march.7 It has been suggested that food allergen recogni-tion via antigen-presenting cells in eczematous skin might actas an important mediator of food sensitization and FA.8,9

This systematic review critically appraises the current body ofevidence according to theBradfordHill principles10 of inferring cau-sality from clinical studies, to further assess a potential causalpathway between AD and IgE-mediated FA. These principles takeinto account (1) the strength of the association between AD andFA in selected and unselected populations; (2) whether a dose-response effect is demonstrated, with more severe AD showing agreater association with FA, and equally whether FA can predictage of onset or chronicity of AD; and (3) the temporal sequence ofevents—inotherwords,whetherADor FAarises first among infants.

METHODSThis systematic review was registered on the International Prospective

Register of Systematic Reviews (PROSPERO),11 and Preferred Reporting

Items for Systematic Reviews and Meta-Analyses (PRISMA) guidance12

was followed throughout. Medline and Embase were searched from inception

to the end of November 2014, with no language limits imposed. Search terms

included the Cochrane Skin Group strategy for AD combined with terms

describing FA, food hypersensitivity, and food sensitization both overall and

to specific foods (including milk, egg, peanut, tree nuts, wheat, sesame, and

seafood). The full search strategy is available in the Methods section in this

article’s Online Repository at www.jacionline.org. This online search was

supplemented by an extensive hand search of the literature and

communication with individual authors where necessary. The Medline search

identified 339 articles, the Embase search identified 1704 articles, and the

1071

niversity of Melbourne April 15, 2016. Copyright ©2016. Elsevier Inc. All rights reserved.

Delta:1_given name

Delta:1_surname

Delta:1_given name

Delta:1_surname

Delta:1_given name

Delta:1_surname

Delta:1_given name

http://www.jacionline.org

mailto:[email protected]

http://dx.doi.org/10.1016/j.jaci.2015.10.049

http://crossmark.crossref.org/dialog/?doi=10.1016/j.jaci.2015.10.049&domain=pdf

J ALLERGY CLIN IMMUNOL

APRIL 2016

1072 TSAKOK ET AL

Abbreviations used

AD: A
topic dermatitis
aOR: A
djusted odds ratio
DBPCFC: D
ouble-blind placebo-controlled food challenge
FA: F
ood allergy
FLG: F
ilaggrin
OFC: O
pen food challenge
OR: O
dds ratio
PPV: P
ositive predictive value
sIgE: S
pecific IgE
SPT: S
kin prick test
UK: U
nited Kingdom
hand search identified a further 34 articles (see Fig E1 in this article’s Online

Repository at www.jacionline.org).

Two authors (T.T. and M.M.) independently screened all abstracts for

suitability, resulting in 164 articles that were read in full. Discrepancies in the

assessment were resolved through discussion with a third author (CF). Article

selection for further analysis was based on the following inclusion criteria: (1)

the majority of subjects being less than 18 years old and (2) at least a

proportion of subjects having AD and an overlapping proportion of subjects

having food sensitization and/or FA, as evidenced by at least one of skin prick

testing (SPT; standard cutoff, 3-mm wheal), serum specific IgE (sIgE;

standard cutoff, 0.35 kU/L), open food challenge (OFC), or double-blind

placebo-controlled food challenge (DBPCFC). Exclusion criteria were as

follows: (1) AD diagnostic criteria including SPT/sIgE sensitization to

common allergens, as well as typical skin findings; (2) FA evidenced only

by parent- or patient-reported symptoms; and (3) data referring to mixed food

and aeroallergen sensitization, with no separate values provided for food

allergens. Furthermore, we excluded management guidance documents,

reviews, intervention trials, and studies using animal models.

Data were extracted from 66 selected articles using a predefined pro forma.

Formal meta-analysis was considered neither feasible nor appropriate because

of marked heterogeneity in study design, participants, and diagnostic criteria

for both FA and AD. Therefore we assigned a quality score to each article

(maximum score, 3 points) based on AD diagnostic criteria (1 5 clear

validated criteria or physician’s diagnosis, 0 5 unclear diagnostic criteria or

process), study size (1 5 >_500 subjects, 0 5 <500 subjects), and a modified

Newcastle-Ottawa score13 (15 >_4 of 5 points, 0 5 <4 points). Articles were

then ranked by quality score. Those with the same quality score were further

ranked by number of participants.

RESULTS

Strength of association between AD and FAPopulation-based studies (see Table E1 in this

article’s Online Repository at www.jacionline.org).

Eleven population-based studies scored 3 points for quality, andtheir results closely reflect the age groups of the children theyinvestigated (Fig 1).

The largest study14 used a case-control design nested withinthe consecutively recruited Barn (Children’s) Allergy MilieuStockholm Epidemiology (BAMSE) birth cohort (n 5 2256),reporting that 27% of patients with AD at 2 years had positiveSPT responses to food (egg, 21%; peanut, 15%; milk, 8%;and cod, 2%). However, the investigators did not compare thiswith sensitization amongst children without AD. The second-largest study15 investigated SPT sensitization among 1501 olderchildren aged 12 to 16 years whowere participating in the DanishAdolescent Odense Cohort (TOACS). Although a stronglypositive association was demonstrated between current AD andpeanut sensitization (16.2% of patients with AD vs 0.5% of

Downloaded from ClinicalKey.com.au at UFor personal use only. No other uses without permission

control subjects, P < .05) during adolescence, no significantdifferences persisted for milk and egg sensitization.

Investigations among younger children reflected the higherprevalence of milk and egg allergies within that age group (Fig 1).

Two high-quality population-based articles were based on theIsle of Wight allergy birth cohort (n5 1456), with Arshad et al16

reporting that at 4 years of age, the highest independentrisk factor for AD was SPT sensitization to peanut (odds ratio[OR], 4.65; 95% CI, 1.02-21.34). There was also a positiveassociation with egg sensitization that just missed statisticalsignificance (OR, 6.08; 95% CI, 0.88-42.01), likely because ofa lack of power while a proportion of children were outgrowingtheir egg allergy.

Meanwhile, the United Kingdom (UK) cross-sectional studyby Peroni et al17 of 1402 children aged 3 to 5 years found that eggsensitization on SPTswas strongly associated with AD (OR, 9.53;95% CI, 2.40-37.82), whereas milk sensitivity was not (OR, 1.26;95% CI, 0.27-6.00).

The Danish Allergy Research Cohort18 (DARC) monitoredchildren from the age of 3 months to 6 years, undertaking foodchallenges in children with food-specific SPT responses of3 mm or greater/sIgE levels of class 1 or greater, and also inchildren with parent-reported adverse reactions to food. Theinvestigators reported that 18 (15%) of 122 children with ADdemonstrated symptoms on food challenge, and that 52% of theremainder were food sensitized. More recently, Flohr et al9

analyzed those undergoing SPTs at 3months within the EnquiringAbout Tolerance (EAT) study UK birth cohort (n 5 619),reporting that children with AD were significantly more likelythan healthy control subjects to be sensitized to at least 1 of 6foods (milk, egg, sesame, peanut, wheat, or cod), with an adjustedodds ratio (aOR) of 8.53 (95% CI, 3.51-20.65; P < .001). Thisassociation was independent of filaggrin (FLG) loss-of-functionmutation inheritance. Equally, there was a positive associationbetween AD and sensitization to individual foods, includingegg (aOR, 9.48; 95% CI, 3.77-23.83; P < .001), milk (aOR,9.11; 95% CI, 2.27-36.59; P 5 .002), and peanut (aOR, 4.09;95% CI, 1.00-13.16.76; P 5 .05).

In Figs 2 to 4, ORs for each of the major food allergens arepresented in order of ascending age of participants atcross-sectional analysis. Pooling of results between allergensand beyond the first 6 years of childhood was not consideredappropriate in view of the contrasting natural history of individualallergens, as well as the changes in association seen for each foodallergen as a child becomes older.

High-risk cohort studies (see Table E2 in this article’s

Online Repository at www.jacionline.org). Eight articleswere based on high-risk cohorts that recruited according toparental atopy. The UK study (n5 497) by Burr et al19 reported asignificant association between AD arising before the patient’sfirst birthday and a positive SPT response to egg at 6 months(OR, 5.57; 95% CI, 2.70-11.5); this was even stronger at 1 year(OR, 7.57; 95% CI, 3.57-16.05).20

Three articles used the Melbourne Atopy Cohort Study(MACS) cohort (n 5 552), including the report by Lowe et al21

that food SPT-determined sensitization at 6 months wasassociated with an increased risk of AD (hazard ratio, 1.63;95% CI, 1.13–2.35) up to 7 years of age. This was in line withother MACS data from Hill et al,22 who found that more patientswith AD had positive SPT responses to milk, egg, or peanut at 6months (22% vs 5%, P < .001) and at 12 months (36% vs 11%,

niversity of Melbourne April 15, 2016.. Copyright ©2016. Elsevier Inc. All rights reserved.




FIG 1. Egg/milk/peanut SPT sensitization among AD and control subgroups in population-based cohorts.


VOLUME 137, NUMBER 4

TSAKOK ET AL 1073

P <.001) than thosewithout AD (see Fig E2 in this article’s OnlineRepository at www.jacionline.org). Four years later, the samegroup reported that 36% of infants with AD had a positivefood-related SPT response compared with 12% of infants withoutAD (P < .001).23

Recently, Brough et al24 investigated whether AD and diseaseseverity modified the strength of the association betweenenvironmental peanut protein exposure and the risk of peanutsensitization andpeanut allergy. Five hundred twelveAmerican in-fants (<15 months old) with egg/milk SPT-determinedsensitization plus either parent-reported egg/milk allergy orparent-reported AD were recruited from the Consortium of FoodAllergy Research (CoFAR) study (n 5 512). Environmental pea-nut exposure was quantified from dust on the living room floor.An association between environmental peanut exposure and thelikelihood of peanut sensitization or allergy was only seen inparticipants with a history of AD. Indeed, a history of AD andincreased environmental peanut exposure in participants wasfound to increase the risk of both peanut sensitization (OR, 1.97;95% CI, 1.26-3.09; P < .01) and peanut allergy (OR, 2.34; 95%CI, 1.31-4.18) per log2 unit environmental peanut exposure in-crease. The effect for peanut sensitization was further augmentedwith a history of severeAD (OR, 2.41; 95%CI, 1.30-4.47;P<.01),although this was not observed for peanut allergy.

Studies investigating FA and food sensitization

among children with established AD (see Table E3 in

this article’s Online Repository at www.jacionline.

org). Thirty-four studies selected groups of patients with estab-lished AD rather than taking a population-based approach, andtherefore are likely to have inflated prevalence estimates of atopy.Two investigations25,26 comprising more than 2000 subjects aged11.8 to 25.8 months from 12 countries were derived from the EarlyPrevention of Asthma in the Atopic Child (EPAAC) study cohort.


Hill et al25 reported that 64% of thosewithADcommencing before3 months of age were sensitized to egg and/or milk and/or peanut,based on sIgE values greater than 95% of positive predictive value(PPV). The following year, de Benedictis et al26 compared theallergic sensitization patterns associated with AD between the 12countries, reporting that positive sIgE levels to egg predominatedin each country (53% in the UK), whereas milk sensitization washighest in Italy (48%) and peanut sensitization was highest inAustralia (45%).

Of note, 16 studies in patients with established AD reportedboth oral food challenge and sensitization data. However, manyfailed to apply systematic criteria in determining whowas eligiblefor food challenges, limiting their insight. One of the largeststudies27 in this category was performed in Finland and reported aprevalence of 54% for milk allergy among 183 children with ADup to 3 years old.

Burks et al28 proposed a standard food SPT panel for patientswith AD recruited from a pediatric allergy clinic in the UnitedStates. This group reported that of 165 participants aged between4 months and 22 years, 60% had at least 1 positive SPT response.The investigators then performed 266 DBPCFCs, and 39% ofpatients reacted to 7 foods (milk, egg, peanut, soy, wheat,cod/catfish, and cashew), accounting for 89% of the positivechallenge results.

Finally, Gray et al29 investigated 100 black or mixed-racechildren who attended an urban Cape Town clinic withmoderate-to-severe AD. The authors reported high rates of foodsensitization (66%) and challenge-confirmed FA (44%), withegg and peanut being the most common allergens.

Studies investigating AD among children with

established FA (see Table E4 in this article’s Online

Repository at www.jacionline.org). Four investigationsselected patients with established FA and then measured the






FIG 2. ORs of egg sensitization in patients with AD, in ascending age order.

FIG 3. ORs of milk sensitization in patients with AD, in ascending age order.


APRIL 2016

1074 TSAKOK ET AL

prevalence of AD. Two of these focused on children with peanutallergy and were particularly enlightening.

Fox et al30 compared 133 patients with peanut allergy, 160 high-risk control subjects with allergy to egg but not peanut, and 150low-risk control subjects with no known allergy. Cases andhigh-risk control subjects were recruited from an FA clinic,whereas low-risk control subjects were recruited from childrenattending a general pediatric clinic with a nonallergic complaint.A parent-reported history of AD in the first year of life was highlyprevalent among both peanut allergic patients (92%) and high-riskcontrol subjects (88%) but significantly less so among low-riskcontrol subjects (42%, P < .001).

More recently, Brown et al8 investigated the role of FLGmutations as a risk factor for peanut allergy. The study designcomprised 71 patients aged up to 18 years from the UK, Ireland,and The Netherlands with challenge-proven peanut allergy, and1000 non–peanut-sensitized population control subjects.Covariate analysis using both AD and FLG as predictors in thelogistic regression model demonstrated a strong associationbetweenAD and peanut allergy in patients with positive challengeresults (OR, 7.4; 95%CI, 4.1-13.7). The association betweenFLGmutation inheritance and peanut allergy remained significant,even after adjusting for a history of AD in early life (OR, 3.8;95% CI, 1.7-8.3; P < .001).

Is FA associated with AD of increased severity?Twenty-two articles addressed the question of whether FA

affects AD severity. A UK population-based study9 among 619exclusively breast-fed 3-month-old infants was the highestquality article in this area, showing that the association betweenAD and sensitization to milk, raw egg, cod, sesame, and peanutwas significantly stronger in AD patients with SCORAD scoresof greater than 20 (aOR, 25.60; 95% CI, 9.03-72.57; P < .001)than in patients with SCORAD scores of less than 20 (aOR,3.91; 95% CI, 1.70-9.00; P 5 .001).

Four articles were awarded 2 quality points; for instance, thecase-control study by Lack et al31 found a trend toward an


association between the severity of a flexural rash (as determinedby topical steroid use) in the first 6months, and a history of peanutallergy and a positive peanut challenge result (P < .001 for bothassociations). When the same research group performedscreening assessments in 834 infants to enroll those with eggallergy and/or severe AD into the Learning Early AboutPeanut (LEAP) trial, a strong association between foodsensitization and AD severity was evident.32 When the infantswere split into 3 SCORAD severity groups, increasing ADseverity was associated with a higher likelihood of sensitizationto peanut (Ptrend < .001), raw egg (Ptrend < .001), pasteurizedegg (Ptrend < .001), milk (Ptrend 5 .001), and ‘‘any food’’(Ptrend < .001).

In a group of Australian infants whose AD did not respond totopical steroid therapy, the equivalent rates of SPT-determinedfood sensitization were 83% at 6 months and 65% at 12 months.The relative risk estimates for SPT-determined food sensitizationalso rose with increasing AD severity as defined by topical steroiduse in the first year of life.23

In a later study by the same group, based on the EarlyPrevention of Asthma in the Atopic Child (EPAAC) cohort of2084 infants with AD, the authors used regression analysis toshow that children with sIgE to milk, egg, or peanut (greater than95% PPV cutoff) were more likely to have severe AD of earlyonset up to 1 year of age (P < .001).25 Lastly, data from the SouthAfrican pediatric cohort (n5 100)29 showed that 15 (30%) of 50children with moderate AD had FA, compared with 25 (50%) of50 of those with severe AD (P 5 .04).

Is FA associated with AD of earlier onset and

greater chronicity?Six articles examined whether FA is associated with AD of

early onset or increased chronicity, including the 2008 EPAACarticle by Hill et al,25 as mentioned above. The best qualitystudy33 assessed the relationship between FLG status, AD, foodsensitization, and FA (inferred from symptom-based diagnosis)in patients up to 18 years of age on the Isle of Wight


FIG 4. ORs of peanut sensitization in patients with AD, in ascending age order.



TSAKOK ET AL 1075

(n 5 1456). AD at 1 and 2 years of age was associated with bothFA (OR, 6.04; 95%CI, 1.25-29.9;P <.001) and food sensitization(OR, 20.1; 95% CI, 4.40-54.50; P < .001) at 4 years andsensitization at 18 years (OR, 18.2; 95% CI, 5.47-65.3;P < .001). In addition, AD at 4 years was linked to foodsensitization at 10 years (OR, 6.9; 95% CI, 3.65-13.04;P < .001) and at 18 years (OR, 3.82; 95% CI, 2.44-5.99;P < .001), supporting the hypothesis that early childhood ADraises the risk of food sensitization as a subsequent complication.

Further support in this area comes from the retrospectiveanalysis by Ricci et al34 of 252 infants with AD, in which eggsensitization was associated with more persistent AD(egg sensitized: 11.1 6 6.9 years mean duration vs non–egg-sensitized: 8.36 6.9 years mean duration; P < .02). As describedabove, the case-control study by Fox et al30 examining peanutallergy showed that parent-reported AD in the first year of lifewas significantly less prevalent among low-risk control subjectsrecruited from general as opposed to FA clinics; furthermore,AD in low-risk control subjects was significantly later in onsetand also less severe (P < .001).

The Belgian Early Treatment of the Atopic Child(ETAC) cohort (n 5 397)35 showed that egg and milk sIgEsensitization at enrollment was associated with persisting AD(SCORAD score > 7) at 18 months’ follow-up (P5 .006). Burkset al28 showed that children with both positive SPTand DBPCFCresponses had a significantly lower age of AD onset than patientswithout positive responses (2 vs 7 months, P5 .003). Finally, theSouth African study by Gray et al29 reported that onset of ADbefore 6 months was a significant risk factor for FA (P 5 .002).

Is there a temporal relationship between the

development of AD and FA?Two articles18,33 described repeatedly examining for AD and

food sensitization throughout childhood, and the DARC cohort18

initiated this early enough to discriminate between the emergenceof AD and FA. Five hundred sixty-two participants were recruitedat birth in the Danish general population and followed up at 3, 6, 9,12, 18, 36, and 72 months of age. All follow-ups except the9-month visit included an interview, clinical examination, andSPT/sIgE measurements. Where food-related symptoms weredeclared, OFCs were undertaken. By 3 months of age, 2% ofthe cohort was found to have signs of AD on examination, andyet none of the participants demonstrated symptoms on OFCs(Fig 5).18

Unfortunately, the Isle of Wight cohort33 only investigatedfood sensitization among children whose families reportedadverse food reactions before 4 years of age, preventing apopulation-based comparison.

Of note, the high-risk birth cohort study by Soderstrom et al36

was unique in focusing solely on low sIgE levels in an attempt to


capture allergic sensitization at the earliest possible stage.Therefore all IgE levels exceeding 0.7 kUA/L were omitted inthe risk analysis. AD at 2 years of age was associated withincreased low-level sensitization to egg, milk, or both at 6 months(OR, 3.07; 95% CI, 1.44-6.55), 12 months (OR, 3.33; 95% CI,1.60-6.96), and 24 months (OR, 1.74; 95% CI, 1.01-3.01).However, 35% of 6-month-old infants already had AD andwere not excluded from the analysis, meaning that low-gradesensitization might relate more to persistence of AD that isalready established, as opposed to newly arising AD. Further-more, no follow-up data on clinical FA status were reported.

Four articles investigated cord blood sIgE, attempting to assessthe possibility that food sensitization precedes the development ofAD. The Avon Longitudinal Study of Parents and Children(ALSPAC) cohort31 followed children until they had challenge-proven peanut allergy, but found no relationship with cord bloodpeanut sIgE levels. Likewise, in the DARC cohort,37 food sIgElevels proved no better than total cord IgE levels in predictingAD within the first 18 months of life. Overall, the key messagefrom these studies is that cord blood sIgE is rarely detectableand not a significant risk factor for FA, in line with the hypothesisthat AD precedes and is likely to drive food sensitization.

DISCUSSIONThis review investigates whether AD causes FA, according to

the established Bradford Hill principles.10 We found a significantassociation between FA, food sensitization, and AD in bothselected and unselected populations. FA is associated with ADof increased severity and chronicity. In addition, there is evidencefrom a Danish population cohort showing that AD arises beforeFA, as well as data from the Isle of Wight demonstratingassociations between early childhood AD and later foodsensitization and allergy.

Our systematic review protocol was registered online inadvance, in keeping with PRISMA guidelines.12 Cognizant ofthe potential for bias in literature searching, we supplementedthe online search strategy with an extensive hand search tomaximize capture of relevant publications. However, this wouldnot overcome inherent limitations, such as publication bias.Sixty-six publications fulfilled our selection criteria, of which12 scored 3 of 3 in our a priori independent quality grading.The 66 selected articles represent 3% of the total number ofarticles identified through our online search and hand search.Among these, substantial methodological differences and studyheterogeneity (different age groups, sIgE/SPT cutoffs, and foodchallenge protocols) precluded formal meta-analysis. Only 49(74%) of 66 studies stated the use of a doctor’s diagnosis,validated diagnostic criteria for AD, or both, and only 26 (39%)of 66 studies used food challenges to determine FA status.Furthermore, the rationale for investigating certain subgroups of


FIG 5. Temporal sequence of AD and FA onset in the Danish Allergy Research Cohort (DARC).18


APRIL 2016

1076 TSAKOK ET AL

participants was often unclear, with numerous studies using foodchallenges only in selected patients.Where food sensitization wasinvestigated, only 4 articles used values equating to the 95% PPVcutoff as part of their methodology for diagnosing FA (see thesummary of study limitations in Table E5 in this article’s OnlineRepository at www.jacionline.org).

It is a challenge for any study to capture the emergence of bothAD and FA, partly because AD characteristically waxes andwanes. In addition, clinical FA, in particular to egg andpeanut, often arises without the child having ever directlyingested the food in question. Therefore the only way toinvestigate the emergence of FA is to prospectively screen forfood sensitization at regular intervals and perform OFCsthroughout infancy. Given the strong association found betweenearly onset and also moderate-to-severe AD and foodsensitization, as well as the lack of sIgE in cord blood, it is likelythat food sensitization occurs primarily across the inflamed skinbarrier in eczematous skin, potentially leading to the developmentof clinical FA. Brough et al24 have recently shown that foodprotein content in dust samples collected from the infant’senvironment (such as the living room floor) predisposes towardearly food sensitization—especially in the presence ofeczematous skin inflammation. Indeed, carrying a FLG mutationalso increases this risk, further supporting the evidence that adisrupted skin barrier predisposes to sensitization. However,detailed work from our own birth cohort demonstrated that itis the presence of AD and its severity, rather than FLGmutation carriage per se, that relates to food sensitization riskin early life.9

When investigating the role of food allergen ingestion as aprevention strategy, the interplay between environmental versusoral food allergen exposure alongside AD status and managementbecomes of crucial importance. Fallon et al38 used theFLG-deficient flaky tail mouse model to demonstrate that the


application of ovalbumin to intact yet FLG-deficient mouseskin was sufficient to induce cutaneous inflammation and increaseovalbumin sIgE levels. Bartnikas et al39 contrasted thephysiologic effect of epicutaneous sensitization with that of oraltolerance induction. BALB/c mice were epicutaneouslysensitized with repeated applications of ovalbumin to tape-stripped skin over 7 weeks, or orally immunized with ovalbuminand cholera toxin over 8 weeks. Both the epicutaneous and orallyexposed groups of mice demonstrated sIgE antibody responses,and yet only those without oral immunization had signs inkeeping with anaphylaxis on oral challenge. Thus thedetermination of FA or tolerance depends not only on a periodof cutaneous sensitization, but also the timing and likely doseof gastrointestinal tract allergen exposure.

The design of methodologically sound studies examining thedynamic and interrelated nature of AD and FA will facilitatethe generation of future therapeutic and preventativeinterventions (see Table E5). Recent data in human subjectshave shown that carrying an FLG mutation disrupts the infantskin barrier by 3 months of age, even before AD emerges.40

Two small intervention studies41,42 suggest that the regularapplication of emollients from birth reduces the risk of ADdevelopment and might thus affect FA, although a pilot studyinvestigating egg sensitization did not find a significantreduction.41

Finally, it is important to remember that FA can develop inpatients without a prior diagnosis of AD. The seminal article byLack et al31 showed that 5 (0.06%) of 824 childrenwithout a historyof a rash over joints or skin creases had peanut allergy onDBPCFCs. However, even in the absence of AD, skin barrieralterations of a different nature might play a role in thedevelopment of FA. For instance, the previously discussed studyby Flohr et al9 also found an independent positive associationbetween skin barrier impairment (increased transepidermal water





TSAKOK ET AL 1077

loss) and food sensitization, in the absence of clinically visibleskin inflammation. Other environmental factors, such as waterhardness, the use of soaps and detergents, and the frequency ofwashing, could further contribute to skin barrier permeability andthus food sensitization.1 These are all areas that warrant furtherresearch.

The evidence presented in this systematic review providesfurther support for skin barrier repair, early proactive ADtreatment, and reduction of environmental food allergen exposurein the prevention of food sensitization and allergy. Largepopulation-based intervention studies using validated diagnosticcriteria and gold standard food challenges are required to test thishypothesis further.

We thank Professor Hasan Arshad for providing data from the Isle ofWight

cohort, Professor Magnus Wickman and Dr Natalia Ballardini for providing

data from the BAMSE cohort, and Professor Katrina Allen and Dr Jennifer

Koplin for providing data from the HealthNuts cohort.

Key messages

d There is a strong association between AD, food sensitiza-tion, and FA.

d AD of increased severity and chronicity is particularlyassociated with FA.

d AD arises before the development of food sensitization inmost cases, supporting a causal relationship.

REFERENCES

1. Flohr C, Mann J. New insights into the epidemiology of childhood atopic

dermatitis. Allergy 2013;69:3-16.

2. ISAAC Steering Committee. Worldwide variation in prevalence of symptoms of

asthma, allergic rhinoconjunctivitis, and atopic eczema: ISAAC. Lancet 1998;

351:1225-32.

3. Roduit C, Frei R, Loss G, Buchele G, Weber J, Depner M, et al. Development of

atopic dermatitis according to age of onset and association with early-life

exposures. J Allergy Clin Immunol 2012;130:130-6.

4. Schmitt J, BauerA,MeurerM.Atopic exzema in adulthood.Hautarzt 2008;59:841-52.

5. Flohr C, Johansson SG, Wahlgren C-F, Williams H. How atopic is atopic

dermatitis? J Allergy Clin Immunol 2004;114:150-8.

6. Werfel T, Breuer K. Role of food allergy in atopic dermatitis. Curr Opin Allergy

Clin Immunol 2004;4:379-85.

7. van der Hulst AE, Klip H, Brand PL. Risk of developing asthma in young

children with atopic eczema: a systematic review. J Allergy Clin Immunol 2007;

120:565-9.

8. Brown SJ, Asai Y, Cordell HJ, Campbell LE, Zhao Y, Liao H, et al. Loss-of-

function variants in the filaggrin gene are a significant risk factor for peanut allergy.

J Allergy Clin Immunol 2011;127:661-7.

9. Flohr C, Perkin M, Logan K, Marrs T, Radulovic S, Campbell LE, et al. Atopic

dermatitis and disease severity are the main risk factors for food sensitization in

exclusively breastfed infants. J Invest Dermatol 2014;134:345-50.

10. Hill AB. The environment and disease: association or causation? Proc R Soc Med

1965;58:295-300.

11. PROSPERO International prospective register of systematic reviews. Available at:

http://www.crd.york.ac.uk/PROSPERO/.

12. PRISMA: transparent reporting of systematic reviews and meta-analyses. Available

at: http://www.prisma-statement.org/index.htm.

13. Wells GA, Shea B, O’Connell D, Petersen J, Welch V, Losos M. The

Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised

studies in meta-analyses. Available at: http://www.ohri.ca/programs/clinical_

epidemiology/oxford.asp.

14. Bohme M, Svensson A, Kull I, Nordvall SL, Wahlgren CF. Clinical features of

atopic dermatitis at two years of age: a prospective, population-based case-

control study. Acta Derm Venereol 2001;81:193-7.


15. Mortz CG, Lauritsen JM, Andersen KE, Bindslev-Jensen C. Type I sensitization in

adolescents: prevalence and association with atopic dermatitis. Acta Derm

Venereol 2003;83:194-201.

16. Arshad SH, Tariq SM, Matthews S, Hakim E. Sensitization to common allergens

and its association with allergic disorders at age 4 years: a whole population birth

cohort study. Pediatrics 2001;108:E33.

17. Peroni DG, Piacentini GL, Bodini A, Rigotti E, Pigozzi R, Boner AL. Prevalence

and risk factors for atopic dermatitis in preschool children. Br J Dermatol 2008;

158:539-43.

18. Eller E, Kjaer HF, Host A, Andersen KE, Bindslev-Jensen C. Food allergy and food

sensitization in early childhood: results from the DARC cohort. Allergy 2009;64:

1023-9.

19. Burr ML, Merrett TG, Dunstan FD, Maguire MJ. The development of allergy in

high-risk children. Clin Exp Allergy 1997;27:1247-53.

20. Burr ML, Dunstan FD, Hand S, Ingram JR, Jones KP. The natural history

of eczema from birth to adult life: a cohort study. Br J Dermatol 2013;168:

1339-42.

21. Lowe AJ, Abramson MJ, Hosking CS, Carlin JB, Bennett CM, Dharmage SC, et al.

The temporal sequence of allergic sensitization and onset of infantile eczema. Clin

Exp Allergy 2007;37:536-42.

22. Hill DJ, Sporik R, Thorburn J, Hosking CS. The association of atopic

dermatitis in infancy with immunoglobulin E food sensitization. J Pediatr

2000;137:475-9.

23. Hill DJ, Hosking CS. Food allergy and atopic dermatitis in infancy: an

epidemiologic study. Pediatr Allergy Immunol 2004;15:421-7.

24. Brough HA, Liu AH, Sicherer S, Makinson K, Douiri A, Brown SJ, et al.

Atopic dermatitis increases the effect of exposure to peanut antigen in dust on

peanut sensitization and likely peanut allergy. J Allergy Clin Immunol 2015;135:

164-70.

25. Hill DJ, Hosking CS, de Benedictis FM, Oranje AP, Diepgen TL, Bauchau V, et al.

Confirmation of the association between high levels of immunoglobulin E food

sensitization and eczema in infancy: an international study. Clin Exp Allergy

2008;38:161-8.

26. de Benedictis FM, Franceschini F, Hill D, Naspitz C, Simons FE, Wahn U, et al.

The allergic sensitization in infants with atopic eczema from different countries.

Allergy 2009;64:295-303.

27. Isolauri E, Turjanmaa K. Combined skin prick and patch testing enhances

identification of food allergy in infants with atopic dermatitis. J Allergy Clin

Immunol 1996;97:9-15.

28. Burks AW, James JM, Hiegel A, Wilson G, Wheeler G, Jones SM, et al. Atopic

dermatitis and food hypersensitivity reactions. J Pediatr 1998;132:132-6.

29. Gray CL, Levin ME, Zar HJ, Potter PC, Khumalo NP, Volkwyn L, et al. Food

allergy in South African children with atopic dermatitis. Pediatr Allergy Immunol

2014;25:572-9.

30. Fox AT, Sasieni P, Toit du G, Syed H, Lack G. Household peanut consumption as a

risk factor for the development of peanut allergy. J Allergy Clin Immunol 2009;

123:417-23.

31. Lack G, Fox D, Northstone K, Golding J. Avon Longitudinal Study of Parents and

Children Study Team. Factors associated with the development of peanut allergy in

childhood. N Engl J Med 2003;348:977-85.

32. Toit du G, Roberts G, Sayre PH, Plaut M, Bahnson HT, Mitchell H, et al.

Identifying infants at high risk of peanut allergy: the Learning Early About

Peanut Allergy (LEAP) screening study. J Allergy Clin Immunol 2013;131:

135-43.

33. Venkataraman D, Soto-Ramirez N, Kurukulaaratchy RJ, Holloway JW, Karmaus

W, Ewart SL, et al. Filaggrin loss-of-function mutations are associated with

food allergy in childhood and adolescence. J Allergy Clin Immunol 2014;134:

876-82.

34. Ricci G, Patrizi A, Baldi E, Menna G, Tabanelli M, Masi M. Long-term follow-up

of atopic dermatitis: retrospective analysis of related risk factors and association

with concomitant allergic diseases. J Am Acad Dermatol 2006;55:765-71.

35. Wolkerstorfer A, Wahn U, Kjellman NI, Diepgen TL, De Longueville M, Oranje

AP. Natural course of sensitization to cow’s milk and hen’s egg in childhood atopic

dermatitis: ETAC. Clin Exp Allergy 2002;32:70-3.

36. Soderstrom L, Lilja G, Borres MP, Nilsson C. An explorative study of low levels of

allergen-specific IgE and clinical allergy symptoms during early childhood.

Allergy 2011;66:1058-64.

37. Johnke H, Norberg A, Vach W, Host A, Andersen KE. Patterns of sensitization in in-

fants and its relation to atopic dermatitis. Pediatr Allergy Immunol 2006;17:591-600.

38. Fallon PG, Sasaki T, Sandilands A, Campbell LE, Saunders SP, Mangan NE, et al.

A homozygous frameshift mutation in the mouse Flg gene facilitates enhanced

percutaneous allergen priming. Nat Genet 2009;41:602-8.

39. Bartnikas LM, Gurish MF, Burton OT, Leisten S, Janssen E, Oettgen HC,

et al. Epicutaneous sensitization results in IgE-dependent intestinal mast cell


http://refhub.elsevier.com/S0091-6749(15)03165-6/sref1
























http://www.crd.york.ac.uk/PROSPERO/

http://www.prisma-statement.org/index.htm

http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp

http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp















































































APRIL 2016

1078 TSAKOK ET AL

expansion and food-induced anaphylaxis. J Allergy Clin Immunol 2013;131:

451-60.

40. Flohr C, England K, Radulovic S, McLean WH, Campbel LE, Barker J, et al.

Filaggrin loss-of-function mutations are associated with early-onset eczema,

eczema severity and transepidermal water loss at 3 months of age. Br J Dermatol

2010;163:1333-6.


41. Horimukai K, Morita K, Narita M, Kondo M, Kitazawa H, Nozaki M, et al.

Application of moisturizer to neonates prevents development of atopic dermatitis.

J Allergy Clin Immunol 2014;134:824-30.

42. Simpson EL, Chalmers JR, Hanifin JM, Thomas KS, Cork MJ, McLean WH, et al.

Emollient enhancement of the skin barrier from birth offers effective atopic

dermatitis prevention. J Allergy Clin Immunol 2014;134:818-23.














does atopic dermatitis cause food allergy? a systematic review · 4/15/2016 · foods,5 ad is...

Documents