Increased Intestinal Permeability in Atopic Eczema

Michael G . Pike, M.B., M .R.C.P., Robert]. Heddle, Ph.D., F.R.A.C.P. , Peter Boulton, B.Sc., Malcolm W. Turner, Ph .D. , M.R.C.Path. , and David J. Atherton , M.B ., F.R.C.P. Depa rtm ent of Immuno logy, Institute of C hild Hea lth , London, and the Hospi tal for Sick C hildren, London, U .K .

We have in vestigated gas trointestina l perm ea bility in chil­dren with ato pi c eczema by m easuring the relative urinary excretion rates of the inert di- and monosaccharides lac­tu lose and rh a mnose fo llowing th e ir o ra l administration. The m edi an lac tulose/rh amn ose rat io was g rea te r in 26 children with atopic eczema than in a contro l g roup of 29

The benefi ci al effect of elimination diets in many chil­dren with ato pic eczema impli es an etio logic ro le for foods [I ,2 1, although the pathogenetic mechanisms in vo lved arc unknown. Fo ll owing the ingestion of egg o r milk , ve ry sm all amoums of protein antigen

ap pear in the circul ation of health y individuals, and preliminary ev idence sugges ts that g reater amounts ma y do so in subjects w ith atop ic eczem a [3,4 [. Possible explanations for such a phe­nomenon include either increased passage of food anti gens across the gut wa ll o r red uced clea r:mcc of such antigens from the cir­cu lati on . The first of these possi bilities has been in ves tig:ned by measu rin g the g:~s trointes tin a l uptake of inert molecules. The gas tro intestinal absorption of po lyethylene g lyco l o f mea n mo­lecular weight 4000 (PEG 41<) is in creased in a proportion of individuals with atopic eczema [51. However, intestin al uptake of PEG is very much more rapid than that of o ther inert molecules of comparable molecular weight [ 6 I, and the hi gh lipid so lubility o fPEG sugges ts that it may be absorbed differently. Furthermore, no data arc ava ilable as ye t regard ing the f.1 te of i. v. inj ected PEG in humans , and va ri ati ons of distribution and meta bolism may occur and be w rongly interpreted as differences in permeabi lity. We, and others 17,8 1 ha ve, therefore, used instead the monosac­charide L-rhamnose (M, 164), and the disa cd1aridc, lactulose (M, 342), both of w hi ch are un charged, met:1bo li ca ll y inert molecules of low lipid so lubili ty [9, lOj. L-Rhamnosc passes fi·ccly across the gu t wa ll , probabl y through pores in the en rcrocyte cell membrane [11 ]. In contrast, lactu losc is no rm all y abso rbed less readil y, prob­ably throug h chann els betwem adjacent cnterocytcs. Virtuall y all o f th e L-rhamnosc and la ctulose absorbed is recovered intact in the urine 1111. 13ecause L-rhamnosc and lactu losc arc given si­mu ltaneously, facto rs such as intraluminal dilution, gastroi ntes­tina l tr:msit tim e, absorpti on rate, distribution after absorption, and renal clea ran ce should affec t both molecules similarl y and any increase in the excreti on ratio should reflect onl y th eir relative intestinal perme3tion rates . It is believed that such increases in­dica te either g reater " lea kiness" at intcreellular junctions, or areas of ente rocytc loss. In this stud y we have also concurrentl y eval­uated lactose absorption as a meas ure of possiblt: small intestinal mucosa l dama ge.

Manuscript received N ovember 26. I'JH4; accepted for publication Jul y 17, 1985.

Rep rint requests to: l)avidJ. Atherton. M .B . . F.R.C. P .. Department o f Immunology, Institute of C hild Hea lth , 30 G uilfo rd Street, London WC1N !El-l, U.K .

Abbreviation: PEG: polyeth ylene glyco l

children which included both health y individuals and o th­e rs with various noneczem ato us dermatoses. This increased p ermea bility may be a prim ary abnormality of the g ut o r m ay reflect intestinal mu cosa l damage ca used b y loc:1 l h y­persensitivity reactions to food antigen s. J hwesr Den11fllo/ 86:101-104, 1986

SUBJECTS

. Twenty-six children of va rious races, with a clini ca l diagnosis of atop ic eczem a, were studied (8 girls, 18 boys; age ran ge 1. 3-16.2 years, mean 5.8 years). All were receiving standard topical treat­m ent , which in m any cases included mild corti costeroid app li­cations. Several were receiving o ral antihista mines at nig ht and some were avo iding certain foods as part of their treatment.

Two contro l g roups were used: 23 healthy children, mostly rc la ti ves and friends of members of the department ( 10 boys , 13 g trls; age ran ge 0.5-14 years, m ea n 8.2 yea rs) and 6 children with genera lized noncczcm atous sk in disorders comprising 2 patients w tth recessive dystrophi c epidermolysis bullosa, 1 w ith bullous ichthyosiform ery throderm a, and 3 with psoriasis (3 boys . 3 gi rl s; age range 5-14 yea rs , mea n 7.3 years). N one of these contro l children had eczema , asthma, o r hay fever and they were not skin tested.

Info rmed co nsent was obtained from all parents, and approval for th L· study was obtained from th e Hosp ital's Standing Com­mittee on Ethi ca l Practice.

METHODS

O n the m orning following an overnight fast of at least 6 h and after voiding and discarding overnight urin e, the subjects were g tvcn L-rh amnose (0.02 g / kg body weight), lactulose (0 .1 g/ kg body wetght), and lactose (0 .2 g/ kg body weight). in aq ueous solution made isotoni c by the add ition of D-g lucosc. N othin g further was taken by mouth for 5 h except water . During thi s penod, all unnc, mcludmg th at passed ar the end of the 5 h. was coll ected w ith m erthiolate as a preservative, and kept frozen at ho me.

The urine volume was measured; aliquots were coded and stored fro zen. Samples were later analyzed for L-rhamnose, lactulose, and lac tose conten t by thin-la yer chromatography, using a mod­tfi ca non of the method described by Mcnzics, Mount. and Wheeler [1 2], on plastic-backed si li ca gel la yers (Schleicher and Schi.ill F1500) with butan-1-ol:ethanol:g la ia l ace tic acid :watcr. 60:30:10:10 v/ v (14 em , ascend ing) using sucrose and fructose as internal standards. T he plates were developed 3 times and were dried between each dev~l opmcnt. Res ults were expressed as excretion ratt os (the ratt o o t the proportion of the ora l lactul ose dose ex­creted to the proportion of the o ral L-rhamnose dose exc reted (lactul osc/rhamnose ratio), and the rati o of the proportion of the o ral lactose dose excreted to the propo rtion of the o ral lactu losc dose excreted (lactosc/ lactul ose r:ltio).

RESULTS

T he distribution of lactulosclrhamnosc excretion ratios in the 26 eczematous children was signifi can tl y eleva ted compared with

0022-202X /86/$03.50 Copy righ t © 1986 by The Society fo r In ves tiga ti ve Dcnmrology. Inc.

101

102 J>I IZE ET t\1.

that o f th e co ntro l g ro u p , w h eth er th e heal th y and skin di sease con t ro l g ro ups were analyzed sep;Jrately o r together (p < 0.0'1, W ilcoxon 's rank ed sum test, no npaired , fo r eac h co m pa ri son) . H owever , in 12 o f th e ecze m ato us children th e suga r excre ti o n rati o fe ll wi thin the ran ge fo r the hea lth y children and th at pub­lished fo r hea lth y adults [1 J J (Fi g 1). There we re n o sig nifica nt diffe rences between boys and g irls o r between the 2 contro l g ro ups.

T he lactost:! la ctu losc excre tio n rat ios of the ecze m ato us chil­d ren d id not di ffe r fro m th ose of e ith er o r bo th con t ro l g ro ups. In no p atien t was the ra tio hi g her th an the published m ean ± 2 SD va lue fo r health y ad u lts [1 4].

A n alysis of cl ini ca l and hi stori ca l data fo r th e ecze m atous p a­tients showed n o associatio n a m o ng in creased perm eability and race, severit y o f ecze m a (cl ass ifed as mild , m o derate, severe, o r ve ry severe), o r o ral antihi sta min e usc. In th e ecze m ato us g ro up th e hig hest lactu lose/rh a mn ose ra tios were fo und in th e yo un ge r p ati en ts an d no ne o f the 6 pa ti ents over R yea rs o f age h ad in creased permea bility (Fi g 2a) . T he contro l g ro ups, th o u g h s li g htl y o ld er , did no t show a simi la r rela tio nship of perm eabili ty to age (Fi g 2/J).

T wen ty-o ne of the children were ski n-pri ck tested to commo n food and airbo rn e alle rgens. N either the anti gens to w hich pos­itive immedi ate res po nses were o btained no r the dia m eter o f th e w h ea ls p rod uced were rela ted to gastro intestinal p erm ea bi lity. In 7 pa tien ts to tal lgE was ava il able and, in these in di v idua ls, there w as no co rrelation b e tween this and th e urin ary suga r ra ti o, 4 patients hav in g no rm al and 3 in creased perm ea bi lity. In 15 pa­tients in who m the eosino phil coun t was m easured there was also no corre lati o n be tween this and th e urin ary su ga r ratio.

Immediate h y persensiti vity reactio ns to food s, such as vo m­itin g, con tact urti ca ria, and an g ioede m a, were repo rted by 8 chi!-

Lactulose/Rhamnose Excretion Ratio

0 .4

0 .3

0 .2

0.1

£

•• •

• .. •• • ... ..... .. ..... ....... .... .................... .... ...... ... .... .. .... .... .

* * • ••• * ••

' " "· ···tgr ···· .. ··············· ··· ······ ···"' " " ''''"'"'" '' '"''' "''''

0 ~---------------------------------Controls

(n • 29)

Eczema subjects

(n •26)

Fig ure l. T he ratio of the proportion o f the o ral dose o fl actulosc excreted in 5 h. to the proportion o f the o ral close o f L-rh amnosc excreted . in 2o child ren w ith atopic eczem a and 29 contro l children. e = C hi ld ren w ith ato pic eczem a; 0 = hea lth y chi ldren; • = children w ith pso ri as is; T = child ren with clrstrophic epiclc rmolrs is bullosa; .t. = ;J chi ld w ith bullous ichth yos iform eryth roderm a. 1-/ol'iz olllnllill es in dica te ± 2 SD fo r heal thy adul ts [1 31.

T H E .JOURNAL O F INVESTIGATIV E DERMAT OLOGY

dren , 4 w ith in creased and 4 w ith n o rm al permea bi li ty. P rovo­ca t io n o f ecze m a b y food without a histo ry o f food-indu ced immedi a te-h y persensitivity reac tio ns w as repo rted in 4 children, 3 with in creased and 1 with n o rm al permeability . O ne chi ld , who had in creased perm eJ bili ty, repo rted b o th immedi ;~ te reactions to food s and food - induced ecze m a .

Thirteen children were o n so m e kind o f exclusio n die t at the tim e o f the stud y and, of th ese, 8 had in creased and 5 no rm al perm ea bility. H owever, o n rev iew, in o nl y 6 o f these pa tients, 3

Lactulose/Rhamnose Excretion Ratio

0 .40

0 .30

0.20

0

Lactulose /Rhamnose Excret ion Rat io

0 .3

0 .2

0 .1

0 0

0

0

0

0

5

0 0

8 0

5

10 15 17

Age in years

0 , ..

0 0 0 8

8 ,

0 0 0

0 0

10 15 Age in years

Figure 2. Ll ctu losc/rhamnosc excreti on ratio in relatio n to age in (n) 2b eczem atous children and (b) hea lth y and noneczematous dermatosis con­t ro l pati ents . Sy mbols arc as in Fig I .

VOL. 86. NO. 2 FEI!HUA itY I <JHr,

fro m each g roup, was the ecze ma judged to h ave been improved by dietary trea tm ent. T he 13 childre n w ho we re n ot o n dieta ry trea tm ent a t th e tim e (6 w ith in creased and 7 w ith no rm a l pt:rm c­ability ) ha ve been put o n diets subsequentl y. Of these, 10 pat ients (4 w ith in c reased and 6 w ith n o rma l pe rm eabi li ty) s howed so m e

benefit. Three patients re ported abn o rm a l gas troinres tinal sy mpto m s

con s is tin g of mild recurren t a bdo min a l di s te nsio n a nd inte rmit­tent loose stools w ith no o bvi o us rela ti ons hip to food. All three had in c reased gas trointest inal perrne:tbility and o ne, w ho had a serum a lbumin o f1 6 g/l iter, was subsequent ly di scovered to h ave a protein-losin g ente ro p ath y o f :llle rg ic o ri g in co nfmn ed b y 51C r­labeled a lbumin s tudies and jejuna l biopsy, a nd trea ted s uccess­full y w ith a strict exclusi o n diet; thi s child is th e s ubj ec t o f a case report 11 5 1. Serum a lbumin was a lso measured in 9 o the r children and in those there was no co rre latio n with the urin ar y su ga r ra ti o , a lth o u g h the o nl y child w ith a lo wer th a n n o rmal serum albumin (30. 5 g/ lite r, n o rm a l ran ge 36-52 g/ lite r) was also I o f the 3 patie nts w ith intermittent gas troi n testin a l sympto m s and an in­creased permeab ility; this child has n o t been in vest iga ted for gas­trointest in al protein loss.

DISCUSS IO N

These res ults indi cate that a pro po rti o n of childre n w ith atop ic ezce m a h a vc in c reased gas tro intestin a l pe rm eability as judged b y the urinary excret ion ratio o f inert s uga r m a rke rs. They a re a t vari a n ce with the findin gs of a smaller s tud y b y DuMom, Beach , and M en zies 17] , but in agreem ent w ith those o f Pa rrilli e t a l [8]. T h e y a re also consis tent w ith the results of Ukaba m , M ann , and Cooper f1 6 l us ing mannito l rath e r th an rhamn ose as the s m alle r s u ga r m a rke r. O ur res ults s how a g rea ter frequen cy of this ab­no rm ality in th e youn ge r ecze m ato us child .

T h ese findin gs do n o t provide d irect evidence for s imilarl y enhanced g astro intestinal upta ke of a nti gen in eczematous pa­t ients, sin ce it cannot be assumed that the m echanis m for in creased lactulose a bso rptio n is th e sa me as th at for food anti gen. The lactulose mol ecule is mu c h s malle r (M,. = 342) than intact food proteins , which gene ra lly ha ve a M,. exceed in g 10\ a lth o ug h it is k n own that s m a ll er p o lypeptide fragments re leased during prote in d iges ti o n m a y be anti genic [17].

The in creased gas tro intest in al perm eability w e have shown mig ht reflect mu cosal dama ge induced by loca l h ypersen sitivi t y reac­tions to foods . Gas t roi ntesti nal sympto ms arc common in chil­dren with atopic ecze m a and a re o ften associated w ith the in ges­tion of s pecifi c fo ods, su ggestin g th at alle rgy m ay be the ca use (1 8-21 J. Furthe rmo re, morpho logic a bno rmalities , in associa ti o n w ith impaired ma x imal g astric acid secretio n [22], h ave bee n reported in gas tri c and j ej un al biopsies fro m so m e chil dren w it h a topic eczema, fllldin gs that were n o t confmed to t hose w ith frank gastro intes tinal symptoms [20,22). It was proposed that these changes were allergi c in o ri g in.

None of o ur patients repo rted gas tro intestin al symptoms at the time of urin e co llec t io n, but so me were intermittentl y subject to su ch symptoms and othe rs w ere avoidin g food s that had prev i­o u sly pro voked them. W e pe rform ed a j ej una l bio ps y in o ne patient in w h o m th e re was s tron g clinical sus picio n of m alab­sorpti on and this confirm ed an allerg ic entero path y; the re were no t felt to be adequate clini cal g rou nds fo r pe rfo rmin g jejunal b io p sies in an y o f the o the r p a ti ents . Lactose/ lac tul osc rati os were norm al in a ll pati e nts, includin g the o n e w ith hi stologic e viden ce of enteropath y, su ggestin g that su ch mucosa l abno rm a lities as were present were in s u fl:ic icnt to cause lactase defi cien cy.

We examined o ur patients for e vidence o f food a ll ergy o n the basis of skin tests, his tory of food-provoked symptom s, and re­sponse to exclus ion di et and found n o associati o n be tween an y of these a nd in creased gas troi ntestin al pe rm ea bilit y . Thi s, and the tenden cy fo r youn ger c hil dren in o ur g ro up to h ave the hi g h es t lactu lose/ rhamnose ra tios, rai ses the possibility that we are d e­scribing a prima r y abno rmality, possibly a d elay in gas trointes­tina l m a tura tio n. Inc reased pe rm ea bility is normall y present at

IN C it EASED INTESTINA L I'E I{MEAIJ ILIT Y IN ATOPIC ECZEM A 103

birth , but fa lls to "'adult" va lues b y th e 9 th da y of life 123]. If this ea rl y period of e nh an ced pe rm ea bility were pro lo n ged , in­c reased exposure to an tigen mi g ht result a nd pe rh aps predis pose an individu al to the subsequent deve lop m ent of ecze m a.

It is, h owever , o ur view t hat food a llergy re m ains th e m ost lik el y cause of the ab nor m a lity we h ave s ho wn. This hypothes is ca n be tes ted b y the m eas ure rn c n t of gas tro intest in al perm eability be fore a nd alter s uccess ful o ral cha lle nge w ith s pecific food a l­le rgen s a nd before and a fter tota ll y effecti ve exclusio n die ts (i. e., d ie ts in w hi ch a ll clini ca ll y re leva nt food a ll e rg en s ca n definite ly be sa id to have b een excluded). \Y/e a rc, the refore, unde rtakin g suc h s tudies a t th e present tim e.

Wt• grarcjil!ly ackn owledge rhcfinancin / SllfJJ><> r! ofrh c Nn riowd Eczc111n So(it•ry (M C P a11d DJ A), rh e Nnrio11al 1-lca lrh m1d Medica l l~escn rch Co 1111Cil 1~( A lls­rralia ( Njl-1), and C /axa Labor,1/orics Lrd. (PB) . Dr. I. Mc11zics a11d his sra[f i11 the Dcptll'tlllent of C hc111iw l Pathology nt Sr. Th onws' Hospital pro11idcd itwnluab lc help and ndFicc. We tl11wk Pro{essMj. Soorltifl(o r hclJ!fiil dismssiom t111d n/1 the children n•hv rook pnl"l in !his study nnd !heir pnrcnrs.

REFERE N CES

I . Atherton OJ . Sewe ll M. Soothill JF, Wells RS, C hil ve rs CD: A double blind , controlled, crossover tri al of an anrigcn avoidance diet in atopic eczem a. Lancet I :40 1-403, 1978.

2. Juto P. Engberg S, Winberg J: Treatm ent o f infant ile atopic der­ma titis w ith a stri ct elimination diet . C lin Allergy 8:493-500, 1978

3. Paga nelli 1{. Levinsk y RJ , Bros ro ff J , Wraith DG: Immune com­plexes containing food prote ins in norm al and ato pic subj ects after oral ch.11i cngc and cfrccr of sod ium cro moglyca rc on anrigen ab­sorption . Lancet I : 1270-1272. !979

4. Paganelli R, Athcn on DJ , Levinsky RJ: D iffe rences between no rmal and milk all ergic subjects in their immune responses afte r milk in ges tion . Arch Dis C hild 58:20 1-206, 1983

5. Jackson PG. Lessof MH . Baker RWR , Fcrrett J, MacDona ld DM : Intestina l permeability in patients w ith eczema and food allergy. Lancet I :285-286, 198 1

6. Wheeler PG , Menzies IS. C ream er B: Effect o fh yperosm o lar stimuli and coe li ac di sease on the permea bility of the hum an gast ro intes­ti nal tract. C lin Sci Mo l Mcd 54 :495-50 1, 1978

7. DuM ont GCL. Beach RC , Menzies IS: Gas tro intes tinal perm ea bi lity in food-allergic eczem atous children . C lin Allergy 14:55-59, 1984

8. Parrilli G, Aya la F, Lembo G. C uomo 11 .• B udillon G. Santoianni 1': Abno rm al intestinal pcrrneabi li ty to lactu lose in patients with atop ic dermatiti s, in lm nHtnodcrmato logy. Edited by D M Ma cDonald. London, Butterworth . 1984, pp 2 1-22

9. Da hlqv ist A, G ryboski J D: Inabili ty of the human smal l intestina l lactase to hydrolyse lactulosc. Biodtim Biophys Acta 11 0:635-636. 1%5

10. Fo rdtran JW, Recto r FC , Ew ton MF, Soter N , Kinney J : Permea­bilit y characteristics o f the human inrcstinc. J C lin In vest 44: I 935- 1944, 1%5

II . Menzies IS , Lak er MF, Po unde r R, Bull]. Heyer S. W heeler PG. C rea mer 13: Abno rmal intestinal perm ea bil ity to suga rs in vi llous atro ph y. Lancet 2: 11 07- 11 09, 1979

12. Menzies IS, M ount JN, Wheeler MJ: Q uantitati ve es timation of cl inica ll y impo rtant m onosaccharides in pla sm a by rapid thin laye r ch romatography. Ann C lin Bi ochcm 15:65-76, 1978

13. Menzies IS: T ransmucosa l passa ge of inert m olecules in hea lth and disc:tsc, lntestin :t ! Absorption and Secretio n. Proceedings of Fa lk Sy m posium 36. Edited by E Skad haugc , K Hcimz. Londo n, MTP Press. pp 527-543, 1983

14. Men zies IS: Absorptio n of intact o ligosaccharide in hea lth and dis­case. Bi ochent Soc T rans 2: I 042- 1047 , 1974

15. Jenkins HI< , Walker-Smith J A, Ath erton DJ : P1·orein- losing cnter­opath y in atopic der matitis. l'cdiatr Derma to!. in press

16. Ukabarn SO, Mann R). Cooper BT: Sm all in testinal permeabili ty to sugars in p:ttients w ith atopic eczema. Dr J Derma toll ! 0:649-652. 1984

17. Spies JR , Stcva n MA. Stein WJ , Gordon WG: C hemistry o f aller­gens. 1m Arch Allergy Appl!mntunol 48:49-7 1, 1975

104 PI KE ET AL

18. Schur S, H yde JS, Wypych Jl: Egg w hite sensitivity and atopi c cczem;L J Allergy C lin lmmuno l 54: 174-179, 1974

19. Bonifazi E, Ga rofa lo L, M onteri si A, Mcncghini C L: Food allergy in atopic derm atiti s: experim ental observations. Acta Derm Ve­nereal (Stockh) 5H:349-352, 197H

20. McCalla R, Savilahti E, Pcrkki ii M, Kuitunen l', Backman A: Mo r­phology of th e jejunum in children w ith eczem a due to food al­lergy. A ll crgy 35:563-571, 1980

THE j OU HNAL O F INVESTI GATIV E DERMATOLOGY

2 1. Sampson HA: Ro le of immediate food hypersensitivity in the path­ogenesis o f atopic dermatiti s. J Allergy C lin lmmun o l 7 1:473-480, 1983

22. Kokkonen J , Similii S, Herva R: Gast rointes tina l findin gs in atopic children. Eur J Pcdiatr 134:249-254, 1980

23. Beach RC: C hanges in gastro-in testina l permea bility in the human neonate. M.D. Thes is, Uni versity o f London, 1984, pp 145-1 58