Change in diagnostic accuracy when tests undertaken in different ways

StudyPopulation testedPrevalenceDiagnostic accuracy for various food allergensComments
Positive test on challengeImmediate reactionDelayed reactionCombined immediate & delayedSensitivity (%)Specificity (%)PPV (%)NPV (%)
Niggemann 2002182

EL=DS III
30 children aged 3–58 months with atopic eczema and suspected food-related symptoms.48% milk, 20% egg, 64% sow, 22% wheat

0 placebo
NRNRNRMilk
60 (12mm)
0 (6mm)
Milk
100 (12mm)
100 (6mm)
Milk
100 (12mm)
0 (6mm)
Milk
73 (12mm)
52 (6mm)
Antihistamines were stopped for at least 72 hours before testing.
TCS were allowed.

Food challenge: 55 challenges undertaken.
Placebo: neocate.
Test: fresh pasteurised milk, raw egg, wheat powder, soya milk. Provocation stopped if clinical symptoms were observed or when the highest dose was reached. Children challenge. Positive test: if one or more of the following: urticaria, angioedema, wheezing, vomiting, diarrhoea, abdominal pain, shock, or exacerbation of eczema

Patch test: using Finn chambers of 6mm or 12mm in diameter. Foods (as in the DBPCFC): 50microlitre or 20microlitre respectively used. Sites checked after 20 hours, read after 20 minutes, and again at 72 hours.

Skin prick testing was also undertaken.
Egg
71 (12mm)
29 (6mm)
Egg
100 (12mm)
100 (6mm)
Egg
100 (12mm)
100 (6mm)
Egg
67 (12mm)
44 (6mm)
Soya
100 (12mm)
0 (6mm)
Soya
100 (12mm)
100 (6mm)
Soya
100 (12mm)
0 (6mm)
Soya
100 (12mm)
82 (6mm)
Wheat
100 (12mm)
0 (6mm)
Wheat
89 (12mm)
100 (6mm)
Wheat 75
(12mm)
0 (6mm)
Wheat
100 (12mm)
75 (6mm)
Heine 2006185

EL=DS Ib
87 children aged 0.5–13.5 years (mean 2.4 years) with atopic eczema and suspected food allergy to cow’s milk, hen’s egg, wheat and soya.45%75%11%15%Mild erythema (39%): 45675359Topical and systemic corticosteroids were discontinued 72 hours before testing.
DBPCFC ‘as per previously published protocol’ 165 were undertaken.
Atopy patch test: one drop fresh pasteurised cow’s milk, fresh soya milk, whisked whole hen’s egg and a wheat gluten flour suspension applied to the skin and covered with 12mm Finn chambers for 48 hours.
Skin changes graded as none, mild, moderate, severe, induration, papule formation, vesiculation, and presence of a crescendo phenomenon (increase in severity of reaction between hours 48 and 72).

Personnel reading the patch test results were blind to the results of the DBPCFC.

Diagnostic accuracy results are for delayed reactions.
Moderate erythema (10%): 15936557
Any severity (49%): 60605664
Minor induration (11%): 15926156
Extensive (5%): 9998857
Any severity (16%): 24916959
Papules (1–3, 16%): 19875456
4–6 (12%): 12894755
7 or more (12%): 21968059
any papules (39%): 60746074
Crescendo (8%): 11935756
Moderate erythema + crescendo (3%): 5998056
Induration + crescendo (3%): 4986055
Papules + crescendo (≥7; 4%) 5986755
Moderate erythema + induration (4%): 810010057
Moderate erythema + papules (≥7; 4%): 8998656
Induration + papules (≥; 7%):1510010058
Moderate erythema + induration + papules (≥7; 4%): 810010057
Moderate erythema or induration (27%): 41867064
Moderate erythema or papules (≥7; 18%): 27906960
Induration or papules (≥7; 21%): 31876660
Moderate erythema or induration or papules (≥, 26%): 36826361
Verstege 2005184

EL=DS III
385 children aged 3 months to 14.5 years with suspected food-dependent symptoms to cow’s milk, egg, wheat, and/or soya. 87% had atopic eczema.

11% also had asthma, 6% recurrent wheezing, and 27% hay fever.
43% (63% egg, 49% milk, 28% wheat, 19% soya)

4% placebo
67%14%19%Hen’s egg 93%598083Antihistamines were discontinued 72 hours before testing.
TCS were allowed twice daily.

Food challenges 735 were undertaken. 75% were DB and 25% were open.
Placebo – 280 challenges (neocate).
Test: cow’s milk, egg, gluten, or soya milk. Doses were titrated. Challenges were positive if objective cutaneous symptoms (urticaria, worsening of eczema), or respiratory or gastrointestinal symptoms were seen.

Skin prick test: one drop of each fresh food applied to the forearm: cow’s milk, native hen’s egg (whisked white and yolk), gluten powder, and soya milk.
Positive (histamine) and negative (saline) controls were also applied. Reactions were read at 15 minutes. A wheal size of 3mm or greater than the positive control was considered a positive reaction.

Other analysis of data was undertaken: wheal diameter 13mm for hen’s egg and 12.5mm for milk, would give a 95% PPV. respectively). Predictive values could not be calculated for wheat and soya.
Cow’s milk 85757683
Wheat 65775285
Soya 21882983
Fiocchi 2002190

EL=DS III
34 children aged 1–4.4 years, median 2.26 years) with atopic eczema and IgE sensitisation for foods. They were enrolled in this study if they reported immediate symptoms attributed to consumption of beef.59%NRNRNRUsing commercial beef extract:: 90100NRNR
DBPCFC: with beef. No further details in this publication.

Skin prick test: using extract of lyophilised skeletal muscle tissue and with raw unfrozen skeletal muscle. Positive (histamine) and negative (saline) controls were also applied. Reactions were read after and unspecified time. A wheal size of 3mm or greater than the positive control was considered a positive reaction.

It is not stated whether the food challenge was undertaken without knowing the results of the prick test.
Using fresh beef extract:: 10078.57
Kim 2002189

EL= DS III
292 children and older people (mean age 12 years) with atopic eczema35% for crude milk
37% crude egg
35% crude soyabean
NRNRNRMilkElimination diet was used for 2 weeks prior to testing.

DBPCFC:
Placebo – the vehicle for DBPCFC (mixed cereal flour).
For the food challenge skimmed milk powder, freeze-dried flour of egg and soybean powder were used.
Determinants of a positive test were the appearance of dryness/scaling, erythema, wheal, excoriation, or papulation. An increase of 20% compared to pre-test score was also regarded as a positive reaction.

Skin prick testing: crude extracts of milk, egg, and soybean were prepared using phosphate buffered saline. Egg and soyabean were boiled for 1 hour prior to extraction. Extracts were centrifuged and supernatants collected, which were then dried. Stock solutions were than prepared. Glycerol was used as a negative control.
Prick testing was done on the left forearm using crude and commercial extracts. Histamine was the positive control. Reactions were read after 15 minutes. The minimum size of positive reaction was 3mm.
44 crude864075
34 commercial702772
Egg
64 crude815963
56 commercial534071
Soya
54 crude654378
33 commercial712638
Bibliographic informationStudy type and evidence levelAim of studyNumber of patientsPatient characteristicsOutcomes and resultsComments
Darsow U;Vieluf D;Ring J;

1995 Mar183
Study Type:
Case series

Evidence Level: 3
To compare the proportion of positive results to an atopy patch test when two different concentrations and vehicles were used.Total No. of Patients = 36

Children and adults with atopic eczema N = 36
Children and adults, aged 3–69 years (mean 29 years) with atopic eczema. Of these 16 reported eczematous reactions after exposure to at least one of the three allergens tested (HDM, cat dander, grass pollen). All were in the stable phase (partial or complete remission).Source of Funding: None declared.

Comments:
Antihistamines and systemic/topical corticosteroids were discontinued for 7 days before testing.
The patch testing use two concentrations, 1000 protein nitrogen units (PNU)/gm, and 10,000 PNU/gm, in two different vehicles (white petrolatum/10% isopropyl myristate and methylcellulose hydrigel/10% propylene glycol).
Lyophilised grass pollen extracts were used.

The patch was applied for 72 hours under 12mm Finn chambers. They were evaluated at 48 & 72 hours and classified as follows: (+) erythema, + erythema, infiltration, none or few papules, ++ erythema, intensive filtration, many papules, occasionally vesicles, and +++ for densely aggregated papules and vehicles.

Skin prick tests were also done (no details, no definition of a positive test), and IgE (total and specific) levels measured. ‘Concordance’ between the tests was also reported (not defined) - data not reproduced here.
Perackis K;Staden U;Mehl A;Niggemann B;

2004188
Study Type:
Case series

Evidence Level: 3
To compare the results of a skin prick test using whole egg and egg white.Total No. of Patients = 45
Children who underwent skin prick testing
N = 45
Children aged 6–113 months with suspected allergy to hen’s egg. 96% had atopic eczema.Source of Funding: None declared.

Comments:
Two drops of egg were applied to the volar forearm (one drop of whisked native whole egg, one drop of native egg white).
Reactions were read after 15 minutes.
A positive test was indicated by a wheal diameter of 3mm or more, without reaction of the negative control (sodium chloride 0.9%). All responded to histamine dihydrochloride (the positive control).
Antihistamines and corticosteroids were prohibited 48 hours before testing.
Bibliographic DetailsStudy type and evidence levelNumber of patientsPatient characteristicsIntervention and comparisonOutcome measures, follow-up and effect sizeComments
Niggemann B;Binder C;Dupont C;Hadji S;Arvola T;Isolauri E;

2001 Apr 200
Study Type: Randomised Controlled Trial

Evidence Level: 1+
Total number of patients = 73
Amino-acid based formula
N = 42

Extensively hydrolysed whey formula
N = 31
Infants aged 1–9 months (median 5.7 months) with atopic eczema and proven cow’s milk allergy/intolerance (on DBPCFC).
Mean SCORAD scores 24.6 (0–72).
Median total IgE 16.0 kU/L (less than 2.0 yo 4710.0).
Amino-acid formula
vs
Extensively hydrolysed whey formula
Outcomes at 6 Months:

SCORAD
No numerical data for each group (data shown in graphs only).
Mean overall score at endpoint: 10.7 (95% CI 71 to 14.2, p<0.0001) vs

Growth
No numerical data (shown only in graphs).
Reported that there was a statistically significant increase in length standard deviation scores in the amino-acid group, p<0.04; weight-for-length scores developments were ‘similar’ in both groups.
Source of Funding: SHS, Liverpool UK.

The amino-acid based formula used was Neocate, and the whey formula Alfare or Pepti- Tutteli. Quantities consumed were not specified.

Energy intake was similar in both groups.
Businco L;Benincori N;Nini G;Businco E;Cantani A;De Angelis M;

1986 Dec 219
Study Type: Randomised Controlled Trial

Evidence Level: 1−
Total number of patients = 31

Sodium cromoglicate + exclusion diet
N = 31

Placebo solution + exclusion diet
N = 31
Children aged 6 months-10 years with severe atopic eczema requiring ‘continuous treatment’ (not defined) but not corticosteroid therapy (not stated whether topical or systemic). The children also had evidence of exacerbation of symptoms caused by eating one or two foods (established by challenge tests at home). They had positive skin tests to ‘a range of allergens’ and serum IgE levels higher than the normal range for their age.
39% also had asthma, 26% allergic rhinitis, 10% conjunctivitis, and 6% urticaria.
Sodium cromoglicate + exclusion diet
vs
Placebo + exclusion diet
Outcomes at 8 Weeks:

Severity*
No numerical data; results shown in graphs only.
p=NS between treatments when the cross-over sequence was sodium cromoglicate followed by placebo.
p<0.05 in favour of sodium cromoglicate when placebo was taken first in the crossover sequence. vs

Parent rating of symptoms
No numerical data; results shown in graphs only.
p=NS between treatments when the cross-over sequence was sodium cromoglicate followed by placebo.
p<0.01 in favour of sodium cromoglicate when placebo was taken first in the crossover sequence. vs

Opinion as to which treatment ‘most effective’
Sodium cromoglicate: 52% parents rating, 68% clinician’s rating
Placebo: 13% and 6% respectively
Both: 16% and 0% respectively
Neither: 3% and 10% respectively. vs

Adverse effects
6% transient worsening at the beginning of the trial
0% dizziness
0% drowsiness
3% bowel disturbance vs

3% transient worsening at the beginning of the trial
3% dizziness
3% drowsiness
6% bowel disturbance
Source of Funding: Fisons Ltd supplied drugs

The study was a DB crossover trial with 2×8- week treatment periods with a 2-week washout period in between.
[EL=1−] because no baseline data and analysis was undertaken on fewer children than were randomised.
Withdrawals: 8, due to lack of response (2 - excluded from the analysis); non-adherence (2); and ineffective treatment (4).

Exclusion diet: based on skin test and IgE results. Cow’s milk and egg was eliminated in 81%, fish in 6%, and wheat in 13%. The diet was taken for weeks 1–4 of the 8-week treatment period, and then foods reintroduced stepwise.

Severity assessed by dividing the body into ten areas which were assessed for redness/weeping/vesiculation/crusting, excoriations, lichenification on a scale of 0–3, none-severe. Maximum total score 240.

Parents recorded daytime itch, sleep disturbance due to itch, weeping, and redness on a 0–3 scale (maximum score 12).

Dose of sodium cromoglicate: 400mg daily for children of 10–20kg bodyweight, 800mg daily for 20–30kg, 1200mg for 30–40kg, 1600mg for >40kg. The daily dose was taken in four divided doses.

No topical or systemic corticosteroids were allowed.
Businco L;Meglio P;Amato G;Balsamo V;Cainelli T;Cantone P;Castro M;Coletta A;Corrias A;Giorgi PL;Grazioli I;Longo- Papadia L;Marcucci F;Masi M;Pavesio D;Scotta S;Seidenari S;Vierucci A;

1996 Mar 216
Study Type: Randomised Controlled Trial

Evidence Level: 3
Total number of patients = 1085

Restricted diet*
N = 505

Oral sodium cromoglicate
N = 506
Children aged 5 months - 14 years (median 2 years) with AE. 58% also had a personal history of atopy, e.g. asthma, rhinitis.Restricted diet*
vs
Oral sodium cromoglicate
Outcomes at 4 Weeks:

Severity

Change in proportion with severe AE: from 43%–13% vs 48%–15%
Change in proportion with mild AE: from 23%–69% vs 19%–61%, p<0.001 from baseline for both groups for both outcomes, ‘no significant differences’ between groups (no p value stated) vs

Extent

Change in % whose extent severe: 35%-17% vs 36%-21%
Change in % whose extent mild: 29–51% vs 27–44% vs

Adverse effects

1% diarrhoea
0.5% lack of appetite
0.2% restlessness
0.2% weight loss vs

Adverse effects
1% diarrhoea
1% vomiting
1% nausea
0.2% lack of appetite
0.2% restlessness
2% abdominal pain
0.4% headache
0.4% pruritus
0.2% rash
0.2% urticaria
0.2% constipation
0.2% joint pain
Source of Funding: none declared

EL=1− because only 80% analysed (overall 93% completed, 82% in the diet group and 91% in the sodium cromoglicate group).

*Restricted diet consisted of rice, lamb, turkey, lettuce, cooked carrots, sweet potatoes, pears, olive oil, mineral water, black tea, salt, brown sugar.
Sodium cromoglicate: 80mg/kg/day in four divided doses - powder diluted in 20ml water. Mean dose used was 71mg/kg.

Children were randomised to treatment irrespective of SPT and RAST results.

Severity: pruritus, erythema, vesiculation, papules, excoriation, scale crusting and lichenification assessed on a 4-point scale (1–4, no symptoms - severe). Total score = score for each area x 20. Global score: 140=absent, 141–170 mild, 171–200 moderate, more than 200 severe.

Extent: according to the number of body areas involved; o=absent, 1–5 mild, 6–10 moderate, more than 10 severe.

Treatment considered effective when at least 40% improvement of the global score occurred.

Significantly more children in the diet group had positive tests to foods on SPT. Results were compared for children with positive or negative tests - ‘no significant differences’ in response were noted (data presented in graphs only).
Ewing CI;Gibbs ACC;Ashcroft C;David TJ;

1991 220
Study Type: Randomised Controlled Trial

Evidence Level: 1−
Total number of patients = 50

Zinc sulphate (sustained release capsules containing 61.8mg zinc sulphate)
N = 25

Placebo
N = 25
Zinc
vs
Placebo
Children aged 1–16 years (mean 8 years) with atopic eczema, being treated with emollients and TCS, and some with trimeprazine.Outcomes at 8 Weeks:

Itch

4.6 vs

3.4, p=0.01 vs

Sleep disturbance

No numerical data, p=0.77 between groups vs

Trimeprazine dose

No numerical data, p=0.11 between groups vs

TCS quantity applied

259.3g (mean) vs

188g (mean), p=0.23 vs

Emollient quantity applied

1159.1g (mean) vs

511.6g (mean), p=0.13 vs

Surface area score (mean change)

+5.5 (29%) vs

+1.6 (11%), p=0.53 vs

Erythema score (mean change)

−0.1 (4%) vs

−0.4 (17%), p=0.10 vs

Combined disease severity score (mean change)

−12.6 (35%) vs

−4.7 (14%), p=0.60
Source of Funding: Smith Kline French supplied drugs

[EL=1−] as only those who completed treatment were analysed. Withdrawal rates were 6% zinc and 10% placebo; reasons were nonadherence (1 each group), diarrhoea (1 placebo), 1 exacerbation of eczema (1 zinc), itchy rash (1 zinc, 2 placebo), Herpes simplex infection (1 placebo).
Usual treatment continued during the study.
Families recorded redness, daytime itch, and night-time sleep disturbance on a 1–10 scale.

Severity: body divided into 14 areas and the surface area affected estimated. Each area score on a scale of 1–5 for severity. Surface area x severity = combined disease severity score.
Graham P;Hall-Smith SP;Harris JM;Price ML;

1984 217
Study Type: Randomised Controlled Trial

Evidence Level: 1−
Total number of patients = 29

Sodium cromoglicate
N = 29

Placebo
N = 27
Children aged 3–12 years (mean 7 years 5 months) with chronic AE requiring regular attendance at outpatient clinics. Children were treated with a ‘tailored diet’ which was not detailed, other than foods were eliminated and re-introduced according to IgE levels.
Baseline severity score 2.09, extent 1.98.
Sodium cromoglicate
vs
Placebo
Outcomes at 27 Weeks:

Severity (mean score change)

−0.69, p<0.01 vs baseline vs

−0.63, p<0.01 vs baseline vs

Extent (area; mean score change)

−0.10 vs

−0.16
Source of Funding: Fisons Ltd provided study medication

DB cross-over RCT. Only 76% completed and were analysed [EL=1−].

Sodium cromoglicate dose: 100mg four times a day before meals (capsules) for 3 weeks increasing to 200mg for next 3 weeks.

Treatment periods were of 6 weeks’ duration, with a 2-week washout in between (usual diet).

Symptoms, severity and extent were measured on a 4-point scale.

All previous medication for AE was stopped and all were given HC 1% (not stated whether cream or ointment), and emulsifying ointment as needed.
Leung TF;Ma KC;Cheung LT;Lam CW;Wong E;Wan H;Hon EK;

2004 Dec 212
Study Type: Randomised Controlled Trial

Evidence Level: 1−
Total number of patients = 15

Amino-acid based elemental diet*
N = 15

Control
N = 15
Infants and young children aged under 3 years (median 1.4 years, IQR 0.6–2.6) with AE. All had a positive SPT to at least one of six food allergens (cow’s milk, soy, whole egg, peanut, wheat, mixed fish), and raised cow’s milk or soya bean specific IgE (35 KaU/L or more).
Median SCORAD 23.9 (IQR 10.5–29.7).
Amino-acid-based elemental diet
vs
Control
Outcomes at 5 Months:

SCORAD

treatment different 3.97, p=0.274
treatment × period interaction 7.23, p=0.012 vs

Parental global health score (VAS 1–9, worst-best)

treatment difference 0, p=0.792
treatment × period interaction 0.01, p=0.958
Source of Funding: Chinese University of Hong Kong

EL=1− because no baseline data reported therefore not known whether groups were similar at baseline. Only completers were analysed (73%). The reasons for the 4 withdrawals were: 2 drank less than required, 1 refused to drink the amino-acid formula, 1 had ‘too mild’ AE.

The amino-acid formula used was Neocate. 500ml or more was advised to be taken (not stated whether this is per day).
No dairy or soya based products were allowed during the study.
The control group continued with their pre- existing formula.
A dietician conducted a nutritional assessment.

All treatments for AE remained unchanged (TCS and sedating antihistamines).

Severity was assessed by a paediatric dermatologist unaware of treatment allocation.

The positive tests to skin prick allergens were: cow’s milk, soya bean, wheat, mixed fish (all 1 each), whole egg (11), mixed peanuts (4), house dust mite (6).

Daily fluid and energy intake did not differ before and after the interventions.

Treatment was given for 6 weeks with a 6 week washout in between.
Tsoureli-Nikita E;Hercogova J;Lotti T;Menchini G;

2002 Mar221
Study Type: Randomised Controlled Trial

Evidence Level: 1−
Total number of patients = 96

Vitamin E (400 units [268mg]) once daily
N = 50

Placebo
N = 46
Children and adults aged 10–60 years with moderate-severe atopic eczema affecting 30–70% of body surface area.Vitamin E
vs
Placebo
Outcomes at 8 Months:

Global assessment (response to questionnaire)

8% worsened
12% no change
20% slight improvement
46% great improvement
14% almost complete remission
vs
78% worsened
11% no change
9% slight improvement
2% great improvement
0% almost complete remission vs

Adverse effects

none
vs
none
Source of Funding: None declared

[EL=1−] because although the study is described as randomised in the abstract the methods would suggest that the treatments were not allocated randomly. No baseline data were reported.

Only petrolatum emollients were permitted during the study.
Viljanen M;Savilahti E;Haahtela T;Juntunen- Backman K;Korpela R;Poussa T;Tuure T;Kuitunen M;
2005 Apr 222
Lactobacillus (5×10 -9 colony forming units)
N = 80

Mixture of probiotics (Lactobacillus, Bifidobacterium, Propionibacterium)
N = 76

Placebo
N = 74
Lactobacillus
vs
Mixture of probiotics (Lactobacillus, Bifidobacterium, Propionibacterium)
vs
Placebo
Outcomes at 1 Months:

SCORAD (all infants - mean score change)

−16.6 (48%) vs

−14.0 (42%) vs

−14.2 (47%), p=NS between groups

SCORAD (in 52% with verified cow’s milk challenge - mean score change)

−15.1 (45%) vs

−14.5 (43%) vs

−15.2 (46%), p=NS between groups
DB RCT. Only 91% completed and analysed; reasons for withdrawals were: moved away (2), did not start diet because symptoms alleviated (11), unable to tolerate diet (4), protocol too difficult (3).

Eczema lesions were treated with emollients and HC 1% (not stated whether cream or ointment).

The probiotics and placebo were given as capsules mixed with food, twice daily.

The dose of Lactobacillus given was 5×10 (−9) colony forming units.
Bibliographic informationStudy type and evidence levelStudy aims/objectivesNumber of patientsPatient characteristicsOutcomesComments
Brouwer ML;

2006 Jul 224

Country: Netherlands
Study Type: Randomised Controlled Trial

Evidence Level: 1−
Total No. of Patients = 50

Hydrolysed whey formula + Lactobacillus rhamnosis
N = 17

Hydrolysed whey formula + Lactobacillus
N = 16

Control: hydrolysed whey formula only
N = 17
Infants (age 1.1 – 5.2 months) with AE and suspected cow’s milk allergy routinely attending a baby health clinic.Outcomes at 3 months:
Reduction in SCORAD index
Funding: Not stated

SCORAD index was reduced in all groups irrespective of treatment implying reasons other then the studied intervention caused this reduction.
Agata H;Kondo N;Fukutomi O;Shinoda S;Orii T;

1993 Feb 205
Study Type: Cohort Study

Evidence Level: 2−
Total No. of Patients = 150

Children 3 months- 13 years with sensitivity to hen’s egg or cow’s milk (on basis of history and food challenges)
N = 43

Nonatopic healthy children without milk or egg sensitivity
N = 64

Children sensitive to egg and milk with urticaria, angioedema, acute gastroenteritis within 1 hour of the food challenge
N = 53
Children aged 3 months - 13 years with atopic eczema with positive food challenge to eggs/milk, or who developed urticaria, angioedema, acute gastroenteritis, and bronchial asthma within 1 hour of the challenge test, and a control group who did not have atopy.Outcomes at 3 Months:

Severity in those sensitive to egg

(n=33)
in 27 who had elimination diet, 23 improved by one category, 4 improved by 2 or more (at baseline 7 were mild, 13 moderate, 7 severe)

in 6 who did not have an elimination diet, 4 had no improvement, 2 worsened by one category (at baseline 1 was mild, 5 moderate)

Severity in those sensitive to milk

(n=21)
in 16 who had elimination diet, 1 was unchanged, 10 improved by one category, 5 improved by 2 or more (at baseline 1 was mild, 9 moderate, 6 severe)

in 5 who did not have an elimination diet, 2 had no improvement, 3 worsened by one category (at baseline 2 were mild, 3 moderate)
Funding: Ministry of public welfare, Japan

Only 43 of 54 ‘randomly selected’ children were treated with elimination diets; the other 11 continued with the ‘offending’ foods.

Severity of AE was graded on the basis of food-challenge symptom scores, where 0=absent, 1=mild, 2=moderate, 3=severe.

DBPCFC was performed if there was a clear-cut history of major allergic skin symptoms after ingestion of a specific food or if there was a chance of systemic anaphylaxis. The food challenge consisted of egg, milk, or placebo.
Devlin J;David TJ;Stanton RH;

1991 Jan 209
Study Type: Case series

Evidence Level: 3
To describe the treatment of children with atopic eczema with a diet eliminating all but six foods.Total No. of Patients = 43

Elemental food (100% free amino acids)
N = 43
Children included in the Devlin 1991 study.208
Those who were subsequently treated with an ‘extreme antigen avoidance regimen’ including an elemental diet (100% free amino acids) in hospital using Vivonex .
At baseline median 70% (20–96%) body surface area was affected. Median erythema score was 3 (2–3), median severity score 210 (60–288).
Severity

Median score: 33% of baseline score (3–134%)
Global success/failure

27% treatment failures (score same or worse than at beginning)
73% treatment success (median reduction to 27% of baseline score [3–67%]; 96% were only using emollients at the end of treatment)
Adverse effects

89% (of n=34) lost up to 17% body weight.
19% loose stools
0% electrolyte disturbance
serum albumin fell in 93% (of n=27), from mean 30.8g/l to mean nadir of 21.2g/l
Source of Funding: North Western Regional Health Authority

Comments:
Of the 43 children, 1 declined the intervention, 2 refused to drink the formula feed, and only 37 who had been followed up for 12 months or longer were analysed.

Severity score = surface area affected x degree of erythema (0–3).

All mammalian and avian pets were removed from the home. Investigators ‘ensured that rigorous measures’ were taken to reduce house dust mite levels in the bedroom.
Corticosteroids (not stated whether topical) were discontinued at the time of hospital admission, but emollient and trimeprazine (night sedation) were continued. All were also given an appetite stimulant (cyproheptadine 2mg twice daily).
All usual food and drink (including water) were excluded, and the child fed exclusively on unlimited quantities of unflavoured Vivonex. A low concentration was used to start, gradually increasing to isotonicity on day 3.

After 28 days if there was little or no improvement (not defined), the diet was abandoned and systemic corticosteroids or TCS used instead. If there was moderate improvement, the elemental diet was extended for 1–2 weeks. If ‘largely unresolved’, open food challenges were commenced. The demographics and clinical features of treatment failures were compared with those for whom treatment was successful, and ‘ no significant differences’ were found.

Children were discharged from hospital once established on three foods, at which point Vivonex was discontinued. After the first week at home, food challenges were continued, and if positive repeated at intervals of 6–12 months. The number of food challenges done and the number of positive tests were reported - data not reproduced here.
Sloper KS;Wadsworth J;Brostoff J;

1991 Aug 196
Study Type: Case series

Evidence Level: 3
To investigate whether food elimination in childhood eczema would improve the condition in at least some patients.Total No. of Patients = 91Children aged 0.42–15 years (median 4.5 years) with AE.
Baseline severity score: median 32 (range 0–80).
76% were breast-fed; cow’s milk had been given to 96%.
Foods exacerbated AE in 56% - mainly egg or cow’s milk (each 28.6%), followed by colourings (13.2%).
Severity (median score change)

−6 (−36 to 10), p<0.001
Source of Funding: Heinz provided tinned foods for challenge test

Comments:
Withdrawal rate 27% (66 provided adequate pre- and post-elimination diet data).

*elimination diet (started when eczema was stable): eggs, cow’s milk, and ‘other foods according to history’. Other foods avoided included: 42% nuts, 36% fish, 26% food colours, 23% tomatoes and wheat, 21% citrus fruit, 18% potato, 17% soya and chocolate.
Dietary advice was given by a dietician.
52% were avoiding at least one food at the start of the study (32% egg, 13% cow’s milk, 10% nuts).

Cow’s milk and egg challenges were undertaken in some - data not reproduced here.
Usual treatment was unchanged.

Severity score reached by dividing body into 20 areas and each assessed for the presence or absence of erythema, vesiculation, excoriation, lichenification (maximum score 80).
Pike MG;Carter CM;Boulton P;Turner MW;Soothill JF;Atherton DJ;

1989 Dec

211
Study Type: Case series

Evidence Level: 3
To investigate the effects of a few foods diet in children with atopic eczemaTotal No. of Patients = 66
Few foods diet*
N = 66
Children aged 0.6–16.8 years (mean 4.2 years) with severe atopic eczema inadequately controlled by standard topical treatment (no further details). 42% had at least one hospital admission for atopic eczema, 91% were woken by itching more than 50% of nights, 53% had previous dietary treatment for their eczema, 52% were already excluding one or more foods at start of the study.‘Worthwhile improvement’ (not defined)

46% parental opinion
35% investigator’s opinion

Reintroduction of foods**

25% deteriorated months 1–3
15% withdrew despite ‘benefit’ (diet too burdensome)
60% persisted (mean 47.9 weeks, range 26.4–71.1)
Source of Funding: none declared

Comments: Median duration of follow-up was 26 weeks (range 19–44).
Some underwent skin prick testing for cow’s milk, egg, dog fur, cat fur, HDM, grass pollen.
*the diet was individually tailored: foods implicated in the exacerbation of eczema (generally or in the child) were excluded. The diet was as strict as the child could tolerate and palatable enough to ensure adherence. Mean number of foods taken: 8.76 (SD 3.76), range 1–19.
Children who responded to the diet, with parental agreement, continued serial reintroduction of individual foods at weekly intervals. Those who successfully completed the food reintroduction underwent DBPCFC (n=10). Those who did not respond either discontinued or proceeded to a second diet, similar in type but with different constituents.

**in children who the investigator thought had improved.

Severity assessed on 20 body areas, each on 0–3 scale for redness, surface damage, lichenification.
Parents recorded itch, redness, and sleep disturbance (0–3).

Adverse effects were not considered.

% improvement - numerical data were only reported for the ‘diet responsive’ group. Characteristics of responders and non-responders were reported (data not reproduced here).
Aoki T;Kojima M;Adachi J;Okano M;

1992 199
Study Type: Case series

Evidence Level: 3
To ascertain the relationship between the effect of egg exclusion and egg allergy.Total No. of Patients = 213

Egg exclusion diet
N = 213
Infants aged under 3 years with infantile or atopic eczema.

Exclusions: purely milk-fed infants, infants already on egg exclusion diet, eating small amounts of egg, infants with severe skin symptoms, and those who needed immediate treatment.
Skin condition ‘better’ (not defined)

48.5% in children aged 3–6 months (n=33)
44% in children aged 7–11 months (n=25)
19.6% in children aged 1 year (n=46)
17.6% in children aged 2 years (n=34)

Results according to positive vs negative test for egg allergy (n=99 only):
70% vs 30%
37.5% vs 62.5%
28.6% vs 71.4%
0 vs 100% respectively (for age groups as listed above)
Source of Funding: none declared

Comments:
RAST test performed for egg white, milk, soybean, wheat, house dust mite (scores of 2 or more regarded as positive).

Condition of skin at follow-up was compared to sketches and photographs taken at the first visit (without knowing the results of RAST). At the first visit the skin symptoms were graded into 3 categories.
Infants considered allergic if either RAST or skin test proved positive.

The study was called a ‘controlled’ trial but there was no control group evident in the paper.

Authors also explore the effects of egg exclusion in those with positive and negative RAST or SPT to egg or other allergens, and between age groups. ‘Correlation’ reported between egg exclusion and allergy in infants aged 3–6 months. Correlation reported between egg exclusion and a positive test to egg allergy for the age group 3–6 months (only).
Withdrawals due to:
non-attendance (11%)
nonadherence (17%)
restricted foods other than egg (36%)
change of symptoms (infections; 16%)
change of treatment (12%)
‘effect no indicated’ (8%)
Businco L;Businco E;Cantani A;Galli E;Infussi R;Benincori N;

1982 Jul 195
Study Type: Case series

Evidence Level: 3
To investigate the efficacy of milk and/or egg free diets in children with severe AE.Total No. of Patients = 59
Cow’s milk elimination (plus/minus egg elimination)
N = 59
Children aged 2–14 years (mean 4 years and 2 weeks) with severe and chronic atopic eczema. They had been referred to the Allergy & Immunology section of a paediatric hospital department after no improvement from usual treatments (antihistamines, TCS and systemic corticosteroids).Global response to treatment

80% ‘cured or improved’
20% unchanged
Source of Funding: None declared

Comments:
*the elimination diet was tailored to the history of suspected allergy. The proportions having either or both foods eliminated was not stated.

Skin tests, IgE levels were undertaken, but children were treated with an eliminiation diet regardless of these test results. Skin tests for cow’s milk proteins were positive in 30, for egg in 5, and for both egg and cow’s milk in 10.

Response to treatment was also examined in terms of the child’s age and age of onset of AE, family history of atopy, duration of breast- feeding, and total and specific IgE - data not reproduced here.
David TJ;

1992 210
Study Type: Case seriesThis study is included only as duplicate/related publication to refs 208 and 209 - there no additional data reported in this paper.
van Asperen PP;Lewis M;Rogers M;Kemp AS;Thompson S;

1983 Sep 204
Study Type: Case series

Evidence Level: 3
To describe the authors’ experience with an elimination diet in children with AE.Total No. of Patients = 29

Elimination diet
N = 29
Children aged 2–12 years with persistent AE despite regular TCS treatment.Parental global assessment

7 improved
3 unchanged
3 deteriorated
Dermatologist’s assessment

5 improved
7 unchanged
1 deteriorated
Source of Funding: none declared

Comments:
*consisting of 19 foods: lamb, chicken, beef, lettuce, carrots, parsley, pears, rice, plain flour, semolina, matzo crackers or Carrs water biscuits, sugar, golden syrup, honey, oils, vinegar, salt and pepper, and coffee.
55% withdrew from the diet, in 28% this was because the diet was too restrictive.

The study design was as follows: 2 weeks of usual diet (baseline), 2 weeks of the elimination diet, then reintroduction of foods at the rate of a new one every 2 days (limited details reported for the latter stage).

Outcomes were assessed using parental diary card, with sleep and itch scores (both using scales of 0–3 none-severe). In addition, a dermatologist assessed severity (inflammation, lichenification, and cracking, all on a grade of 1–2); a change of 2 or more was considered significant.

The authors also reported that there were significant improvements in itch score and in the area of eczema affected, and no significant difference in sleep score or severity (data shown in graphs only).
Martino F;Bruno G;Aprigliano D;Agolini D;Guido F;Giardini O;Businco L;
1998 Nov 214
Study Type: Case series

Evidence Level: 3
To investigate the effectiveness of a home- made meat-based formula and its adequacy as a diagnostic tool for children with food- induced atopic eczema.Total No. of Patients = 16
Home-made meat-based formula (the ‘Rezza-Cardi’ diet)
N = 16
Children aged 5–24 months (mean 9.1) with severe atopic eczema, suspected to be ‘multiple food-induced’. Severity score of more than 15 (maximum score 30) and more than three positive skin prick test responses to food allergens (14 were positive to cow’s milk, egg and wheat, and 2 were positive to cow’s milk, egg, and soya).Severity (median score change)

−21 (no p value reported) Growth
No numerical data; ‘all gained weight normally according to Italian standards’, and the body weight centile increased in 38% children.
Lipids (mean change)

Total cholesterol +1.5mg/dl (0.04 mmol/l)

High density lipoprotein +8.5mg/dl (0.22 mmol/l)

Low density lipoprotein −0.7mg/dl (0.02 mmol/l)
Triglycerides −40.6mg/dl (0.46 mmol/l)

(no change was statistically significant from baseline)
Source of Funding: None declared

Comments:
Severity was measured on a scale of 0–3 for 10 areas.
The diet consisted of lamb meat, olive oil, pre-cooked rice flour, water, and sodium chloride.
Calcium 300mg and vitamin D 400 units were given daily as a supplement.
Fruit and age were also given according to the patient’s age.

Topical betametasone dipropionate was allowed for the first week of the study only.

Adverse effects were not considered.
Broberg A;Engstrom I;Kalimo K;Reimers L;

1992 Sep 206
Study Type: Case series

Evidence Level: 3
To report the authors’ experience of using an elimination diet to treat atopic eczemaTotal No. of Patients = 13

Elimination diet*
N = 13
Children aged 10 months-4 years with severe atopic eczema in spite of ‘adequate’ topical treatment (emollients, hydrocortisone, intermittent triamcinolone, antihistamines, and antibiotics) and elimination of the food items to which the child was suspected to be allergic.Proportion improved

6 based on the investigator’s scores
8 based on parents’ scores
Source of Funding: Grants from two institutions

Comments:
*elimination diet: casein hydrolysate, lamb, rice, corn, corn oil, potato, cucumber, melon, bilberries, salt, sugar, and gluten and milk-free bread.

The children’s usual treatment for atopic eczema was continued during the study.

One child withdrew due inability to keep to the diet.

Not all the children who improved according to the investigator improved according to the parents, however the scoring system used was different. Investigator’s rating: intensity of erythema, lichenification, vesiculation, excoriation, papules, dryness scored on scale of 0–4, none-severe, and distribution measured on scale of 0–4 (maximum total score 96). Parent’s rated eczema and pruritus on a scale of 0–4, and disturbed nighttime sleep on 0–3.
Sporik R;Hill DJ;Hosking CS;
2000 187
Study Type: Case series

Evidence Level: 3
To determine the minimum wheal size that rules in a diagnosis of food allergy.Total No. of Patients = 467Children referred for suspected food allergy. Median age 3 years.Results:
Wheal sizes of 8mm for cow’s milk, 7mm for egg and 8mm for peanut are the minimums required to predict allergy on open food challenge in this high-risk population.
Source of Funding: Not stated

Comments: Data are not independent as some children had more than one SPT result. Open food challenge used as reference for these wheal size data but this is not the gold standard.
Bibliographic DetailsStudy type and evidence levelNumber of patientsPatient characteristicsIntervention and comparisonOutcome measures, follow-up and effect sizeComments
Atherton DJ;Sewell M;Soothill JF;Wells RS;Chilvers CE;

1978 Feb 25 193
Study Type: Randomised Controlled Trial

Evidence Level: 1−
Total number of patients = 36

Egg and cow’s milk elimination diet, with soya-based milk substitute
N = 36

Control: egg and cow’s milk elimination diet, with a preparation containing a mixture of dried egg and cow’s milk as milk substitute
N = 36
Children aged 2–8 years (median 6 years) attending a dermatology clinic with ‘clinically typical’ atopic eczema. Of the 20 who completed the study, 3 had a previous history of exacerbation of skin symptoms after ingestion of eggs or cow’s milk.Egg and cow’s milk elimination diet with soya-based milk substitute vs egg and cow’s milk elimination diet, with a mixture of dried egg and cow’s milk as milk substituteOutcomes at 4 Weeks:

Activity scores (treatment effect*)

2.06, p<0.001 vs

Area scores (treatment effect*)

2.73, p<0.005 vs

Pruritus (treatment effect*)

4.49, p=NS vs

Sleeplessness (treatment effect*)

4.95, p<0.05 vs

Antihistamine usage (not explained further; treatment effect*)

14.15, p<0.025 vs

Activity scores (order effect*)

1.34, p<0.01 vs

Area scores (order effect*)

2.02, p<0.05 vs

Pruritus (order effect*)

4.74, p=NS vs

Sleeplessness (order effect*)

8.83, p<0.01 vs

Antihistamine usage (order effect*)

3.99, p=NS
Source of Funding: Cow & Gate provided milk powders

DB cross-over RCT, withdrawal rate 44% (25% were due to ‘dietary lapses’, defined as drinking less than a pint of milk substitute per day or eating excluded food, i.e. non- adherence).

EL=1− because only completers analysed, and lack of baseline data regarding comparability of intervention and control groups.

Dietary advice given by a dietician. The elimination diet also excluded chicken and beef.

Treatment/control was given for 4 weeks followed by a 4-week ‘washout’ during which the usual diet was resumed. Children were asked to drink at least a pint of the milk substitute per day.

Usual treatment for atopic eczema was continued (daily bath with emulsifying ointment, HC ointment 1%, oral trimeprazine).

Parents recorded daytime itch and sleep disturbance on a scale of 0–3. Two dermatologists scored 20 body areas as affected or unaffected, plus ‘activity’ of eczema (+2 for major improvement, +1 minor improvement, 0 no change, −1 minor deterioration, −2 major deterioration).
Lichenification and ichthyosis alone were ignored.

*treatment effect = mean difference between groups. Order effect = difference between mean scores in the first and second treatment periods using the intervention.

Prick tests were performed at the start of the second treatment period with 10 allergen solutions (including house dust mite, grass pollen, cat fur, egg, milk, control). It was also reported that there was no correlation between positive prick test to egg and cow’s milk antigens and response to diet.
Mabin DC;Sykes AE;David TJ;

1995 Sep 207
Study Type: Randomised Controlled Trial

Evidence Level: 1−
Total number of patients = 85

Few foods diet with whey hydrosylate as milk substitute
N = 27

Few foods diet with casein hydrosylate as milk substitute
N = 32

Control (continued usual diet)
N = 26
Children aged 0-3-13.3 years (median 2.3 years) with AE that persisted despite conventional treatment and involved 12% or more of body surface area.

Exclusions: if breast-fed, had unstable or infected AE, intolerance to casein or whey hydrosylate formulas, received oral corticosteroids within 4 weeks.
Few foods diet with whey hydrosylate as milk substitute
vs
few foods diet with casein hydrosylate as milk substitute
vs
control (continued usual diet)
Outcomes at 6 Weeks:

Body surface area (median change in score, 95% CI)

−4.9 (−12 to −1.5), p=0.49 between groups vs

−5 (−21.2 to −1.6) vs

−4.9 (−12 to −1.5), p=0.49 between groups vs

Skin severity score (median change in score, 95% CI)

−21.8 (−30.2 to −12.8) vs

−13.5 (−38 to −13.4) vs

−15.9 (−22.5 to −5), p=0.88 between groups vs

Sleep disturbance score (median change in score, 95% CI)

−0.4 (−1.4 to 0.3) vs

−0.2 (−0.7 to 0.1) vs

−0.1 (−0.2 to 0.2) vs

Daytime itch score (median change in score, 95% CI)

−0.1 (−1.72 to 0) vs

Daytime itch score (median change in score, 95% CI)

−0.6 (−1 to −0.21) vs

0 (−0.4 to 0.14), p=0.08 between groups
Source of Funding: Two authors supported by Cow and Gate

Single-blind RCT.
The parents and dietician who advised on the diet were blind to the identity of the milk (but not to which diet). A single observer was blind both to whether the child was receiving a diet and to which milk the child was receiving.
A dietician gave advice to parents regarding the few foods diet over a 6-day period. The diet consisted of one meat, rice, potato, one of the brassicas, one fruit, and whey or casein hydrosylate formula milk. Up to three additional foods were allowed if it was judged by the dietician that compliance with the diet would otherwise be poor. Tap water and pure fruit juice (the juice of whichever fruit chosen as a food) were also permitted.
Severity was measured on a scale of 0–3 for each of 32 areas (extent of area affected and degree of erythema). Sleep and itch were also assessed on a 0–3 scale.

Criteria for withdrawal from the study were defined a priori (withdrawal rates were 46% overall, 67% of the whey arm, 53% of the casein arm, and 15% of the control arm. 37% and 25% in the whey and casein arms respectively withdrew due to failure to adhere to the diet, compared to none for this reason in the control arm).
EL=1− because only those who completed the 6- eek treatment period were analysed.

15% in the whey group, 28% in the casein group and 42% of the control group used antihistamines.
Tan BB;Weald D;Strickland I;Friedmann PS;

1996 Jan 6 227
Study Type: Randomised Controlled Trial

Evidence Level: 1−
Total number of patients = 60

House dust mite avoidance
N = 30

Placebo
N = 30
Children and adults aged 7–65 years with AE (defined as atopic on the basis of a positive 15 minute response to a prick-test challenge with a range of aeroallergens). 30 (50%) were aged under 17 years.

Exclusions: pets, house dust mite avoidance measures, or systemic treatment for AE in the previous 6 weeks.
House dust mite reduction vs placeboOutcomes at 6 Months:

Body surface area (mean difference between groups)

10% (3–17), p=0.006 (8.3, 95% CI 2.5 to 19.1, p=0.13 accounting for mattress dust weight and carpet Der p1 concentrations) vs

SASSAD (mean score change)

−12.6 vs

−4.2 (no baseline data) vs

SASSAD (mean difference in score change)

4.2 (95% CI 1.7 to 6.7, p=0.008), accounting for differences in initial eczema scores, mattress dust weight, and bedroom carpet concentrations vs

Reduction in geometric mean dust load in mattress

−98% vs −16%, p=0.002 vs

Median reductions in Der P1 (antigen) concentrations in bedroom carpet

−91% vs

−89%, p=0.94 vs

Median reductions in Der P1 (antigen) concentrations in living room carpet

−76% vs

−38%, p=0.27 vs

Mean difference in final severity scores

4.3 (95% CI 1.3 to 7.3, p=0.006, accounting for differences in initial eczema scores, mattress dust weight, and bedroom carpet concentrations (11.1, 95% CI −3.1 to 25.3, p=0.019 in children aged under 17 years)
Source of Funding: not declared

House dust mite reduction consisted of a Goretex bedding system, benzyltannate complex spray for carpets, and a high-filtration vacuum cleaner.

The placebo group used light cotton bedcovers, water with a trace of alcohol to spray on carpets and a standard upright vaccum cleaner with a poor filtration performance.
A trained nurse applied the bedcovers and spray in all households.

In both groups, treated carpets were vacuumed daily and the rest of the house 2–3 times per week. Soft toys were excluded from bedrooms.
Use of usual range of treatments was permitted.
A physician unaware of treatment group allocation examined all patients monthly.
Dust sampling was not done after 3 months for any beds used by the intervention group because the reduction in dust seen earlier in the trial meant there was insufficient dust to sample.

Withdrawal rates were 3% in the intervention group and 17% in the placebo group - of the total of 12 who withdrew, 10 were due to moving house or acquiring pets or changing carpets; 2 were due to breaking the protocol.

[EL=1−] because only those who completed treatment were analysed.

Analysis of variance to investigate what the treatment effect could be due to was also undertaken - data not reproduced here.
Ricci G;Patrizi A;Specchia F;Menna L;Bottau P;D’Angelo V;Masi M;

2000 Aug226
Study Type: Randomised Controlled Trial

Evidence Level: 1−
Total number of patients = 41

House dust mite avoidance
N = 21

Control
N = 20
Children aged 2–10 years (mean 3.9 years) with AE associated with high total and/or specific IgE serum levels (specific to foods or inhalant allergens). Baseline SCORAD scores 33 in the intervention group and 27 in the control group.House dust mite avoidance vs ControlOutcomes at 2 Months:

SCORAD (mean score change)

−76%, p=0.025 vs baseline vs

−11%, p value vs baseline not stated. No between-group analysis. vs

Change in geometric mean dust load in beds (mg/m2)

−54%, p=0.014 vs baseline vs

−43%, p=NS vs baseline. No between-group analysis. vs

Change in geometric mean concentration of Der p1 and Der f1 in beds (ng/m2)

−76%, p=0.025 vs baseline vs

−58%, p=NS vs baseline. No between-group analysis.
Source of Funding: National Research Council, Italy

[EL=1−] because no baseline data therefore unclear whether groups similar at baseline.

House dust mite avoidance consisted of: encasing mattresses and pillows with mite microfine fibres or Goretex bedding system, a hot weekly wash of bedding, living room and bedroom vacuumed at least twice a week, soft toys washed once a week or excluded from bedrooms, carpets removed or vacuumed once a week or more. No pets allowed. Advice on mite avoidance was given by a person not involved in the later assessment.
The control group continued with previous house cleaning strategies (no specific mite avoidance measures were used).
No dietary restriction was used during the study.
Cant AJ;Bailes JA;Marsden RA;Hewitt D;

1986 Jul 26

202
Study Type: Randomised Controlled Trial

Evidence Level: 1−
Total number of patients = 19

Exclusion diet* plus soya as milk substitute
N = 19

Exclusion diet* plus milk substitute containing cow's milk and egg
N = 19
Exclusively breast-fed infants aged 6 months - 6 years with atopic eczema. All underwent skin prick testing for eggs, cow's milk, chocolate, cod, mixed nuts, and wheat; 8 (42%) tested positive at entry to trial 1 (see comments), and 9 (50%) at entry to trial 2.
Mean activity score 17.3 (SD 9.7), and mean area score 12.4 (SD 5.8).

Exclusions: seborrhoeic dermatitis.
Exclusion diet plus soya milk substitute vs Exclusion diet plus cow's milk and egg milk substituteOutcomes of trial 2

Activity scores at weeks 2, 4, 6: 17.2, 13.2, 14.1 (p<0.001 for difference between weeks 2 and 4).

Area scores at weeks 2, 4, 6: 13.2, 10.7, 10.7 (p<0.01 for difference between weeks 2 and 4) vs

Outcomes at 12 Weeks:

Area score

9.1 on usual diet, p<0.01 vs baseline vs

Activity score

11.8 on usual diet, p<0.01 vs baseline vs

Outcomes at 4 Weeks:

Area score

9.0 vs

8.9 vs

Activity score

10.4 vs

12.6 vs

Outcomes at 8 Weeks:

Activity score

11.2 vs

11.8 (SE for difference between means at week 8: 1.62, p=NS) vs

Area score

8.3 vs

9.9 (SE for difference between means at week 8: 0.98, p=NS)
Source of Funding: Wyeth supplied milk substitutes

Trial 1 (DB RCT):
*excluding cow's milk, egg, chocolate, wheat, nuts. fish, beef, chicken, citrus fruits, colourings, preservatives.
Design: 3x4-week periods; the first two consisting of cross-over randomised treatment with the milk substitutes, the last period consisting of usual diet.
The milk substitutes were supplied as powders for reconstitution with water; the equivalent of a pint a day was consumed. Diets were supervised by a dietician.

Area score: presence or absence of eczema on 20 body areas. Activity: severity on scale of 0–3 for each body area.
Of 17 completers, 12 were exclusively breast-fed. Reasons for withdrawal: mother vomiting on soya substitute (n=1), baby developed eczema and bloody diarrhoea within 24 hours of cow’s milk/egg substitute (n=1).
It was also reported that the quantity of TCS used did not differ significantly between groups (no numerical data).

Trial 2:
Design: open trial for 6 weeks; 2 weeks usual diet (containing cow’s milk and egg), 2 weeks exclusion diet (as in trial 1), 2 weeks usual diet. If activity scores fell by more than 20% during the exclusion diet and increased by more than 20% on reintroducing the usual diet, the mothers were invited to take part in a further randomised crossover phase of 2 other milk substitutes. However only 2 infants qualified for this and only 1 underwent the trial (data not reproduced here).
Neild VS;Marsden RA;Bailes JA;Bland JM;

1986 Jan194
Study Type: Randomised Controlled Trial

Evidence Level: 1−
Total number of patients = 53

Egg and cow’s milk exclusion, soya milk substitute
N = 53

Egg and cow's milk exclusion, dried mixed preparation used as substitute
N = 53
Children and young people aged 1–23 years with AE requiring regular treatment with emollients, TCS, and oral antihistamines.

Positive skin prick tests to egg and/or milk were seen in 26%, and to two or more inhalant allergens in 79%. Overall 59% had an IgE level of more than 100 (units not stated).
Egg and cow’s milk exclusion, soya milk substitute
vs
Control: preparation containing mixture of dried egg and cow's milk as substitute
Outcomes at 6 Weeks:

Area score (treatment difference*)

−1 (95% CI −6 to 3.4) vs

Total (day & night) itch score (treatment difference*)

15 (95% CI −21 to 51) vs

Total TCS consumption (treatment effect*)

fluorinated TCS: 5.8 (95% CI 1 to 10)
HC 1%: 6.0 (95% CI 0.1 to 12)
[i.e. greater use when treated with the trial diet]
Source of Funding: South West Thames Regional Health Authority

DB cross-over RCT; 25% withdrew, due to non-adherence.
EL=1- only completers analysed.

*treatment effect = mean difference between trial and control diets.

Chicken and beef were also excluded from the diet as they may contain proteins common to egg and milk.
Dietician gave the dietary advice.
The two 6-week treatment periods were separated by a 6-week washout where the usual diet was consumed.

Usual treatment for AE was continued.
Parents recorded day and night itch and sleep disturbance on a 10cm scale.
Two dermatologists assessed extent and activity of AE.
At the start of the trial a skin prick test for house dust mite, grass pollen, cat fur, egg or cow's milk was undertaken.
Lever R;MacDonald C;Waugh P;Aitchison T;

1998 Feb198
Study Type: Randomised Controlled Trial

Evidence Level: 1−
Total number of patients = 62

Egg exclusion diet
N = 28

Control
N = 27
Children of mean age 11–17 months (across both groups), with AE and suspected egg sensitivity, optimally controlled with conventional topical treatment and on stable maintenance treatment using mild-moderate TCS at the time of entry into this study.

All had a raised IgE to eggs (RAST test). Results to DBPCFC: positive in 69% in the egg exclusion group vs 67% control; and negative in 13% and 10% respectively (the remaining patients defaulted).
Egg exclusion diet
vs
Control (no specific dietary advice)
Outcomes at 4 Weeks:

Body surface area affected (mean change)

−8.7% vs

−3.2%, p=0.04 (95% CI for mean difference in score between groups 0.1 to 10.9) vs

Severity score (mean change)

−9.4 (SD 12.3) vs −3.3 (SD 10.5), p=0.05 (95% CI for mean difference between groups −0.1 to 12.3)
Source of Funding: none declared

Children were not eating eggs as such, but in hidden forms such as pasta and cakes.
All continued with topical treatment for AE during the study.

Egg exclusion diet = exclusion of all foods containing eggs.
Control = no specific advice on any particular item of food.

Severity score = six clinical features assessed on a scale of 0–3, on 16 body sites (extent, erythema, oedema/papulation, oozing/crusts, dryness, lichenification).

Body surface area was calculated using the rule of nines.

EL=1− because only the 89% who completed the treatment period were analysed.
Majamaa H;Isolauri E;

1997 Feb

223
Study Type: Randomised Controlled Trial

Evidence Level: 1−
Total number of patients = 27

Extensively hydrolysed whey formula + Lactobacillus*
N = 13

Extensively hydrolysed whey formula
N = 14
Children aged 2.5–15.7 months with AE and a history suggestive of cow’s milk allergy, confirmed by DBPC cow’s milk challenge. Duration of exclusive and total breast-feeding were 2.8 months (range 2.1–3.5) and 5.9 months (4.5–7.2) respectively. Baseline median (IQR) SCORAD scores were 26 (17–38) in the intervention group and 21 (14–31) in the control group, p=0.33.Extensively hydrolysed whey formula + Lactobacillus*
vs
Extensively hydrolysed whey formula
Outcomes at 1 Months:

SCORAD (median score change)

−11 (42%), p=0.008 vs baseline vs

−2 (10%), p=0.89 vs baseline (no between-group analysis)
Source of Funding: Academy of Finland, Medical Research Fund

EL=1- because of limited baseline data therefore cannot determine whether groups similar at baseline.

Eczema lesions were treated with emollients and TCS.

*5x10 (−8) colony forming units per gram, added to the whey formula. The quantity varied from 500–1000ml, depending on the age of the child. Otherwise, diet was ‘normal for age’.

Other parameters were also measured (data not reproduced here): faecal concentration of eosinophil cationic protein, alpha- antitrypsin, tumour necrosis factor alpha.
Isolauri E;Sutas Y;Makinen-Kiljunen S;Oja SS;Isosomppi R;Turjanmaa K;

1995 Oct

201
Study Type: Randomised Controlled Trial

Evidence Level: 3
Total number of patients = 45

Cow’s milk substitutes
N = 45
Infants aged 4–8 months (mean age 6 months) who had AE and a positive reaction to a DB challenge with cow’s milk, had not been breast-fed, and had needed a cow milk substitute formula for at least 6 months.
Baseline SCORAD 17 in those receiving the whey substitute and 21 in the amino-acid formula group.
Cow’s milk substitute (hydrolysed whey, n=22; or amino-acid derived formulae, n=23, as desired)Outcomes at 8 Months:

SCORAD (mean score change)

−12 (71%) in those receiving whey and −17 (81%) with the amino-acid formula, p=0.001 vs baseline vs

Weight 'increased similarly' in both groups (no numerical data). The data were shown in graphs; weight increased in both groups in the first month of treatment, and continued to increase in the amino-acid group over the 9-month follow-up period. The pattern in the whey substitute group was less consistent, but weight at 9 months was about the same or worse than at baseline. In terms of statistical significant there is overlap of the 95% CI for the groups for weight. vs

Length

increased in amino-acid formula group but not in whey group, p=0.006 (no numerical data). Data were shown in graphs which show that length increased in both groups in the first month of treatment, and continued to increase in the amino- acid group over the 9-month follow- up period. The pattern in the whey substitute group was less consistent, but length at 9 months was about the same or worse than at baseline. In the graphs there is no overlap of the 95% CI for the groups for length which indicates statistically significant differences between groups.
Source of Funding: Academy of Finland

Additional dietary restrictions were made on the basis of history, skin tests, RASTs, and clinical challenge. These included no eggs and no cereal for 68% of the whey group and 65% of the amino-acid group.

Plasma amino-acid concentrations were compared with corresponding results in healthy age-matched infants who were breast-fed. Fasting morning plasma amino acid concentrations were taken and mean essential and branched amino-acid concentrations quoted. Data not reproduced here.

Energy intake was similar in both groups.

[EL=1−] Although described as randomised in the abstract, randomisation is not described elsewhere in the document. Additionally the interventions were given 'as desired', implying some degree of choice in the milk substitute given, which would nullify randomisation.
Glover MT;Atherton DJ;

1992 Apr

229
Study Type: Randomised Controlled Trial

Evidence Level: 1−
Total number of patients = 26

House dust mite sensitisation
N = 13

Placebo
N = 13
Children aged 5–16 years (mean 10.26 years) with severe AE unresponsive to adequate treatment with emollients, mild TCS, ichthammol paste bandages, systemic antihistamines, and 'appropriate' elimination diets. All had positive skin prick reaction to Dermatophagoides pteronyssinus (a weal at least 4mm in diameter). 78% also had asthma, and 92% allergic rhinitis.Tyrosine adsorbed glycerinated extract of D. pteronyssinus*
vs
Control (tyrosine suspension alone*)
Outcomes at 6 Months:

Severity of erythema (mean score change)

−45% vs

−49%, p=0.643 for difference between scores at endpoint vs

Severity of surface damage (mean score change)

−47% vs

−32%, p=0.907 for score difference at endpoint vs

Severity of lichenification (mean score change)

−43% vs

−48%, p=0.685 for score difference at endpoint vs

Parents assessment

62% better
31% same
8% worse vs

82% better
18% same
0% worse vs

Adverse effects

6 redness at injection site vs

Adverse effects

6 redness at injection site
1 faintness and dizziness 4 hours after injection
Source of Funding: National Eczema Society/Beecham's Pharmaceuticals

*given by subcutaneous injection once a week for 6 weeks (dose increasing to a maximum of 400 Noon units); then given once a month for up to 6 months. Injections were given in hospital and followed by medical supervision for at least 2 hours.

Parameters measured fro severity were assessed using a scale of 0–3.
Scores for erythema and lichenification were 'slightly higher' at the start in patients receiving active treatment.
Skin prick tests were done at baseline and after 12 injections to variety of allergens.
IgE also measured at baseline and after 12 injections.
Bibliographic informationStudy type and evidence levelStudy aims/objectivesNumber of patientsPatient characteristicsOutcomes and resultsComments
Galli E;Chini L;Nardi S;Benincori N;Panei P;Fraioli G;Moschese V;Rossi P;

1994 Jan230
Study Type: Cohort Study

Evidence Level: 2−
To evaluate the efficacy of an oral specific hyposensitisation therapy (to house dust mite) in children with AE and positive skin prick tests and/or RAST to house dust mite.Total No. of Patients = 60

Children with AE and respiratory allergy (asthma or rhinitis), all given oral hyposensitisation therapy; had previously had a 6- week free diet of cow's milk and/or eggs
N = 26

Children with AE only; all given oral hyposensitisation therapy
N = 16

Children with AE exclusively, treated with conventional therapy
N = 18
Children aged 0.5–12 years (mean 4.6 years) with AE and positive skin prick tests to house dust mite solutions and/or positive RAST for anti- house dust mite IgE.Outcomes at 3 Years:

'Dermatitis score' (mean score change)

−8.4 (54%), p=NS between groups
Funding: none declared

Oral hyposensitisation therapy was randomised to the two groups who received this intervention.
EL=2− because baseline characteristics were not given therefore it is not possible to determine whether groups were similar other than in the intervention.

Clinical features of erythema, vesicles, fissuration, lichenification, and itching gives a score of 0–3 (absent- severe).

Oral hyposensitisation therapy contained major (Der p1 and Der p11) and minor antigens of house dust mite. The dose was increased up to a final dosage of 250 'STU' (not defined) administered three times a week. Duration of immunotherapy was: mean 18.7 vs 16.3 months in the AE plus allergy group vs AE only group respectively.

All children were treated with conventional therapy when needed, and used preventive measures to avoid the exposure to house dust mite.

The % improvement was also reported but this was not defined.
Sanda T;Yasue T;Oohashi M;Yasue A;

1992228
Study Type: Cohort Study

Evidence Level: 2−
To investigate the effectiveness and mechanism of action of an air- cleaning system (clean room therapy) in HDM allergen-sensitive patients with atopic eczema.Total No. of Patients = 30

Clean room therapy (patients had HDM allergen-specific IgE RAST score of 3 or higher)
N = 30

Clean room therapy (HDM allergen- specific IgE RAST score of 0)

N = 11

Control (common sickroom; and HDM allergen-specific IgE RAST score of 3 or higher)
N = 10
Children and adults aged 9–75 years (mean early 20s) with atopic eczema (score of 4.5 [scale not specified]) covering at least 18% of body surface area.
All had a score of 0 for mold-specific (Candida) animal dander-specific and pollen-specific IgE RAST scores.
Time to recurrence of symptoms (unspecified; mean [range])

8.4 (2–34) vs

1.7 (1–4) vs

1.6 (1–3) months
Funding: none declared

Patients were hospitalised for treatment. Clean room therapy consisted of an air- cleaning system incorporating a HEPA filter; ventilation exchange of inside for outside air was conducted for about 10 minutes a day. Patients in the clean rooms were not allowed out except to go to the washroom/toilet.
The ordinary 'sick room' used by the control group was identical in design to the clean room but without the air- cleaning system; patients were allowed free movement in and out of the room.
There were two patients to every room.
Use of hydrocortisone butyrate 0.1% and/or beclometasone dipropionate 0.025% was permitted.

The statistical significance of changes in laboratory parameters and HDM particle counts were also reported (no numerical data) - not reproduced here.

Patients were hospitalised for 3–4 weeks.
Devlin J;David TJ;Stanton RH;

1991208
Study Type: Case series

Evidence Level: 3
To describe the treatment and follow-up of children with AE treated at home with a diet eliminating all but six foods.Total No. of Patients = 63
Few foods diet
N = 63
Children aged 0.4–14.8 years (median 2.9 years) with AE, selected for the study either because of extensive (more than 30% skin surface area) skin involvement poorly responsive to conventional therapy or because of a 'clear history' of food intolerance (those with a food intolerance were already avoiding the foods concerned). 73% had a history of intolerance to 1–8 (median 3) foods, usually manifested as urticaria/angio- oedema, or exacerbation of AE.Severity

change in median score: from 60 (20–240) to 40 (4–270), −33%, p<0.001.

39% had 'little or no improvement'.

52% had 20% or greater reduction in disease severity score
Source of Funding: North Western Regional Health Authority

Comments:
The diet consisted of six foods: lamb, potato, rice, rice krispies, carrot and pear. Only water was permitted to drink. If a child had a history of intolerance or dislike of one of the foods, a small number of alternatives were given instead. 20 were given a casein hydrosylate milk formula. All diets were supervised by a dietician. Parents asked to continue with usual treatment (although some changes were permitted if there was marked improvement or deterioration in the skin condition).
If there was a 20% of greater improvement in disease severity score, then foods were reintroduced singly.

Severity score = surface area x erythema score (0–3).

It was also noted that 68% were followed up for '12 months or more.

Regardless of the response to treatment, at one year the final outcome was very similar'.
Ehlers I;Worm M;Sterry W;Zuberbier T;

2001 Aug

215
Study Type:
Case series

Evidence Level: 3
To investigate whether sugar exacerbates atopic eczema.Total No. of Patients = 30
Sugar (sucrose) elimination*
N = 30
Children and adults aged 2–47 years (mean 25 years) with atopic eczema.

Exclusions: those with diabetes and phenylketonuria.
Changed from 31.7 (13, 65) to 29.4 (8, 60) after the 1-week elimination diet, then +3.1 (−9, 15) after sucrose challenge, and −4.4 (−22, 2) after the placebo challenge.Source of Funding: Charite Research Foundation

Comments:
*1-week elimination of sugar, sweets, and avoid regarding alternative sweeteners such as honey, maple syrup, and fruits. Aspartame was offered as a replacement sweetener.

In the DBPCFC, 100g sucrose (40g for children aged under 6 years) was given +200mg aspartame (so that both active and placebo challenges tasted similar). The placebo was 500mg aspartame (+200mg to ensure the same taste).
Foods were added to a dessert.
Mehl A;Verstege A;Staden U;Kulig M;Nocon M;Beyer K;Niggemann B;

2005 Aug191
Study Type: Case series/diagnostic

Evidence level 3
To investigate whether a higher ratio of specific to total IgE would result in higher probability of symptomatic food allergyTotal No. of patients = 501
Children with suspected food allergy N = 501
Children aged 3 months to 16 years (median 13 months) with suspected food allergy (88% had atopic eczema)Proportion of positivie tests results on food challenge: 49% to cow’s milk, 66.5% egg, 35% wheat, 6% soya. Delayed 6% milk, 3% egg, 10% wheat, 8% soya. Both early and delayed reactions: 8% milk, 12% egg, 9% wheat, 5% soya.

Total IgE ranged from 0.3–13.525 ku/l (median 94.3).
Ratio of specific to total 0–91% (median 0.3%) for milk, 69.4% (median 1.7%) for egg, 70.7% (median 0) for wheat, and 15% (median 0) for soya.
Significant correlation was reported for the outcome of food challenges for milk, egg, and wheat but not for soy.

At the 95% predictive probability, for hen’s egg a ratio of specific to total IgE of 19.1% had sensitivity of 10%, NPV 35.7%, and specificity and PPV of 100%.

No predictive probabilities could be calculated for cow’s milk, wheat or soya.
Source of Funding: None declared

Comments: Antihistamines were withdrawn 72 hours before testing.
TCS were allowed twice daily. Testing was only undertaken when eczema was controlled.
Elimination diets were used 1 week before testingh.

Food challenges (n=992) 74% were DBPCFC (placebo = neocate). Open challenges were used for children younger than 12 months who had a clear history of immediate type reactions. Increasing doses of foods were used (fresh milk, soyamillk, wheat powder, and raw hen’s egg).
Test positivie if 1 or more of the following: urticaria, angiodema, wheezing, vomiiting, diarrhoea, shock or exacerbation of eczema.

IgE: to cow’s milk, egg, wheat and soya. The lower detection limit was 0.35ku/l.
Hill DJ; Lynch BC

1982 May 213
Study Type:
Case series

Evidence level: 3
To investigate the effects of an elemental-based diet in children with severe atopic eczema.Total No. of Patients = 10

Elemental diet (Vivonex-Eaton) for 2 weeks, then pumpkin, potatoes, zucchini, apples, pears, and pure vegetable margarine added for weeks 3–6. N = 10
Children (age not stated) with severe generalised atopic eczema which was persistent and unresponsive to topical treatment.Eczema scores fell in the 8 children who completed 6 weeks’ treatment, and were significantly lower than at baseline, p<0.001.
After resuming their usual diet for 6 weeks, scores increased towards baseline, p>0.05 vs baseline. Adverse effects were not considered.
Source of Funding: None declared.

Comments:
2 children withdrew in the first week.
The eczema score was calculated by adding severity (0–3) with extent (0–3, non- involving trunk and flexures), and use of TCS.
Bibliographic informationStudy type and evidence levelAim of studyNumber of patients and patient characteristicsPopulation characteristicsOutcome measuresResults and commentsStudy summary
Niggemann B, Reibel S, Roehr C, et al.; 2001547Cohort
Evidence level = 2
To identify the number of patients with food allergy but without IgE sensitivity.n=139
Age 2 months – 11.2 years, median 13 months
All children had atopic eczema.
Children with atopic eczema referred for food allergy investigation.Positive reaction to DBPCFC208 DBPCFCs undertaken, 111 were positive, of which 59 were early, 25 were late and 27 were combined early and late.

46 early reactions included urticaria
All late and combined reactions were related to atopic eczema

Allergens tested were cow’s milk, egg and wheat.
There were 52 +ve challenges to cow’s milk, 38 to egg and 21 to wheat.
There were 12 +ve tests with allergens that did not display high sIgE.
Retrospective review of consecutive referrals to allergy clinic.
Sampson H. 1983548Cohort
Evidence level = 2
To determine whether immediate reactions to food play a part in the pathogenesis of atopic eczeman=26
age 16 months – 19 years, median 11 years

all children with atopic eczema, serum IgE concentrations > 1000U/ml, history of possible food hypersensitivity, capable of cooperating with challenge procedures.
Diagnosis of food allergy15 children had +ve challenge tests (23/104 tests)
21 tests provoked cutaneous reactions

+ve challenges or convincing history of anaphylaxis were to egg (10), milk (4), peanut (3), wheat (3), soya (2), chicken (2) fish (1), chocolate (1), potato (1), rye (1)
Small study investigating concordance between skin tests and food challenge.
Sampson H, McCaskill C.; 1985549Cohort
Evidence level = 2
As in study 548n=113, age 4 months – 24.5 years, median 6 years

all had atopic dermatitis
Children referred for investigation of severe atopic eczemaFoods inducing reactions in DBPCFC101/370 +ve food challenges in 56% of patients. Main allergens as derived by DBPCFC or convincing history were egg (44 challenges), peanut (20 challenges) and milk (11 challenges). Other allergens were soya, wheat, fish, chicken, pork, beef and potato.
Burks A, James J, Hiegel A, et al.; 1998550Cohort Evidence level = 2To determine if screening for food allergy by skin prick testing can identify food allergy.n=165, age 4 months – 21.9 years, median 49 months)Patients attending allergy clinic with atopic eczemaFood allergy identified by DBPCFC266 DBPCFC tests carried out, 83 were positive plus 12 identified by convincing history of anaphylaxis. All reactions occurred within 2 hours. Main allergens were peanut (27/44 challenges), milk (14/28 challenges). Other foods producing reactions were wheat, soya, cod, catfish, cashew, chicken, kidney bean, tomato, beef, other pulses and shrimp.
Eigenmann P, Calza, A-M.; 2000142Cohort Evidence level = 2To report how food allergy diagnosis is made.n=74, age 6 months – 16.3 years, median 2.5 yearsPatients attending paediatric allergy or dermatology clinics with atopic eczemaFood allergy identified by CAP or DBPCFC6 children underwent DBPCFC of whom 3 were allergic to milk and 2 to soya.Retrospective review of consecutive referrals

NR=not reported

From: Evidence tables

Cover of Atopic Eczema in Children
Atopic Eczema in Children: Management of Atopic Eczema in Children from Birth up to the Age of 12 Years.
NICE Clinical Guidelines, No. 57.
National Collaborating Centre for Women's and Children's Health (UK).
London: RCOG Press; 2007 Dec.
Copyright © 2007, National Collaborating Centre for Women’s and Children’s Health.

No part of this publication may be reproduced, stored or transmitted in any form or by any means, without the prior written permission of the publisher or, in the case of reprographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency in the UK [www.cla.co.uk]. Enquiries concerning reproduction outside the terms stated here should be sent to the publisher at the UK address printed on this page.

The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant laws and regulations and therefore for general use.

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.