Vandenplas 1992.

Methods	Single‐centre randomised controlled trial in Belgium.
Participants	Inclusion criteria: infants with at least 2 first‐degree relatives with allergic disease, whose mothers intended not to breast feed.
Interventions	Exclusive formula feeding for 6 months with: Treatment (n = 32): pHWF (Nan HA, Nestle). Control (n = 35): CMF (Nan, Nestle). Co‐interventions (both groups): grated apple from 4 months. 'Normal' diet after 6 months.
Outcomes	Primary outcome(s): atopic disease. Other outcomes: blinded physician assessment for allergic disease monthly for first year of life. Total IgE, specific RAST, IgG4 antibodies, SPTs. Definitions Atopic dermatitis: at least 3 of 4 criteria including typical rash, recurrence or chronicity and specific IgE. Urticaria: no definition given. Allergic wheezing: cough without infection ≥ 24 hours. Chronic rhinitis: clear nasal discharge. CMPI: confirmed by unblinded elimination/challenge. Allergic diarrhoea: infection excluded. Infants with diarrhoea had jejunal biopsy performed. Follow‐up to 12 months (infant allergic disease incidence).
Notes	Trial of prolonged sole pHWF vs CMF in high‐risk infants 3‐ and 5‐year results excluded owing to excess losses. Data for cumulative specific allergic disease manifestations up to 12 months not extractable separately. Data for CMA and sensitisation not able to be extracted separately. Data included in previous version of review now considered to include infants with CMPI and infants without symptoms and with sensitisation. Conflict of interest: Nestle provided formula and performed statistical analysis.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Infants postnatally allocated; method not reported.
Allocation concealment (selection bias)	Unclear risk	Allocation process not reported.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Infants were given 1 of 2 coded formulas... or an adapted formula with native cow's milk proteins, delivered in an unlabelled package.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "Follow‐up was blinded as much as possible: neither the parents nor the physician(s) involved in the follow‐up were informed about the nature of the formula".
Incomplete outcome data (attrition bias) All outcomes	High risk	12 months: 8/75 (11%) post‐randomisation losses. At 3 and 5 years: 17/75 (23%) lost to follow‐up.
Selective reporting (reporting bias)	Unclear risk	Reported prespecified definitions of specific allergic disease but not time of reporting.
Other bias	Low risk	Groups similar at baseline for risk factors for allergic disease.