This prospective study investigated patients with a clinical diagnosis of bullous pemphigoid (BP) who presented to a tertiary dermatology referral centre in Singapore. All patients had blood samples and skin biopsies taken for histology, immunofluorescence (IF) and immunoblot analysis prior to initiation of treatment. We analysed 23 new cases of BP during the 1-year study period. Seventeen of 22 biopsy specimens showed subepidermal blister formation, and 12 of the 17 (71%) had a predominance of eosinophils (>50%) in the blister cavity. The dermal inflammatory infiltrate of 22 biopsy specimens was predominantly lymphocytic in nine (41%) and eosinophilic in eight (36%). The histological picture was highly suggestive of BP in 15 of 22 patients (68%), suggestive in two (9%) and poorly suggestive in five (23%). Twenty-one of 23 (91%) patients had linear deposits of IgG and C3 along the dermo-epidermal junction. Serum indirect IF was positive in 22 of 23 (96%) patients, all showing antibody binding to the roof of the induced blister on salt-split skin. All of the 23 serum samples demonstrated positive immunoblot reactivity to BP180 and/or BP230 from epidermal extracts of normal human skin. Immunoblot reactivity with BP180 and BP230 was 78% (n=18) and 52% (n=12), respectively. The BP180 NC16A antibody could be detected in 22 of 23 (96%) sera using the enzyme-linked immunosorbent assay (ELISA) technique. The sensitivity of traditional diagnostic techniques, i.e. direct IF (91%) and indirect IF (96%), was comparable with that of the newer techniques, i.e. immunoblot analysis (100%) and ELISA (96%). ELISA in combination with routine indirect IF may be a useful diagnostic tool in patients with suspected BP who refuse a skin biopsy but consent to give a serum sample.