Background: A multicentre study was carried out in Ghana and southern India to determine the aetiology of suppurative keratitis in two regions located at similar tropical latitudes. Studies of fungal keratitis from the literature were reviewed. Methods: Patients presenting at rural and urban eye units with suspected microbial keratitis were recruited to the study. Corneal ulceration was defined as loss of corneal epithelium with clinical evidence of infection with or without hypopyon. Microscopy and culture were performed on all corneal specimens obtained. Results: 1090 patients were recruited with suspected microbial keratitis between June 1999 and May 2001. Overall the principal causative micro-organisms in both regions were filamentous fungi (42%): Fusarium species and Aspergillus species were the commonest fungal isolates. Pseudomonas species were most frequently isolated from cases of bacterial keratitis in Ghana but in India the commonest bacterial isolates were streptococci. Conclusion: Infections of the cornea due to filamentous fungi are a frequent cause of corneal damage in developing countries in the tropics and are difficult to treat. Microscopy is an essential tool in the diagnosis of these infections. A knowledge of the "local" aetiology within a region is of value in the management of suppurative keratitis in the event that microscopy cannot be performed.

In total, 122 patients (61.3%) had their treatment either determined or altered based on the results of the microbiological diagnosis; in 87 of these solely on the basis of direct microscopic examination. Conclusions-Infection by filamentous fungi accounted for more than half of the ulcers from which cultures were obtained. Both training in technique and experience in interpretation are necessary for microscopy based diagnosis by staff in the clinic to be of greatest value. Direct microscopy was particularly useful for detecting fungi.

Using World Health Organization definitions of visual loss and a standardised methodology, 905 children were examined in Chile, West Africa and South India. Of these 806 (89%) suffered from blindness (BL) or severe visual impairment (SVI). Causes of SVI and BL were classified anatomically and aetiologically, and avoidable causes identified. In W. Africa (n = 284) the major anatomical cause of SVI/BL was corneal scar/phthisis bulbi (35.9%). Retinal disease accounted for 20.4%, cataract 15.5% and glaucoma 13.0%. Aetiologically 33.8% of SVI/BL was due to childhood factors and 21.1% to hereditary disease. In S. India (n = 305) the major anatomical cause of SVI/BL was corneal scar/phthisis bulbi (38.4%). Retinal disease accounted for 22.6%, cataract 7.4% and glaucoma 3%. Aetiologically 37.0% of SVI/BL was due to childhood factors and 29.8% to hereditary disease. In Chile (n = 217) the major anatomical cause of SVI/BL was retinal disease (47.0%). Cataract accounted for 9.2%, glaucoma 8.3% and 6.9% was due to corneal pathology. Aetiologically 30.4% of SVI/BL was due to hereditary factors, and 20.8% to perinatal factors of which four-fifths (16.6%) was due to retinopathy of prematurity. Avoidable conditions accounted for 70%, 47% and 54% of cases in W. Africa, S. India and Chile respectively.

An attempt was made to assess the true public-health importance of onchocercal skin disease throughout the African region and hence provide an objective basis for the rational planning of onchocerciasis control in the area. The seven collaborative centres that participated in the study (three in Nigeria and one each in Ghana, Cameroon, Tanzania and Uganda) were all in areas of rainforest or savannah-forest mosaic where onchocercal blindness is not common. A cross-sectional dermatological survey was undertaken at each site following a standard protocol. At each site, the aim was to examine at least 750 individuals aged 5 years and living in highly endemic communities and 220-250 individuals aged 5 years and living in a hypo-endemic (control) community. Overall, there were 5459 and 1451 subjects from hyper-and hypo-endemic communities, respectively. In the highly endemic communities, the prevalence of itching increased with age until 20 years and then plateaued, affecting 42% of the population aged 20 years. There was a strong correlation between the prevalence of itching and the level of endemicity (as measured by the prevalence of nodules; r=0.75; P<0.001). The results of a multivariate logistic regression analysis showed that, at the individual level, the presence of onchocercal reactive skin lesions (acute papular onchodermatitis, chronic papular onchodermatitis and/or lichenified onchodermatitis) was the most important risk factor for pruritus, with an odds ratio (OR) of 18.3 and 95% confidence interval (CI) of 15.19-22.04, followed by the presence of palpable onchocercal nodules (OR=4.63; CI=4.05-5.29). In contrast, non-onchocercal skin disease contributed very little to pruritus in the study communities (OR=1.29; CI=1.1-1.51). Onchocercal skin lesions affected 28% of the population in the endemic villages. The commonest type was chronic papular onchodermatitis (13%), followed by depigmentation (10%) and acute papular onchodermatitis (7%). The highest correlation with endemicity was seen for the prevalence of any onchocercal skin lesion and/or pruritus combined (r=0.8; P<0.001). Cutaneous onchocerciasis was found to be a common problem in many endemic areas in Africa which do not have high levels of onchocercal blindness. These findings, together with recent observations that onchocercal skin disease can have major, adverse, psycho-social and socio-economic effects, justify the inclusion of regions with onchocercal skin disease in control programmes based on ivermectin distribution. On the basis of these findings, the World Health Organization launched a control programme for onchocerciasis, the African Programme for Onchocerciasis Control (APOC), that covers 17 endemic countries in Africa.