Re-examination of the diagnostic criteria of tropical pulmonary eosinophilia

There is no agreement on the minimum absolute eosinophil count essential fo r the diagnosis of tropical pulmonary eosinophilia (TPE) at present. The ai m of this study was to determine this figure as well as to evaluate the oth er diagnostic criteria of TPE. The response to diethylcarbamazine (DEC) was tested in 98 patients [of whom 79 (80.6%) completed the study] by means of clinical scores, lung function tests and the absolute eosinophil counts. The minimum absolute eosinophil count necessary for the diagnosis of TPE wa s found to be 3300 for two reasons. Firstly there was a marked fluctuation in the mean percentage change of the absolute eosinophil count after treatm ent with DEC, when it was below 3225 cells mm(-3), while the mean percentag e reduction showed a remarkable stability when the eosinophil count exceede d 3600. Secondly there was a marked difference in the response to DEC in pa tients whose eosinophil counts were above and below these values. All patients who had eosinophil counts greater than 3600 responded to DEC a nd were diagnosed as cases of TPE. All of them were from filarial endemic a reas. The total eosinophil count decreased by a mean of 92.5%, 3 months aft er administration of DEC. The sensitivities of the following tests in TPE w ere as follows: filarial antibody test (FAT) 30%, radiological changes 45.5 %, erythrocyte sedimentation rate (ESR) 80%. The radiological changes and t he ESR, but not the FAT, were helpful in differentiating TPE from those pat ients with TPE-like symptoms but with lower eosinophil counts, e.g. those w ith asthma. Patients with cough who had eosinophil counts of between 53 and 2000 cells mm(-3), showed elevated filarial antibody levels in a significant number of cases when compared to asymptomatic subjects. (P < 0.001). Five of them re sponded to DEC. Three of these had filarial antibody in their serum and one had bilateral mottling on chest Xray. These results suggest that atypical cases of TPE may exist. Our study has shown that the diagnosis of TPE rests on the following criter ia: cough worse at night; residence in a filarial endemic area; the eosinop hil count greater than 3300 cells mm(-3), clinical and haematological respo nse to DEC. The diagnosis is supported by radiological changes and elevated ESR. The FAT is of little value. The clinical benefit and the improvement in lung function which follows the administration of DEC was sustained up to a minimum period of 15 months. ( C) 1999 HARCOURT PUBLISHERS LTD.