Authors' reply to acid steatocrit.

Abstract

To the Editor: We are very grateful to Dr. G. M. Addison for his interest and criticism of the acid steatocrit (1) (Letter to the Editor, J Pediatr Gastroenterol Nutr 1996;22:227). Dr. Addison suggests the correlation we found between acid steatocrit but not classical steatocrit results and chemically measured fecal fat concentration to be due to the misuse of the Pearson correlation coefficient, which, due to the apparent nonnormal distribution of the data, should not have been used in our study. Although we quite agree with this criticism, the relationship was also evaluated by the nonparametric Kendall's and Spearman's correlation tests. All three correlation tests pointed in the same direction: significant correlations were found between chemically measured fecal fat concentrations and acid steatocrit results, while no correlations were found with classical steatocrit results. Furthermore the data shown in our paper concerned a very unhomogeneous patient group, which can partly explain the distribution of the data. We recently studied the same relationship in a homogeneous group of 50 children with cystic fibrosis. Again all three correlation coefficients between chemically measured fecal fat concentrations and acid steatocrit were above 0.8 (p < 0.0001). We do agree with Addison that the confidence intervals for the prediction of fecal fat from the acid steatocrit are too large and that is why we do not advise the use of the acid steatocrit for this purpose. Furthermore, our published data were used by Addison in order to try to categorize our patients into diseased and nondiseased groups. This should, in our opinion, only be done on wellknown and well-described patient groups, which was not the case for our published data. We have performed such a study recently and have shown that when compared to 3-day fecal fat excretion results, the acid steatocrit results correctly classified most (>90%) of the 40 evaluated children (see abstracts of the ESPGAN meeting 1995). Addison also suggests that, due to fecal inhomogeneity, results from small samples might not be representative of samples taken from large daily homogenized stool collections, which would obviate the potential of the acid steatocrit. This problem has been studied in our laboratory. We compared steatocrit results from 60 small samples taken from unhomogenized versus homogenized daily stool collections. Results showed not only the mean values to be very similar (11.6% for small unhomogenized samples versus 14.1% for samples from homogenized 1-day collections) but the standard deviations of results were also very similar (11.8% for unhomogenized versus 14.7% for homogenized samples), which indicates that the variability of the results is not significantly influenced by the homogenization procedure. The latter findings indicate that the potential of the acid steatocrit is not decreased by the use of small fecal samples. Lastly, Addison advises against the use of the acid steatocrit for diagnostic purposes. For 3 years we have been using this test for the detection and monitoring of pancreatic insufficiency in children. We know, from more than 500 determinations in these patients, that acid steatocrit results are higher than 20% in all cystic fibrosis patients with pancreatic insufficiency. Further, in our clinical experience, we have never seen acid steatocrit results higher than 20% in sick controls without fat malabsorption. In spite of Addison's criticism, we are still convinced that the acid steatocrit is a “much improved method” and is useful for detecting and monitoring pancreatic insufficiency in children. M. Tran; P. Forget; A. Van den Neucker Department of Pediatrics; Division of Pediatric Gastroenterology; University of Maastricht; Maastricht, The Netherlands