Determinations of blood glucose levels by an enzyme strip test and the AutoAnalyzer were compared in 542 samples. The enzyme strip test tended to overestimate glucose values in the lower ranges and underestimate them in the upper ranges. Although agreement between the methods was reasonably good in the 60 to 130, and 130 to 200 mg. per 100 ml. ranges, underestimation of values by the strip test occurred in 10 per cent of the samples at the latter range. A study of the sources of variability of the enzyme strip test revealed that the color-matching step was most likely the source of most of the technical error. The latter was high for the test, the coefficient of variation of blind independent determinations on ten samples by four technicians being 27.5 per cent. The clinical usefulness of the test is limited, and it should not be employed if other methods of determining glucose levels are easily available.

This content is only available via PDF.