Accuracy of low electronic platelet counts using platelet distribution curves.

At low platelet levels electronic counts are inaccurate because background counts are an increasingly important contribution to the total count. In this study the background contribution was selectively eliminated by coupling a multichannel particle size analyzer (Coulter Channelyzer II) to an electronic counter (Coulter ZBI). Platelet counts below 50 x 10(9)/l required an adjustment of the lower threshold in 11 of 15 specimens. Non-platelet counts averaged 18 percent of the platelet particle count with a maximum value of 50 percent. In contrast, platelet counts between 50 to 100 x 10(9)/l required only a minimum adjustment of the lower threshold and extraneous counts averaged only 2.7 percent of the platelet particle count. Phase platelet counts ranging from 4.0 x 10(9)/l to 115 x 10(9)/l were compared in ZBI-Channelyzer counts in 30 patients. The electronic counts showed good agreement with the reference method. Results were statistically analyzed using linear regression analysis and paired t tests. Technical errors involved in electronic and visual platelet counts are discussed.