The inability to process haemolysed and underfilled specimens has been a perennial problem in routine coagulation testing causing delays, inconvenience and increased costs. While there has always been a degree of ambiguity surrounding the true impact of these pre-analytical variables on test results, the general consensus is that they should not be tested. The controversy has been amplified in recent years with studies suggesting that some test results from haemolysed and underfilled specimens may in fact be reportable. However to date, these findings have not been incorporated into best practice guidelines.
This study evaluated the effect of both haemolysis and inadequate volume on routine coagulation testing to provide laboratories with more scope to reduce the number of recollected specimens than the guidelines currently allow.
From November 2016 to March 2017 at Wellington Hospital New Zealand, the Prothrombin Time (PT), Activated Partial Prothrombin Time (APTT) and Fibrinogen results of the original rejected haemolysed / underfilled specimens were compared with those of the recollected specimens. Specimens were rejected if they were >10% underfilled or contained >0.40 g/L free Haemoglobin (Hb). All testing was performed using photo-optical measurements on a Sysmex® CS-2100i coagulation analyser. Statistical analysis was performed using regression analysis and 2-tailed paired t tests.
A total of 83 consecutive paired patient specimens included in the study were either underfilled (n = 41) or haemolysed (n = 42). In volume studies, we showed that accurate results are still achieved for PT if a specimen is >73% filled (R2 = 0.9879, y = 0.9964x; P < 0.05), APTT if a specimen is >79% filled (R2 = 0.9665, y = 0.9752x; P < 0.05) and fibrinogen if the specimen is >59% filled (R2 = 0.9587, y = 0.977x; P < 0.05). It should be noted however that our data was limited below a fill volume of 80%. In haemolysis studies, there was no statistically significant difference in PT results (R2 = 0.9879, y = 0.9637x; P < 0.05) or fibrinogen results (R2 = 0.9607, y = 1.0162x; P < 0.05) at any level of haemolysis. APTTs showed a statistically significant difference overall (R2 = 0.87, y = 0.9684x; P < 0.05), however the only clinically meaningful differences were observed in three paired samples with >1 g/L free Hb. Incorporation of these findings into our processes at Wellington Hospital reduced the number of recollected specimens by approximately 90%.
This study has validated a minimal fill volume of 80% and a maximum haemolysis level of 1 g/L free Hb for PT, APTT and Fibrinogen. Evidence-based acceptance criteria was developed using these findings resulting in significant cost savings, improved turnaround times and increased satisfaction across clinical and allied health teams. This review offers a clinically appropriate solution to a problem which has always hindered routine coagulation testing.