Abstract
BACKGROUND: This study aimed to compare the prothrombin time (PT), international normalised ratio (INR) and activated partial thromboplastin time (aPTT) values obtained using the photo-optical method and to assess these values according to the reference method, which was the mechanical method. METHODS: Plasma samples from 340 patients, submitted to our hospital's biochemistry laboratory for PT, INR, and aPTT analyses, were assayed using the mechanical coagulometric measurement method in a Stago Compact Max3 automated coagulation analyser, which served as the reference device. The same samples were also analysed using the Sunbio UP5500 automated analyser with a simultaneous optical method. There were 30 turbid samples analysed in both devices without exclusion from the study. Correlation coefficient analysis was carried out using SPSS to assess intervariable correlations. Passing-Bablok regression analysis was performed in R software version 3.6.0 to compare PT, INR, and aPTT values between the two devices. Bland-Altman plots were used to analyse the agreement. RESULTS: A good level of statistically significant agreement was found between the PT and INR values measured by the Stago Compact Max3 and Sunbio UP 5500 devices (Interclass Coefficient Correlation (ICC): 0.627, p=0.001; p<0.01 and ICC: 0.653, p=0.001; p<0.01, respectively). Additionally, there was an excellent level of statistically significant agreement for the aPTT values (ICC: 0.902, p=0.001, p<0.01). The Bland-Altman analysis revealed the mean 95% limits of agreement values as 2.46 (lower limit: -2.44, upper limit: 7.37) for PT, 0.07 (lower limit: -0.32, upper limit: 0.46) for INR, and 2.45 (lower limit: -1.67, upper limit: 6.58) for aPTT. The Passing-Bablok regression results indicated a systematic difference for PT measurement but no proportional difference. No systematic or proportional differences were found for the measured INR and aPTT values between the Stago Compact Max3 and Sunbio UP 5500 devices. The intra-assay and interassay coefficient of variation (CV) values from level 1 and 2 controls of the optical method were below 5%. CONCLUSIONS: The results from the optical method were consistent and reliable compared to the mechanical method. PT and INR results showed statistically good agreement, while aPTT results demonstrated excellent agreement. Larger multicenter studies are needed to evaluate turbid samples.