Comparison of automated total nucleated cell counts vs manual cell counts in the amniotic fluid for the classification of disease status in healthy and diseased pigs.

Background: An increase in the total nucleated cell count (TNCC) of amniotic fluid can detect inflammation in the amniotic cavity. Traditionally, a manual count is employed, but the technique suffers from inaccuracy and is labor intensive. Automated cell counting could enhance precision and timeliness.

Objectives: We aimed to analyze the cell counts in the amniotic fluid of preterm pigs using an automated method and compared the results with manual cell counts. We also tested if increased TNCC is associated with an inflammatory group in a porcine model of chorioamnionitis.

Methods: Amniotic fluids from 34 preterm pigs were analyzed blindly using two automatic settings of the ADVIA 2120i hematology analyzer (whole blood [WB] and cell poor [CP] settings) and manually by two observers using Neubauer chambers. Inter- and intra-variability were calculated. The correlation analysis of TNCC/μL in amniotic fluid was determined between the methods, including Deming and Bland-Altman analyses. Fischer's exact tests compared the known health status of the pigs to the outcomes of the automatic and manual TNCC.

Results: Inter- and intra-observer variability of the manual TNCC were high, although the correlation of TNCC between (r = .95, P < .0001) and within observers (r = .98, P < .0001; r = .89, P < .0001) was good. Correlation between the manual and CP TNCC was moderate and significant (r = .50, P = .014). Manual and WB TNCC were not correlated. The CP (P = .003) and manual (P = .0001) analyses accurately classified the disease state of the pigs.

Conclusions: An acceptable correlation between automatic CP TNCC and manual counting was demonstrated. Both methods could accurately classify the disease state of the pigs.