The diagnosis of periprosthetic infection of large joints is constantly being improved, and more effective markers are being searched for. We offer new diagnostic approaches based on automated blood analysis.
OBJECTIVE
To study the use of research parameters (WDF channel of the XNL-550 hematology analyzer, Sysmex, Japan) as diagnostic markers of periprosthetic infection of large joints at the preoperative stage.
MATERIAL AND METHODS
Patients with revision surgical interventions were examined: aseptic (group 1, n=83) and for periprosthetic infection of large joints (group 2, n=77), operated on in 2022-2023. Traditional parameters of hematological examination (number of leukocytes, platelets, hemoglobin concentration, relative number of granulocytes, relative number of immature granulocytes (IG, %)) and 25 research parameters of CPD, performed at the preoperative stage of diagnosing periprosthetic infection of large joints, were studied. A comparative analysis of the prognostic value of research and traditional parameters of the hematology analyzer was carried out.The quality of prognostic parameters was evaluated based on the values of the area under the ROC curve (AUC). The level of statistical significance was taken to correspond to p<0.05. Statistical processing was carried out using computer programs MedCalc, Statistica 12.0 (StatSoft) and Microsoft Office Excel 2017.
RESULTS
In an intergroup comparison of traditional hematological study parameters, only the platelet count showed significant differences in the study groups and was lower in patients with aseptic revision surgery. The values in the first and second groups were 253.0·109/l (213.5; 289.50) and 307.41·109/l (237.00; 349.00) (p=0.002), respectively. The optimal criterion for diagnosing periprosthetic infection of large joints was obtained as 281.07·109/l with an area under the characterological curve (AUC) of 0.65 with a sensitivity of 68.8% and specificity of 63.8%. Of the 25 research parameters, NEUT# and LY-WY showed significant intergroup differences. In patients of the first group, the values of NEUT# and LY-WY were significantly lower compared to patients of the second group: 4.1±0.15, 4.78±0.2 (p=0.023) and 896±9.5, 937.06±11.01 (p=0.016). The optimal criterion values for diagnosing of periprosthetic infection of large joints were calculated: NEUT# — 4.79 (4.45-5.13) and LY-WY — 937.06 (918.73-955.4). Areas under the characterological curve (AUC) for NEUT# and LY-WY — 0.62 (0.54-0.695) and 0.648 (0.568-0.721), respectively. The sensitivity and specificity of NEUT# were 71.4% and 53.01%, and LY-WY — 77.9% and 51.81%, respectively. When NEUT# and LY-WY were isolated into diagnostic model (DM) with regression equation DM=0.244·NEUT#+0.005·LY-WY–5.307, an increase in AUC to 0.659 (0.565-0.685) was observed, as well as an increase in the specificity of the marker for diagnosing periprosthetic infection large joints up to 84.7% with a sensitivity of 53.2%.
CONCLUSIONS
The predictive value of DM with the regression equation DM=0.244·NEUT#+0.005·LY-WY–5.307 is comparable to the predictive value of the platelet count in the diagnosis of periprosthetic infection of large joints, which shows that the resulting model can be used to predict the absence of periprosthetic infection in patients with revision surgical interventions, as it has a high specificity with an average predictive value.
