PURPOSE OF THE STUDY: The primary aim of the study was to identify characteristics predicting the blood loss associated with primary total hip (THA) and knee (TKA) arthroplasty surgery. The other objective was to find out which characteristics were important for peri-operative allogeneic blood transfusion in the same group of patients. MATERIAL AND METHODS: This prospective study comprised 210 consecutive patients who underwent primary THA (n = 115) or primary TKA (n = 95) at our department. In each patient, 21 pre-operative and peri-operative characteristics were recorded. Of them, the following characteristics were selected for statistical evaluation: age, gender, BMI, primary diagnosis, Charlson co-morbidity score, type of prophylaxis for deep-vein thrombosis, type of implant fixation (in THA), pre-operative INR value, haematocrit, haemoglobin (Hb) and platelet levels, amount of autologous blood donated by the patient, ASA score, operative time, use of tourniquet (in TKA), type of anaesthesia used, blood recuperation and patient's smoking habits. Multivariate analysis was used as the statistical method. For hypothesis testing, a statistical significance level of 0.05 was stated and, for enclosing (removing) characteristics to (from) multivariate models, the significance level was set at 0.11. RESULTS: The group included 81 men and 129 women; the mean age at the time of surgery was 65.5 ± 11.97 years (mean±SD) in the THA patients and 68.5 ± 8.52 years in the TKA patients. Primary osteoarthritis was the most frequent surgical diagnosis (THA, 64.35%; TKA, 82.1%). The mean amount of blood loss was 1258 ± 402.6 ml in THA and 1580 ± 475.5 ml in TKA. The mean amount of allogeneic blood required was 130 ± 202 ml when all THA patients were considered, and 371.95 ± 159.3 ml when only those who received it were involved. For the TKA patients, the corresponding values were 160.1 ± 278.8 ml for all patients and 507 ± 264.5 ml for blood recipients only. The characteristics that affected the amount of blood loss in THA included BMI, ASA score, blood recuperation, type of anaesthesia, and smoking habits; in TKA these were BMI, pre-operative platelet count, INR and type of anaesthesia. High pre-operative Hb levels made the probability of allogeneic blood requirement lower in both THA and TKA. Autotransfusion decreased the probability of allogeneic blood requirements only in THA. DISCUSSION That the pre-operative Hb level is the strongest predictor for the probability of allogeneic blood transfusion during both THA and TKA is a logical and well-known fact. What remains to be established is the optimal protocol for pre-operative preparation of the patients with low Hb levels undergoing elective replacement (hip and knee) surgery. This study clearly showed that, in THA patients, pre-operative autologous blood donation decreased the probability of allogeneic blood transfusion. Other results of our multivariate analyses were not clinically unambiguous and therefore further research on a larger patient group is warranted. Such studies will also require the development of a more exact method for the assessment of blood loss at the operating theatre. CONCLUSION: The patients with low pre-operative Hb levels have a high probability that they will require allogeneic blood transfusion during primary THA and TKA. Autologous blood donation can decrease this probability significantly (here proved only for THA patients). The multivariate model of blood loss published here could assist in estimation of peri-operative blood loss and potential risk of blood transfusion requirements.

Ortopedická klinika FN a Univerzita Palackého v Olomouci Czech Republic

Univerzita Palackého v Olomouci Přírodovědecká fakulta Katedra geoinformatiky

Univerzita Palackého v Olomouci Přírodovědecká fakulta Katedra matematické analýzy a aplikací matematiky

Multivariate analysis of blood loss during primary total hip or knee arthroplasty