Tourniquet use in routine primary total knee arthroplasty is associated with a higher transfusion rate and longer postoperative length of stay: a real-world study

A prospective multicenter study, which was sponsored by the Ministry of Health of the People’s Republic of China (project number: 201302007) and enrolled in 26 university teaching hospitals in China (included 10 national and 16 regional hospitals), was carried out to assess the efficacy and safety of perioperative management measure and improve the development enhance recovery after surgery of TKA and total hip arthroplasty in China from 2013 to 2016. The large database generated from the study was analyzed secondarily for evaluating the effect of tourniquet use in TKA on transfusion rate and PLOS. Whether to use a tourniquet during surgical procedure in TKA was determined by the leader of each medical team. The transfusion criteria of allogeneic blood for each patient followed the specification for clinical transfusion technology issued by Ministry of Health of the People’s Republic of China in June 2000. Those were (1) hemoglobin level < 70 g/L and (2) according to the anemia degree of the patient, cardiopulmonary compensatory function, whether or not metabolic rate increased and age and other factors when hemoglobin level between 70 and 100 g/L. In addition, the PLOS was defined as the date of discharge minus the date of surgery. The database analyzed in this study included patient-level hospital discharge data provided by enrolled 26 hospitals. The accuracy and completeness of the data were inspected and verified carefully by comparison with information from hospital electronic record. This study was deemed exempt by the hospital’s institutional review board. And all patients gave their informed consent.

The Clinical Modification procedure codes of the International Classification of Diseases, Tenth Revision (IDC - 10) was used to identified the patients who underwent TKA from the large database (n = 7789). Patients who had undergone revision hip and/or knee arthroplasty or bilateral TKA, with metastatic and/or bone cancer, with lower-extremity fractures, younger than 18 years, with bleeding diathesis, and received an autologous blood transfusion rather than an allogeneic blood product were excluded according to their electronic medical records (n = 1464). Finally, 6325 patients were included in this study, including 4902 (77.50%) patients for whom a tourniquet was used during the TKA procedure and 1423 (22.50%) patients who did not have a tourniquet applied. Tourniquet was used before skin incision for 4531 patients, while before placing the prosthesis for the remaining 371 patients in tourniquet use group. And the pressure of tourniquet was determined by the leader of each medical team. All the operations were performed under the supervision of the medical team leaders, chief surgeons or associate chief surgeons come from Grade-A Tertiary Hospital. The demographic characteristics, which including age, sex, body mass index (BMI), and diagnoses, comorbidities which including hypertension, diabetes, coronary heart disease, and chronic obstructive pulmonary diseases, preoperative drug use, results of preoperative laboratory examination, operative variables, such as anesthesia methods, American Society of Anesthesiologists (ASA) class, TXA use applicated intravenously or intravenously combined with locally, drain use, intraoperative blood loss and operative time of the included patients who underwent primary unilateral TKA were enrolled and analyzed, and all subjects also were shown in Table 1. The total dose of TXA was 1-3 g for almost all of the included patients who received TXA. The drainage tubes were removed on the first day postoperatively. The first discharge diagnoses of the included patients were classified into two types: (1) degenerative arthritis (osteoarthritis) and (2) inflammatory arthritis, which consisted of rheumatoid arthritis and ankylosing spondylitis. The ASA classification of each patient evaluated by anesthetist before surgery was divided into two class: < 3 or class ≥3. In addition, the anesthesia methods were divided into two types: general or regional anesthesia.

Based on comparison with data from the hospital information system, the completeness of reporting was nearly 94% for unilateral primary TKA. The portion of missing or incorrect information of enrolled variables was generally less than 0.5%. Spearman’s rank correlation analyses showed that the absolute value of rank correlation coefficient between covariates enrolled in logistic and linear regression analyses ranged from 0.002 to 0.278, which indicated that correlations between the included variables were low and met the requirements of logistic and linear regression analyses Fig. 1.

The total transfusion rate was 12.71% for all enrolled patients who underwent unilateral primary TKA, while 14.57% for the tourniquet group and 6.47% for the no-tourniquet group. The average PLOS was 7.41 ± 3.57 days for all patients after TKA, while 7.72 ± 3.54 days for the tourniquet group and 6.44 ± 3.48 days for the no-tourniquet group (Table 2 and Fig. 2).

As shown in Table 3, several statistical models including different covariates were built to evaluate the influence of tourniquet use in TKA on transfusion rate and PLOS. First, univariate analyses (crude model) showed that tourniquet use in TKA was associated with a higher transfusion rate (RR = 2.458, 95%CI 1.962–3.081, P < 0.001) and longer PLOS (B = 1.280, 95%CI 1.066–1.494, P < 0.001). After adjusting for age, sex, BMI, and orthopedic diagnoses using logistic and linear regression analyses, our results showed that tourniquet application remained correlated with a higher transfusion rate (RR = 2.604, 95%CI 2.075–3.267, P < 0.001) and longer PLOS (B = 1.294, 95%CI 1.080–1.508, P < 0.001) (model 1).

After further adjustment for preoperative comorbidities (hypertension, type 2 diabetes, coronary heart disease, chronic obstructive pulmonary disease), preoperative analgesic use, and preoperative hemoglobin and albumin based on model 1, using logistic and linear regression analyses tourniquet use was still correlated with a higher transfusion rate (RR = 2.319, 95%CI 1.841–2.922, P < 0.001) and longer PLOS (B = 1.278, 95%CI 1.061–1.495, P < 0.001) (model 2). Finally, After further controlling for all covariates (covariates in model 2 plus the operative variables including anesthesia method, ASA class, intraoperative blood loss, operative time, and TXA,and drain use), tourniquet use during unilateral primary TKA remained associated with a higher transfusion rate (RR = 1.888, 95%CI 1.449–2.461, P < 0.001) and longer PLOS (B = 0.923, 95%CI 0.690–1.156, P < 0.001), as presented in model 3.