Comparison between oral and intravenous application of tranexamic acid for total hip arthroplasty: a meta-analysis

Electronic databases, including PubMed, Cochrane CENTRAL, ScienceDirect and Embase, were searched, in order to identify the studies comparing oral or intravenous TXA administrations on blood conservation in unilateral primary THA; the databases were searched with a range from the origination of the electronic databases to July 2018. Structured search strategies were used in combination, according to Boolean logic: ‘tranexamic acid or TXA,’ ‘hip arthroplasty or replacement,’ ‘oral’ and ‘intravenous’. In addition, a manual search of reference lists was conducted for other trials that could potentially be included. The meta-analysis was based on the acknowledged guidelines of the prioritized reported items for systematic review and meta-analysis.

Articles were included in the further meta-analysis when they met the following criteria: the patients were treated with THA; the intervention used oral versus intravenous applications of TXA; the outcomes included blood loss, blood transfusions, hemoglobin reductions, transfusion rates, hospitalizations, complications and medical costs; and the study was a published comparative trial (RCTs or non-RCTs).

We excluded articles that were: repeat published articles or articles having the same content and result; case report, theoretical research, conference report, systematic review, meta-analysis, expert comment and economic analysis; the outcomes were not relevant.

Two of the reviewers independently reviewed and extracted information from the identified studies by including the first author name, the publication year, the sample size, the study design and outcomes. The main outcomes were a decrease in hemoglobin, blood loss and transfusion requirements. Postoperative thromboembolic events and lengths of hospitalization were secondary outcomes to be reported.

The methodological index for nonrandomized studies was used to evaluate the quality of the nonrandomized studies. The quality assessment of the randomized trial was conducted by using the Cochrane Collaboration. The literature quality was evaluated separately by two reviewers. If there was a divergence, a consensus was reached via consultation.

The RevMan 5.1 (The Cochrane Collaboration, Oxford, UK) was used for the analysis of the pooled data. Heterogeneity was estimated by the p-value and I², based on the standard χ² test. If there was significant heterogeneity (I² > 50%, p < 0.1), a random-effects model was used for the data analysis. When no significant heterogeneity was found (I² < 50%, p > 0.1), a fixed-effects model was used for the data analysis. In cases of significant heterogeneity, a subgroup analysis was performed, in order to investigate the sources. The mean differences (MDs) and 95% CIs were determined for the continuous outcomes. The dichotomous data were calculated by the use of the odds ratio (OR) and the 95% CIs.

The searching process is illustrated in Figure 1; the original database search yielded 52 records, of which 45 trials were excluded by their titles and abstracts. In a study by Irwin et al. [7], data were reported together from TKA and THA, and no available data were able to be used for further meta-analysis. Another study compared oral versus intravenous administrations of TXA; however, intravenous TXA was offered to all of the patients before the incisions, even though TXA was postoperatively applied either orally or intravenously [10]. Both of these studies were excluded. Finally, a total of five articles were included in the current study, among which four of the studies were randomized trials [9,11–13] and one was a retrospective study by Gortemoller et al. [14].

The demographic characteristics and details concerning the included studies are summarized in Table 1. The study by Gortemoller et al. included both THA and TKA patients, and we only pooled the data of patients who received THA [14]. All of these studies that were included in the meta-analysis were published in the past 2 years and included 229 patients who were treated by oral TXA and 222 patients who were treated by intravenous TXA. The smallest sample size was 29, whereas the largest sample size was 60. All of the patients received a unilateral primary THA for hip osteoarthritis or femoral head necrosis. No drains were used, and the total blood loss was calculated based on the method that was presented in the previous study [15]. Different TXA protocols were applied in the four studies. Oral TXA administration of 2 g for 2 h prior to the incision was used in three of the studies, whereas an administration of 20 mg/kg at 2 h before and 3 h after THA was used in the study by Zhao et al. [12]. For the intravenous TXA administration, 1 g of TXA administered immediately prior to the incision was applied in two of the studies, and 20 mg/kg of TXA administered prior to the incision was applied in the study by Luo et al. [11]; alternatively, 15 mg/kg of TXA administered 10 min before and 3 h after THA was applied in the study by Zhao et al. [12]. Furthermore, one dose of 1 g TXA before the incision and two times after the operation was adopted by Wu et al. [13].

The methodological quality score of the retrospective study was 18, based on the methodological index for nonrandomized studies score (Table 2). The data from this study were not prospectively collected, and the patients of the oral TXA group were measured from January 2015 to August 2015, whereas the patients of the intravenous group were measured from the same time period of 2016. The other four randomized trials were well designed and conducted (Figure 2).

The hemoglobin levels were tested in all of the studies, and hemoglobin reductions were not significantly different between the two groups (I² = 0, p = 0.59; MD = -0.06, 95% CI: -0.22–0.10, p = 0.45; Figure 3). Total blood loss and transfusion rates were reported in the three randomized trails. No additional blood loss was reported from the drain. Total blood loss was similar in both of the groups, according to the modification of the Gross formula (I² = 0, p = 0.92; MD = -6.33, 95% CI: -55.81–43.14, p = 0.80; Figure 4). Hemoglobin levels that were less than 7 g/dl without significant clinical symptoms or higher hemoglobin levels with cardiac problems or symptoms were set as transfusion criteria. Six patients in the oral TXA group and five patients in the intravenous TXA group needed transfusions for total arthroplasty without separate hip replacements in the study by Gortemoller et al. [14]. Based on the number of patients from the four RCTs, the transfusion rates were calculated, and the pooled data showed similar transfusion rates in both of the groups (I² = 0%, p = 0.61; OR = 0.91, 95% CI: 0.36–2.29, p = 0.84; Figure 5).

Four randomized studies reported the time that patients spent in the hospital. No significant differences were observed. Intramuscular venous thrombosis was reported in the study by Zhao et al. [12], including three patients in the oral TXA group and four patients in the intravenous TXA group. Three patients in the oral TXA group and five patients in the intravenous TXA group, in the study by Wu et al., suffered calf muscular vein thrombosis [13]. The pooled data showed no significant difference between the groups (I² = 64%, p = 0.1; OR = 0.02, 95% CI: -0.10–0.14, p = 0.73). No deep venous thrombosis or pulmonary embolisms were noticed in any of the four studies. One wound problem was also reported by Luo et al. [11]. One venous thromboembolism was reported in the oral TXA group, without a clear explanation, from knee or hip replacement by Gortemoller et al. [14].