1 Introduction

Total hip replacement (THR) is one of the most successful orthopaedic operations with a total of 101,651 operations performed in the UK in 2016.¹ As the prevalence of osteoarthritis (OA) increases with age, the coexistence of significant patient comorbidies become more common.² There has been an associated increase in the number of patients undergoing hip arthroplasty surgery who are on chronic oral anticoagulants. The most common indications for oral anticoagulation are atrial fibrillation, mechanical heart valves and venous thromboembolism.³

The management of anticoagulation in patients undergoing surgical procedures remains a challenge. Consensus guidelines for the perioperative management of these patients recommend discontinuation of warfarin preoperatively. In high risk patients, bridging with low-molecular weight heparin, and then resuming therapy after surgery.^4,5 In Low risk patients undergoing elective surgery, warfarin is stopped 5 days preoperatively and the International Normalised Ratio (INR) is checked on the day of surgery.⁶ The most common practice is to restart warfarin together with low molecular heparin as soon as the risk of surgical haemorrhage is considered low, usually 24–48 h postoperatively. There is an expected delay before a therapeutic anticoagulant effect is achieved, typically 5–7 days after restarting warfarin.^7–9

Recent publications demonstrated the safety of continuation of anticoagulants perioperatively in patients undergoing elective total knee replacement and cardiac device implantation.^4,10–12 The interruption of anticoagulants perioperatively could mean that patients would require longer hospital stay until their INR is within the desired range. We hypothesised that patients who had their warfarin continued during the perioperative period would have shorter hospital stay and no significant increase in the risks of surgical complications compared to patients who had their warfarin interrupted. To our knowledge, this is the first study to assess the safety of hip arthroplasty surgery in orally anticoagulated patients without interrupting their warfarin therapy.

2 Materials and methods

This was a retrospective review of 20 consecutive patients receiving long-term warfarin (Warfarin group) who underwent primary total hip replacement between 2016 and 2018 with no interruption to warfarin therapy. As a control group, we collected the same data from 20 age and gender matched patients also on long term warfarin but their warfarin was stopped prior to their surgery and restarted postoperatively (Control group). Data was extracted from inpatient medical notes, laboratory results, drug charts, operation notes, and anaesthetic reports. Patients in the control group were operated on by 3 surgeons. The indication for primary total hip replacement was osteoarthritis in all patients involved in this study.

In the warfarin group, warfarin daily maintenance dose was reduced for 2 days preoperatively aiming for an INR of ≤2.5 on the day of surgery. The INR is checked on the morning of surgery and patients with an INR of ≤2.5 underwent their planned procedure. Patients were given their routine warfarin daily dose postoperatively following an INR check (12–24 h postoperatively). Prophylactic dose of low molecular weight heparin (Enoxaparin 40 mg subcutaneously) was also given until INR ≥ 2.0.

In the control group, warfarin was stopped 5 days preoperatively, INR checked on the morning of surgery and patients underwent surgery if INR was ≤1.5.¹³ All patients in the control group were assessed preoperatively for the risk of venous thromboembolism and were found to have low risk, hence no bridging therapy started. Postoperatively, patients were started on a prophylactic dose of low molecular dose (Enoxaparin 40 mg subcutaneously) and warfarin restarted 48 h after surgery. Low molecular weight heparin was stopped when INR was within therapeutic range. All patients were routinely prescribed mechanical prophylaxis for deep vein thrombosis unless contraindicated.

Patients received either general or spinal anaesthetic and Tranexamic acid on induction. Total hip replacement was performed via a posterior approach with meticulous haemostasis. Local anaesthetic infiltration of the soft tissues and the hip capsule was performed in all patients to help with postoperative analgesia and reduced the need for opioid analgesia. All surgical wounds were closed in a water-tight subcuticular Monocryl suture. A second dose of Tranexamic acid (1 g) was given to all patients 8 h postoperatively. All patients were allowed to full weight bear immediately post-operatively and a standardised physiotherapy rehabilitation protocol started on the first postoperative day. Postoperative blood transfusion was based on the National Institute for Health and Care Excellence (NICE) transfusion guidelines depending on the volume of the blood loss, fluid status, haemoglobin (Hb) concentration and the patient's clinical condition.¹⁴ Patients who were symptomatically anaemic with Hb of <80 were transfused red blood cells and their Hb rechecked. Table (1) shows the characteristics of the two groups of patients.

3 Statistical analysis

Statistical analysis was performed using SPSS version 14.0 (SPSS Inc,. Chicago, Illinois) and Microsoft Excel 2016 (Microsoft, Redmond, WA, USA). Statistical significance was set at p < 0.05. Linear variables were summarised by the mean, standard deviation (SD), median and 95% confidence interval (CI). Linear variables were assessed for change over time using the Wilcoxon signed-rank test. The differences between the two groups were examined using the Mann-Witney U test. The null hypothesis for the Wilcoxon test is that there is no difference between the two groups.

4 Results

4.2

4.2 International Normalised Ratio (INR)

In the warfarin group the mean preoperative INR was 1.85 (SD 0.245, range; 1.5–2.4) and the mean INR on postoperative day 1 was 1.99 (SD 0.375, range; 1.4–2.7). The mean preoperative and day 1 postoperative INR in the control group were 1.23 (SD 0.156, range; 1–1.6) and 1.41 (SD 0.32, range; 1–2.3) respectively. There was a statistically significant difference between the two groups in the preoperative INR (P < 0.0001) and the postoperative INR (P < 0.0001). We have also calculated the mean INR for postoperative day 4 for the 2 groups. The mean INR for postop day 4 was 2.23 (SD 0.04, range; 1.9–2.7) for the warfarin group and 1.72 (SD 0.26, range; 1.4–2.3) for the control group. The difference was also statistically significant (P < 0.0001). Figures (1&2) are box plots showing the mean and 95% confidence interval for the daily INR for the both groups.

Close

Fig. 1 Box plot showing the mean and 95% confidence interval for the daily INR for the warfarin group.

Export to PPT

Close

Fig. 2 Box plot showing the mean and 95% confidence interval for the daily INR for the cohort group.

Export to PPT

4.3

4.3 Haemoglobin (Hb)

In the warfarin group the mean preoperative haemoglobin was 134.7 g/L (SD 8.93, range; 117–157) and the mean postoperative haemoglobin was 117.1 g/L (SD 9.74, range; 95–137). In the control group the mean preoperative haemoglobin was 134.4 g/L (SD 13.39, range; 98–152) and the mean postoperative Hb was 108.3 g/L (SD 14.7, range; 73–128). In this study, the drop in haemoglobin (difference between the pre and postoperative haemoglobin levels) has been calculated as an indirect measure of blood loss. The mean drop in Hb postoperatively was 25.95 g/L in the warfarin group (SD 2.51, range; 5–48) and 35.7 g/L in the control group (SD 14.94, range; 7–57) which was statistically significant (P = 0.0066). 95% CIs for pre and post op Hb for both groups is shown in figure (3).

Close

Fig. 3 Means plot showing the mean daily haemoglobin changes in both groups. The error bars show the standard deviation.

Export to PPT

5 Discussion

To our knowledge, this is the first study to assess the safety of continuation of warfarin therapy in patients undergoing elective total hip replacement. The study compared two different strategies in managing warfarin perioperatively. The consensus guideline in low risk patients is to stop warfarin 5 days preoperatively with postoperative bridging to warfarin 48 h postoperatively.⁶ This is mainly to reduce the risk of postoperative bleeding and wound complications. The main findings of our study support the continuation of warfarin in patients on long term anticoagulation therapy as it appears to be a safe practice with no significant increase in complications or need for blood transfusion. Discontinuation of warfarin (control group) was associated with longer hospital stay owing to the time required for the INR to reach therapeutic levels.

Recent published studies assessed the safety of continuation of warfarin in knee arthroplasty.^12,15,16 Chana et al.,¹⁵ Phillips et al.,¹² and Rhodes et al.¹⁶ investigated the continuation of warfarin in patients undergoing total knee arthroplasty and found no increase in the rate of haemorrhage for patients on continuous warfarin. In our study, there was a statistically significant difference (P < 0.0001) in INR levels between the warfarin group (1.85 preop, 1.99 day 1, and 2.23 day 4) and the control group (1.23 preop, 1.41 day 1, 1.72 day 4). The control group had significantly larger fluctuations between pre and postoperative INR levels. This would explain the longer hospital stay observed in the control group as it took significantly longer for therapeutic levels of warfarin to be achieved.

In the warfarin group the mean hospital stay was 4.8 days (SD 2.14, range; 2–9), compared to the 6.1 days (SD 1.80, range; 3–9) in the control group. This was statistically significant with shorter hospital stay observed in the warfarin group (P = 0.0447). We found no significant increase in the incidence of postoperative complications in the warfarin group. Chana et al., Rhodes et al., and Phillips et al. also found no statistically significant difference in complication rates.^12,15,16

There was no significant increase in blood transfusion requirement observed in the warfarin group. There was also no difference in preoperative Hb between the two groups and the expectation would be that a larger drop in Hb would be observed in the warfarin group due to the risk of bleeding but this was not the case. With regards to haemoglobin levels, our study showed a statistically significant (P = 0.0066) mean drop in Hb postoperatively with larger mean drop in Hb on postoperative day 1 in the control group (35.7 g/L) compared to the warfarin group (25.95 g/L).

Interruption of anticoagulation in patients on chronic warfarin therapy poses a major risk of morbidity and mortality.^16–18 There is an increased risk of thromboembolism when long term oral anticoagulation is interrupted prior to surgery.^16,19 Several studies have reported increased thrombin activity after warfarin withdrawal as a result of rebound hypercoagulability.^15,20,21 In addition to this, achieving a therapeutic INR and restoring the patient to their preoperative levels is often associated with extended hospital stays and increased costs of treatment.^15,22,23 There have been several reports of large INR fluctuations during the first few days following surgery.^15,16 Inconsistent anticoagulation during this time could increase the risk of DVT and PE.^15,24,25

6 Conclusion

The results of this study support the hypothesis that it is safe to continue warfarin in patients on long term anticoagulation undergoing elective total hip replacement. The continuation of warfarin was associated with significantly shorter hospital stay and less fluctuations in INR levels. There was no significant increase in the risk of complications or blood transfusion. We acknowledge the limitations of this study; however, we believe that our results contribute to the growing body of evidence that supports the continuation of warfarin in patients on long-term anticoagulants undergoing elective arthroplasty surgery. Perhaps a randomised study powered to show a difference between the two strategies could provide a more robust evidence and guidance for clinical practice.

Funding

None.

Conflicts of interest

All authors declare that they have no conflict of interest.