Abstract
Introduction
The optimal management of patients taking DOAC medications in the perioperative trauma setting is based on limited evidence. Current guidelines recommend withholding DOAC medications 48-72 hours pre-operatively. The objective of this trial was to determine the utility of measuring DOAC levels prior to surgery, evaluate the safety of a cut-off level of <50 ng/mL and to compare the outcomes with time parameters.
Materials and Methods
We performed a cohort study of patients aged 50 years and older admitted with a hip fracture who underwent surgery between January 12 017 and December 31 2019. Primary outcome was time to surgery (TTS) in hours. Secondary outcomes include inpatient transfusion and thromboembolism rates and 30-day mortality.
Results
1579 patients underwent operative management of their hip fracture. The mean TTS in the DOAC group was 33.0 hours. This was significantly longer when compared to patients not on anticoagulation whose mean TTS was 24.4hours (95% CI -13.78: −8.71, P value <.05). It was also significantly higher than patients on warfarin whose mean TTS was 26.4hours (95% CI -12.41: −4.18, P value <.05). There was no significant difference in the transfusion rate and 30-day mortality between the groups (X2 = 2.086, df = 2, P value = .352)]. There was no significant difference in transfusion rates and 30-day mortality between the patients with a DOAC level <50 ng/mL compared with the patients not on any anticoagulation. There was no significant difference in 30-day mortality or transfusion rates between those patients on a DOAC operated within 48 hours compared with those operated after 48 hours (P value = .67).
Discussion and Conclusion
DOAC therapy delays surgery for patients with a hip fracture. Using a DOAC level <50 ng/mL is a safe level to proceed with surgery and reduces the TTS compared to following current guidelines. If DOAC levels are not available the data still supports operation at 48 hours.
Introduction
The burden of osteoporotic hip fractures on healthcare systems is projected to increase worldwide due to the impact of our ageing population. 1 A large body of evidence demonstrates that a delay in the time to corrective surgery leads to an increased risk of morbidity and mortality in the perioperative period.2,3 National Institute of Clinical Excellence (NICE) guidelines recommend performing hip fracture surgery within 48 hours due to increased mortality rates associated with delayed surgery beyond this time. 4
Part of the issues related to aging and increased frailty relate to the increase in medical comorbidities associated with aging. 5 It has been found that up to 30% of patients who present with fracture neck of femur are also anticoagulated. Direct oral anticoagulants (DOAC) are becoming the preferred anticoagulant having proven efficacy in both venous thromboembolism treatment and prevention of systemic embolism in patients with non-valvular atrial fibrillation.6,7 They do not have a direct reversal agent available in Australia currently and there is evidence that coagulation factor concentrates are not completely effective in the emergency reversal of these agents. 8 There is limited evidence available to guide the optimal management of patients taking DOAC medications in the perioperative trauma setting.9–11 The evidence is particularly lacking in the setting of urgent but not immediately life-threatening emergency surgery such as surgery for hip fracture. Current guidelines recommend withholding DOAC medications 48-72 hours prior to major elective orthopaedic surgery.12–14 The recommendation to delay emergency hip fracture surgery beyond 48-hours in patients taking DOAC medications does not account for the associated morbidity and mortality with delaying hip fracture surgery beyond 48 hours in the emergency setting. 15 Therefore, our institution developed a local guideline to measure serum DOAC levels to guide surgical decision-making surrounding hip fracture surgery; using a cut-off DOAC level of <50 ng/mL for surgery to proceed.
The goals of this study were to determine the utility of measuring DOAC levels prior to surgery, to evaluate the safety of the cut off level of <50 ng/mL and to compare the outcomes with time parameters. We hypothesised that by implementing a DOAC level measurement protocol in hip fracture patients we could reduce the time to surgery for hip fracture patients without increasing the rate of peri-operative complications and mortality.
Materials and Methods
We performed a cohort analysis on prospectively collected data from a consecutive series of hip fracture patients treated at a single tertiary referral trauma centre from January 12 017 – December 31 2019. This data was obtained from the Australia and New Zealand Hip Fracture Registry (ANZHFR) after Institutional Review Board approval and Ethics Board approval.
Inclusion criteria included patients aged 50 years or older who underwent surgical management of their hip fracture via the following surgical treatment options: cannulated screws, sliding hip screw, intramedullary nail, hemiarthroplasty or total hip arthroplasty (THA). Exclusion criteria were patients managed non-operatively. All hip fracture patients at our institution are followed up by phone at 120 days post admission.
The primary outcome was time to surgery according to anticoagulation status and DOAC level. Secondary outcomes were inpatient perioperative transfusion rate, inpatient arterial and venous thrombotic rate and 30day mortality.
Demographic information was collected and included age, sex, place of residence at admission, time of admission, type of fracture, type of surgery, type of anaesthesia and time to surgery. The American Society of Anaesthesiologist’s (ASA) score for each patient was collected by the anaesthetist at the time of surgery and we used this score as a surrogate marker of comorbidity. Perioperative complications were retrieved from the electronic medical record (eMR) and included: blood transfusion rate and quantity, cardiovascular complication (acute myocardial infarction), neurological complications (transient ischaemic attack, stroke) and venous thrombotic complications (deep vein thrombosis, pulmonary embolus).
Details of anticoagulation included type of anticoagulant, indication for anticoagulation, and associated antiplatelet agents and other medication. The eMR was reviewed to determine the time of last DOAC dose. If it was documented as either a non-specific morning or evening time, this was recorded as either 08:00 or 20:00 for analysis. A DOAC level was performed on patients taking a DOAC at the time of admission to the operative hospital (Fiona Stanley Hospital). Time to surgery was measured according to the ANZHFR definition which is from time of first presentation to a hospital to the start of the anaesthetic. The time from the last dose of DOAC to surgery was also calculated. A reduced dose of a DOAC was considered to be apixaban 2.5 mg twice daily, Rivaroxaban 10 mg daily and Dabigatran 110 mg twice daily.
Hospital protocol suggested surgery was delayed until the DOAC level was <50 ng/mL. DOAC levels were therefore repeated on consecutive days until the level was less than 50 ng/mL.
Apixaban and rivaroxaban drug levels were assayed via automated chromogenic analysis of their direct factor Xa activity, using apixaban or rivaroxaban as a calibrator (STA-R Stago®). The assay is limited to a minimum measurable drug concentration of 23 ng/mL for apixaban and 25 ng/mL for rivaroxaban. Dabigatran drug levels were measured with the Dilute Thrombin Time assay (Haemoclot, Hyphen-Biomed®).
Statistical Analysis
Descriptive statistics were performed on all data. Frequency distributions were used to describe categorical outcomes and means, standard deviations and ranges were used to describe continuous outcomes. The significance level to compare the mean time to surgery between the two groups was computed using a Students T Test or Mann-Whitney U test as appropriate. Linear regression used for the analysis of eGFR levels and to assess 30day mortality based on time to operation. Statistical analysis was performed using IBM SPSS 26.0 (Armonk, NY). P-values <.05 were considered statistically significant.
Results
Patient baseline characteristics.
DOAC breakdown.
Overall, 96.4% (135/140) of patients had DOAC levels performed at some stage during their admission. 69 had levels <50 ng/mL at the time of admission to the operative hospital. 86 had levels of <70 ng/mL, 101 had levels <100 ng/mL at the time of admission and on the corollary 34 had levels ≥100 ng/mL at the time of admission. 115 of patients had levels <50 ng/mL at the time of surgery and then proceeded to surgery. 20 patients had levels >50 ng/mL at the time of surgery and still underwent surgery as the risk of delayed surgery was thought to be high (Table 2).
The mean time to surgery from first hospital admission in the DOAC group was 33.0 hours. This which was significantly higher compared to patients not on anticoagulation whose mean time to surgery was 24.4 hours (95% CI -13.78: −8.71, P value <.05). It was also significantly higher than patients on warfarin whose mean time to surgery was 26.4 hours (95% CI -12.41: −4.18, P value <.05) (Table 1).
There was no statistical difference in the rate of transfusion in the anticoagulant groups compared with patients not on anticoagulation (Table 1). There was a 19.3% (27/140) transfusion rate in the DOAC group and 19.4% (14/72) rate in the warfarin group compared with 21.6% (198/917) rate in the group not on anticoagulants [X2 = 2.524, df = 2, P value = .283].
Perioperative complications of arterial and venous thrombosis were low and comparable between the 3 groups. There was a total of 27 arterial and venous thrombotic events; 2 (1.4% event rate) occurred in the DOAC group, 0 in the warfarin group and 25 (2.7%) in the group not on anticoagulants.
The 30-day mortality rate was 8.1% in the group not on anticoagulants or antiplatelet agents compared with 6.4% in the DOAC group and 12.5% in the warfarin group. There was no significant difference in the proportion of deaths within 30 days between the groups [X2 = 2.086, df = 2, P value = .352].
When comparing the patients according to DOAC level at the time of surgery, there was a statistically significant difference in the time to surgery between the patients with a DOAC level <50 ng/mL compared with the patients not on any anticoagulation. The mean time to surgery in patients with a level <50 ng/mL was 32.3 hrs compared with 24.4 hrs for patients who were not on any anticoagulation (P value <.05, mean time for DOAC<50 ng/mL = 29.5 hours 95% CI = 26.5:32.4, mean time for not on anticoagulation = 23.4 hours 95% CI = 20.36:26.45). There was however no statistically significant difference in these two groups for transfusion rates and 30 day mortality. This suggests that the DOAC level of <50 ng/mL is safe for operation but does delay surgery. There were insufficient numbers of patients who did not have a level to allow for meaningful comparison in outcomes between these groups.
Outcomes stratified by DOAC level.
Outcomes stratified by TTS amongst patients on a DOAC.
There was a statistically significant difference in the renal function as assessed by eGFR in the patients not on any anticoagulant compared with those on warfarin and those on a DOAC (No anticoagulant eGFR 66.23, CI = 65.05:67.41 vs warfarin and DOAC group eGFR 60.47, CI = 56.71:64.230; P = .004). However amongst the patients on a DOAC, there was no significant difference in eGFR for those whose time to surgery was <24hours compared with time to surgery of 24-48hours compared with >48 hours [(P value for greater than 48 hours = .75, P value for under 24 hours = .52) The estimated eGFR = 59.9 (LCI = 55.9, UCI = 64.2)].
Discussion
Due to the large number of patients requiring oral anticoagulation, most frequently to prevent strokes in the setting of atrial fibrillation, it is essential to determine a safe and practical approach to managing these medications in the perioperative period. 14 A recently published study demonstrated a safe approach to perioperative management on DOACS in elective surgery for most patients anticoagulated for atrial fibrillation, but there remains no guideline for urgent surgery. 12
Current evidence for patients presenting with a fractured hip suggest surgery performed beyond the first 24 hrs is associated with an increased morbidity, mortality and overall length of stay. Additionally, there is evidence in the literature that patients on DOAC have a delay to surgery compared with no anticoagulation.16–20 This was confirmed in our study which showed time to surgery significantly longer in the DOAC group at 33.0 hrs compared with the group of patients not on anticoagulation which was 24.4 hrs. Time to surgery in the warfarin arm was comparable to those not on anticoagulation so it is likely the delay to surgery in the DOAC arm was due to the mandate of DOAC level <50 ng/mL.
There was no difference in transfusion or mortality rates or complication rates in the no anticoagulation arm compared with the group where the DOAC level was <50 ng/mL supporting the safety of this level for surgery.
Current guidelines however suggest considering surgery after a 48 hour period off the DOAC.12–14 By performing levels and using a level of <50 ng/mL to guide surgical timing, the time to surgery can be reduced from this guideline recommended 48hour wait period. This may lend itself to an improvement in overall morbidity, mortality and length of hospital stay.
However, our data suggests that the 48 hour wait period recommended by current guidelines is likely safe as there was no significant difference in 30 day mortality or transfusion rates between those operated within 48 hours compared with those operated after 48 hours. A recent retrospective trial comparing outcomes in patients operated within a 48 hour period and after 48 hours also did not show any significant difference in perioperative haemoglobin loss or transfusion rate. However a higher 90 day mortality rate was found in the group operated on after 48hours. This may have been confounded by a higher ASA score and lower mobility score in the late DOAC group however. 21 One trial compared patients on DOAC vs no anticoagulation both operated within the first 24hour period. It found increased intraoperative blood transfusion rates in the DOAC group with lower admission haemoglobins. There was no difference in perioperative blood loss or mortality. 22
In the PAUSE trial assessing perioperative DOAC management in elective surgery, over 97% of patients had a DOAC level <50 ng/mL at 48 hrs and 87% of patients had a DOAC level <50 ng/mL at 24 hours. 12 In our trial the vast majority of patients were suitable for surgery (as defined by a DOAC level <50 ng/mL) at 48 hours so if levels were not available, our data would support operating on patients at the 48 hour period.
A level of 50 ng/mL was initially chosen when developing the protocol as levels below this threshold was not thought to exacerbate haemorrhage. 23 It may be that we could have a higher cut off for DOAC levels as we saw no increase in transfusions or bleeding complications in our patients taking DOACs with levels >50 ng/mL at the time of surgery.
Recent studies from different countries have published retrospective reviews of time to surgery and transfusion rates for patient with a hip fracture taking a DOAC.24–27 None involved the measurement of DOAC levels, but some demonstrated similar blood transfusion rates even with no delay to surgery to await DOAC clearance.
The current cost of a DOAC level in Australia is A$37.50. Repeated levels can certainly add to the overall cost of care supporting the management plan of operating without a DOAC level.
In our study, patients with poorer renal function were not anticoagulated as frequently either with warfarin or with a DOAC. Despite the known pharmacokinetics however, renal function did not affect the clearance of DOAC with no difference in renal function between the DOAC patient group whose time to surgery was less than 24 hours compared with the group whose TTS was 24-48 hours compared with greater than 48 hours.
Strengths of this study include the relatively large number of patients taking DOACs. Limitations of our study include the retrospective analysis of prospectively collected data. Complications were collected from the eMR in the acute setting and more thrombotic complications may have occurred after discharge from the acute orthopaedic ward. Ideally a longer follow up to gather 90 day mortality data would also have strengthened the study. Additionally, the sample size for patients with a DOAC level ≥50 ng/mL at the time of surgery was small limiting conclusions.
Conclusion
This study has demonstrated that DOAC usage does delay surgery for fracture neck of femur. The strength of this study is that the majority of patients did have a DOAC level allowing us to conclude that a DOAC level of <50 ng/mL is a safe level to proceed with surgery. By using levels to guide timing, patients can be operated on earlier than the current guidelines recommendation of 48 hour wait period. If DOAC levels are not available the data still supports operation at 48 hours. Complications weren’t different between the groups so perhaps surgery could also proceed without any delay. As familiarity with DOAC usage increases with experience, the time to surgery and their outcomes for this group of patients will consistently improve. Having a standard approach with multidisciplinary care to high volume presentations such as hip fracture will only improve and enhance the outcomes for this difficult and increasing patient group.
Footnotes
Author Contributions
HS, SP and SR were responsible for study idea and design. YWT, SR, HS performed data collection. YWT, SP, GB, SR, CJ and HS were responsible for writing the manuscript. CJ and GB performed statistical analysis.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Trial Registration
RGS0000000009.
