Increased relative risk of delayed hemorrhage in patients taking anticoagulant/antiplatelet medications with concurrent aspirin therapy: implications for clinical practice based on 3-year retrospective analysis in a large health system

Abstract

PURPOSE:

The incidence of delayed posttraumatic intracranial hemorrhage (DH) in patients on anticoagulant (AC) and antiplatelet (AP) medications, especially with concurrent aspirin therapy, is not well established, with studies reporting disparate results with between 1–10% risk of DH and 0–3% mortality. The purpose of this three-year retrospective study is to evaluate the true risk of DH in patients on AP/AC medications with or without concurrent aspirin therapy.

METHODS:

1046 patients taking AP and AC medications presenting to network emergency departments with head trauma who had repeat CT to evaluate for DH were included in the study. Repeat examinations were typically performed within 24 hours (average follow-up time was 21 hours and 99% were within 3 days). Mean time to DH was 20 hours. All positive studies were reviewed by two board certified neuroradiologists. Patients were excluded from the study if hemorrhage was retrospectively identified on the initial examination. Cases were reclassified as negative if hemorrhage on the follow-up examination was thought to be not present or artifactual. Cases were considered positive if the initial examination was negative and the follow-up examination demonstrated new hemorrhage.

RESULTS:

Overall, there was 1.91% incidence (20 patients) of DH and 0.3% overall mortality (3 patients). The group of patients taking warfarin or AP agents demonstrated a significantly higher rate of DH (3.2% compared to 0.9%) and higher mortality (0.9% compared to 0.0%) compared to the DOAC group (p<0.01). The risk of DH in patients taking AC or AP agents with aspirin (13/20 cases) was significantly higher (RR 3.8, p<0.01) than those taking AC or AP alone (7/20 cases).

CONCLUSION:

The risk of DH was significantly higher in patients taking aspirin in addition to AC/AP medications. Repeat imaging should be obtained for trauma patients taking AC/AP agents with concurrent aspirin. The rate of DH was also significantly higher in patients taking warfarin or AP agents when compared to patients taking DOACs. Repeat examination should be strongly considered on patients taking warfarin or AP agents without aspirin. Given the relatively low risk of DH in patients taking DOACs alone, repeat imaging could be reserved for patients with external signs of trauma or dangerous mechanism of injury.

INTRODUCTION:

The incidence of post-traumatic intracranial hemorrhage (ICH) in the setting of aspirin, anticoagulant (AC) and/or antiplatelet medications other than aspirin (AP) has been widely studied, with most studies finding a 4–6% risk of hemorrhage in this patient population [1–2]. Direct oral anticoagulant medications (DOACs), in this study including apixiban, rivaroxaban, and dabigatran, have been found to have lower risk of ICH when compared to warfarin or AP medications [3].

The risk of delayed hemorrhage (DH), defined as patients with initially negative imaging, followed by hemorrhage found on subsequent examinations, is less well established. The literature to date consists primarily of retrospective studies and meta-analyses mostly in the surgical and emergency medicine literature with variable findings. Numerous studies have found a relatively low risk of DH ranging from 0.6–1.2% without mortality [4–15] while others found moderate risk of DH around 2% [16–17], with higher mortality from DH of around 1%. Several investigators found an extremely high risk of DH from 7–30% and 2% or higher mortality [18–20].

Studies which recommended against routine repeat imaging for patients on AP/AC medications who presented with head trauma primarily cited the potential cost of the intervention as well as increased radiation exposure for the patients. Most of these studies were retrospective with variable study design. They differed in follow-up period, whether patients received follow-up CT or were followed clinically, and whether the determination of DH was made from radiology reports or reinterpreting each study for presence of ICH not identified on initial interpretation. Most studies also had a very short follow-up period of around 4 hours.

Only a few studies reviewed the risk of DH in patients taking AP medications and most investigators did not assess the effect of aspirin and/or AP medications taken concurrently with AC agents. Relatively few studies analyzed the risk of DOACs [21–25] with divergent results demonstrating risk of DH ranging from 1–10%, although most studies did demonstrate a lower rate of DH when compared to warfarin or AP agents. Some of these trials were very limited with few patients, variable follow-up intervals, and many patients not receiving repeat imaging. Studies did not demonstrate a significant difference between the various DOACs in rate of hemorrhage, mortality, or risk of DH [21–25].

Our previous work [26–27] demonstrated much lower risk of DH and morbidity/mortality in patients taking DOACs. Given the large number of patients taking AP/AC medications, the high number of patients taking concomitant aspirin or AP agents, and increasing numbers of patients taking DOACs (>50% in a recent survey of nursing home patients) [28], accurate characterization of the true incidence, morbidity, and mortality from DH, especially of patients taking DOACs, has high clinical utility. In this study, we present our three-year retrospective analysis of patients taking AP/AC medications, including analysis of the risk of patients taking AC medications along with AP medications or aspirin. Based on our preliminary data, we hypothesize that taking AP/AC medications with concurrent aspirin will significantly increase the risk of DH and that warfarin and AP agents will demonstrate increased risk of DH when compared to DOACs.

MATERIALS AND METHODS:

Compliance with Ethical Standards:

This retrospective study was performed under the supervision of the local Institutional Review Board. Informed consent was not required as the study was retrospective and IRB exempt. There were no conflicts of interest from any of the authors.

Patients were considered candidates for the study if they were taking AP or AC agents, including DOACs, suffered head trauma and their initial head CT from the emergency department was interpreted as negative for ICH. For all potential candidates, the radiology reports for the initial examinations included a recommendation for repeat imaging to evaluate for DH. Approximately 50% of patients for which repeat imaging was recommended received repeat imaging. Repeat examinations were typically performed within 24 hours (average follow-up time was 21 hours and 99% were within 3 days).

1046 patients who had a negative initial scan, had no additional instances of trauma between examinations and had repeat imaging to evaluate for DH were included in the study. Patients were included if their initial examination was interpreted as negative and there were no additional instances of trauma between the examinations. All positive studies were reviewed by two board certified neuroradiologists. Patients were excluded from the study if ICH was retrospectively identified on the initial examination. Cases were reclassified as negative if ICH on the follow-up examination was thought to be absent or thought to be artifactual on retrospective review. Cases were considered positive if the initial examination was considered negative and the follow-up examination demonstrated new ICH. Cases where the two readers did not agree would have been adjudicated by a third board certified neuroradiologist, however, this did not occur during this study. The patients’ medical records and medication lists were reviewed for the presence of AP/AC agents and the types and dosages of these medications were recorded and reconciled. Chi-squared and Fisher’s Exact Probability tests were used for statistical analysis using SPSS 28.

RESULTS:

1046 patients were included in the study. Patient demographics and medications are presented in Table 1 and morbidity and mortality with each AC or AP agent are presented in Table 2. 539 patients were in the DOAC group; 352 taking apixiban, 168 taking rivaroxaban, and 19 taking dabigatran. 287 patients were taking warfarin, and 183 were taking clopidogrel. 23 patients were taking other AP agents, and 14 patients were taking other AC agents (heparin, enoxaparin). Of patients taking aspirin, 340 were taking 81 mg baby aspirin and 5 were taking full dose 325 mg aspirin. All incidences of DH with aspirin occurred with patients taking an 81 mg dose.

Table 1:

Number of patients taking each anticoagulant/antiplatelet medication, sex, average age, and number of patients in each group taking aspirin. The other group includes patients taking dabigatran (19), ticagrelor (18), enoxaparin (10), heparin (4), prasugrel (4) and fondiparinux (1).

Medication	Total Patients	Male	Female	Average age (years)	Patients taking Aspirin
Apixiban (Eliquis)	352	149	203	78.5	119
Warfarin (Coumadin)	287	136	151	78.7	85
Rivaroxaban (Xarelto)	168	77	91	75.0	44
Clopidogrel (Plavix)	183	109	74	78.1	91
Other	56	28	28	76.1	6
Total	1046	499	547	77.5	345

Table 2:

Patients taking each anticoagulant/antiplatelet medication, incidence of DH, percent incidence of DH in each group, mortality from DH in each group, percent mortality in each group, and number of patients and incidence of DH in patients taking anticoagulants or antiplatelet agents along with concurrent asprin therapy.

Medication	Total Patients	Negative	Positive	% Positive	Deaths	% Death	Aspirin	Positive w/aspirin
Apixiban (Eliquis)	352	352	0	0	0	0	119	0
Warfarin (Coumadin)	287	277	10	3.48%	2	0.7%	85	5
Rivaroxaban (Xarelto)	168	163	5	2.97%	0	0	44	4
Clopidogrel (Plavix)	183	178	5	2.73%	2	1.1%	91	4
Other	56	56	0	0	0	0	6	0
Total	1046	1026	20	1.91%	3	0.3%	345	13

Overall, there was 1.91% incidence (20 patients) of DH and 0.3% mortality related directly to DH (3 patients). Average time to CT scan showing DH was 20 hours. 10 occurred in patients taking warfarin, 5 of which were also taking aspirin. 5 occurred in patients taking clopidogrel (4 also taking aspirin), and 5 occurred in patients taking rivaroxaban (4 also taking aspirin). All the DOAC patients with DH were taking rivaroxaban. No patients taking apixiban or dabigatran had DH. Of the 1046 patients, 345 were concurrently taking aspirin, and 13 of the 20 cases of DH were taking either AC or AP agents with aspirin. Patients taking both AC/AP agents with aspirin had increased risk of DH relative to AC or AP medications alone (RR 3.8, p <0.01). The group of patients taking warfarin or AP agents demonstrated a higher rate of DH (3.2% compared to 0.9%) and higher mortality (0.9% compared to 0.0%) compared to the DOAC group. The differences were statistically significant (p <0.01) for both incidence of DH and mortality. Of 576 patients taking DOACs, 163 were also taking aspirin. Of 5 patients taking DOACs who demonstrated DH, 4 (80%) were taking aspirin (RR 9.8), p<0.03 compared to patients taking DOAC without aspirin). These data are summarized in Table 3.

Table 3:

a) Incidence of DH in all patients in the study taking AC or AP agents alone compared to those taking AC or AP agents along with aspirin with relative risk and p value, b) Incidence of DH in patients taking DOACs alone compared to those taking DOACs with aspirin, with relative risk and p value, c) Incidence of DH in patients taking warfarin and AP agents along compared to those taking warfarin and AP agents along with concurrent aspirin therapy with relative risk and p value.

All Patients				DOACs				Warfarin + AP
	DH	No DH			DH	No DH			DH	No DH
Aspirin	13	332		Aspirin	4	159		Aspirin	9	167
No Aspirin	7	694		No Aspirin	1	375		No Aspirin	6	310
p < 0.01, RR = 3.77				p < 0.03, RR = 9.80				p < 0.05, RR = 2.69

The 20 cases of DH included 5 cases of subdural hemorrhage, 7 cases of subarachnoid hemorrhage, 6 cases of parenchymal hemorrhage, and 2 cases of intraventricular hemorrhage. All 20 patients required admission after the hemorrhage was detected (average length of stay: 8.5 days), 12 required admission to the intensive care unit and 3 died from complications of DH.

DISCUSSION:

The risk of DH in patients taking AC/AP medications is an area of active investigation, especially after the wider adoption of DOAC agents. The literature to date has demonstrated a wide range of results for the risk of DH, in many cases due to selection bias or methodology. In this study, we present three-year data with retrospective imaging review on more than 1,000 patients taking AC/AP agents who received repeat imaging after head trauma, demonstrating nearly 2% risk of DH, higher than many published reports, and 0.3% mortality from complications of DH in these patients, higher than some published reports. The rate of DH was significantly higher in patients taking AC/AP agents with concurrent aspirin, compared to patients taking AC/AP agents alone. The relative risk was much higher in DOAC patients taking aspirin compared to patients taking DOACs alone (RR 9.8). Patients taking warfarin or AP agents had a higher rate of DH when compared to patients taking DOACs, even controlling for the effects of aspirin therapy.

We noted a higher risk of hemorrhage in patients taking rivaroxaban (2.97%, 5/168 patients) when compared to apixiban (0%, 0/352 patients) and this difference was statistically significant (p<0.003), although not when controlling for patients also taking aspirin (p=0.31). Prior studies on DOACs did not demonstrate significant differences in primary hemorrhage, DH, or mortality [21–25] so further investigation may be warranted.

Some studies have analyzed the cost-effectiveness of repeat head CT in the ED setting [12, 21]. However, these studies did not quantify the efficacy of repeat imaging, its effect on clinical outcomes, or costs associated with missing DH (including both treatment related and medicolegal expenses). Several studies have found that only prompt early reversal of anticoagulation in anticoagulated patients with ICH will improve outcomes [29] so early intervention is essential in this patient population. Patients discharged from the emergency department after negative initial imaging will remain on anticoagulation, increasing the risk of worse outcomes in the setting of DH [29]. Management of severe ICH carries significant costs, as the average cost of ICU admission was $75,969 in a recent study [30], increasing to $152,373 in patients on AC medications. Additionally, ICH results in high mortality, as high as 36% in a recent study [30]. Costs of long-term care for patients with severe ICH requiring ICU admission are substantial, averaging more than $4,850 a month in another recent study. [31] Therefore, despite the costs incurred in repeat imaging and retaining the patient in the emergency department or observation area, repeat imaging and early intervention could potentially decrease health care costs and improve outcomes. This is especially true in patients on both AC/AP agents and aspirin where one case of DH was detected in every 26 CT scans. Patients on DOACs alone without aspirin had much lower risk (1 DH for every 376 CT scans).

Our study found that the majority (13/20) of cases of DH occurred in patients taking AC or AP agents with aspirin, significantly increasing the rate (3.8% compared to 1.0% in patients without aspirin) and relative risk (RR 3.8) of DH, especially in the DOAC group (9.8 RR for patients on DOACs + aspirin, 4.1 RR for patients taking AP agents and aspirin, and 2.4 RR for patients on warfarin and aspirin), with rate of DH of 5.9% (1/16) in patients on warfarin and aspirin and 4.4% (1/23) taking AP agents with aspirin.

Given the high rate of DH in patients taking both AC/AP agents and aspirin, we recommend repeat CT examination in these patients. The rate of DH was higher in patients taking AP agents or warfarin when compared to DOACs (3.1% vs 0.9%, RR 3.3, p<0.03). This rate was higher both with and without aspirin (4.9% vs 2.4% with aspirin, 2.0% vs 0.3% without aspirin. Given patients taking AP agents or warfarin was still 2.04%, nearly as high as the rate of patients taking DOACs with aspirin (2.4%), indicating repeat CT examination to evaluate for DH should be strongly considered in this group. Patients taking DOACs without aspirin had only a 0.27% rate of DH (1/376); repeat examination is likely unnecessary in this group unless there are concerning findings such as dangerous mechanism of injury or external signs of trauma, as this was associated with higher risk of ICH [5]. Our proposed recommendations are presented in Figure 1.

Figure 1.

Proposed recommendations

One limitation of this study is that only approximately 50% of the patients presenting with head trauma on AC/AP medications received repeat imaging, which could result in selection bias as repeat imaging may have been more likely in patients with external signs of trauma or more dangerous mechanism of injury, which has been shown to increase the risk of ICH and mortality/morbidity [5]. In future studies, given the significant risk of DH, routinely performing repeat imaging on all patients per the above recommendations may result in a more accurate assessment of the risk of DH. Another area of future investigation may involve quantifying the cost effectiveness of repeat imaging by comparing the overall costs of patients with DH on repeat exams to historical cost data from patients taking AC/AP medications who had head trauma and returned with ICH after ED discharge. Although aspirin has been shown to increase the risk of ICH in the setting of head injury, repeat imaging is not routinely performed in patients taking aspirin alone. Analysis of the risk of DH in this patient population could have value given the findings of this study, but could be challenging with the ubiquity of aspirin therapy in older adults. Our recommendations as described in Figure 1 are proposed recommendations given the results of this study, but future clinical validation is necessary to confirm these results.

CONCLUSION:

The rate of DH in our patient population was 1.9%, with 0.3% mortality from DH. The risk of DH was also significantly higher in patients taking aspirin in addition to AC/AP medications (RR 3.8), with one case of DH for every 26 scans in this patient population compared to 1 case per 100 scans without aspirin. Given the significantly increased risk of DH, repeat imaging should be obtained for patients with head trauma taking AC or AP medications with concurrent aspirin. The rate of DH was significantly higher in patients taking warfarin or AP agents when compared to patients taking DOAC medications, even correcting for aspirin therapy (4.9 % vs 2.4% with aspirin, 2.0% vs 0.3% without aspirin). Repeat examination should be strongly considered in p patients taking warfarin or AP agents even without aspirin. Given the relatively low risk of DH in patients in DOACs alone (1 case of DH per 376 scans), repeat imaging could be optional in this patient population and reserved for patients with external signs of trauma or dangerous mechanism of injury.