Real-World Data on Characteristics and Management of Community Patients Receiving Anticoagulation Therapy Who Presented with Acute Bleeding to the Emergency Department at a Regional Australian Hospital: A Prospective Observational Study.

OBJECTIVE: To study patients receiving anticoagulants with or without antiplatelet therapy presenting at a regional Australian hospital with bleeding. The main aims are to explore: (1) patients' characteristics and management provided; (2) association between the type of anticoagulant and antiplatelet agent used and the requirement of reversal; (3) and the length of hospital stay (LoS) in conjunction with bleeding episode and management.
METHODS: A prospective cross-sectional review of medical records of all patients who presented at a tertiary referral centre with bleeding while receiving anticoagulation therapy between January 2016 and June 2018. Data included: patients, demographics, investigations (kidney and liver function tests, coagulation profile, FBC), LoS, bleeding site, type of and reason for anticoagulation therapy, and management provided. Data analysis included descriptive statistics, χ2 association, and regression models.
RESULTS: Among the 144 eligible patients, 75 (52.1%) were male, and the mean age was 76 years (SD=11.1). Gastrointestinal tract bleeding was the most common (n=48, 33.3%), followed by epistaxis (n=32, 22.2%). Atrial fibrillation was the commonest reason for anticoagulation therapy (n=65, 45.1%). Warfarin was commonly used (n=74, 51.4%), followed by aspirin (n=29, 20.1%), rivaroxaban (n=26, 18.1%), and apixaban (n=12, 8.3%). The majority had increased blood urea nitrogen (n=67, 46.5%), while 58 (40.3%) had an elevated serum creatinine level, and 59 (41.0%) had a mild reduction in eGFR. Thirty-five of the warfarinised patients (47.3%) had an INR above their condition's target range despite normal liver function. Severe anaemia (Hb<80g/L) was reported in 88 patients (61.1%). DOACs were associated with a reduced likelihood of receiving reversal (B= -1.7, P=<.001), and with a shorter LoS (B= -4.1, P=.046) when compared with warfarin, LMWH, and antiplatelet therapy.
CONCLUSION: Warfarin use was common among patients who presented with acute bleeding, and the INR in many warfarinised patients exceeded the target for their condition. DOACs were associated with a reduced likelihood of receiving reversal and a shorter LoS than warfarin, LMWH, which might support a broader application of DOACs into community practice.

Introduction

One in every 20 patients will suffer from venous thromboembolism (VTE), either in the form of DVT alone or in combination with PE.1 VTE is associated with high morbidity and mortality rates and significantly harms the quality of life. Furthermore, it has negative financial consequences. All of this highlights the need for proper prevention2 and treatment. VTE management requires a challenging risk assessment,2 and measures which may be pharmacological, mechanical, surgical, or a combination.3 Pharmacological anticoagulation therapy is most common and includes anticoagulants4 and antiplatelet agents.5 Anticoagulants are classified into vitamin K antagonists (such as warfarin), unfractionated heparin, low molecular weight heparin (LMWH, e.g., enoxaparin), and the novel direct oral anticoagulants (DOACs), including rivaroxaban, apixaban, and dabigatran.4 Anticoagulants work by targeting steps in the coagulation cascade.4 DOACs achieve an equivalent anticoagulant effect to classical anticoagulants (warfarin, heparin, and its derivatives) with equal or reduced bleeding risk.6–8 While more specific,9 the DOACs are more costly than warfarin, which may hinder widespread use in the community, even though they do not need a specific monitoring test.10 Balancing the benefits of anticoagulants against the associated risks is a concern for clinical practice and requires further real-world evidence to support decision-making.

The rate of major bleeding resulting from receiving anticoagulants in Australia is high (seven out of every 100 patients per year),11 suggesting the need for pragmatic, evidence-based guidelines for their use. While the DOACs have relatively low bleeding risk when compared with warfarin,8,12,13 clinicians do not tend to use DOACs because they are difficult to monitor and no standard reversal agent is available.14,15

The treatment of patients presenting with bleeding while receiving anticoagulants with or without antiplatelet agents is based on many factors, such as the source of bleeding, hemodynamic stability of the patient, and the severity of blood loss.16 In major bleeding, the management provided might include interventions such as reversing the effect of a therapeutic agent, a surgical achievement of homeostasis, or a combination of both.17,18 Since reversal is indicated for severe and life-threatening haemorrhage among such patients,19 it might be acceptable to consider reversal-receiving as an indicator of a severe bleeding episode. There are currently limited studies of the real-world association between pharmacological anticoagulation therapy and reversal being implemented in severe and life-threatening bleeding events. Further, patients who receive a reversal of anticoagulant therapy often require hospitalisation and recommencement of anticoagulation therapy. This decision could be challenging, given the lack of evidence-based guidelines in the selection of therapy.20,21

LoS is used extensively in the literature to indicate the severity of a condition and the efficacy and cost of treatment.22 Moreover, LoS is used as an outcome measure for health services,23 including quality improvement.24 It is worth noting that the use of LoS as an outcome measure should be taken into account for other individual factors as an indicator of both bleeding severity and management cost.24 Currently, only a few studies have explored the LoS associated with bleeding events among patients receiving anticoagulation therapy in Australia.

The aim of this study was to explore the gap between VTE assessment and management guidelines on the one hand and clinical practice on the other. This purpose was achieved by investigating patients receiving anticoagulants with or without antiplatelet agents who presented with an acute bleeding episode at a regional tertiary referral hospital, Launceston General Hospital Emergency Department (LGH-ED). This exploration included the patients’ characteristics, organ function in correlation with their bleeding presentation and the management provided, type of anticoagulation therapy, the severity of bleeding, and LoS. Translating the findings into real-world clinical practice might bridge the knowledge–practice gap in this field.

Methods

Ethics approval

The project was approved by the Human Research Ethics Committee, the University of Tasmania (H0016734). Because this study was a clinical audit where patient management was not affected, and patients were not actively participating, consent was not required from patients; thus, the Ethics Committee agreed to waive the consent requirement for this low-risk audit.

Data sampling

Data sampling was limited to patients who presented to the LGH-ED with acute bleeding and at the same time were receiving anticoagulation therapeutic agent(s). The LGH is a tertiary regional referral centre in Northern Tasmania, and it has the only Emergency Department within a 100-kilometer radius in the region. The LGH has an electronic/digital medical record (DMR) for all patients presenting to the Emergency Department. Thus, we conducted an electronic search for all patients who presented with bleeding in the period between January 2016 and June 2018. The Pharmacy Department then checked the records at the LGH to determine whether those patients were receiving anticoagulation-therapeutic agent(s) in the form of an anticoagulant with or without antiplatelet agents. Accordingly, only records that satisfied our selection criteria – presentation with acute bleeding while receiving anticoagulation therapeutic agent in the form of DOACs plus/minus antiplatelet agents – were included in the analysis.

Data collection

Data were extracted from the LGH patients’ DMR. The LGH electronic patient file and computerised records provided basic demographics such as age, gender, ethnic group, and language. Further, the system data provided information about the bleeding episode, including the source of bleeding, admission/discharge details, management provided, and LoS in days. The DMR offers data about the indications for administering anticoagulation therapeutic agent(s) and their doses and results of routine blood tests carried out on admission for each patient presenting with bleeding. These tests included full blood count (FBC), renal function (blood urea nitrogen, serum creatinine, eGFR), liver function (ALT, AST albumin, bilirubin), bleeding and coagulation profile (INR, APTT, PT, platelet count), and intervention provided.

Data analysis

Data were analysed using SPSS V26.1.25 The values of laboratory tests were categorised after adjusting for gender, as per the reference intervals published by the Royal College of Pathologists of Australasia or the World Health Organization.26 The anticoagulation therapeutic agent(s) were re-categorised based on the mechanism of action (antiplatelet or anticoagulants). Participants’ characteristics, count, and valid percentages (for non-missing values) were calculated for categorical variables, and means with standard deviation (SD) were calculated for continuous variables. The association between the reason for administering anticoagulation therapeutic agent(s) and the medication used was evaluated by a multinomial logistic regression. A Firth logistic regression model was used to overcome the small sample size to explore the association between the type of anticoagulation therapeutic agent(s) and receiving reversal. Finally, an adjusted linear regression model was used to explore the association between the type of anticoagulation therapeutic agent(s) used and LoS.

Results

Participant characteristics

Among the 1501 patients presenting to the ED at the LGH in the period between January 2016 and June 2018 with a diagnosis of acute bleeding, only 144 (14.4%) were identified by the Pharmacy Department as receiving anticoagulation therapeutic agent(s) in the form of anticoagulants or antiplatelet agents and therefore were eligible for inclusion in our study. Just over half of patients were males n=75 (52.1%), and the mean age was 76 years (SD=11.1). Gastrointestinal tract (GIT) bleeding was the most common site of bleeding in 48 patients (33.3%), followed by epistaxis (n=32, 22.2%), while haematuria was present in 14 patients (9.7%). The most common reason for administering anticoagulation therapeutic agent(s) was atrial fibrillation (AF) (n=65, 45.1%), while PE/VTE treatment was documented in 19 patients (13.2%), and other reasons for anticoagulation therapy were also given (n=27, 18.8%). In a multiple response descriptive analysis for the type of anticoagulation therapeutic agent(s) used among the patients presented with acute bleeding, warfarin was the most common (n=74, 51.4%), followed by aspirin, which was used by 29 patients (20.1%), then rivaroxaban (n=26, 18.1%). Patients’ characteristics are detailed in Table 1.

Table 1

Characteristics of patients receiving anticoagulation therapy who presented to the LGH-ED with acute bleeding during the study period between January 2016 and June 2018.

		Count	Percent
Gender
	Female	69	(47.9%)
	Male	75	(52.1%)
Site of bleeding
	CNS	3	(2.1%)
	GIT	48	(33.3%)
	Epistaxis	32	(22.2%)
	Haematuria	14	(9.7%)
	Skin	9	(6.3%)
	Mucous membrane	2	(1.4%)
	Internal	6	(4.2%)
	Others	22	15.3%)
Reason for anticoagulation therapy
	AF	65	(45.1%)
	valvular	7	(4.9%)
	PE/VTE	19	(13.2%)
	Stroke	3	(2.1%)
	Prophylactic	27	(18.8%)
Anticoagulation therapeutic agent(s)		Multiple Response Count	Response %
	Warfarin	74	(51.4%)
	Rivaroxaban	26	(18.1%)
	Apixaban	12	(8.3%)
	Dabigatran	8	(5.6%)
	Clexane	11	(7.6%)
	Clopidogrel	12	(8.3%)
	Aspirin	29	(20.1%)
	Other anticoagulants	1	(.7%)
		Mean	(SD)
Age		76	(11.1)
Lab investigations on admission
	Creatinine μmol/L	105	(57.7)
	Urea mmol/L	10.66	(8.3)
	eGFR mL/min	60	(22.3)
	AST U/L	25	(18.6)
	ALT U/L	22	(27.6)
	Bili μmol/L	13	(12.0)
	Hb g/L	117	(26.2)
	Platelet /mL	274	(302.1)
	INR	2.2	(3.5)
	aPTT (sec)	38	(14.0)
	PT (sec)	15.14	(5.4)
Length of stay (days)		5	(5.0)

According to a multinomial logistic regression model for the type of anticoagulation therapeutic agent(s) used and the reason for administration, DOACs were more likely to be used with AF patients (OR=9.6, P=.016) than warfarin (OR=6.1, P=.044). Also, for DVT/PE treatment, LMWH was significantly used (OR=30.0, P=.003) than DOACs when compared with warfarin.

Patient management

Laboratory investigations

On patients’ presentation at the LGH-ED with acute bleeding, routine laboratory investigations were carried out, including coagulation profile, FBC, and kidney and liver function for all patients. It was found that a majority (n=67, 46.5%) had an increased blood urea nitrogen level (>3.0–8.0 mmol/L), while 59 (41.0%) had a mild reduction in eGFR (60–89 mL/min). The liver function test showed that most patients had normal AST (n=103, 71.5%), all of them had a normal ALT test (n=144, 100%), and normal bilirubin levels (<20 mmol/L) (n=131, 91.0%). Among those under the vitamin K antagonist warfarin (n=74), many (n=35, 47.3%) had their INR above the target range (adjusted for the reason of administration),27 despite a minority having elevated AST (n=14, 18.9%), while all (100%) had normal ALT without a known liver disease. It is worth noting that no data were available on assays used for measuring DOACs activities. Most patients (n=112, 77.8%) had a normal platelet count (150–400 ×109/L), but a few (n=18, 12.5%) had thrombocytopenia (<150 ×109/L), and among them 13 (72.2%) had severe thrombocytopenia (less than 30 ×109/L). Thrombocythemia (>400 ×109/L) was reported in 14 (9.7%) patients. However, most patients (n=88, 61.1%) were found to have severe anaemia (Hb: female <80g/L, male <80g/L), which was based on the PenaRosas, et al.26 guidelines on haemoglobin concentration for diagnosis and assessment of anaemia severity published by the World Health Organization. Details of laboratory investigations are found in Table 2.

Table 2

Laboratory profile of patients presenting with acute bleeding receiving anticoagulation therapy using lab reference ranges.

		Count	(%)
Blood urea
	Reduced (<3.0mmol/L)	10	(6.9%)
	Normal (3.0–8.0 mmol/L)	67	(46.5%)
	Increased (>8.0mmol/L )	67	(46.5%)
Serum creatinine on admission
	Lower than the normal range	8	(5.6%)
	Normal (male 59–104μmol/L, female: 45–84μmol/L)	78	(54.2%)
	Higher than the normal range	58	(40.3%)
eGFR on admission
	Normal (>90 mL/min)	19	(13.2%)
	Mild reduction (60–89 mL/min)	59	(41.0%)
	Moderate reduction (30–59 mL/min)	49	(34.0%)
	Severe reduction (15–29 mL/min)	8	(5.6%)
	Renal failure (<15 mL/min)	9	(6.3%)
AST on admission
	Normal AST range (female 6–34 IU/L or male 8–40 IU/L)	103	(71.5%)
	Elevated AST	14	(9.7%)
ALT on admission
	Normal ALT range <35U.L	144	(100.0%)
	Elevated ALT >35 U/L	0	(0%)
Bill Categories on admission
	Normal (<20 umol/L)	131	(91.0%)
	Elevated(>20 umol/L)	13	(9.0%)
Anaemia status on admission (based on WHO Hb concentration for anaemia diagnosis and adjusted for gender)
	Non-anaemia (Hb: Female>120 g/L, male>130g/L)	32	(22.2%)
	Mild anaemia (Hb: female 110–119g/L. male 110–129g/L)	9	(6.3%)
	Moderate anaemia (Hb: female 80–109g/L, male 80–109g/L)	15	(10.4%)
	Severe anaemia (Hb: female <80g/L, male <80mg/L)	88	(61.1%)
Platelet count on admission
	Normal platelet count (150–400 ×109 per litre)	112	(77.8%)
	Thrombocytopenia (<150 ×109 per litre)	18	(12.5%)
	Thrombocythemia (>400 ×109 per litre)	14	(9.7%)
INR target range adjusted for the administration reason in warfarinised patients27 (n=74)
	Below target INR range	15	(20.3%)
	At target INR range	23	(31.1%)
	Above target INR range	35	(47.3%)
	INR range in warfarinised patients (n=73)
	< 2.00	15	(20.3%)
	2.00–5.00	40	(54.1%)
	5.01–10.00	18	(24.3%)

Treatment provided

Among those patients who presented with bleeding while receiving anticoagulation therapy, 128 patients (88.9%) were admitted for management, and 81 patients (56.3%) received an intervention. Among those patients who received an intervention, medical management was the most common (n=44, 54.3%) followed by surgical intervention (n=28, 34.6%), such as ligation/cautery of the bleeding vessel, while a few received a combined medical and surgical management (n=9, 11.1%). Among those who received reversal (n=47), by using multiple response descriptive, vitamin K was the most frequent (n=23, 48.9%), while 17 (36.2%) received prothrombin (Table 3).

Table 3

Management provided to patients.

		Count	(%)
Admitted for management (Yes)		128	(88.9%)
Intervention provided (n=81, 56.3%)
	Medical	44	(54.3%)
	Surgical	28	(34.6%)
	Medical and surgical	9	(11.1%)
Reversal received	Yes	47	(58.0%)
Medication used for reversal		Multiple response count	Multiple response %
Vitamin K		23	(48.9%)
Prothrombin		17	(36.2%)
Fresh Froze Plasma		11	(23.4%)
Platelet		3	(6.4%)
Other reversal medications		18	(38.3%)
Novoseven		0	(0.0%)
Idrucizamab		0	(0.0%)
Protamine sulphate		0	(0.0%)

The choice of anticoagulation therapeutic agents on recommencement was similar to the pre-admission agent: warfarin (OR=17.5, P=<.001), rivaroxaban (OR=60.7, P=<.001), apixaban (OR=22.2, P=<.001), clexane (OR=8.1, P=<.033), clopidogrel (OR=61.9, P=<.001), and aspirin (OR=64.0, P=<.001). For more detail, see Table 4.

Table 4

Bivariate association between the anticoagulant therapeutic agent(s) used before and those recommenced after bleeding event.

VTE therapeutic on presentation	Recommenced anticoagulants Therapeutic agents	OR	P-value	95% CI for OR
				Lower	Upper
Warfarin	Warfarin	17.5	<0.001	6.3	48.3
Rivaroxaban	Rivaroxaban	60.7	<0.001	14.1	261.3
Apixaban	Apixaban	22.8	<0.001	5.1	101.0
Dabigatran	Dabigatran	8.4	0.999	0.000	-a
Clexane	Clexane	8.167	0.033	1.1	56.2
Clopidogrel	Clopidogrel	61.9	<0.001	9.0	422.2
ASA	ASA	64.0	<0.001	16.8	242.6

Not calculated due to observation being <=5;

ASA: acetylsalicylic acid; CI: confidence interval; OR: Odds Ratio.

Type of anticoagulation therapeutic agent(s) associated with receiving reversal

Based on a χ2 association for receiving reversal and the type of anticoagulation therapeutic agent(s), vitamin K antagonist (χ2 =24.2, P=<.001) and DOACs (χ2=12.7, P=<.001) were significantly associated with receiving reversal (Table 5). Using a Firth logistic regression, DOACs use was associated with a reduced likelihood of receiving reversal compared with vitamin K antagonists (B=−1.7, P=<.001), as shown in Table 5. It is worth noting that idarucizumab is the approved reversal for dabigatran in Australia. We observed the use of other options19 for reversing the effect of DOACs in some cases, such as prothrombin complex and fresh frozen plasma.

Table 5

Bivariate association between receiving reversal and anticoagulation therapeutic agent(s).

Anti VTE therapeutic agent			Reversal received (YES)
			N	(%)	X2	df	P-value
Vitamin K antagonists		Yes	39	82.7	24.199	1	<.001
Directly oral anticoagulants		Yes	6	12.7	12.756	1	<.001
LMWH		Yes	2	4.2	1.426	1	0.232b
Antiplatelet agents		Yes	12	25.5	0.304	1	0.582
Firth Logistic Regression of receiving reversal
Coefficients
			95% CI for B
	B estimate	Std. Error	Lower	Upper		X2	Sig.
Vitamin K antagonists	Ref	-	-	-		-	-
Directly acting oral anticoagulants	−1.712	0.484	−2.728	−0.827		15.618	0.000
LMWH	−1.260	0.786	−2.976	0.102		3.257	0.071
Antiplatelet agents	0.091	0.455	−0.805	0.968		0.041	0.840

Association between the type of anticoagulation therapeutic agent and LoS due to the acute bleeding event in an adjusted linear regression model

In an adjusted linear regression model for LoS in days, DOACs were associated with a significantly shorter LoS (B=−4.1, 95% CI: −8.177, −0.082, P=0.046) when compared with vitamin K antagonist (warfarin); additionally, a higher haemoglobin concentration on admission was associated with a shorter LoS (B=−0.083, 95% CI: −0.150- −0.016, P=0.016) (Table 6).

Table 6

Linear regression model for LoS in days

				95% CI for B
	B estimate	Std. Error	P-value	Lower	Upper
Constant	1.119	17.11	0.948	−33.87	36.11
Age	0.152	0.074	0.051	−0.001	0.304
Gender (male)	3.773	2.150	0.090	−0.624	8.170
Anticoagulation therapeutic agent(s)
Vitamin K antagonist	Ref	-	-	-	-
Directly acting oral anticoagulants	−4.129	1.979	0.046	−8.177	−0.082
LMWH	−2.180	3.750	0.566	−9.850	5.490
Antiplatelet agents	1.296	1.940	0.509	−2.672	5.263
Intervention required	0.765	1.480	0.609	−2.261	3.792
creatinine μmol/L	−0.017	0.065	0.800	−0.151	0.117
Urea mmol/L	−0.032	0.119	0.788	−0.275	0.210
eGFR mL/min	−0.011	0.110	0.919	−0.237	0.214
AST U/L	−0.010	0.056	0.857	−0.125	0.104
ALT U/L	0.086	0.068	0.214	−0.053	0.225
Bili μmol/L	−0.101	0.093	0.286	−0.292	0.089
Hb g/L	−0.083	0.033	0.016	−0.150	−0.016
Platelet /mL	0.002	0.012	0.878	−0.022	0.026
INR	−4.749	2.806	0.101	−10.489	0.991
APTT (sec)	0.220	0.130	0.101	−0.046	0.486
PT (sec)	0.150	0.212	0.486	−0.284	0.584

ALT: alanine transferase; APTT: artificial partial prothrombin time; AST: aspartate transferase; Bili: bilirubin; eGFR: glomerular filtration rate; INR: international normalizing ratio; PT: prothrombin time; VTE: venous thromboembolism.

Discussion

This study illustrates the characteristics and profile of patients receiving different anticoagulation therapy – in the form of oral anticoagulants including DOAC and antiplatelet agents – who presented with acute bleeding at a regional tertiary hospital in Tasmania, Australia. The associations between the type of anticoagulation therapeutic agent on the one hand and the severity of bleeding and receiving reversal agent(s) on the other, in conjunction with LoS, were studied. The study showed that warfarin was a frequent anticoagulation therapeutic agent among patients who presented with bleeding. Additionally, many of those warfarinised patients had INRs above the desired target range for the condition being administered. While conventional coagulation profile tests were requested for most patients, no agent-specific laboratory tests were requested for patients receiving DOACs. When compared with warfarin, DOACs use was more common in patients with AF. It is worth noting that the Therapeutic Goods Administration (TGA) approves dabigatran in non-valvular AF patients only; rivaroxaban and apixaban are approved for both non-valvular AF and anticoagulation-treatment and prophylaxis. While most patients were admitted for management, many had already received medical management to reverse the effect of the anticoagulation therapeutic agent(s). The reversal agents were less likely to be used with DOACs than warfarin and other anticoagulation therapeutic agents (s). On exploring the association between LoS and individual agents (in an adjusted analysis), the use of DOACs was associated with a shorter LoS than LMWH or antiplatelets compared to warfarin. It is worth mentioning that LoS was longer when the patient had lower haemoglobin concentration on admission.

The study findings might enhance the use of DOACs,28 which were introduced about ten years ago in Australia and had a steady prescription pattern at the present study time. However, because the study did not weigh the prevalence of these medications’ prescription rates, caution should be exercised. Overall, the recommended target INR range was not achieved in many patients who received warfarin and presented with bleeding.29 This suggests the need for continued educational development on pharmacological anticoagulation therapy and clear guidelines and decision aids for medical professionals. While most patients had global coagulation tests, it is argued that these tests are not reliable in patients receiving DOACs.30 Some assays are currently available for DOACs, such as ecarin clotting time (ECT) and chromogenic anti-FXa,18 but they were seldom requested by the ED physicians in this study. This finding might suggest the need to improve medical practitioners’ knowledge about more reliable tests for measuring DOACs activity.

The majority (n=128, 88.9%) of patients who had bleeding because of anticoagulation therapeutic agents were admitted for management. Thirty-one patients (24%) needed reversal. However, this study was able to identify that patients on DOACs were less likely to receive reversal when compared with those who were on warfarin or LMWH. This finding supports the wider implementation of DOACs28 when compared with warfarin and other anticoagulation therapeutic agents. The real-world association between receiving reversal in patients who presented with life-threatening bleeding due to anticoagulation therapy is very difficult to obtain using other research designs, considering that prolonged cohort studies require substantial resources. However, the present study arrived at the same inference using a cross-sectional design.

Furthermore, the present study was able to find a significant association between pre-and post-bleeding pharmacological anticoagulation therapeutic agents. In contrast, dabigatran and clexane were less likely to be used on the resumption of pharmacological anticoagulation therapy when compared with other agents. It is worth noting that, in Australia, dabigatran and antiplatelet agents are not indicated for the treatment of VTE. Accordingly, this finding ought to be explored in future research.

Using an adjusted analysis24 for LoS, it was found that DOACs were associated with a shorter LoS (P=0.046) compared with warfarin, LMWH, and antiplatelets. This finding was consistent with two recent studies.31,32 These studies have concluded that DOACs were significantly associated with a shorter LoS compared to warfarin.31,32

Furthermore, there is evidence that DOACs cost significantly less than warfarin for hospitalisation due to a specific bleeding event with blunt traumatic intracranial hemorrhage.22 However, what is novel in the current study was the wide variety of the bleeding sites and the wide range of anticoagulation therapeutic agents used for various reasons, such as VTE prophylaxis or treatment, in correlation with coagulation profile and kidney and liver function and management and or interventions that were conducted at the time of presentation. Although it might be argued that upfront costs for warfarin administration are cheaper when compared with DOACs,10 our finding suggests that DOACs are more cost-effective overall in the long run when compared with warfarin or other agents, considering the reduced likelihood of patients’ presentations to ED and receiving reversal and the significantly shorter LoS.

Recent literature showed that DOACs have a better safety profile than warfarin, particularly intracranial and subarachnoid haemorrhage.33 Moreover, rivaroxaban appears to be better than warfarin in limitation of blood-brain barrier disruption after intracranial haemorrhage.34 In addition to the VTE prophylaxis effect, DOACs show non-inferior results and superior results compared to warfarin in the management of non-valvular atrial fibrillation and prevention of stroke, especially after the availability of reversal agents such as idarucizumab.35 In this regard, Coons et al. demonstrated in an extensive study of 1840 patients with morbid obesity (BMI>40 kg/m2 ) and VTE that DOACs are more effective and less risky than warfarin.36 In another study, there was no advantage of warfarin over DOACs as VTE prophylaxis in patients who have cancer or atrial fibrillation.37

It is worth noting that in our hands that the use of DOACs in the studied cohort with renal impairment was not associated with excessive bleeding as occurs with warfarin. In comparison to warfarin, the safety of DOACs in case of chronic kidney disease (CKD) was always a concern among clinicians. A recent study by Weber, found that apixaban is safer than warfarin in CKD.38 Nonetheless, careful consideration of anticoagulation’s desired level and anticoagulant dose to achieve the best possible anticoagulation effect and outcome is warranted.39 It is worth noting that there are no reliable, up-to-date guidelines for recommending DOACs in different doses in case of impaired renal function.40 However, in practice, DOACs are considered to have a similar or safer profile compared to warfarin in mild to moderate renal impairment, but this is not the case in severe renal impairment, especially in the renal transplant setting.41

This study’s main limitation was the small sample size yielded from our perspective cross-sectional sampling of patients during the sampling timeframe. However, the same approaches were used to overcome the small sample size by re-categorising anticoagulation therapeutic agents and using statistical methods such as the Fisher’s exact test and Firth logistic regression. It may be worth noting that some studies, such as the one conducted by Lamb et al.22 in the USA, have investigated a closely-related topic but relied on smaller sample size. On the other hand, the present study has several strengths: it included all patients who had bleeding secondary to anticoagulant with or without antiplatelet agents in an entire regional population in mid and north of Tasmania. The LGH is the only tertiary referral hospital in this area, and any patient with acute bleeding would be referred to it. The study contributes to clinical practice by showing the need for better control and effective monitoring of patients on pharmacological anticoagulation therapeutic agents based on administration.

Additionally, our study contributes to research on health services cost-effectiveness by showing that the use of DOACs is associated with a reduced likelihood of receiving reversal and shorter LoS in the absence of life-threatening bleeding compared with warfarin. Furthermore, this study contributes to clinical decision-making with respect to selecting anticoagulation therapeutic agents by showing that reduced morbidity was associated with the use of DOACs compared with warfarin. This study contributed to translational medical research by obtaining real-world evidence on risk assessment and management in patients receiving anticoagulation therapy who presented with bleeding while considering the available guidelines and practice information.

Conclusions

Despite the limitations of our study, it is suggested that the application of DOACs is associated with fewer bleeding complications compared to warfarin. Further, bleeding in DOACs was shown to be less severe in our cohort study with reduced LoS that encourage DOACs’ use, although the costs due to the fact of their pharmacodynamic reversal is not often required.

DOACs were associated with a reduced likelihood of receiving reversal, a shorter LoS, and better overall clinical outcomes. The guidelines should probably address and include better indicators for DOACs bleeding risk, such as ECT and Chromogenic anti-FXa. Therefore, ECT and chromogenic anti-FXa should be better understood and utilised in the context of bleeding associated with DOACs among clinicians, especially in the Emergency Department.