Introduction
Worldwide, atrial fibrillation (AF) is the most common sustained cardiac arrhythmia in adults and poses a significant burden to patients, physicians and healthcare systems (Hindricks et al., Reference Hindricks, Potpara, Dagres, Arbelo, Bax, Blomström-Lundqvist, Boriani, Castella, Dan, Dilaveris, Fauchier, Filippatos, Kalman, La Meir, Lane, Lebeau, Lettino, Lip, Pinto, Thomas, Valgimigli, Van Gelder, Van Putte and Watkins2021). Approximately 1 in 3 European adults over the age of 55 will develop AF (Hindricks et al., Reference Hindricks, Potpara, Dagres, Arbelo, Bax, Blomström-Lundqvist, Boriani, Castella, Dan, Dilaveris, Fauchier, Filippatos, Kalman, La Meir, Lane, Lebeau, Lettino, Lip, Pinto, Thomas, Valgimigli, Van Gelder, Van Putte and Watkins2021). Data from the Irish Longitudinal Study on Ageing (TILDA) estimate the prevalence of AF in Ireland to be 3.2% in those over the age of 50 years and nearing 20% in men over 80 years of age (Frewen et al., Reference Frewen, Finucane, Cronin, Rice, Kearney, Harbison and Kenny2013). They have predicted this prevalence to increase three-fold by 2040 (Frewen et al., Reference Frewen, Finucane, Cronin, Rice, Kearney, Harbison and Kenny2013). This rising prevalence is due to an aging population, more vigorous searching for undiagnosed AF and the increasing burden of comorbidities including hypertension, diabetes, heart failure, coronary artery disease (CAD), chronic kidney disease, obesity and obstructive sleep apnea (OSA) (Hindricks et al., Reference Hindricks, Potpara, Dagres, Arbelo, Bax, Blomström-Lundqvist, Boriani, Castella, Dan, Dilaveris, Fauchier, Filippatos, Kalman, La Meir, Lane, Lebeau, Lettino, Lip, Pinto, Thomas, Valgimigli, Van Gelder, Van Putte and Watkins2021). Modifiable risk factors are potent contributors to AF development and progression. (Hindricks et al., Reference Hindricks, Potpara, Dagres, Arbelo, Bax, Blomström-Lundqvist, Boriani, Castella, Dan, Dilaveris, Fauchier, Filippatos, Kalman, La Meir, Lane, Lebeau, Lettino, Lip, Pinto, Thomas, Valgimigli, Van Gelder, Van Putte and Watkins2021).
Despite its prevalence, management of AF remains suboptimal both in Ireland and across the world (Hannon et al., Reference Hannon, Sheehan, Kelly, Marnane, Merwick, Moore, Kyne, Duggan, Moroney, McCormack, Daly, Fitz-Simon, Harris, Horgan, Williams, Furie and Kelly2009; Frewen et al., Reference Frewen, Finucane, Cronin, Rice, Kearney, Harbison and Kenny2013).
The European Society of Cardiology (ESC) highlights that anticoagulation is a key priority when providing AF care in their Atrial Fibrillation Better Care pathway (Hindricks et al., Reference Hindricks, Potpara, Dagres, Arbelo, Bax, Blomström-Lundqvist, Boriani, Castella, Dan, Dilaveris, Fauchier, Filippatos, Kalman, La Meir, Lane, Lebeau, Lettino, Lip, Pinto, Thomas, Valgimigli, Van Gelder, Van Putte and Watkins2021). There is a five-fold increased risk of stroke with AF, and AF-related strokes are associated with higher levels of both morbidity and mortality compared to other stroke subtypes (Jones et al., Reference Jones, Hobbs and Taylor2018; Hindricks et al., Reference Hindricks, Potpara, Dagres, Arbelo, Bax, Blomström-Lundqvist, Boriani, Castella, Dan, Dilaveris, Fauchier, Filippatos, Kalman, La Meir, Lane, Lebeau, Lettino, Lip, Pinto, Thomas, Valgimigli, Van Gelder, Van Putte and Watkins2021). In Ireland, stroke is the leading cause of acquired disability (Irish Heart Foundation, 2010) and the second leading cause of death (Institute for Health Metrics and Evaluation, 2019).
Thankfully, appropriate use of anticoagulation for AF can reduce the risk of stroke by up to 64%; one of the highest impacts of any medical intervention for a common long-term condition (Hart et al., Reference Hart, Pearce and Aguilar2007). However, we know that patients do not receive adequate monitoring of anticoagulants (Kassianos et al., Reference Kassianos, Arden, Hogan, Dew and Fuat2013; Murphy et al., Reference Murphy, Brosnan, McCarthy, O'Raghallaigh, Bradley and Kirby2020; Wijtvliet et al., Reference Wijtvliet, Tieleman, van Gelder, Pluymaekers, Rienstra, Folkeringa, Bronzwaer, Elvan, Elders, Tukkie, Luermans, Van Asselt, Van Kuijk, Tijssen and Crijns2020) and are undertreated with anticoagulation (Hannon et al., Reference Hannon, Sheehan, Kelly, Marnane, Merwick, Moore, Kyne, Duggan, Moroney, McCormack, Daly, Fitz-Simon, Harris, Horgan, Williams, Furie and Kelly2009; Frewen et al., Reference Frewen, Finucane, Cronin, Rice, Kearney, Harbison and Kenny2013; Shantsila et al., Reference Shantsila, Wolff, Lip and Lane2015; Lacoin et al., Reference Lacoin, Lumley, Ridha, Pereira, McDonald, Ramagopalan, Lefèvre, Evans and Halcox2017; Robson et al., Reference Robson, Homer, Ahmed and Antoniou2018; Murphy et al., Reference Murphy, Brosnan, McCarthy, O'Raghallaigh, Bradley and Kirby2020). One of the main barriers to appropriate prescribing is the underestimation of stroke risk in young patients and an overestimation of bleeding risk in elderly patients (Lacoin et al., Reference Lacoin, Lumley, Ridha, Pereira, McDonald, Ramagopalan, Lefèvre, Evans and Halcox2017). Recent figures from the United Kingdom show that only 67% of patients with AF are treated with oral anticoagulation (Lacoin et al., Reference Lacoin, Lumley, Ridha, Pereira, McDonald, Ramagopalan, Lefèvre, Evans and Halcox2017). Those treated with Warfarin are only in the therapeutic range 40% of the time (López-López et al., Reference López-López, Sterne, Thom, Higgins, Hingorani, Okoli, Davies, Bodalia, Bryden, Welton, Hollingworth, Caldwell, Savović, Dias, Salisbury, Eaton, Stephens-Boal and Sofat2017), and patients are frequently prescribed a suboptimal non-vitamin K antagonist oral anticoagulant (NOAC) dose (Steinberg et al., Reference Steinberg, Shrader, Pieper, Thomas, Allen, Ansell, Chan, Ezekowitz, Fonarow, Freeman, Gersh, Kowey, Mahaffey, Naccarelli, Reiffel, Singer, Peterson and Piccini2018). In the United States of America (USA) nearly 60% of reduced-dose NOAC regimens do not follow Food and Drug Administration (FDA) recommendations (Steinberg et al., Reference Steinberg, Shrader, Pieper, Thomas, Allen, Ansell, Chan, Ezekowitz, Fonarow, Freeman, Gersh, Kowey, Mahaffey, Naccarelli, Reiffel, Singer, Peterson and Piccini2018). Inappropriate dosing may be associated with significant adverse events such as bleeding, thromboembolic events and all-cause mortality (Steinberg et al., Reference Steinberg, Shrader, Pieper, Thomas, Allen, Ansell, Chan, Ezekowitz, Fonarow, Freeman, Gersh, Kowey, Mahaffey, Naccarelli, Reiffel, Singer, Peterson and Piccini2018).
Combination antithrombotic therapy, which refers to an anticoagulant in addition to an antiplatelet agent, after an event such as a myocardial infarct or coronary artery bypass graft, is rarely indicated for more than 1 year in duration, and for many, the recommended duration of combination antithrombotic therapy is 6 months (Xie et al., Reference Xie, Robson, Williams, Carvalho, Rison and Raisi-Estabragh2022). If a patient is taking an antiplatelet agent for secondary prevention and develops AF; if they are > 1 year since event and have stable disease then they should be suitable for oral anticoagulant (OAC) monotherapy.
It has been shown that there is no difference in cardiovascular mortality or stroke with prolonged use of combination antithrombotic therapy compared to OAC monotherapy (Shakir et al., Reference Shakir, Khan, Agarwal, Clifton, Reese, Munir, Nasir, Khan, Gopinathannair, DeSimone, Deshmukh, Jackman, Stavrakis and Asad2022). However, combination antithrombotic therapy (which includes an anticoagulant and an antiplatelet agent) significantly increases the risk of fatal and non-fatal bleeding; upper gastrointestinal bleeding (UGIB) in particular (van Rein et al., Reference van Rein, Jørgensen, Lijfering, Dekkers, Sørensen and Cannegieter2019; Yasuda et al., Reference Yasuda, Kaikita, Akao, Ako, Matoba, Nakamura, Miyauchi, Hagiwara, Kimura, Hirayama, Matsui and Ogawa2019; Shakir et al., Reference Shakir, Khan, Agarwal, Clifton, Reese, Munir, Nasir, Khan, Gopinathannair, DeSimone, Deshmukh, Jackman, Stavrakis and Asad2022).
In Ireland, general practitioners (GPs) can initiate OAC in patients with stable, new-onset atrial fibrillation based on their CHA2DS2-VASc score and are encouraged to start NOAC therapy rather than Warfarin (Irish College of General Practitioners Quick Reference Guide (ICGP QRG), 2020). The use of NOACs in Ireland is increasing (Kennedy et al., Reference Kennedy, Gabr, McCormack, Collins, Barry and Harbison2022). Once a patient is on OAC (even if this is initiated in secondary care), the GP will be responsible for the ongoing monitoring and prescription of these drugs. Certain patients will attend both primary and secondary care and require an integrated approach to management.
A ‘General Medical Service’ (GMS) card or a ‘Doctor Visit Card’ (DVC) entitles patients in Ireland to free GP care. These are awarded to all patients over 70 years, under 6 years of age and those below a certain income level based on means testing. These patients are eligible for a biannual structured chronic disease management (CDM) review for AF. This provides an opportunity to manage comorbidities, control risk factors and ensure that the treatment strategy for the patient is optimal; paying particular attention to the presence of anticoagulation and the dose of anticoagulant based on renal function. This was introduced in March 2020.
The success of the CDM program thus far can be seen in the improvements made in the risk factor profile of AF patients (HSE, 2023). This program has resulted in increased physical activity levels, reduced alcohol intake, reduced blood pressure and reduced BMI (HSE, 2023). We know that alcohol increases the bleeding risk in anticoagulated patients and increases the incidence of AF. The recent Alcohol-AF trial demonstrated that even moderate drinkers have a reduction in AF recurrence, symptoms and time spent in AF with abstinence (Voskoboinik et al., Reference Voskoboinik, Kalman, De Silva, Nicholls, Costello, Nanayakkara, Prabhu, Stub, Azzopardi, Vizi, Wong, Nalliah, Sugumar, Wong, Kotschet, Kaye, Taylor and Kistler2020).
Tackling obesity in addition to other CVD risk factors can reduce AF incidence, progression, recurrence and symptoms. Weight loss of ≥10% has been shown to reverse the progressive nature of AF, with those who achieve this being more likely to revert to sinus rhythm or paroxysmal AF (Middeldorp et al., Reference Middeldorp, Pathak, Meredith, Mehta, Elliott, Mahajan, Twomey, Gallagher, Hendriks, Linz, McEvoy, Abhayaratna, Kalman, Lau and Sanders2018).
The aim of this pilot study is to determine the pattern of OAC and antiplatelet use in patients with permanent AF in Irish general practice.
Methods
Study design
A descriptive, cross-sectional observational study was undertaken.
Setting
The 14 practices in the Ireland East practice-based research network were invited to participate in this pilot study via a letter of invitation. The GPs were invited to a virtual information evening run by the investigation team in January 2022 to distribute written and oral information on the study and provide a platform to answer questions. The research team then contacted the GPs via email to provide participant information leaflets and consent forms in April 2022.
The GPs were asked to return a signed consent form and the total number of patients with permanent AF in their practice if they were happy to participate. We used proportionate sampling to generate a sample size for each practice.
The research team then provided report forms to complete and a notice to be displayed in their practices informing their patients that this study was taking place.
Each participating practice used a random number generator to generate a list of patients from the practice to include in the study.
Data were collected retrospectively on these patients in May and June 2022 and recorded on a report form provided by the research team. This was undertaken by the GP and data controller in each participating practice to comply with general data protection regulations.
The GPs recorded anonymous data only. Data collected included demographics, risk factors for AF, medications, investigations and AF management details. The soft copy report forms were returned to the research team at University College Dublin (UCD). This data was then cleaned, pooled and inputted into a secure database.
Participants
All practices in the Ireland East practice-based research network were invited to participate.
They identified patients in the practice with permanent AF.
Inclusion criteria:
-
– Age ≥ 18 years
-
– Permanent AF detected on electrocardiogram, Holter recording or event recorder
-
– Active in practice software and has attended the practice in the last 24 months
Exclusion criteria:
-
– No electrocardiographic objectified AF
Data sources
The GPs undertook a retrospective chart review of the electronic patient record in the GP management software system. Anonymous details were recorded manually on a report form provided by the research team.
Bias
Each practice used a random number generator to create a list of patients to include to reduce sampling bias.
Study size
As this was a pilot study, a formal sample size was not calculated (Billingham et al., Reference Billingham, Whitehead and Julious2013). Based on previous work, it was estimated that 12 per group would be sufficient (Julious, Reference Julious2005; van Belle, Reference van Belle2008). We estimated that 80% of the practices would complete the study (11 practices). Therefore we sought to recruit a minimum of 132 patients using proportional sampling.
Statistical methods
The secure database RedCap was used for data storage and analysis. The Statistical Package for Social Sciences program was used for statistical analysis. The Chi square test of independence was used to analyse categorical variables. Missing data were either excluded from analysis or calculated based on information provided, for example, creatinine clearance.
Ethical approval
Granted by Irish College of General Practitioners Research Ethics Committee.
Results
Participants
Eleven practices consented to participate in this pilot study. These 11 practices had a total number of 1855 patients with AF. We received data on 153 patients proportionally sampled across these 11 practices.
Descriptive data
The demographic details can be seen in Table 1. Most patients – 68% were male (N = 104) and the mean age was 76 years.
Patient demographics
GMS = General Medical Services; DVC = Doctor Visit Card; CDM = Chronic Disease Management; CAD = Coronary Artery Disease; TIA = Transient Ischaemic Attack; PVD = Peripheral Vascular Disease; OSA = Obstructive Sleep Apnoea.
The majority of patients had a GMS card or a DVC (83% N = 127). 115 (75%) patients had CDM review in their practice.
Hypertension (52%), cardiovascular disease (37%), diabetes (27%) and cerebrovascular disease (15%) were the most common comorbidities in this group.
The mean BMI was 29.2 kg/m2 (Range 17–45 kg/m2) with 36.4% categorized as obese with a BMI > 30 kg/m2. The mean weekly alcohol intake was 4.8 units.
Main results
Of the records analysed, 14% of AF patients (21 patients) were not on any anticoagulation. Based on the patient’s CHA2DS2-VASc score of 1 in a male or 2 in a female, OAC was indicated in the majority of these patients (17 patients). Of the patients without OAC at high risk of stroke (CHA2DS2-VASc ≥ 2), 11 out of 17 were on an antiplatelet agent. Nine of these were on aspirin for secondary prevention and two did not have a clear indication.
There were 132 patients on OAC. The most common OAC was Apixaban (52.3%), followed by Rivaroxaban (18.2%), Edoxaban (16.7%), Warfarin (8.3%) and Dabigatran (4.5%). See Figure 1.
Oral anticoagulants.
There were 121 patients on NOACs. We analysed these patient records to determine if they were on the correct dose of the NOAC. Sixteen records were incomplete and did not have key information recorded, such as age/weight/renal function, so we could not determine if the dosage was correct or not.
Of the 105 records analysed, 84 patients (80%) were on the correct dose and 21 patients (20%) were on an incorrect dose. Thirteen patients should have been on a higher dose of OAC and eight met the criteria for dose reduction. The most common error was the inappropriate prescription of the lower dose of Apixaban; 2.5 mg BD. See Figure 2.
Non-vitamin K antagonist oral anticoagulant doses.
Of those who were on the correct dose of NOAC, 61 were male and 23 were female. Of those on an incorrect dose, 10 were male and 11 were female. A chi-square test of independence was performed to evaluate the relationship between correct NOAC dose and gender. The relationship between these variables was significant, X 2 (1, N = 105) = 4.795, P < 0.05 (P = 0.029). Males were more likely to be on the correct NOAC compared to females.
In this pilot study, 41 patients were taking an antiplatelet agent. In total, 32 of these patients were taking an antiplatelet agent for secondary prevention and 9 patients did not have a clear indication. There were no patients taking dual antiplatelet therapy. The majority of these patients were taking aspirin 75 mg (90.2%). Of these, 29 were taking both an antiplatelet agent and OAC. Only 1 of these patients had percutaneous coronary intervention (PCI)/coronary artery bypass graft (CABG) within the last year meaning that 28 patients were on combination antithrombotic therapy without an indication. Combination antithrombotic therapy was defined as concomitant prescription of an antiplatelet agent (aspirin, clopidogrel, prasugrel or ticagrelor) and an anticoagulant (warfarin, apixaban, rivaroxaban, edoxaban or dabigatran). Coronary artery disease was present in 64% (N = 18) of patients who were inappropriately prescribed combination antithrombotic therapy.
A chi-square test of independence was performed to evaluate the relationship between potentially inappropriate combination antithrombotic therapy and a history of coronary artery disease. The relationship between these variables was significant, X 2 (1, N = 153) = 10.713, P = < 0.001. Those with CAD were more likely to be on inappropriate combination of antithrombotic therapy.
Discussion
Key results
The key results from this project are that:
-
1. 11% of patients were undertreated with OAC
-
2. 20% of patients were on an incorrect NOAC dose
-
3. 28 patients (18%) were inappropriately prescribed combination antithrombotic therapy
Limitations
There are several limitations to this retrospective observational study.
The report form that we devised did not ask explicitly if there was an absolute contraindication to OAC, a history of major bleeding or another indication for cessation of therapy. As the information was inputted manually, some of the data were not complete.
We did not ask if there specific reason for an antiplatelet agent in addition to OAC, for example, high-risk disease/recommendation from interventional cardiology/hematological indication.
We also did not assess the international normalised ratio value for those on Warfarin, therefore, we were unable to determine if the dose was therapeutic.
Interpretation
Patients undertreated with OAC
This study found that 11% of patients at high risk of stroke were not anticoagulated. Of the patients without OAC at high risk of stroke (CHA2DS2-VASc ≥ 2), 11 out of 17 were on an antiplatelet agent. Nine of these were on aspirin for secondary prevention and two did not have a clear indication. This suggests that these patients may have been started or maintained on antiplatelet monotherapy as stroke prevention in the setting of AF. However, there is no role for antiplatelet agents as monotherapy in AF for stroke prevention (Hindricks et al., Reference Hindricks, Potpara, Dagres, Arbelo, Bax, Blomström-Lundqvist, Boriani, Castella, Dan, Dilaveris, Fauchier, Filippatos, Kalman, La Meir, Lane, Lebeau, Lettino, Lip, Pinto, Thomas, Valgimigli, Van Gelder, Van Putte and Watkins2021; NICE, 2021) and aspirin is associated with a significant risk of bleeding.
Undertreatment with anticoagulation is a common problem in AF care across the world and the rates of this pilot study are comparable to other work in this area. A cross-sectional study from the UK in 2017 demonstrated that only 67% of patients with AF were receiving anticoagulation (Flaker et al., Reference Flaker, Eikelboom, Shestakovska, Connolly, Kaatz, Budaj, Husted, Yusuf, Lip and Hart2012; Lacoin et al., Reference Lacoin, Lumley, Ridha, Pereira, McDonald, Ramagopalan, Lefèvre, Evans and Halcox2017). The GARFIELD – AF registry reports that 71.1% of patients with AF are on OAC and a recent randomised trial from the Netherlands reported an anticoagulation rate of 81% (Wijtvliet et al., Reference Wijtvliet, Tieleman, van Gelder, Pluymaekers, Rienstra, Folkeringa, Bronzwaer, Elvan, Elders, Tukkie, Luermans, Van Asselt, Van Kuijk, Tijssen and Crijns2020; Bassand et al., Reference Bassand, Apenteng, Atar, Camm, Cools, Corbalan, Fitzmaurice, Fox, Goto, Haas, Hacke, Jerjes-Sanchez, Koretsune, Heuzey, Sawhney, Oh, Stępińska, Cate, Verheugt, Kayani, Pieper and Kakkar2021).
We also know that patients with known AF who present with acute ischemic stroke (the most serious consequence of AF) have a high prevalence of inadequate anticoagulation preceding their stroke (Xian et al., Reference Xian, O'Brien, Liang, Xu, Schwamm, Fonarow, Bhatt, Smith, Olson, Maisch, Hannah, Lindholm, Lytle, Pencina, Hernandez and Peterson2017). This was shown in an Irish context where a Dublin-based prospective study on stroke in 2009 showed that only 32% of patients with known AF and prior stroke were anticoagulated at the time of their recurrent ischemic stroke (Hannon et al., Reference Hannon, Sheehan, Kelly, Marnane, Merwick, Moore, Kyne, Duggan, Moroney, McCormack, Daly, Fitz-Simon, Harris, Horgan, Williams, Furie and Kelly2009). Echoing these findings, a systematic review in 2010 of 54 studies from 11 countries found a consistent pattern of OAC underuse in AF patients with an elevated stroke risk (Ogilvie et al., Reference Ogilvie, Newton, Welner, Cowell and Lip2010).
One of the main barriers to appropriate prescribing is the overestimation of bleeding and falls risk in elderly patients (Lacoin et al., Reference Lacoin, Lumley, Ridha, Pereira, McDonald, Ramagopalan, Lefèvre, Evans and Halcox2017; Xian et al., Reference Xian, O'Brien, Liang, Xu, Schwamm, Fonarow, Bhatt, Smith, Olson, Maisch, Hannah, Lindholm, Lytle, Pencina, Hernandez and Peterson2017). The absolute contraindications to OAC are quite limited. They include serious active bleeding, severe thrombocytopenia <50 platelets/uL or recent high-risk bleeding, for example, intra-cranial hemorrhage (Hindricks et al., Reference Hindricks, Potpara, Dagres, Arbelo, Bax, Blomström-Lundqvist, Boriani, Castella, Dan, Dilaveris, Fauchier, Filippatos, Kalman, La Meir, Lane, Lebeau, Lettino, Lip, Pinto, Thomas, Valgimigli, Van Gelder, Van Putte and Watkins2021). Age and falls risk are certainly not contraindications in their own right.
Patients on incorrect NOAC dose
Of those who were on a NOAC, 20% (21 patients) were on an incorrect dose. Interestingly, 13 patients should have been on a higher dose and eight met the criteria for dose reduction.
It is a common research finding that patients are on an incorrect NOAC dose; in particular inappropriate dose reduction. A retrospective trial in Boston demonstrated that 14.7% of patients on a reduced-dose NOAC had no indication for such a reduction (Barra et al., Reference Barra, Fanikos, Connors, Sylvester, Piazza and Goldhaber2016). An Australian study of NOAC prescribing showed that 34% of patients were on an incorrect NOAC dose; of these 22% were overdosed and 38% were underdosed (Pattullo et al., Reference Pattullo, Barras, Tai, McKean and Donovan2016). In the USA, nearly 60% of reduced-dose NOAC regimens do not follow FDA recommendations (Steinberg et al., Reference Steinberg, Shrader, Pieper, Thomas, Allen, Ansell, Chan, Ezekowitz, Fonarow, Freeman, Gersh, Kowey, Mahaffey, Naccarelli, Reiffel, Singer, Peterson and Piccini2018).
Inappropriate dosing can be associated with significant adverse events such as bleeding, thromboembolic events and all-cause mortality (Barra et al., Reference Barra, Fanikos, Connors, Sylvester, Piazza and Goldhaber2016; Steinberg et al., Reference Steinberg, Shrader, Pieper, Thomas, Allen, Ansell, Chan, Ezekowitz, Fonarow, Freeman, Gersh, Kowey, Mahaffey, Naccarelli, Reiffel, Singer, Peterson and Piccini2018). Undertreating these patients exposes them to the increased risk of bleeding and yet does not offer them protection against embolic strokes (Barra et al., Reference Barra, Fanikos, Connors, Sylvester, Piazza and Goldhaber2016; Steinberg et al., Reference Steinberg, Shrader, Pieper, Thomas, Allen, Ansell, Chan, Ezekowitz, Fonarow, Freeman, Gersh, Kowey, Mahaffey, Naccarelli, Reiffel, Singer, Peterson and Piccini2018). In a large retrospective study from the USA, we saw that inappropriate under dosing was associated with a nearly 5-fold increased risk of stroke in apixaban-treated patients with bleeding rates comparable to the standard dose (Yao et al., Reference Yao, Shah, Sangaralingham, Gersh and Noseworthy2017).
It is a paradox that women are deemed at higher risk of embolic stroke and will score an extra point on the CHA2DS2-VASc score and yet are less likely to be on an appropriate dose of OAC as illustrated in this pilot study. This finding is not unique and has been highlighted in several other research studies (Steinberg et al., Reference Steinberg, Shrader, Thomas, Ansell, Fonarow, Gersh, Kowey, Mahaffey, Naccarelli, Reiffel, Singer, Peterson and Piccini2016; Lee et al., Reference Lee, Kim, Cha, Park, Park, Kang, Shim, Uhm, Kim, Park, Lee, Choi, Kim, Joung and Kim2018). It has also been shown that women are less likely to be prescribed OAC in AF and less likely to be maintained on this treatment (Thompson et al., Reference Thompson, Maddox, Lei, Grunwald, Bradley, Peterson, Masoudi, Turchin, Song, Doros, Davis and Daugherty2017; Ferroni et al., Reference Ferroni, Gennaro, Costa, Fedeli, Denas, Pengo and Corti2019; Tzeis et al., Reference Tzeis, Savvari, Skiadas, Patsilinakos, Stamatelopoulos, Kourouklis, Kyrikos, Tsatiris, Menegas, Hahalis and Giannakoulas2021).
The reasons for this discrepancy are not entirely clear but we do know that a history of minor bleeding is associated with NOAC underdosing (Reference Caso, de Groot, Sanmartin Fernandez, Segura, Blomström-Lundqvist, Hargroves, Antoniou, Williams, Worsley, Harris, Caleyachetty, Vardar, Field and RuffCaso et al., 2023) and women tend to present more frequently with adverse drug reactions such as this (Giner-Soriano et al., Reference Giner-Soriano, Prat-Vallverdú, Ouchi, Vilaplana-Carnerero and Morros2023). It may also be due to renal function assessment; using eGFR rather than creatinine clearance can overestimate renal function in females (Howard et al., Reference Howard, Lipshutz, Roess, Hawes, Deyo, Burkhart, Moll and Shilliday2017).
The four NOACs available to us have different dosages and different criteria for dose reduction making identification of the correct dose complicated and challenging (Steffel et al., Reference Steffel, Collins, Antz, Cornu, Desteghe, Haeusler, Oldgren, Reinecke, Roldan-Schilling, Rowell, Sinnaeve, Vanassche, Potpara, Camm and Heidbüchel2021). All NOACs have some degree of renal clearance. Apixaban includes age, weight and serum creatinine in its dose reduction criteria, whereas reduction of the other three NOACs (Rivaroxaban, Edoxaban and Dabigatran) is based on creatinine clearance. This variation in dosing guidelines likely contributes to prescribing error (Steinberg et al., Reference Steinberg, Shrader, Pieper, Thomas, Allen, Ansell, Chan, Ezekowitz, Fonarow, Freeman, Gersh, Kowey, Mahaffey, Naccarelli, Reiffel, Singer, Peterson and Piccini2018). A guide to correct dosing is included in the appendix. Creatinine clearance needs to be calculated using the patient’s age and weight using the Cockcroft-Gault formula. Most laboratories report renal function as an estimation of the glomerular filtration rate (eGFR ml/min) using the Modification of Diet for Renal Disease formula, which can overestimate renal function in older patients with low body weight (Johnson et al., Reference Johnson, Jones, Mathew, Ludlow, Doogue, Jose, Langham, Lawton, McTaggart, Peake, Polkinghorne and Usherwood2012). Estimated GFR produces higher values than Cockcroft & Gault's estimated creatinine clearance at older ages (Schwartz, Reference Schwartz2016) and may fail to identify those who should receive reduced-dose NOACs (Dowling et al., Reference Dowling, Wang, Ferrucci and Sorkin2013). Having renal function reported using a different formula to what is in the prescribing guidelines creates a barrier to appropriate dosing.
Inappropriate co prescribing of an antiplatelet with an OAC
Most patients who are on an antiplatelet agent for secondary prevention and then develop AF should be suitable to stop their antiplatelet and continue on OAC alone. The current ESC and National Institute for Health and Care Excellence (NICE) guidelines support the health care provider to stop their antiplatelet agent if they have stable coronary artery disease (>1-year post-event, CABG or PCI) and no symptoms of CAD such as angina. These patients will then remain on OAC monotherapy (Yasuda et al., Reference Yasuda, Kaikita, Akao, Ako, Matoba, Nakamura, Miyauchi, Hagiwara, Kimura, Hirayama, Matsui and Ogawa2019; Hindricks et al., Reference Hindricks, Potpara, Dagres, Arbelo, Bax, Blomström-Lundqvist, Boriani, Castella, Dan, Dilaveris, Fauchier, Filippatos, Kalman, La Meir, Lane, Lebeau, Lettino, Lip, Pinto, Thomas, Valgimigli, Van Gelder, Van Putte and Watkins2021; NICE, 2021).
In some cases where a patient is unstable or has high-risk disease, they may remain on combination antithrombotic therapy. This should be a specialist-led decision given that combination antithrombotic therapy increases the risk of fatal and non-fatal bleeding (van Rein et al., Reference van Rein, Jørgensen, Lijfering, Dekkers, Sørensen and Cannegieter2019; Yasuda et al., Reference Yasuda, Kaikita, Akao, Ako, Matoba, Nakamura, Miyauchi, Hagiwara, Kimura, Hirayama, Matsui and Ogawa2019).
The combination of antiplatelet therapy with OAC in patients who have stable coronary artery disease highlights a gap between guidelines and clinical practice.
In our pilot study, 18% of patients (28 people) were inappropriately prescribed an antiplatelet in addition to their OAC putting these patients at higher risk of bleeding.
Using multiple antithrombotic therapies results in an increased risk of serious bleeding complications; in particular, UGIB (Xie et al., Reference Xie, Robson, Williams, Carvalho, Rison and Raisi-Estabragh2022). UGIB is the most common cause of hospital admission for an adverse drug-related event (Pirmohamed et al., Reference Pirmohamed, James, Meakin, Green, Scott, Walley, Farrar, Park and Breckenridge2004). Antiplatelet, anticoagulant use and especially the use of both in combination with older age accounts for a large proportion of these admissions (Pirmohamed et al., Reference Pirmohamed, James, Meakin, Green, Scott, Walley, Farrar, Park and Breckenridge2004). A recent observational study in primary care in the UK has shown that most patients (65.9%) were prescribed combination antithrombotic therapy for an inappropriate prolonged duration of > 12 months (Xie et al., Reference Xie, Robson, Williams, Carvalho, Rison and Raisi-Estabragh2022).
A US-based quality improvement study showed that reducing inappropriate co-prescription of aspirin with warfarin was associated with a reduction in bleeding events, major bleeding events and attendance at an emergency department with bleeding (Schaefer et al., Reference Schaefer, Errickson, Gu, Alexandris-Souphis, Ali, Haymart, Kaatz, Kline-Rogers, Kozlowski, Krol, Shah, Sood, Froehlich and Barnes2022). Those on combination antithrombotic therapy were identified and their primary care physician was asked to consider if this co-prescription was necessary and if not then to modify the prescription (Schaefer et al., Reference Schaefer, Errickson, Gu, Alexandris-Souphis, Ali, Haymart, Kaatz, Kline-Rogers, Kozlowski, Krol, Shah, Sood, Froehlich and Barnes2022). Aspirin de-prescribing increased by 50% (Schaefer et al., Reference Schaefer, Errickson, Gu, Alexandris-Souphis, Ali, Haymart, Kaatz, Kline-Rogers, Kozlowski, Krol, Shah, Sood, Froehlich and Barnes2022).
It has been repeatedly shown that there is no difference in cardiovascular mortality or stroke with prolonged use of combination antithrombotic therapy compared to OAC monotherapy (Schaefer et al., Reference Schaefer, Errickson, Gu, Alexandris-Souphis, Ali, Haymart, Kaatz, Kline-Rogers, Kozlowski, Krol, Shah, Sood, Froehlich and Barnes2022; Shakir et al., Reference Shakir, Khan, Agarwal, Clifton, Reese, Munir, Nasir, Khan, Gopinathannair, DeSimone, Deshmukh, Jackman, Stavrakis and Asad2022). In summary, prolonged combination antithrombotic therapy results in more harm than good.
Generalisability
The results of this pilot study reflect other similar research work from around the world and are likely representative of the type of care provided in general practice as a whole.
Conclusion
This pilot project has highlighted several gaps between guidelines and clinical practice. By identifying these areas, we hope to develop a targeted quality improvement intervention to improve the care that those with AF receive.
Financial support
This project was funded by internal UCD School of Medicine/College of Health & Agricultural Sciences research fund and Ireland East Hospital Group. The lead author, Dr. McErlean, receives funding from the Health Service Executive as an Academic Clinical Fellow.
Competing interests
None.
Ethical standards
Ethical approval granted by the Irish College of General Practitioners. Written consent was obtained from all participating GPs.
Appendix A. Dosing table for non-vitamin K antagonist oral anticoagulants (NOACs)
All NOACS not recommended CrCL < 15 ml/min.
Open access
