An update on hip fracture risk associated with anticoagulant therapy: warfarin versus direct oral anticoagulants

KEYWORDS:

• Apixaban
• betrixaban
• bone
• dabigatran
• DOAC
• edoxaban
• fracture
• NOAC
• osteocalcin
• rivaroxaban
• vitamin K
• warfarin

1. Introduction

It has been recognized that several drugs used for cardiovascular health might influence bone health beneficially or harmfully [1]. Among these agents, warfarin is a vitamin K antagonist prescribed to millions of patients to decrease their risk of blood clots by inhibiting γ-carboxylation of clotting factors II, VII, IX, and X. Osteocalcin, the most abundant non-collagenous protein in the skeleton, also has the same vitamin K-dependent domain and warfarin use results in higher circulating level of undercarboxylated osteocalcin.

The latest meta-analysis of observational studies [2] indicated that use of vitamin K antagonists does not change hip fracture risk (pooled odds ratio 0.91 and 95% confidence interval 0.69–1.20) but significantly increases fracture risk in older patients (≥65 years: 1.07 and 1.01–1.14, respectively) and female patients (1.11 and 1.02–1.21, respectively). Notably, the latter findings are compatible with earlier evidence that higher circulating level of undercarboxylated osteocalcin is an indicator of higher hip fracture risk in elderly women [3].

In addition to vitamin K antagonists such as warfarin, an increasing number of patients have been prescribed alternate direct oral anticoagulants (DOACs) or non-vitamin K antagonist oral anticoagulants (NOACs) including a factor IIa inhibitor (dabigatran) and factor Xa inhibitors (apixaban, betrixaban, edoxaban and rivaroxaban). Until recently, several observational studies investigated the risk of bone fractures associated with use of DOACs/NOACs versus vitamin K antagonists but had not reported fracture risk specifically for the hip; among these, for example, Lau et al. [4] found lower hip plus vertebral fracture risk in users of dabigatran versus warfarin and Norby et al. [5] showed lower hip plus pelvic fracture risk in users of rivaroxaban versus warfarin. Although limited, the first meta-analysis of randomized controlled trials [6] suggested that use of NOACs as well as warfarin does not affect hip fracture risk; relative risks (95% confidence interval) of NOACs versus warfarin regarding hip fracture and femoral neck fracture were 0.99 (0.72–1.34) and 1.00 (0.66–1.51), respectively. Considering that hip fracture is becoming one of the major causes of morbidity and mortality due to the world’s population aging, the present invited editorial concisely updates the topic during the past one year.

2. Three recent observational studies

In the first cohort study using the Danish national registries that included 37,350 patients with atrial fibrillation, Binding et al. [7] reported significantly lower risk of any fracture (adjusted hazard ratio 0.85 and 95% confidence interval 0.740.97) and major osteoporotic fractures (0.85 and 0.72–0.99, respectively), but not hip fracture (Table 1), associated with use of DOACs versus vitamin K antagonists.

Table 1. Hip fracture risk associated with use of DOACs/NOACs versus warfarin in three recent cohort studies.

	adjusted hazard ratio	95% confidence interval	p value
Binding C, et al. (2019) [7]	0.91	0.74–1.13	0.39
Lutsey PL, et al. (2020) [8]	0.91	0.78–1.07	0.27
Without osteoporosis	1.06	0.86–1.30
With osteoporosis	0.74	0.58–0.96	0.03*
Huang HK, et al. (2020) [9]	0.85	0.70–1.04	0.12
Overall of the three studies [10]	0.89	0.80–0.99	0.04

*P value for interaction

In the second cohort study using the MarketScan administrative claims database of 167,275 patients with atrial fibrillation in the USA, Lutsey et al. [8] found significantly lower risk of all clinical fractures (0.93 and 0.88–0.98, respectively) and fractures requiring hospitalization (0.87 and 0.79–0.96, respectively), but not hip fracture (Table 1), associated with use of DOACs versus warfarin. Of importance, the authors [8] noted relation between lower hip fracture risk and use of DOACs versus warfarin in patients with osteoporosis but not those without osteoporosis (Table 1). By contrast, no significant interaction was observed between age <75 and ≥75 years or men and women [8].

In the third cohort study using the Taiwan’s national health insurance research database of 19,414 patients with atrial fibrillation, Huang et al. [9] showed significantly lower risk of vertebral fracture (0.75 and 0.65–0.86, respectively), but not hip fracture (Table 1) or humerus/forearm/wrist fractures (0.88 and 0.73–1.06, respectively), associated with use of NOACs versus warfarin. Similarly to the meta-analysis [2], the authors [9] noted connection between lower fracture risk and use of NOACs versus warfarin in older patients (65–79 years: 0.76 and 0.65–0.88, respectively; ≥80 years: 0.81 and 0.70–0.94, respectively) but not relatively younger patients (≤64 years: 0.90 and 0.67–1.20, respectively). In agreement with the second study [8], however, there was no significant sex-related interaction [9].

Finally, a meta-analysis of the three aforementioned cohort studies by Huang et al. [10] detected significantly lower hip fracture risk (Table 1), with minimal heterogeneity (I² 0%), in adult users of DOACs/NOACs versus vitamin K antagonists such as warfarin, although the results of each DOAC/NOAC were somewhat different. Taking the limitation of confounding adjustment into account, the authors [10] concluded that further research, especially interventional studies, would be needed to clarify the causality.

3. Expert opinion

Significantly lower hip fracture risk associated with use of DOACs/NOACs versus vitamin K antagonists [10] appears to be closely linked to a diagnosis of osteoporosis [8] (Table 1) and it is important for cardiovascular physicians to note that major risk factors for osteoporosis include sedentary lifestyle as well as aging and female. The mechanisms involved remain unclear but habitual physical inactivity is likely to play a key role because use of vitamin K antagonists could impair bone quality while resultant skeletal fragility in the hip can be potentially compensated if mechanical loading engendered by physical activity is sufficient [11,12].

The current consensus is that, when selecting oral anticoagulants, the risk of bone fractures would be less considered generally [13,14]. Among elderly osteoporotic female patients with physical inactivity, however, DOACs/NOACs rather than vitamin K antagonists seem to be a safer choice [11,12]; the pharmacological management of osteoporosis [15] is recommended to reduce hip fracture risk. Given that follow-up period of the three cohort studies [7–9] was a few years, future research on the topic is expected to focus on longer-term hip fracture risk. In addition, clarification of hip fracture risk associated with use of different DOACs/NOACs could be clinically meaningful.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.