The elderly are more prone to arterial as well as venous thromboembolism, and thrombotic disease is the commonest cause of hospital admission, disability and death in patients over 50 years of age in the developed world.

The incidence of deep vein thrombosis (DVT) and pulmonary embolism (PE) and thus venous thromboembolism (VTE) increases with increasing age. Between 65 and 69 years of age, annual incidence rates per 1000 for DVT and PE are 1.3 and 1.8 respectively, and rise progressively to 2.8 and 3.1 in individuals aged between 85 and 89 years. PE is more likely to occur in older men than women of similar age. About 1.7% of older people develop PE and 8% develop recurrent PE within one year of treatment for DVT.1 The diagnosis of PE is often missed in elderly people and it is more often diagnosed only after death.2

The elderly are particularly likely to have concomitant conditions that substantially increase the risk of VTE. Thus the incidence of hip fracture increases from 116 (women) and 0 (men) per 100 000 person-years in the 60–64 year age group to 2597 (women) and 1187 (men) per 100 000 person-years in the 85+ year age group. The elderly are about eight times more likely to develop VTE in hospitals, nursing homes or other chronic care facilities. About 3% of patients with DVT and 21% of those with PE die in hospital. One-year mortality associated with DVT is 21% and that with PE is 39%. Elderly patients are more likely to have a new DVT at the time of admission to hospital with an acute non-surgical illness.3, 4

As a result of the factors described above, the effective prevention of this common and life-threatening condition could produce major benefits. In 2005, the UK House of Commons reported that more than 25 000 patients in England died every year from VTE acquired in hospital; which is more than the combined annual total of deaths from breast cancer, AIDS and traffic accidents, and more than twenty-five times the number who die from MRSA each year.5

In the arterial circulation, the risk of thrombosis and embolism also increases with increasing age. Thus, thrombotic and embolic strokes are more frequent with increasing age. One reason for this is that non-valvular atrial fibrillation (AF) is much more common in the elderly and is associated with a fivefold increased risk of stroke.6 The prevalence of AF rises from 0.5% in the fifth decade of life to 8.8% in the eighth.6 Such individuals who do have an acute stroke are two- to threefold more likely to die than those with acute stroke who are in sinus rhythm.7 The association with age is sufficiently important that the three commonly applied risk algorithms used to consider the risk–benefit of antithrombotic therapy in AF (CHADS2, NICE and ACCOP) all incorporate age >75 years into the risk calculation.8

Over 100 years ago, Virchow identified three main factors for development of thrombosis. The first was reduction in blood flow, which may be a factor in heart failure, a condition seen more frequently in the elderly. The second factor was changes in the vessel wall, which occurs with atherosclerosis and therefore with increasing age. Finally, changes in blood coagulability were cited by Virchow. Increases in clotting factor concentration, in platelet and clotting factor activation and a decline in fibrinolytic activity are all seen in the elderly.9

Several studies have confirmed the original findings of Eccles that oral anticoagulant requirements decline with age.10 In one study, Routledge and co-workers showed that warfarin requirements fell with increasing age so that patients aged less than 35 required a mean of 8.1 mg per day, more than twice as much to maintain the same international normalized ratio (INR) as in patients aged 75 years or over. The relationship between age and warfarin requirements was rather weak, however, indicating that other factors may be more important in determining warfarin dose than chronological age itself.11 These findings have been confirmed in a retrospective longitudinal study in 104 patients aged between 31 and 74 years when warfarin was started and followed for a median of 10 years. This suggests that the decline in dose is not related to a birth cohort effect but does occur in individuals as they grow older.12 The average decline in dose (1.4% per year) in the study was very similar to the rate of decline observed in the much larger cross-sectional study discussed earlier.11

A similar reduced dose requirement in the elderly has also been reported with other coumarin anticoagulants such as acenocoumarol, bishydroxycoumarin and phenprocoumon. The reason for the increased sensitivity to oral anticoagulants is still not fully known, although most studies indicate a pharmacodynamic sensitivity rather than any major change in the pharmacokinetics of warfarin.13, 14 However, one study has reported an age-related decline in warfarin clearance.15

Increased age was also associated with an increased risk of haemorrhage in patients receiving unfractionated heparin (UFH) therapy intravenously with age greater than 70 years being associated with a significant increased risk of major bleeding.16 It is not known what contribution pharmacokinetic and pharmacodynamic differences make to this increased risk. Low molecular weight heparins (LMWHs) are predominantly eliminated by renal excretion, which on average declines with increasing age. Age appears to be one factor predicting the occurrence and severity of bleeding or major bruising events in association with the widely used LMWH, enoxaparin, when given in therapeutic doses for the treatment of thromboembolic disease.17 However, the manufacturers state that no dosage adjustments are necessary in the elderly unless kidney function is impaired.

Two new oral antithrombotic agents, dabigatran etexilate (an orally active inhibitor of both free and clot-bound thrombin) and rivaroxaban (an orally active inhibitor of both the ‘free’ and prothrombinase complex-bound forms of activated factor X) have received marketing authorization in Europe for the prevention of thromboembolic events following major orthopaedic surgery (e.g. total hip and knee replacement). Both are cleared in significant part by renal elimination (around 80% and 65% respectively), and as stated earlier, renal function does decline overall in the elderly. The manufacturers of dabigatran (but not rivaroxaban) recommend a reduced dose in patients aged over 75 years.18

The elderly are more prone to develop haemorrhage, even when not receiving anticoagulants. Life-threatening haemorrhage is most likely to occur in the gastrointestinal tract or intracranially. Although duodenal ulcer is more common at younger ages, gastric ulcer and gastrointestinal bleeding due to non-steroidal anti-inflammatory drugs (NSAIDs) are both more common with increasing age. In addition, the mortality from gastrointestinal bleeding is higher in the elderly. Intracranial bleeding (intracerebral and subdural) in the absence of anticoagulant therapy is also more frequent in the elderly.

Schulman has comprehensively summarized the increasing body of evidence indicating that age is an independent risk factor for major bleeding in patients receiving oral anticoagulant therapy with an average twofold increase in major bleeding among the elderly.19

Increased age also appears to be a risk factor specifically for intracranial haemorrhage. A case-control study compared 170 patients (median age, 78 years) with non-valvular AF who developed intracranial bleeding while on warfarin with 1020 matched controls (median age, 75 years) who were also receiving warfarin for non-valvular AF but did not develop intracranial bleeding.20 The risk of intracranial bleeding was found to be increased in patients aged 85 years or over (adjusted odds ratio, 2.5; 95% CI, 1.3–4.7) and at INRs between 3.5 and 3.9 (adjusted odds ratio, 4.6; 95% CI, 2.3–9.4).The odds ratio for intracranial bleeding increased to 8.8 (95% CI, 4.6–17) at INRs greater than 4 but INRs less than 2 were not associated with a decreased risk (adjusted odds ratio, 1.3; 95% CI, 0.8–2.2). The incidence of bleeding in the elderly has been shown to be higher in the first 90 days of anticoagulation both due to poor control of anticoagulation and unmasking of an occult lesion.21, 22

However, the incidence of intracranial bleeding in patients receiving oral anticoagulant therapy is still very low. Meta-analysis of six AF trials of warfarin versus placebo, involving 2900 patients, showed an intracranial bleeding rate of 0.3% per year during anticoagulation and 0.1% per year in association with placebo.23 Intracerebral haematomas account for 70% of intracranial bleeds associated with anticoagulation and are associated with a mortality rate of up to 60%. Most of the remaining episodes are subdural haematomas, which are less often fatal. The elderly are more likely to have cerebrovascular disease, leukoaraiosis and cerebral amyloid angiopathy which increase the risk of intracerebral bleeding. Poor mobility may predispose the elderly for recurrent falls that may increase the risk of both subdural and intracerebral bleeding.