Page 394 - ILAE_Lectures_2015

Additional risks to patients with epilepsy

There are many reasons why patients with epilepsy might be at increased risk of bone disease,
including reduced exposure to sunlight (housebound/institutionalised), frequent falls, and
lower physical activity levels in patients with active epilepsy.

Fractures
Many of the early studies showing an increased incidence of fractures in patients with epilepsy
were carried out in institutionalised patients18,19, in whom low activity levels and poor sunlight
exposure are important confounders. However, studies in the community and in ambulatory
patients have confirmed a 2–3-fold increase in fractures in patients with epilepsy20-23. The most
recent of these, from the UK GP database, though retrospective, was population based, and
included over 40,000 epilepsy patients and 80,000 controls23, and found the overall incidence
of fractures to be doubled in epilepsy patients compared to age and sex matched controls.

This was also the conclusion from recent meta-analysis studies24,25, with the highest relative
risks for osteoporosis-related fractures (hip and spine), as might be predicted if metabolic bone
disease is contributory. Others have reported that up to one-third of the increased risk21,26
appears to be a direct result of injury during seizures, again something supported by the meta-
analysis24 and more recent studies27. Thus optimum seizure control, especially where there are
convulsive seizures and/or falls, remains a primary goal when considering bone health.
Avoiding the motor complications of AED treatment that might further predispose to falls
(although this has been little studied to date28,29) and being aware of general fall prevention
strategies (good lighting, appropriate correction of refractive errors, etc) is also important in
this context, as patients on AEDs also seem to be at higher risk of non-seizure falls than controls
in at least one study27.

However the main concern is whether AEDs in themselves confer additional risks. Clearly this
is important to establish, both in order to advise patients with mild/infrequent seizures and
those in remission on the risks/benefits of continued AED treatment, and in terms of
prevention/detection for patients continuing on AEDs. Some epidemiological studies in
ambulatory patients21,26 found the increased risk of fractures in patients on AEDs was barely
significant, once seizure-related fractures were excluded. Similarly in a recent population-
based case-control study, the relative increase in fracture for patients on enzyme inducing
AEDs was modest (OR 1.38, 95% CI 1.311.45) after adjustment for cofounders (steroids,
comorbidity, social variables, prior fracture), using any fracture as outcome, and use of AEDs
as exposure variable30. For non enzyme-inducers the risk was still statistically increased (OR
1.19, 95% CI 1.111.27), though less markedly. However, this was a huge study, totalling
nearly 125,000 fractures, and a dose-response relationship could be shown for carbamazepine,
phenobarbitone, oxcarbazepine and valproate, supporting that this is a biological drug effect,
though not huge. Amongst individual drugs, a significant risk was not shown for any of the
other newer drugs, but the authors acknowledge the study had insufficient power in this
context. That any AED is associated with an increased risk, albeit higher with enzyme inducers,
is supported by meta-analysis studies25. Older AEDs show a clearer association, though this
may reflect the inevitable bias of duration of exposure and cumulative AED load, with
refractory patients often having multiple exposures, as well as the fact that more are enzyme
inducers.

A large case-control study based on the UK GP database cited previously, which took account
of many other confounders (though not diet and exercise), went a step further, also attempting
to control for disease severity (using number of drugs/medical contacts as surrogate markers)
and has shown a clear cumulative association with duration of AED use, each year of exposure
being associated with a 9% increase in fracture risk31. This translates into up to an additional