Rivastigmine and QT interval prolongation
F. Inglis

Walsh & Dourish ( 2002) reported that a 78-year-old man, receiving a number of medications and with a history of myocardial infarction and hypokalaemia, developed an abnormal QTc interval a week after starting rivastigmine treatment. I have performed an extensive review of the tolerability and safety of cholinesterase inhibitors ( Inglis, 2002), in which I described the favourable cardiac safety profile of rivastigmine. Therefore, I contacted Novartis for more information. This case, which was initially submitted to the authorities in June 2001, included further clinically relevant information.

Primarily, the patient's pre-rivastigmine QTc (3 weeks before starting treatment) was 431 ms rather than 397 ms as suggested by Walsh & Dourish (C. Videbaek (Novartis), personal communication, 2002). The reported QTc of 397 ms was obtained a week after starting rivastigmine treatment, indicating that during this week the patient's electrocardiogram (ECG) ‘normalised’. The following week (2 weeks post-rivastigmine) it increased to 477 ms. The QTc prolongation (pre-rivastigmine to 2 weeks post-rivastigmine) was less than 11%. Nevertheless, since this change was above the 30 ms usually considered relevant, it is important to assess in an unbiased manner whether it was drug-induced.

The patient was already at risk of cardiac abnormalities owing to: previous increased QTc; hypokalaemia (a risk factor for QTc change; De Ponti et al, 2002) 2 weeks before starting rivastigmine treatment (no potassium values were reported at the time of the ECG finding); concomitant use of diltiazem, which is known to cause atrio-ventricular blockade and brady-cardia (risk factors for QTc change; De Ponti et al, 2002); a history of hypertension, ischaemic heart disease, myocardial infarction and cerebrovascular accident, reflecting the presence of clinically significant heart disease (another risk factor for QTc change; De Ponti et al, 2002): concurrent Lewy body dementia, which is associated with autonomic failure ( McKeith, 2000) and frontal lobe deficits that may influence QT intervals ( Kubota et al, 2001).

My review of the cholinesterase inhibitors ( Inglis, 2002) included an analysis of 2791 patients involved in pivotal studies of rivastigmine in Alzheimer's disease ( Morganroth et al, 2002). About 30% and 10% of these patients had cardiovascular disorders and heart rate/rhythm disorders, respectively. About 35% were receiving concomitant cardiovascular treatments. Even in this relatively at-risk population, heart rate, PQ, PR, QT and QRS intervals were very similar in rivastigmine- and placebo-treated patients, indicating that rivastigmine did not produce adverse effects on cardiac function as assessed by ECG. The lack of cardiac effects associated with rivastigmine may be explained by its selectivity for central over peripheral cholinesterases, and an apparent brain-region selectivity that may avoid areas such as the medullary cardiorespiratory nucleus (Enz et al, 1993).

Case reports are an important means of communicating clinical observations. However, it is important that the facts are presented clearly to allow a balanced judgement on the available evidence. I would suggest that the prolonged QTc described in this single case report is more likely to be due to the confounding factors described above than to a causal association with rivastigmine treatment. The cholinesterase inhibitors form an invaluable part of our limited armamentarium in managing patients with dementia. It would be unfortunate if patients who might benefit from these treatments were deprived of them because of false-positive associations with cardiotoxicity.


Author's reply

Prolonged QTc interval is defined as a QTc longer than 440 ms ( Khan, 2002); therefore, by this definition, the patient did not have a documented prolonged QTc interval prior to the introduction of rivastigmine.

As detailed in the original report of this case to Novartis, the patient had been admitted a number of weeks previously to a medical ward where he developed diarrhoea which was deemed responsible for the lowering of his potassium. As a result he received potassium supplements while the diarrhoea was ongoing and once the diarrhoea stopped the potassium was rechecked and the potassium supplements were discontinued. The patient had no diarrhoea at any stage during his treatment with rivastigmine that could have led to a further development of hypokalaemia. The patient had been receiving his other medications on a long-standing basis, including diltiazem for 5 years, and electrolytes checked intermittently had not shown previous problems with hypokalaemia. It is therefore unlikely that the patient was hypokalaemic at the time of the prolonged QTc interval.

The patient had no recent history of cardiac abnormalities apart from a myocardial infarct 6 years previously and long-standing hypertension. The patient had been on long-standing medication and there was no evidence of a prolonged QTc while on these medications. Although the patient had symptoms suggestive of dementia with Lewy bodies he did not fulfil the criteria for a diagnosis of probable dementia with Lewy bodies ( McKeith et al, 1996).

In conclusion, this patient had evidence of a normal QTc interval prior to the introduction of the rivastigmine and developed a prolonged QTc while on the treatment which reverted to normal on discontinuation of the drug. His concomitant medication had been long-standing, he had no recent history of cardiac abnormalities and his previous hypokalaemia secondary to diarrhoea had been corrected. Therefore, we suggest there is a possibility of a causal relationship between rivastigmine and prolonged QTc interval. Independently, Novartis have received two isolated reports of QT interval prolongation, which the company have attributed to confounding factors such as co-medication and electrolyte abnormalities as well as insufficient/discrepancies in documentation (J. Collins (Novartis), personal communication, 2001).

I agree with Dr Inglis that the cholinesterase inhibitors are an invaluable part of our limited armamentarium in managing people with dementia but as with any new treatment only when a large number of patients are treated, many of whom will be taking multiple medications, have different comorbidities and be subject to other conditions that were not represented in the original trial population, will adverse effects become manifest that were otherwise not recognised, appreciated or expected. It is important that clinicians monitor, document and report adverse events. Unfortunately, experience demonstrates that this is frequently lacking and can result in the delayed recognition of potentially serious side-effects and interactions.