Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Polyneuropathy and statins


Myopathy is a recognised risk associated with the use of lipid lowering drugs. In general practice in the UK one estimate is that the incidence of myopathy in users of lipid lowering drugs is 2.3 per 10,000 person years, with a relative risk compared with nonusers of 42 for fibrates and 8 for statins [1]. A new study [2] tells us that polyneuropathy is also likely to be a problem, and that it needs looking at.


This was conducted in a county of Denmark with a population of 465,000. Residents have a civil registration number that is used in discharge prescription registries, so that it is possible to find all residents with a particular disorder, and find out what drugs they have been prescribed.

In a five-year period to the end of 1998, all patients with a discharge of polyneuropathy were examined. Some lived elsewhere, some were diagnosed before the study period, some had predisposing conditions (renal failure, diabetes, thyroid, and others had no proper diagnosis or were wrongly diagnosed. Clinical diagnostic features were distal symmetric sensory symptoms or symmetric motor symptoms and no upper motor neurone signs, or both. Neurophysiological criteria were abnormal conduction in two or more peripheral nerves, with at least one being a leg nerve.

A diagnosis of peripheral neuropathy was only accepted if both clinical and nerve conduction criteria were compatible with the diagnosis. Several levels of confidence were defined for idiopathic polyneuropathy (Table 1).

Table 1: Definition of diagnosis of polyneuropathy



Definite Adequate work up and tested for exclusion diagnoses and conditions, and no apparent cause of neuropathy established
Probable Only sufficient information to rule out alcohol overuse, diabetes and renal insufficiency
Possible Information not sufficient to ascertain presence or absence of any exclusion diagnosis

For each case, all inhabitants of the same sex and age were used to randomly choose 25 control subjects per case.


There were 166 cases (mean age 59 years) of first time diagnosis of polyneuropathy in the five years, of which 35 were definite, 54 probable, and 77 possible. Of these nine (5.4%) had a previous exposure to statins (eight current users), with a median duration of 2.8 years. There were 4,150 controls, of whom 66 (1.6%) had exposure to statins (49 current users).

The relative risk of polyneuropathy for current users was 4.6 (2.1 to 10) for all cases with current use, and 16 (5.7 to 45) for definite cases with current use (Table 2). Odds ratios were higher for more than two years of use compared with less than two years, and for larger numbers of doses than smaller numbers.

Table 2: Statin exposure in all cases and definite cases of polyneuropathy

Statin exposure



Odds ratio

All cases

Never use 157 4084 1
Current use 8 49 4.6 (2.1 to 10)

Definite cases

Never use 27 854
Current use 7 17 16 (5.7 to 45)

The number needed to harm (NNH) based on all patients was calculated as 5,500 (2,200 to 18,500). In those over 50 the incidence of polyneuropathy in the background population was 1.7 per 10,000 person years, with an excess rate of 4.5 per 10,000 person years among those exposed to statins. That is roughly one excess case of polyneuropathy for every 2,200 (880 to 7,300) person years of statin use.


In 1998 about 1% of the Danish population used a statin. It's probably more now, both in Denmark and elsewhere. If a PCO with 100,000 inhabitants had 1% taking statins, a case of polyneuropathy might be expected every second year. That's twice as frequent as a case of myopathy.

In primary care in England in 2001 there were about 13 million prescriptions for statins, at a cost of about £420 million. It is not possible to extrapolate too much from that, other than to conclude that with so much statin use, these rare adverse events will occur and should be noticed. Awareness of that may be important in limiting their impact.

This is a powerful and interesting paper, demonstrating how good government information systems can be used for the good of its population. It is important in understanding risk, though not of causation nor mechanism. Papers in the journal Neurology are available free on the Internet. If you think you want to read this one, then read also an thoughtful accompanying editorial [3].


  1. D Gaist et al. Lipid-lowering drugs and risk of myopathy: a population-based follow-up study. Epidemiology 2001 12: 565-569.
  2. D Gaist et al. Statins and risk of polyneuropathy. A case-control study. Neurology 2002 58: 1333-1337.
  3. M Donaghy. Assessing the risk of drug-induced neurologic disorders. Neurology 2002 58: 1321-1322.
previous or next story in this issue