Three publications, including two recently in Diabetes Care, have indicated that a combination of rosiglitazone and fibrates may lower serum HDL cholesterol in some patients (1–3). It is known that a paradoxical fall in HDL cholesterol may occur with fibrates.
These series of cases include only one case of HDL cholesterol lowering in a patient taking rosiglitazone without a fibrate (1). There are no published case reports of this reaction occurring with other thiazolidinediones, where the reaction resolved when the latter, rather than the fibrate, was withdrawn. Thus, confirmatory evidence of this reaction occurring with rosiglitazone alone and whether it is unique to rosiglitazone or a therapeutic class effect is needed.
The New Zealand Pharmacovigilance Centre recently received a report of HDL cholesterol lowering with rosiglitazone in the absence of a fibrate. A fall in HDL cholesterol from 0.8 to 0.2 mmol/l was described in a 64-year-old patient 3 weeks after starting rosiglitazone. When discontinued, recovery was rapid. Of note is that this patient had experienced a similar reaction to bezafibrate previously with the HDL cholesterol falling from 1.1 to 0.3 mmol/l.
Vigibase, the database of the WHO Collaborating Centre for International Drug Monitoring held at the Uppsala Monitoring Centre was searched for further cases.
There were 14 reports from three countries of lowered HDL cholesterol concentrations associated with rosiglitazone. The reporting of this drug/adverse reaction combination was disproportionately high compared with the background data (4). Five of the patients were not taking a fibrate. Outcome data were provided for all but three patients. Two patients recovered when rosiglitazone was withdrawn, and one of the patients was not taking a fibrate. In two others, rosiglitazone had not been discontinued at the time of reporting. Seven other patients recovered, but no information on drug withdrawal was given. There were eight males and five females aged 31–66 years (sex not stated in one report). Duration of use where stated was 1–8 months, and doses, stated in four reports, were 4 mg or 8 mg daily. Five patients had an associated hypertriglyceridemia, two had elevated cholesterol levels, and one a fall in total cholesterol levels. Two patients had angina or aggravated angina, and diabetes was aggravated in one. Concomitant medicines other than fibrates occurring more than once included sulfonylureas, metformin, ACE inhibitors (4), statins (4), and aspirin (2).
Spontaneous adverse reaction reports thus support the single published observation that rosiglitazone can lower HDL cholesterol whether or not a fibrate is prescribed. The occurrence of lowered HDL cholesterol with rosiglitazone in a patient who had experienced a similar reaction with bezafibrate suggests a common mechanism. Although small changes in HDL cholesterol measurements must be interpreted with caution, the New Zealand case report and others published indicate that the fall can be profound (1).
Vigibase also contains five reports of lowered HDL cholesterol with troglitazone and one with pioglitazone. Only one patient was also taking a fibrate. These reports contain incomplete information about response to medicine discontinuation but provide evidence for a therapeutic class effect that requires further confirmation.
