BACKGROUND

The glitazones (thiazolidinediones) bind to a nuclear receptor called peroxisome proliferator activated receptor gamma (PPARγ), mainly expressed in adipocytes and (at lower levels) in the liver and skeletal muscle.2 The resulting activation of the PPARγ affects insulin-sensitive genes, which regulate glucose and lipid metabolism and other metabolic functions.5 Resulting effects include increased insulin sensitivity and glucose uptake, and reduced hepatic gluconeogenesis.6^,7 Rosiglitazone is associated with dyslipidaemia (e.g. raised LDL and triglyceride levels).3^,8 Glitazones do not directly stimulate insulin secretion.

GLYCAEMIC CONTROL

Glitazones as monotherapy

A meta-analysis of 23 randomised placebo-controlled trials comparing monotherapy with pioglitazone or rosiglitazone for 12–26 weeks in a total of 8,550 patients with type 2 diabetes found that each drug similarly reduced HbA_1c levels more than did placebo, in a dose-dependent manner (pioglitazone 30mg daily by 0.99% and 45mg daily by 1.21%; rosiglitazone 4mg daily by 0.90% and 8mg daily by 1.50%), and increased body weight by about 3.0kg.9 We know of no published randomised controlled trials showing that glitazones affect diabetic retinopathy, nephropathy or neuropathy.

Pioglitazone vs. other oral hypoglycaemic drugs

A systematic review, published in 2006, pooled 22 randomised controlled trials of pioglitazone alone, or used either with any other oral hypoglycaemic medication or insulin.10 The studies lasted at least 24 weeks and included a total of around 6,200 patients with type 2 diabetes randomised to pioglitazone. The reviewers concluded that the studies “did not provide convincing evidence that patient-oriented outcomes like mortality, morbidity, adverse effects, costs and health-related quality of life are positively influenced by this compound [pioglitazone].” They also found that “metabolic control measured by...HbA_1c as a surrogate endpoint did not demonstrate clinically relevant differences to other hypoglycaemic drugs”. Oedema was more likely with pioglitazone (relative risk [RR] 2.86, 95% CI 2.14 to 3.18; p<0.00001).

Two subsequent randomised controlled trials have compared pioglitazone monotherapy and pioglitazone plus metformin.11^,12 In one, a double-blind study involving 283 patients with type 2 diabetes given either pioglitazone 30–45mg daily or gliclazide 80–320mg daily for 1 year, HbA_1c levels fell by 0.79% in both groups.11 In the other, non-blinded, study, which lasted 28 weeks and involved 203 adults with poorly controlled type 2 diabetes (HbA_1c 7.5–10%) on metformin alone, the addition of pioglitazone or glimepiride produced similar overall falls in HbA_1c (by 1.23% and 1.30%, respectively, p = 0.4825).12

Triple therapy was assessed in a non-blinded randomised controlled trial involving 62 patients with type 2 diabetes (HbA_1c >8.0%) given either pioglitazone or bedtime NPH (intermediate-acting) insulin, in addition to metformin and sulphonylurea, for 16 weeks.13 HbA_1c levels fell similarly in each group (by 1.9% and 2.3%, respectively), although fewer patients on pioglitazone developed hypoglycaemia (37% vs. 68%, p = 0.02, respectively).

The combination of pioglitazone and insulin was assessed in a 28-week double-blind controlled trial, in which 690 patients on a stable insulin dose (above 30 units daily for at least 30 days) were randomised to pioglitazone either 30mg or 45 mg once daily for 24 weeks.14 Dose-dependent decreases in HbA_1c from baseline were seen in both groups (by 1.17% with 30mg and 1.46% with 45mg). Around 13% of participants developed peripheral oedema, and 4 patients developed heart failure.

We know of no published randomised trials comparing insulin plus pioglitazone with insulin plus metformin.

Rosiglitazone vs. other oral hypoglycaemics

A systematic review pooled data from 18 randomised controlled trials of rosiglitazone alone, or used either with any other oral hypoglycaemic medication or insulin.15 The studies lasted at least 24 weeks and included a total of 3,888 patients with type 2 diabetes randomised to rosiglitazone. The reviewers found that the studies “did not provide evidence that patient-oriented outcomes like mortality, morbidity, adverse effects, and health-related quality of life are positively influenced by this compound [rosiglitazone].” They also found that “metabolic control measured by...HbA_1c as a surrogate endpoint did not demonstrate clinically relevant differences to other hypoglycaemic drugs”. They highlighted evidence of an increased prevalence of cardiovascular events with rosiglitazone, as well as an increased fracture rate in women. Oedema was more likely with rosiglitazone (RR 2.10, 95% CI 1.72 to 2.55).

Triple therapy was assessed in a 24-week non-blinded randomised trial, in which 217 patients on 50% or more of maximal-dose sulphonylurea and metformin received add-on insulin glargine 10 units daily (force-titrated to a target fasting plasma glucose of 5.5–6.7mmol/L or less) or rosiglitazone 4 mg daily (increased to 8mg daily any time after 6 weeks if fasting plasma glucose was above 5.5 mmol/L).16 HbA_1c levels fell similarly in each group (by 1.7% and 1.5%, respectively).

Combination formulations

Products comprising pioglitazone plus metformin (▼Competact) or rosiglitazone plus metformin (▼Avandamet) are being promoted as improving adherence to therapy. However, we know of no published evidence that they improve either adherence or HbA_1c concentrations compared to giving the two drugs separately. Also, they are not licensed in the UK as initial therapy.17^,18

CARDIOVASCULAR EVENTS

The efficacy of pioglitazone (15–45mg daily) in secondary prevention of macrovascular events was investigated in a double-blind randomised placebo-controlled trial involving 5,238 patients with type 2 diabetes and pre-existing macrovascular disease.19 Average observation time was 34.5 months. The treatment groups did not differ significantly on the primary outcome measure (a composite of all-cause mortality, non-fatal myocardial infarction, acute coronary syndrome, cardiac intervention, stroke, major leg amputation, bypass surgery or revascularisation of the leg). Sub-group analyses suggested that pioglitazone reduced the recurrence of myocardial infarction by 28% (hazard ratio [HR] 0.72, 95% CI 0.52–0.99; p = 0.045)20 and the recurrence of stroke by 47% (HR 0.53, 95% CI 0.34–0.85; p = 0.0085).21

Large randomised controlled trials have not individually found an effect on serious cardiovascular events with rosiglitazone, but were underpowered to do so.22^,23 A recent meta-analysis of 42 randomised trials comparing rosiglitazone with control therapy, lasting at least 24 weeks and including a total of 27,843 patients, found an increased likelihood of myocardial infarction with rosiglitazone (86 infarctions vs. 72 in the control group, odds ratio 1.43, 95% CI 1.03 to 1.98; p = 0.03).24 There was no significant increase in the likelihood of cardiovascular death. The study had limitations, however, including its pooling of data from trials with different populations (e.g. patients with type 2 diabetes and those at high risk of such disease) and including some trials lasting only 24 weeks. The meta-analytical approach used has also been criticised and the data re-analysed by reviewers, who thereby concluded that the “risk for myocardial infarction and death from cardiovascular disease for diabetic patients taking rosiglitazone is uncertain”.25

In response to the initial meta-analysis,24 other researchers conducted an unscheduled interim analysis (mean follow-up 3.75 years) of a non-blinded randomised controlled trial involving 4,447 patients with type 2 diabetes given adjunctive therapy comprising rosiglitazone plus either metformin (if already on sulphonylurea) or a sulphonylurea (if already on metformin).26 This analysis revealed no difference in a composite primary outcome of hospitalisation or cardiovascular death (HR 1.08, 95% CI 0.89–1.31) but was underpowered to do so.27 Also, if this analysis is intergrated into the initial meta-analysis, there is still a substantially increased risk for myocardial infarction (HR 1.33).27

Another meta-analysis, including only randomised controlled trials with a follow-up of 1–4 years (four trials with a total of 14,291 patients) found that, compared with control therapy to prevent or treat type 2 diabetes, rosiglitazone increased the likelihood of myocardial infarction (1.46% vs. 1.05%, RR 1.42; 95% CI 1.06 to 1.91; p = 0.02) and heart failure (1.59% vs. 0.79%, RR 2.09; 95% CI 1.52 to 2.88; p<0.001).28 There was no significant increase in likelihood of cardiovascular death.

A meta-analysis of 19 double-blind randomised controlled trials, lasting 4 months to 3.5 years and including a total of 16,390 patients with type 2 diabetes found a reduction for the composite primary outcome of death, myocardial infarction or stroke with pioglitazone compared to control (placebo or active comparator) therapy (4.4% vs. 5.7%, HR 0.82, 95% CI 0.72 to 0.94; p = 0.005).29 The advantage was greatest when the comparator was placebo (HR 0.09, 95% CI 0.01–0.84) or rosiglitazone (HR 0.39, 95% CI 0.08–2.00), and was not seen where the comparator was a sulphonylurea (HR 1.02, 95% CI 0.7–1.48).

A meta-analysis pooled data from seven double-blind randomised controlled trials of pioglitazone (two trials) or rosiglitazone (five trials) that had cardiovascular death and congestive heart failure as outcomes.30 A total of 20,191 patients with pre-diabetes or type 2 diabetes were included. Compared with controls, glitazones increased the likelihood of congestive heart failure (RR 1.72, 95% CI 1.21–2.42; p = 0.002), but not cardiovascular death.

Regulatory advice

On reviewing pioglitazone and rosiglitazone, the European Medicines Agency Committee for Medicinal Products for Human Use (CHMP) concluded in 2007, that both medicines’ benefits continue to outweigh their risks.31 However, the CHMP has since advised the inclusion of a new warning in the product information for rosiglitazone-containing medicines stating that rosiglitazone is not recommended in patients with ischaemic heart disease or peripheral arterial disease, and is contraindicated in patients with an acute coronary syndrome, such as angina or some types of myocardial infarction.32 The Medicines and Healthcare products Regulatory Agency (MHRA) has also recently advised that rosiglitazone might be associated with a small increased risk of cardiac ischaemia, particularly in combination with insulin, and should be used in patients with previous or current ischaemic heart disease “only after careful evaluation of individual risk”.33 The manufacturer of rosiglitazone has subsequently announced it is to revise its summaries of product characteristics (SPCs) for Avandia and Avandamet to state that rosiglitazone may be associated with an increased risk of myocardial ischaemic events and that, as a precaution, the drug is not recommended in patients with ischaemic heart disease or peripheral arterial disease.34

The US Food and Drug Administration has stated that rosiglitazone product information must carry a ‘black box’ warning of the potential increased risk for myocardial infarction.35

FLUID RETENTION AND HEART FAILURE

There is an increased likelihood of fluid retention and congestive heart failure with glitazones,2-4^,30 particularly when insulin is also used. In one randomised controlled study (in 5,238 patients at high risk of developing cardiovascular disease), heart failure was more likely with pioglitazone (11% vs. 8% with placebo, p<0.0001).19 However, the rates of death due to heart failure were comparable. In another randomised controlled study (in 4,360 patients with type 2 diabetes), the incidence of heart failure with rosiglitazone (1.5%) was comparable to that with metformin (1.3%), but over double that with glibenclamide (0.6%, p⩽0.05).23 In a third randomised controlled study, comparing rosiglitazone and placebo in the prevention of type 2 diabetes (in 5,269 people at high risk of developing the condition), heart failure was more likely with rosiglitazone (0.5% vs. 0.1%, p = 0.01).22

The SPCs of pioglitazone and rosiglitazone contraindicate their use in patients with current or previous heart failure, and recommend that all patients taking the drugs should be monitored for adverse reactions relating to fluid retention, including weight gain and heart failure.3^,4 They also state that the drugs should be discontinued if cardiac status deteriorates. The MHRA has recently reinforced these warnings,33 and stated that heart failure is more likely when rosiglitazone or pioglitazone is combined with insulin.33

BONE DENSITY AND FRACTURE RISK

In an observational cohort study involving 61 people with type 2 diabetes (aged 70–79 years), use of pioglitazone or rosiglitazone was associated with bone loss from the greater trochanter, lumbar spine and whole body in women but not in men.36 Moreover, a double-blind randomised controlled trial, including 4,360 patients with recently diagnosed type 2 diabetes and followed for a median of 4 years, found an increased incidence (p<0.01) of fracture in women taking rosiglitazone as monotherapy (9.30%) compared with metformin (5.08%) and glibenclamide (3.47%).23 This was an unexpected finding. Most of the additional fractures were in the humerus, hand or foot. The three treatment groups did not differ significantly in the overall incidence of fractures in men, or of fractures of the hip or spine in women.

The manufacturer of pioglitazone subsequently pooled trial and surveillance data from over 15,500 patients to calculate a fracture incidence of 1.9 fractures per 100 patient years in women treated with pioglitazone compared to 1.1 fractures per 100 patient years in women on comparator therapy.37 Most of the excess fractures were in the forearm, hand and wrist, foot, ankle, fibula and tibia (i.e. not the hip and spine, the typical sites in post-menopausal osteoporosis). No increased risk of fracture was identified in men. The SPCs for pioglitazone and rosiglitazone now state that the risk of fracture should be considered in the care of patients, especially females, treated with these drugs.3^,4

MACULAR OEDEMA

New-onset or worsening diabetic macular oedema with decreased visual acuity has been reported with both pioglitazone and rosiglitazone. Suggestion (but not proof) of an association with such therapy comes from a retrospective review identifying 30 patients with type 2 diabetes using a glitazone who had both lower extremity oedema that increased since starting the glitazone and clinically significant macular oedema; there were no controls.38 The SPCs recommend that prescribers should consider macular oedema where patients report disturbances in visual acuity (especially with concurrent peripheral oedema).3^,4

WHAT GUIDELINES SAY

In 2003, the National Institute for Health and Clinical Excellence (NICE) advised that the use of a glitazone as second-line therapy added to either metformin or a sulphonylurea was not recommended except in patients who are unable to tolerate metformin and sulphonylurea in combination therapy, or patients in whom either metformin or a sulphonylurea is contraindicated.39 New NICE guidance is expected in May 2008, for which preliminary recommendations40 state that, after discussion with the patient, a glitazone should be considered when glucose concentrations are not adequately controlled (to HbA_1c of <7.5% or other higher level individually agreed), adding it to:

• the combination of metformin and a sulphonylurea where insulin is contraindicated or is likely to be poorly tolerated, or

• a sulphonylurea where metformin is not tolerated, or

• metformin as an alternative to a sulphonylurea where occupation or other issues make the hypoglycaemia risk with sulphonylureas particulary significant.

In an update of the Joint American Diabetes Association and European Association for the Study of Diabetes (ADA/EASD) consensus, glitazones are recommended as a potential second-line alternative to basal insulin or sulphonylurea therapy.41

*As recommended in the BNF.

**Based on information in the Drug Tariff.

CONCLUSION

Evidence supports the use of a glitazone in combination with metformin or a sulphonylurea in patients with type 2 diabetes mellitus who are unsuited to one or other of these drugs. However, there is no convincing evidence that they offer any benefits over metformin or a sulphonylurea in terms of improved clinical outcomes when used as monotherapy. Evidence for their use in triple therapy is also weak, and they should be reserved for patients in whom insulin is contraindicated or is likely to be poorly tolerated. The glitazones increase weight, can cause heart failure and peripheral oedema (especially when given with insulin), and increase fracture rates in women. There is also evidence to suggest that rosiglitazone increases the likelihood of myocardial infarction and cardiovascular disease. The drug is now contraindicated in patients with an acute coronary syndrome and not recommended in those with ischaemic heart disease or peripheral arterial disease.

If a glitazone is thought to be necessary, ▼pioglitazone is probably safer, but should only be used as an adjunct to other hypoglycaemic drugs where there is a contraindication or intolerance to metformin or sulphonylureas. As with rosiglitazone, pioglitazone should not be used if there is any risk of heart failure and should probably be avoided in women at high risk of fracture. Although pioglitazone is licensed for use with insulin therapy, this combination can lead to a higher incidence of weight gain, oedema and potentially heart failure.