If you who have chosen to spend inordinate amounts of your precious time reading these pages over the last year, you will remember the February 2003 (TCR, 1:2) issue in which TCR slammed Zyprexa for causing diabetes. Well, we're going to slam Zyprexa again this year, but not quite as viciously. Things have become a little murkier since then, and it's not as easy to distinguish the good guys from the bad guys.

Now the FDA has jumped into the fray, having officially asked all atypical antipsychotic drug-makers to insert a warning about diabetes in their package inserts. The makers of Zyprexa, Risperdal, and Seroquel have agreed, but both Pfizer (Geodon) and Bristol Myers Squibb (Abilify) are resisting, and are busy preparing data to convince the FDA that their drugs do not mess around with glucose or insulin.

When last year’s article was researched, only one large retrospective study had been published examining this controversial issue (BMJ 2002;325:243). Since then, eight additional studies have hit the stands, most of which have been funded by one or the other of the interested parties.

The nine available studies are charted below:

Table: The Nine Good Published Studies of Diabetes Risk with Zyprexa vs. Risperdal (as of February 2004)



As you can see, seven of these nine studies have fingered Zyprexa as being more likely to cause diabetes than Risperdal. How much more likely? The answer varies, depending on how the results are reported. Generally, though, schizophrenic patients on Zyprexa incur double to triple the risk of diabetes than patients on Risperdal.

Two studies found Zyprexa to be no worse than Risperdal. One of these focused specifically on elderly patients in nursing homes who usually receive very low dose antipsychotics for behavioral issues, and rarely for schizophrenia per se. This study (Etminan, Pharmacotherapy 2003;23:1411-1415) found that atypicals as a group actually were less likely to cause diabetes than benzodiazepines, and that there were no differences between Zyprexa, Risperdal, and Seroquel. Thus, while this study does not speak to the issue of which is the safest atypical for patients with schizophrenia and bipolar disorder, it does reassure us that low dose atypicals probably don't cause hyperglycemia in the elderly.

The other Zyprexa-friendly study was the only study out of nine to be funded by Eli Lilly (Buse, J Clin Epidemiol 2003;56:164-170). While it was well-designed overall, psychiatric diagnoses were not ascertained, and since diabetes risk varies by diagnosis, their reported rates could be confounded in a variety of ways. And even this Lilly-funded study did not give Zyprexa a clean bill of health: Zyprexa-treated patients were 3 times more likely than untreated patients to develop diabetes, but Risperdal was non-significantly worse, 3.4 times the baseline risk.

Warning: You may be plied by drug reps with other study "results" not covered here. Look at the fine print and you'll see that these are unpublished conference abstracts, which have yet to be subjected to the rigors of peer review. The famous "Cunningham VA study" being peddled by Lilly reps falls into this category; if it ever achieves publication, you can be assured that it will receive its moment of fame in these pages as well.

Lest we come across as Risperdal boosters, by the way, let there be no mistake that Risperdal can cause diabetes, though at a lower rate than Zyprexa. A recent paper authored and funded by the FDA identified 131 cases of Risperdal-asociated hyperglycemia from 1993-2002 as reported to the Medwatch program. While less than half of the 289 similar cases reported for Zyprexa over this period, it is still concerning. They noted that Risperdalinduced diabetes usually begins within the first 3 months of treatment, and that it seems to be a disease primarily of the young, with the mean age of onset 34.8 years.

So here’s the TCR bottom line:

TCR VERDICT
Recommendations for Monitoring Atypicals in Schizophrenics

• For Clozaril and Zyprexa: Obtain baseline glucose (fasting preferable, random OK), lipid panel (cholesterol, LDL, HDL, triglycerides), weight. Repeat random glucose at 1 month and 3 months, then yearly. Obtain more frequently as needed in response to significant weight gain or symptoms of diabetes (like urinary frequency, disproportionate fatigue).


• For Risperdal and Seroquel: Baseline labs (as above), repeat at 3 months, then as needed for weight gain and symptoms.


• For Geodon and Abilify: No labs needed, but patients should be hooked up with a primary care doctor for routine health maintenance.