Schizophrenia Research Blog: Olanzapine vs. Haloperidol?

December 24, 2004

Olanzapine vs. Haloperidol?

Effectiveness and Cost of Olanzapine and Haloperidol in the Treatment of Schizophrenia: A Randomized Controlled Trial

Robert Rosenheck; Deborah Perlick; Stephen Bingham; Wen Liu-Mares; Joseph Collins; Stuart Warren; Douglas Leslie; Edward Allan; E. Cabrina Campbell; Stanley Caroff; June Corwin; Lori Davis; Richard Douyon; Lawrence Dunn; Denise Evans; Ede Frecska; John Grabowski; David Graeber; Lawrence Herz; Kong Kwon; William Lawson; Felicitas Mena; Javaid Sheikh; David Smelson; Valerie Smith-Gamble

JAMA, Nov 2003; 290: 2693 - 2702.

Article 2:
Robert A. Rosenheck
Open Forum: Effectiveness Versus Efficacy of Second-Generation Antipsychotics: Haloperidol Without Anticholinergics as a Comparator
Psychiatr Serv 56:85-92, January 2005

Article 1 is a study that came out last year which challenges the conventional wisdom that the newer medications are inherently better than properly dosed older medications. While many studies are funded by pharamaceutical companies and tend to favor newer medications in their outcomes, this study was designed to give both the newer medications (olanzapine/Zyprexa) and an older medication (haloperidol/Haldol) a fair chance to be dosed as the clinician in the study wanted to and was given a liberal ability to utilize medications to treat the common side effects. The authors wanted to determine if the newer medications were more effective, had fewer side effects and what the difference would be in overall costs to the Veterans Affairs System for patients on either treatment. Olanzapine costs over $4000/year versus approximately 20 dollars per year for haloperidol. However, if it is true that there is better efficacy and/or fewer side effects from the newer medication (i.e. if it reduced the number of hospital days for the average patient per year) it would make fiscal sense as well as clinical sense to prescribe it automatically and as a first line treatment. However, if that was not the case, then perhaps the conventional wisdom required challenging.

The data prior to the study was already controversial. A leading authority on the quality of research, the Cochrane group, did an analysis where they looked at all the data regarding the differences between olanzapine and haloperidol. While most studies published favored olanzapine, the Cochrane group found that there were sufficient flaws in enough of the studies that they advocated restraint and a more cautious approach. They found that with the number of patients who left studies early, it was hard to draw a strong conclusions that olanzapine was in fact better.

To put their question to the test, the authors recruited approximately 300 patients in the VA system. Patients were randomly assigned to receive either 5-20mg of olanzapine or 5-20mg of haloperidol. The pills looked identical so that the study could be blinded (to help eliminate potential bias.) Also, patients on haloperidol were given 1-4mg of benztropine (Cogentin) automatically, while the olanzapine patients received a placebo. However, both groups could add more active benztropine if there were side effects that warranted such treatment. The authors recorded data on clinical response, side effects, quality of life and overall costs to the VA and to society for the patients. They also tested patients for memory and other tests of their ability to think.

Ultimately, the authors found that the group that received olanzapine and the group that received haloperidol were nearly identical. There were statistically significant differences in a few areas though. The group that received olanzapine performed better on 2 memory tests and also reported less akathisia (severe restlessness) than the group on haloperidol. However, the olanzapine group had a significantly higher amount of weight gain, problems with cholesterol and with diabetes than did the haloperidol group. Lastly, the costs for the olanzapine were significantly higher than the haloperidol group.

The authors conclude that the difference between the groups, being so minimal, means that it is likely that starting a patient with haloperidol and well managed benztropine is an acceptable alternative to starting directly with the more expensive medication. However, they note a few limitations to this study. First, it was conducted exclusively at the VA and with a predominantly male population. This may limit the ability to generalize the results to the overall population. Second, there were many patients recruited who did not participate and many that dropped out of the study after they started. While they followed many of the dropouts, there were some lost to follow-up. This can skew the data towards finding less of a difference between the groups, though the patients that dropped out were equally in the olanzapine and haloperidol group. Also, there were overall about half of the number of patients they initially wanted to recruit. While they did have enough to make their statistics meaningful, they would have been able to make stronger arguments with more patients. Lastly, the overall cost/benefit analysis is still unknown. While the costs were higher with olanzapine, it is hard to determine the overall costs to society and to the patients or the VA. Additionally, the side effect profiles are different between the medications. For some, weight gain is not an issue or can be managed more by diet than for others. While for other patients, having movement or cognitive effects can be more debilitating. It is also possible that the use of benztropine may have contributed to some of the cognitive flattening seen in the haloperidol group as benztropine can have a negative effect on cognition (ability to think.) However, these are all reasons why the debate rages strongly as to the best way to start a patient on antipsychotic medication. It is not right or wrong to start on haloperidol or any other medication first, as long as the patient and physician work closely to assess that the started medication is in fact right for that patient. If it isn’t, it is important to know that it is ok to switch to something that works better, especially if a full trial does not work on an original medication. However, once a medication works, it is useful to stick with it rather than change based on these results especially. Lastly, it goes without saying that this article was/is quite controversial. However, it is an interesting perspective and one that is not frequently presented.

Funding/Support: This study was supported by Lilly, which provided study drug and placebo, and the VA Cooperative Studies Program.

Role of the Sponsor: Employees of Lilly (Alan Breier, MD, Robert Obenchain, PhD, and John Kreuger) participated in the study design and commented on the analyses and on the manuscript. The analyses and writing of the manuscript were carried out by the authors independent of the sponsor.

Article 2 is a very interesting discussion of the factors that may have caused the results of studies that compared haldol with newer antipsychotics to come out so strongly in favor of the newer drugs. While there is always excitement, and also financial pressures from pharmaceutical companies, when new drugs come out, perhaps the excitement got in the way of the best science.

The second article describes a research concept called "effectiveness vs. efficacy." When a drug has "efficacy" it means that it can be demonstrated under controlled circumstances to work better than a comparator drug. To demonstrate "effectiveness" a drug must demonstrate that it works better than a comparitor in regular use, in essence, how it would be used in the general community. In order to perform cleaner reserch, many patients or other treatments must be excluded when doing a trial such that there are only few variables being manipulated and an explanation can therefore be lent to those manipulations. What Rosenheck argues in article 2 is that most trials that compared new antipsychotics with haldol failed to utilize cogentin (an anti-side effect medication) appropriately and therefore patients placed on haldol often had unnecessarily, and unfairly, high levels of side effects compared with what they might have had if they were treated more appropriately. On the flip side, the use of Cogentin (benztropine) is associated with mental slowing, often a side effect that patients do not tolerate. However, as noted above, newer medications have greater cost and have other side effect (weight gain, etc.) Overall, this research is helpful as we have more experience now with the newer medications. Maybe it is not such a bad thing to use older medications, if they are used appropriately? What will work for some patients will not for others and so it is important to keep the factors in mind that will make one medication, and its inherent possible side effects, a better choice for the individual, rather than contemplating the benefits for an entire population at once.

Click here to look up this article on PubMed

Click here for article 2 abstract on PubMed

Posted by Jacob at December 24, 2004 05:49 AM | TrackBack


Post a comment

Please enter this code to enable your comment -
Remember Me?