4 May 2006. Common sense says that early treatment of a disease provides a better chance of success. In schizophrenia, early intervention seems promising enough—studies show that getting patients under care early leads to better outcomes, and the longer psychosis goes untreated, the worse the outcome once treatment begins. But can this strategy be extended to the schizophrenia prodrome, the period of diminished functioning and mild symptoms that typically precedes a psychotic break? Several new papers highlight the promise and pitfalls of pre-psychosis treatment, which remains fraught with practical problems of recognizing impending illness, and controversy over the potential for unnecessarily medicating people who may not be on the road to psychosis. (Ed. Note: For a recent overview of this field, see the British Journal of Psychiatry’s August 2005 supplement, subscription required until August 2006.)
Thomas McGlashan of Yale University in New Haven, Connecticut, and collaborators in the U.S. and Canada, report in the May issue of the American Journal of Psychiatry that treatment of adolescents at high risk for developing schizophrenia with the antipsychotic drug olanzapine can relieve prodromal symptoms, and may delay a first episode of psychosis. Despite an impressive halving of the rate of progression to psychosis in the drug-treated group, the study was inconclusive because of a high dropout rate, with nearly half the subjects leaving the study during the year-long treatment phase. Fatigue and significant weigh gain were major adverse side effects.
The current issue of the American Journal of Psychiatry also features a report from Ingrid Melle of the University of Oslo and collaborators in the U.S. and Denmark on a community education program aimed at helping people identify the early symptoms of schizophrenia. Their results show that the program led to reduced suicidal behaviors in patients reporting to clinics. Both studies point to the benefits of getting patients into treatment at the earliest possible point, and will spur further research on the type and timing of interventions that yield the optimum risk-benefit ratio. A large part of that optimization requires the sensitive and accurate identification of adolescents on the verge of developing schizophrenia, and a study published online recently in Schizophrenia Research, from Alison Yung and Patrick McGorry’s group in Melbourne, Australia, shows that there is still some way to go on that count.
Message from olanzapine trial is promising, but incomplete
The olanzapine trial, undertaken at four centers in Connecticut, North Carolina, Ontario, and Alberta, and funded by a grant from Eli Lilly and Company (manufacturers of olanzapine, or Zyprexa®) recruited help-seeking individuals who were defined as possibly prodromal for schizophrenia. The original framework for such determinations was established by Yung and McGorry (Yung and McGorry, 1996), who identified three “syndromes”: mild positive symptoms, intermittent psychosis, or having a close relative with disease (genetic risk) plus deteriorating function. McGlashan and colleagues have subsequently modified these criteria in creating a data collection instrument (the Structured Interview for Prodromal Symptoms, or SIPS, version 4.0 available from McGlashan) to highlight recent change and to be able to track symptoms over time (see Miller et al., 2003). Most of the individuals in the current study met the attenuated positive symptoms syndrome criteria, meaning that they were experiencing positive symptoms of recent onset that were detectable and ratable but had not reached the clinical definition of psychosis.
Even though the prodrome condition is relatively uncommon (as is schizophrenia), the investigators managed to recruit 60 subjects who were randomized to receive olanzapine or placebo daily for one year, and then observed for an additional year of follow-up off the drug. The power of the study to detect differences in conversion to psychosis was seriously compromised, however, by a 45 percent dropout rate in the first year. Coupled with an average of 27 percent conversion to psychosis, that left only 17 patients completing the 2-year study, eight from the placebo group and nine from the olanzapine group.
Disregarding dropouts, in the first year, the rate of conversion to psychosis, the primary study measure, was 38 percent (11 patients) in the placebo group and 16 percent (five patients) in the drug group. The group difference barely missed statistical significance (p = 0.08), but the trend agrees with earlier studies showing that prodromal interventions, whether pharmacological or cognitive behavioral, do reduce the conversion to psychosis, at least in the short term (McGorry et al., 2002; Morrison et al., 2004). The drug seemed to delay, rather than prevent, conversion, since the rate of conversion in both groups equalized after treatment was discontinued, although the numbers of subjects left in the study by that time were too small to permit any statistical analysis. Consistent with a delay effect, olanzapine treatment significantly reduced some prodromal symptoms, and when the drug was stopped, symptoms worsened for all the patients still in the study.
Interestingly, the five conversions in the treatment group occurred within the first 4 weeks of treatment, and none were seen in the remaining 48 weeks. The five converters were also the most symptomatic, suggesting that perhaps they were too close to converting for the drug treatment to take effect.
According to McGlashan and colleagues, when patients in the study did convert to psychosis, the researchers were “surprised at how operationally benign the process was.” By that, they meant that whether coming from placebo or treatment group, in all cases the patients accepted open-label drug treatment, and complied without hospitalization, missing work or school, or disrupting their social networks. The smooth transition was attributed to the established physician-patient relationship, and the ability to start needed treatment as soon as psychosis appeared.
On the risk side of treatment, the most serious side effects of olanzapine were fatigue and weight gain. Olanzapine-treated patients gained an average of nearly 20 pounds, although this was not accompanied by metabolic disturbances in the first year. The possible long-term health consequences of weight gain and the sensitivity of adolescents to weight gain make this side effect troubling. Based on this study, the number needed to treat is 4.5, which means that between four and five people meeting the criteria for prodromal schizophrenia would need to be treated to prevent one conversion per year of treatment.
Despite the difficulties in recruitment and retention in this trial, the authors write that, “Overall, we feel that this clinical trial demonstrates that the benefits of pre-onset identification and treatment outweighs the risks to a degree sufficient to endorse future clinical trials (in academic centers with programs for early detection and intervention in psychosis) so that definitive recommendations on the use of antipsychotics to treat the prodromal phase of schizophrenia can be made.” (Ed. Note: McGlashan’s group is participating in the NIMH-funded North American Prodromal Longitudinal Study (NAPLS), described in the minutes of the September 2005 meeting of the National Advisory Mental Health Council.)
Testing the test Down Under
One way to improve the risk-benefit equation is to refine the identification of at-risk adolescents to home in on those most likely to progress to psychotic disease. That’s precisely the goal of Alison Yung, Patrick McGorry and colleagues, who started the first early intervention clinic in Melbourne a decade ago, and conducted the first randomized trial of intervention in a high-risk population (McGorry et al., 2002). They have worked to define a symptomatic profile of the “ultra-high risk”(UHR) population—patients with a high likelihood of progressing to a first psychotic episode within a short time. Like McGlashan and colleagues, they have developed a study tool, the Comprehensive Assessment of At Risk Mental States (CAARMS), to evaluate possibly prodromal patients. (For a comprehensive discussion of SIPS, CAARMS, and a third tool, the Bonn Scale for the Assessment of Basic Symptoms [BSABS], see Phillips et al., 2005) Studies employing these different assessment tools have found rates of conversion to psychosis of between 10 and 50 percent. However, as Yung and colleagues discuss in their current paper, published online April 19 in Schizophrenia Research, these studies have been done with subjects who were highly self- and/or clinician “selected.” That is, subjects often came from families with histories of mental illness, or were thought to be “prodromal” by someone—a mental health worker, general practitioner, school counselor, family member or even themselves. This contrasts with people who are not thought to be “prodromal” but who are found to have attenuated or subthreshold psychotic symptoms on routine enquiry.
In the general population or in other research populations, the positive predictive value of prodromal symptoms may be at the lower end of this range, according to new work from the Melbourne group. Using the CAARMS, Yung and coworkers identified UHR young people from a group of 292 15- to 24-year-olds who were referred to a general psychiatric service, not all of whom had primary complaints related to prodromal schizophrenia. They found that about 10 percent of subjects the researchers determined to be at ultra-high risk proceeded to psychosis over 6 months.
Somewhat more reassuring was the observation that only one person out of 173 judged to be at low risk progressed during the study. The authors conclude that UHR status is a sensitive (92 percent) predictor of impending progression, as most progressors (12 out of 13) were in the UHR category. But the specificity was only 62 percent, as the majority of UHR adolescents did not progress, at least in the 6-month time frame of the study. The study further identified low functioning as a factor that increased the risk of progression in the UHR group and the sample as a whole.
The results emphasize the need for additional screening criteria that accurately identify the appropriate treatment candidates for early intervention, given the side effects of medication and the potential for personal distress and lasting stigmatization that might occur from an inaccurate “prodromal” designation.
Communities that educate
Even without additional research, opportunities are all around for effective early intervention, if only communities can get the word out, according to a study from Ingrid Melle and colleagues in Oslo, Norway. Their study, also in the May American Journal of Psychiatry, shows that an educational campaign aimed at identifying and destigmatizing the early symptoms of schizophrenia resulted in a drop in the suicidal behaviors of patients in two communities that implemented the program compared to two that did not. The public information campaign resulted in patients appearing at clinics who were significantly younger, had milder symptoms, and a shorter duration of untreated psychosis. The patients reported significantly lower levels of suicidal behavior both in their lifetimes and in the month preceding first treatment contact. The difference was particularly apparent in reduced suicidal plans and attempts in the early-detection communities.
These results put hard numbers to the life-and-death struggle of many with schizophrenia, as outlined by Jill Harkavy-Friedman of New York State Psychiatric Institute in New York City, in an editorial accompanying the study by Melle and colleagues. She notes that suicide accounts for most of the excess mortality and premature death in schizophrenia, and the Melle study represents “an important first step in identifying individuals with psychotic disorders and implementing treatment earlier in the course of illness." Along with McGlashan’s work, she says, the studies “highlight the need for comprehensive universal public mental health interventions” with the goal of early detection and treatment.—Pat McCaffrey.
McGlashan TH, Zipursky RB, Perkins D, Addington J, Miller T, Woods SW, A Hawkins K, E Hoffman R, Preda A, Epstein I, Addington D, Lindborg S, Trzaskoma Q, Tohen M, Breier A. Randomized, double-blind trial of olanzapine versus placebo in patients prodromally symptomatic for psychosis. Am J Psychiatry. 2006 May;163(5):790-9. Abstract
Melle I, Johannesen JO, Friis S, Haahr U, Joa I, Larsen TK, Opjordsmoen S, Rund BR, Simonsen E, Vaglum P, McGlashan T. Early detection of the first episode of schizophrenia and suicidal behavior. Am J Psychiatry. 2006 May;163(5):800-4. Abstract
Harkavy-Friedman JM. Can early detection of psychosis prevent suicidal behavior?
Am J Psychiatry. 2006 May;163(5):768-70. Abstract
Yung AR, Stanford C, Cosgrave E, Killackey E, Phillips L, Nelson B, McGorry PD. Testing the Ultra High Risk (prodromal) criteria for the prediction of psychosis in a clinical sample of young people. Schizophr Res. 2006 Apr 19; [Epub ahead of print] Abstract