The Danish have comprehensive data bases that have allowed wide-ranging epidemiologic findings. A group from Denmark and the UK interrogated Danish national registries to create a cohort of 8624 people with schizophrenia born after 1955 and who were at least 18 years old as of 1996. The registries allow for determination of ICD diagnosis and inpatient treatment for schizophrenia, but not treatment-resistant schizophrenia (TRS), which is not a diagnosis recognized in ICD or DSM. To determine which patients were treatment resistant, they applied 3 criteria:
- Treatment with clozapine
- An 18-month period during which a patient was treated non-concurrently for 6 weeks with 2 distinct nonclozapine antipsychotics and subsequently required admission to hospital (clozapine eligible)
- 90 days of polypharmacy, i.e. treatment with two or more antipsychotics
Although the use of clozapine may be the best proxy criterion for treatment resistance, the investigators estimate that only a third of Danish patients with TRS receive clozapine, hence the need for other proxy criteria.
The total number of treatment-resistant patients as fulfilled by any of the 3 criteria was 5900. The largest group was those who received 90 days of polypharmacy (n = 3773), which is the least specific criterion. The researchers used a Cox proportional hazards regression to compare the patients who met any of the TRS criteria to those who met none and found the following characteristics more prevalent in the TRS group:
- Younger age at onset
- Residence in less urban area
- Hospital admission at time of diagnosis
- Comorbid mental disorders
- Paranoid subtype
- Early parental loss
- History of substance use disorder
- Receiving disability benefit
- Any psychotropic medication prescribed during the year preceding diagnosis
- At least one suicide attempt
Sensitivity analyses examined the subcohorts of clozapine-treated plus clozapine-eligible patients (n = 1703). This combined group was more likely to have educational attainment beyond primary level but had the same likelihood of receiving a long-term disability benefit as the non-TRS group.
Characteristics that did not differ in the broad TRS cohort (n = 5900) compared with the non-TRS cohort included:
- Season of birth
- Paternal age
- Living with a partner
- History of a violent offense
The recognized risk factors for schizophrenia such as sex and season of birth did not distinguish TRS in this Danish cohort, but urban residence is a well-known risk factor which surprisingly was less common in the TRS subcohort. Other known risk factors that could not be explored in the registry-based design include birth complications, prenatal infections, and duration of untreated psychosis. The data presented however point to a more severe illness from the outset as reflected by medication use prior to, younger age, and hospital admission at the time of diagnosis.
The strength of the study is its size, which the population-based method permits, but this is also a basis for its weakness, namely the inability to precisely define a cohort of people with TRS. Unfortunately, none of the factors identified here would allow a prospective determination of TRS, i.e. who should receive clozapine. For that we require patient-specific biological markers.
Wimberley T, Støvring H, Sørensen HJ, Horsdal HT, MacCabe JH, Gasse C. Predictors of treatment resistance in patients with schizophrenia: a population-based cohort study. Lancet Psychiatry. Published online Feb 24, 2016. Abstract