Sample Selection in Finnish Adoptive Family Study

To the Editor: The article by Pekka Tienari, M.D., et al. (1) made an important contribution to our understanding about the genetics of schizophrenia spectrum disorders. The higher prevalence of schizophrenia among the offspring of mothers with spectrum disorders other than schizophrenia (9.4%) than among the offspring of mothers with schizophrenia (5.1%) indicates a strong genetic relationship between the spectrum disorders and schizophrenia. However, certain methods issues need to be kept in mind while interpreting the results of the study.

The study examined the prevalence of schizophrenia spectrum disorders among the offspring of mothers with schizophrenia spectrum disorders. In their original sample of 170 hospitalized index mothers, 75.9% had schizophrenia and only 1.2% had personality disorders that were considered included in schizophrenia spectrum disorders (2). The authors reported that they would not have missed any community patients with schizophrenia because office and ambulatory clinical care of psychotic persons is rare in Finland. However, this is not true of patients with other disorders that form the putative spectrum, particularly the personality disorders. Their findings among the control mothers, as reported in their earlier article (2), suggest that their sampling technique would have missed a large number of mothers with such disorders. The 201 control mothers were unscreened and formed an epidemiological control group that had a full array of psychiatric illnesses, as found in the community. One percent had schizophrenia, and 1% had affective psychosis; 5% had personality disorders within the putative spectrum. In other words, in the community that they studied, the prevalence of such personality disorders was several-fold higher than that of schizophrenia. However, this group was seriously underrepresented compared to those with schizophrenia in their sample of index mothers (1.2% versus 75.9%). This large difference remained even after they included the control mothers who had spectrum disorders in their final index group. Thus, their final sample of high-risk offspring consisted of children born to seriously ill mothers (median hospitalization of 567 days spanning a median of seven hospitalizations, with a median Global Assessment Scale score of 20) rather than to a representative sample of mothers who had other disorders of the spectrum, especially the personality disorders.

In the light of this, we suggest that the conclusions from the study are modest: as a group, schizophrenia spectrum disorders are more prevalent among offspring born to mothers with spectrum disorders of a severe nature.