Caught in the trio trap? Potential selection bias inherent to association studies using parent-offspring trios

• Published in: American journal of medical genetics 2001-05, Vol.105 (4), p.351-353
• Main Authors: SCHULZE, Thomas G, MÜLLER, Daniel J, HELD, Tilo, PROPPING, Peter, NÖTHEN, Markus M, MAIER, Wolfgang, RIETSCHEL, Marcella, KRAUSS, Harald, GROSS, Magdalena, BAUER, Irina, FANGERAU-LEFEVRE, Heiner, ILLES, Franciska, OHLRAUN, Stephanie, FIMMERS, Rolf, CICHON, Sven
• Format: Article
• Language: English
• Subjects: Adult; Adult and adolescent clinical studies; Age of Onset; Biological and medical sciences; Bipolar Disorder - genetics; Bipolar Disorder - psychology; Bipolar disorders; Family; Family Health; Female; General aspects. Genetic counseling; Humans; Interviews as Topic; Male; Medical genetics; Medical sciences; Middle Aged; Mood disorders; Psychology. Psychoanalysis. Psychiatry; Psychopathology. Psychiatry; Selection Bias; Suicide, Attempted - statistics & numerical data
• ISSN: 0148-7299; 1096-8628

During the last years, the validity of classic case control studies in psychiatric genetic research has been increasingly under question due to the risk of population stratification problems inherent to this type of association study. By consequence, the application of family-based association studies using parent-offspring trios has been strongly advocated. Recently, however, in a study comparing clinical characteristics between index patients from parent-offspring trios and singleton patients with bipolar affective disorder, the question was raised whether a systematic neglect of case control association studies could lead to a selection bias of susceptibility genes. In a similar approach, we compared demographic and clinical characteristics of 122 singleton bipolar patients with those of 54 bipolar patients derived from parent-offspring trios. The singleton patients did not only present with a higher age of onset, but also with a higher frequency of suicidal behavior and a higher familial loading for suicidality. These findings suggest that the genetic mechanism for disease might be different between trio-based and classic case control samples, where patients are examined whose parents are not available for genetic studies. Thus, giving up case control designs for the sake of family-based association studies could be at the risk of selecting against several genetically determined factors.