American Journal of Medical Genetics (Neuropsychiatric Genetics) 114:509–511 (2002) Brief Research Communication No Evidence for Phenotypic Variation Between Probands in Case-Control Versus Family-Based Association Studies of Schizophrenia Anil K. Malhotra,* John A. Bates, Judith Jaeger, Georgios Petrides, Delbert G. Robinson, Robert M. Bilder, and Katharine W. Nassauer Hillside Hospital, Psychiatry Research, Glen Oaks, New York Traditional case-control genetic association studies utilizing unrelated probands are often used interchangeably with family- based designs to detect genes for complex psychiatric disorders. This strategy may be limited, however, if significant phenotypic variation exists between probands enrolled in these two types of studies. The present study compared 37 probands enrolled in a case-control study of schizophrenia with 37 age-, sex-, and ethnically matched pro- bands enrolled in a family-based study of schizophrenia. Age of onset of illness was compared as well as performance on a battery of cognitive tests assessing atten- tion, working memory, executive function, and verbal memory. Results revealed no significant differences in age of onset be- tween the two groups or on any measure of cognitive performance. These data do not support reports of significant phenotypic variation between probands in case-control and family-based studies, and suggest that studies utilizing family-based approaches may be used to replicate reports of asso- ciation made with case-control designs in schizophrenia. ß 2002 Wiley-Liss, Inc. KEY WORDS: phenotype; genotype; asso- ciation; schizophrenia; cog- nition INTRODUCTION Traditional case-control association study designs involving comparison of unrelated probands to healthy volunteers are often used interchangeably with family- based designs in the search for susceptibility genes for complex psychiatric disorders. Case-control studies offer the advantages of: 1) increased feasibility of re- cruitment of subjects, as family members are not required to be available; 2) greater power than family- based studies for the study of gene-gene and gene- environment interactions; 3) the genotyping require- ment is two-thirds of that required in family-based association designs that involve two parents and a proband (such as the transmission disequilibrium test (TDT) [Spielman et al., 1993]; and 4) in most cases, greater statistical power than in family-based techni- ques [Risch and Teng, 1998]. Family-based studies, however, are valuable because they are not susceptible to spurious associations resulting from population stratification between case and control groups [Mal- hotra and Goldman, 1999]. Therefore, it is not uncom- mon that family-based studies are utilized to replicate reports of case-control associations between a genetic locus and psychiatric disorder. A major assumption of such replication attempts is that there are no significant differences in the pheno- type being studied with case-control and family-based designs. However, Brunn and Ewald [1999], in a study of 61 probands derived from a case-control study (CCPs) and 19 probands derived from a family-based study (FBPs) of bipolar disorder, suggested that that there may be important demographic or clinical variation, such as age of onset of illness, between the two groups. In a larger study, Schulze et al. [2001] recently compared demographic and clinical characteristics of 122 CCPs and 54 FBPs with bipolar disorder and reported that FBPs had a significantly younger age of onset, as well as decreased frequency of suicidal behavior. However, subjects in the two groups were not matched for age at interview (47 years for CCPs vs. 31 years for FBPs) or gender (65% female for CCPs *Correspondence to: Anil K. Malhotra, M.D., Associate Direc- tor, Psychiatry Research, Hillside Hospital, 75-59 263rd St., Glen Oaks, NY 11004. E-mail: malhotra@lij.edu Received 30 October 2001; Accepted 6 March 2002 DOI 10.1002/ajmg.10479 ß 2002 Wiley-Liss, Inc.