Familial risk ratios for extreme obesity: implications for mapping human obesity genes JH Lee, DR Reed and RA Price Center for Neurobiology and Behavior, Department of Psychiatry, University of Pennsylvania School of Medicine, Philadelphia, PA 19104, USA OBJECTIVE: To determine familial risk ratios for extreme obesity to aid in the design of obesity linkage studies. DESIGN: Family study of obesity SUBJECTS: 2349 ®rst-degree relatives (parents and siblings) of 840 probands who are members of the National Association to Advance Fat Acceptance (NAAFA) and 5851 participants of the ®rst phase of the National Health and Nutrition Examination Survey III. METHODS: Computed age±gender standardized risk ratios (SRRs) for obesity in relatives categorized by the level of obesity in the index case (proband). MEASUREMENT: Body mass index (BMI) (kg/m 2 ) RESULTS: The risk of extreme obesity (BMI  40) in relatives of extremely obese women (BMI  40) was more than ®ve times greater than in the population; furthermore, the risk of obesity in relatives was approximately linearly associated with the degree of obesity in the proband. The risk of thinness in relatives of obese individuals was substantially lower than in the general population. CONCLUSION: Because the familial risk ratio for extreme obesity is higher than for moderate levels of obesity, the number of families required to achieve adequate statistical power in gene mapping studies of obesity can be reduced substantially by focusing on family members of extremely obese individuals (BMI  40). Keywords: familial risk ratio; lambda; body mass index; obesity; linkage; family study Introduction Traditional methods in genetic epidemiology, such as family, twin and adoption studies have identi®ed genetic susceptibility as a major contributing factor to the familial aggregation of obesity; 1 these studies have provided the impetus for a new area of investi- gation designed to map and characterize genes which predispose humans to be obese. The identi®cation of such genes will be challenging, in large part because the expression of obesity may depend on numerous factors: multiple obesity-susceptibility loci (and their potential interactions), reduced penetrance and envir- onmental variables such as availability of a calorically dense diet. Several segregation studies report results that are consistent with major gene in¯uence on obesity. 1,2,3 Although some genes eventually may be found to have large effects on some obesity pheno- types in humans, already more than 70 genes or chromosome regions have been implicated as having some role in obesity in humans or animal models. 4 The dif®culties facing studies designed to identify obesity-susceptibility genes in humans are evident from the inconsistencies in results across studies. Differences may be due to: (1) actual locus heterogeneity (for example, differ- ent genes or combinations of genes leading to obesity in different populations); (2) inadequate statistical power to detect poly- genic loci with small effects (high rates of false negatives); (3) low prior probability that any particular gene will have a measurable effect in a particular sample (high rates of false positives) or (4) environmental heterogeneity (for example, differences in exposure to risk and protective factors). Furthermore, it has been shown that replication of susceptibility loci for a complex trait requires a longer period of time (for example, a larger sample size) than the initial detection of linkage. 5 Given these challenges facing linkage studies of obesity, it is necessary to employ a study design that maximizes the feasibility of localizing obesity-sus- ceptibility loci and that allows for replication of those ®ndings. Risch 6 has shown that, for both monogenic and polygenic traits, models of inheritance can be examined by computing a risk ratio, which compares the prevalence of a trait in relatives with the preva- lence of the trait in the population. The more extreme the risk ratio, the more likely a gene for a trait will be Correspondence: R Arlen Price, PhD, Center for Neurobiology and Behavior, Department of Psychiatry, 415 Curie Blvd, CRB 135b, University of Pennsylvania, Philadelphia, PA 19104, USA. 12 February 1997; 5 June 1997; 12 June 1997 International Journal of Obesity (1997) 21, 935±940 ß 1997 Stockton Press All rights reserved 0307±0565/97 $12.00