Objective. To describe the differences in knee cartilage defects between offspring of subjects with at least one parent with a total knee replacement for severe primary knee osteoarthritis (OA) and controls; and to estimate the heritability of knee cartilage defects in sib-pairs. Methods. Population based, case-control study of 186 matched pairs (mean age 45 yrs, range 26-61) and sib-pair study of 128 subjects from 51 families (115 sib-pairs) within the case-control study. Knee cartilage defect scores (0-4) and prevalence (a cartilage defect score ≥ 2) were assessed at the patellar, tibial, and femoral sites by processing images acquired using T1 weighted fat-saturated magnetic resonance imaging. Heritability was estimated using the SOLAR genetic analysis program. Results. The prevalence of knee cartilage defects was surprisingly high (50% scored ≥ 2 in any site). Compared to controls, offspring had higher knee cartilage defect scores and prevalence in tibiofemoral (4.39 vs 4.01, p = 0.003; 41% vs 28%, p = 0.009), patellar (1.32 vs 1.10, p = 0.031; 35% vs 26%, p = 0.075), and whole (5.71 vs 5.10, p = 0.002; 57% vs 42%, p = 0.007) compartments. These all became nonsignificant after adjustment for knee pain and radiographic OA. In the sib-pair component, knee cartilage defects had heritability for scores and prevalence, respectively, of 38% (p = 0.072) and 47% (p = 0.082) for tibiofemoral, 52% (p = 0.009) and 78% (p = 0.025) for patellar, and 43% (p = 0.038) and 68% (p = 0.072) for the whole compartments. These estimates became weaker at tibiofemoral and whole compartments after adjustment for bone size, knee pain, and radiographic OA. Conclusion. Knee cartilage defects are common, have a genetic component that is linked to the genetic contribution to knee pain and bone size, and may have a role in the genetic pathogenesis of knee OA.