Unlike in the present study, adjusting for BMI resulted in only minimal attenuation of the association between HBM and hip OA compared with age and gender adjustment alone, consistent with evidence that BMI is less strongly associated with hip than knee OA. However interestingly, following age, gender and BMI adjustment, overall odds

ratios for OA in HBM cases vs. combined controls were similar at the hip (1.52 [1.09,2.11]) [12] and the knee (1.62 [1.22,2.16]), suggesting that the increased risk of OA conferred as a direct result of HBM (independent of BMI) is similar at both joint sites. These findings suggest firstly that increased BMD is an important risk factor for OA at both the hip and knee, and secondly that increased

bone formation, as evidenced by osteophytosis, drives this association at both joint sites. Extreme BMD elevation, as seen Ipilimumab research buy in our HBM cases, is likely to be primarily genetically determined. Therefore an important consideration is the extent to which HBM individuals may be predisposed to “standard” (previously termed “common garden-variety” [43]) OA, as opposed to a distinct OA subtype arising from the pleiotropic effects of rare genetic variants. The former would have greater implications for our understanding of OA in the general population. We explored this question by examining the compartmental distribution of knee OA in our study population; whereas knee OA is expected to predominantly affect the medial tibiofemoral joint (subject to greater loading [44]), many rarer inflammatory, erosive or genetic forms of OA have a predilection for the lateral Trichostatin A cell line compartment [43]. Our observation that predominantly medial compartment knee OA was by far the most common pattern in both the

HBM Ribonuclease T1 and control groups supports the view that HBM is associated with an increased risk of “standard” OA, and that the mechanisms underlying this relationship are applicable to the wider population. Plausible mechanisms that might contribute to a bone-forming phenotype in HBM include upregulation of the Wnt signalling pathway. Activating mutations of this pathway are known to result in HBM [22], and evidence is accumulating for a role of altered Wnt signalling in OA [45], [46] and [47]. Wnt signalling is also known to play a key role in the anabolic response of bone to mechanical loading, as evidenced by animal studies [48] and [49], and blockade of the Wnt signalling pathway inhibitor DKK-1 has been shown to promote osteophytosis in mice [50]. While the precise genetic basis of HBM in the majority of cases remains to be determined [51], and is the subject of ongoing studies, it is interesting to note that a genome-wide association study in this population showed overrepresentation of SNPs associated with BMD in the wider population including loci in Wnt pathway genes [10]. Our study has a number of limitations.