Genome-wide association studies have had a generally poor track record for yielding actionable information, but a new study in older adults at risk for severe osteoarthritis (OA) might allow early intervention to avoid joint replacement.
Risk scores based on analysis of 47 single-nucleotide polymorphisms (SNPs) identified in previous association studies significantly predicted eventual knee and hip replacement among more than 12,000 individuals, according to Paul Lacaze, PhD, of Monash University in Melbourne, Australia, and colleagues.
Odds ratios they reported in Arthritis & Rheumatology for each standard deviation in the novel risk score were as follows:
- Knee replacement: 1.13, 95% CI 1.07-1.20
- Hip replacement: 1.23, 95% CI 1.16-1.30
Some 9.5% of participants with low genomic risk scores (in the bottom quintile) underwent knee replacement during a median of 4.7 years of observation, compared with 13.5% of those scoring at high risk (top quintile). For hip replacement, the corresponding percentages were 8.3% versus 14.6%.
Calling the associations “robust,” Lacaze and colleagues wrote that genomic risk scoring has “the potential to improve prevention of severe knee and hip OA by providing a personalized approach and identifying individuals who may benefit from early intervention.”
The study was a post-hoc analysis of participants in the so-called ASPREE trial, which tested aspirin’s ability to prevent or delay several of the most common medical events afflicting older people, including dementia, cancer, heart disease, and depression. As part of the trial, most patients were comprehensively genotyped, and Lacaze and colleagues had access to hospitalization records for all participants in Australia.
Some participants said they had undergone joint replacement before enrolling in ASPREE, and they were included too; in fact, about 60% of replacement procedures counted in the study were self-reported. Other inclusion criteria were non-Finnish European ancestry and age of at least 70 when joining ASPREE.
For developing the genomic risk score, Lacaze’s group picked 10 SNPs associated with knee replacement and 37 linked to hip replacement that were identified in a 2021 Cell paper. Mean risk scores for knee and hip replacement were 0.25 (SD 0.13) and 0.51 (SD 0.34), respectively.
Overall predictive power of the genomic risk score was modest. A model based on age, sex, BMI, educational attainment, and socioeconomic status had areas under the receiver-operating characteristic curve of 0.666 for knee replacement and 0.570 for hip replacement; adding the risk score to this model raised these values to 0.668 and 0.589, respectively.
Despite this minimal improvement, Lacaze and colleagues noted that the risk score was a statistically significant and independent predictor of joint replacement, with “meaningfully different risks” associated with high versus low scores.
Limitations included the use of self-reported data on previous joint replacement and the age restriction. Also, ASPREE sought participants in relatively good health, and hence the results in the current study (as well as for other ASPREE analyses) may not be generalizable to people with significant comorbidities.
The study was funded by Australian and U.S. government grants.