Quantification of baseline hand joint space width does not differentiate individuals who develop incident knee osteoarthritis from controls: data from the osteoarthritis initiative

Abstract

Purpose: Previous studies that used semi-quantitative scoring of hand radiographs suggest that hand osteoarthritis (OA) puts a person at risk for knee OA. One possible mechanism to explain a relationship between hand and knee OA is a systemic cartilage loss. Therefore, we used a quantitative method to measure hand joint space width (JSW), an imaging marker of cartilage loss, to test whether individuals that develop incident knee OA have less finger joint cartilage. Methods: Participants of the OAI were included in this case-control study if they had knee radiographs at the baseline and 48-month visit as well as quantitative and semi-quantitative radiographic knee readings (as of October 2011). Cases were defined as individuals with an incident knee osteophyte (baseline OARSI osteophyte score < 2 and 48-month OARSI osteophyte score 2). Controls were individuals with no incident knee osteophyte (baseline and 48-month OARSI osteophyte scores < 2). If a participant had a case and control knee, then the control knee was excluded. If the participant had two case knees or two control knees, then the right knee was selected. 276 knee-, cohort-, age-, gender-, and body mass index-matched participants met this criteria. Two readers used a semi-automated, custom software to delineate the joint margins metacarpophalangeal (MCP), proximal interphalangeal (PIP), and distal interphalangeal (DIP) joints of digits 2 to 5 on the dominant hand. The software divided each joint into 5 regions to derive a region-specific JSW measurement (JSW1 to JSW5; inter-tester ICC [2,1 model] 1⁄4 0.82 to 0.92; see Figure). Due to anatomical considerations the outermost regions (JSW1 and JSW5) of the MCP joints could not be reliably replicated and were excluded from calculations. Mean MCP JSW, mean PIP JSW, and mean DIP JSW were calculated by averaging the joint JSW measurements across regions within joint and across fingers. Mean hand JSW was averaged across all regions, joints, and fingers to develop a composite hand measurement. Four paired sample t-tests (p < 0.05) were performed to compare the four mean JSW measurements between cases and controls. Results: Of the 276 participants, 121 matched case-control pairs had readable hand radiographs. Case participants were 62.5 8.6 years of age, had a body mass index of 29.7 4.1 kg/m, and 8 (6.67%) were left handed. Control participants were 62.5 8.4 years of age, had a body mass index of 29.4 5.0 kg/m, and 5 (4.13%) were left handed. Both groups were 68.6% female, included 65 (54%) right knees, and 63 (52%) participants from the Incidence cohort. Paired sample t-tests revealed no significant differences in all mean JSW measurements between the case and control groups (see Table). Conclusions: We found that baseline mean hand JSW is not significantly different among those who develop a definite knee osteophyte over 48 months and those who do not. Our results indicate that systemic loss of cartilage is not an explanation of the relationship of hand and knee OA. Ă