Background/Purpose: Spine osteoarthritis (OA) has been defined by the presence of: 1) disc space narrowing (DSN) and vertebral osteophytes (OST), or 2) facet joint OA (FOA). However, anatomical differences of these structures used in these two definitions suggest they may have different etiologies. Our aim was to determine if differences in associations exist between demographic (age, race, sex) and clinical characteristics (body mass index [BMI]), coincident knee, hip or hand OA, and low back symptoms) and OA involvement in the lumbar spine, defined as 1) FOA only and 2) Spine OA only (DSN plus vertebral OST) or 3) FOA and Spine OA, when compared to those with neither Spine OA nor FOA.

Methods: Data were collected from 2003-2010 in the Johnston County OA Project. Each lumbar spine level was graded for OST and DSN in a semi-quantitative fashion (0-3) while FOA was graded as present or absent, according to the Burnett Atlas. Spine OA was defined as the presence of DSN and OST grade ≥1 at the same lumbar level. Participants reported the presence of low back symptoms (pain, aching, and/or stiffness) on most days of any one month in the last 6 months and their history of back injury (yes/no). Knee OA and hip OA were both defined by a Kellgren-Lawrence score (K-L) of 2-4. Hand OA was defined as K-L 2-4 in a minimum of one distal interphalangeal (IP) joint and 2 other joints (IP or carpometacarpal). Age, race and sex were collected by self-report and BMI measured at clinical examination. Adjusted odds ratios (OR) and 95% confidence intervals (CI) were estimated with multinomial logistic regression.

Results: Complete data were available for 1,874 participants. The table presents the counts and percentages for spine OA groups, demographics and clinical characteristics.

Table 1a presents demographic adjusted associations. All age categories were strongly associated age categories compared to none with the strength of association increasing with greater age. African Americans were 32% less likely to have FOA only and 49% less likely to have both FOA and Spine OA. Women were 64% more likely to have FOA alone.

Table 1b presents clinical characteristic adjusted associations. Those with FOA only were 74% more likely to have BMI ≥30 kg/m². Similar associations were found between knee OA and both FOA only and Spine OA only, though stronger with both FOA and Spine OA. Back injury was strongly associated with Spine OA only. Symptoms, hip OA, and hand OA were not significantly associated with any spine OA involvement.

Conclusion: The differences in demographic characteristics (race and sex) and clinical characteristics (BMI and back injury) suggest pathophysiologic processes may vary for these definitions lumbar spine OA involvement. Increasing associations across categories of age suggests changes over time, however longitudinal studies may elucidate whether those with FOA only or Spine OA only remain an isolated lumbar spine OA phenotype over time.