Table 3.

Factors associated with opioid usage based on multivariable longitudinal mixed-effects logistic regression model (n = 706).

VariablesAdjusted OR (95% CI)p
Antidepressant usage (use vs no use) by muscle relaxants usage*
  When muscle relaxants taken0.9 (0.3–2.9)0.85
  When muscle relaxants not taken2.2 (1.3 3.8)0.004
PtGA ≥ 232.6 (1.5–4.6)0.0006
CES-D total > 161.6 (1.1–2.3)0.02
BASFI ≥ 40/100 mm1.9 (1.2–2.9)0.004
BASDAI ≥ 4/10 cm2.0 (1.4–2.9)0.0005
Exercise ≥ 3×/wk0.7 (0.5–1.1)0.13
> 2 comorbidities1.2 (0.7–2.1)0.46
Medication usage
  Hypnotics5.4 (2.6–11.3)< 0.0001
  Prednisone2.5 (1.4–4.7)0.003
  Anxiolytics4.0 (2.2–7.1)< 0.0001
  TNFi agents0.7 (0.5–1.0)0.05
  NSAID1.1 (0.7–1.5)0.76
  DMARD1.3 (0.8–2.1)0.33
  Antipsychotics21.8 (0.4–8.4)0.32
BASRI baseline ≥ 6**0.8 (0.5–1.4)0.41
Elevated CRP1.5 (0.9–2.5)0.13
Education (> gr 12)1.1 (0.5–2.2)0.84
Marital status: married1.9 (1.1–3.2)0.02
Race: white (compared to others)1.5 (0.7–2.9)0.29
Male sex1.2 (0.7–2.2)0.48
Age ≥ 40 yrs1.9 (0.7–3.4)0.03
Smoking (> 100 packs within lifetime)1.8 (1.1–3.0)0.02
  • * OR for antidepressant usage were calculated separately by muscle relaxants usage (i.e., when muscle relaxants were taken or not taken) based on an interaction effect between muscle relaxants and antidepressant usage in the multivariable model.

  • ** Because of a high correlation between BASRI and mSASSS and the greater completeness of the data in the former, it was decided to adjust BASRI instead of mSASSS in the final multivariable model. BASFI: Bath Ankylosing Spondylitis Functional Index; BASRI: Bath Ankylosing Spondylitis Radiological Index; CES-D: Center for Epidemiologic Studies Depression Scale; DMARD: disease-modifying antirheumatic drugs; mSASSS: modified Stokes Ankylosing Spondylitis Scoring System; NSAID: nonsteroidal antiinflammatory drugs; PtGA: patient’s global assessment; TNFi: tumor necrosis factor inhibitors; BASDAI: Bath Ankylosing Spondylitis Disease Activity Index; CRP: C-reactive protein.