TABLE 3.
HRs and 95% CIs for the association between baseline and longitudinal pain scales and death, IRONMAN Registry 2017–2023 (N = 879; 137 deaths in White participants, 37 deaths in Black participants)
| Pain at enrollment | Longitudinal pain | ||||
|---|---|---|---|---|---|
| Pain scale | Comparison | Age-only model HR (95% CI) | Fully-adjusted modela HR (95% CI) |
Age-only model HR (95% CI) | Fully-adjusted modela HR (95% CI) |
| EORTC scale | 10 points on 0–100 scale | 1.12 (1.06–1.17) | 1.10 (1.03–1.17) | 1.30 (1.20–1.42) | 1.29 (1.19–1.40) |
| Average pain | 1 point on 1–10 scale | 1.22 (1.10–1.35) | 1.19 (1.08–1.32) | 1.34 (1.21–1.48) | 1.32 (1.20–1.46) |
| Worst pain | 1 point on 1–10 scale | 1.17 (1.08–1.26) | 1.16 (1.08–1.25) | 1.31 (0.90–1.91)b | 1.31 (1.20–1.43) |
| Bone pain | Some vs. none | 1.62 (1.12–2.33) | 1.61 (1.10–2.37) | —c | —c |
| Bone pain | A lot vs. none | 2.70 (1.65–4.42) | 2.47 (1.44–4.22) | —c | —c |
aCox model for pain and death adjusted for potential confounders of disease burden including age at enrollment, first PSA level on-study, Gleason score, disease state at enrollment (mHSPC vs. CRPC), de novo metastatic disease at baseline, and sites of metastases at baseline.
bSE is inflated as all imputed datasets led to Hessian matrices that were not positive definite for the worst pain scale.
cLongitudinal models for bone pain not fit because categorical outcomes are currently incompatible with joint longitudinal survival model capabilities in JM R package.