Table 5.

Exercise intervention effects on inflammatory biomarkers, stratified by potential moderators

Potential moderatoraTERd (95% CI)
Baseline levelncCRPIL-6TNF-α
Physical fitness
(VO2max)
 ≤27.5 mL/kg/min82/800.98 (0.83–1.16)1.03 (0.92–1.14)0.99 (0.93–1.05)
 >27.577/790.77 (0.66–0.91)0.95 (0.84–1.08)1.01 (0.95–1.07)
Pe = 0.040Pe = 0.373Pe = 0.439
Age, y
 ≤6073/810.85 (0.71–1.01)0.94 (0.83–1.06)0.98 (0.93–1.03)
 >6087/780.90 (0.77–1.05)1.03 (0.92–1.15)1.02 (0.95–1.09)
Pe = 0.615Pe = 0.256Pe = 0.476
Self-rated healthb
 Low88/760.94 (0.80–1.11)0.93 (0.82–1.05)1.00 (0.93–1.06)
 High71/810.79 (0.67–0.92)1.05 (0.93–1.17)1.00 (0.95–1.06)
Pe = 0.119Pe = 0.153Pe = 0.905
Past year recreational activity
 <7 MET-h/wk87/740.81 (0.69–0.96)0.93 (0.83–1.04)1.03 (0.97–1.09)
 ≥773/850.94 (0.80–1.11)1.04 (0.92–1.17)0.97 (0.92–1.04)
Pe = 0.227Pe = 0.187Pe = 0.193
BMI
 ≤25 kg/m234/300.88 (0.70–1.11)1.16 (0.94–1.44)0.93 (0.81–1.08)
 >25–≤3063/670.91 (0.76–1.10)1.02 (0.91–1.14)1.01 (0.95–1.06)
 >3063/630.84 (0.69–1.03)0.87 (0.76–1.00)1.03 (0.98–1.09)
Pe = 0.735Pe = 0.053Pe = 0.030
CRP
 <3.0 mg/L119/1250.90 (0.79–1.02)
 ≥3.040/320.78 (0.60–1.01)
Pe = 0.354
IL-6
 <1.45 pg/mL77/851.08 (0.98–1.19)
 ≥1.4583/750.90 (0.78–1.03)
Pe = 0.061
TNF-α
 <1.4 pg/mL78/831.03 (0.98–1.08)
 ≥1.482/770.97 (0.90–1.04)
Pe = 0.163

NOTE: Top 1% of CRP measures were excluded from analyses.

  • aLevel of potential moderator at baseline.

  • bSelf-rated health measured by self-administered questionnaire using SF-36 scale (range, 0–100) where low was a score of x < 82 and high was x ≥ 82.

  • cNumber of exercisers/number of controls.

  • dThe TER was defined as the geometric mean ratio, estimated from least square means, for the difference in treatment effect between exercisers and controls averaged across the entire study period, and then back log transformed. A ratio less than 1.00 indicates lower inflammatory marker levels in exercisers relative to controls at 6 and 12 months; a ratio more than 1.00 indicates higher inflammatory marker levels in exercisers; and a ratio equal to 1.00 indicates no difference between exercisers and controls.

  • eStatistical significance of the interaction term between intervention group and the potential moderator. Note that all moderators were treated as continuous variables when calculating this P value for heterogeneity.