Table 3

Risk estimates [n (%)] for smoking related-SPTa risk in relation to IGF-I, IGFBP-3, and smoking status in logistic regression analysis

No SPT (n = 173)SPTb (n = 58)Univariate OR (95% CI)Multivariate OR (95% CI)c
Smoking status
 Never24 (13.9)4 (6.9)1.001.00
 Former100 (57.8)26 (44.8)1.56 (0.50–4.89)1.48 (0.43–5.14)
 Current49 (28.3)28 (48.3)3.43 (1.08–10.89)3.22 (0.91–11.38)
Treatment regimen
 Placebo88 (50.9)35 (60.3)1.001.00
 13 cRA85 (49.1)23 (39.7)0.68 (0.37–1.25)0.71 (0.37–1.40)
IGF-I (ng/ml)
 < 104.2598 (57.7)19 (32.8)1.001.00
 ≥ 104.2572 (42.3)39 (67.2)2.79 (1.49–5.23)2.90 (1.37–6.13)
IGFBP-3 (ng/ml)
 ≥ 2276, < 414091 (53.2)25 (43.1)1.001.00
 < 227674 (43.3)20 (34.5)0.98 (0.51–1.91)1.61 (0.74–3.51)
 ≥ 41406 (3.5)13 (22.4)7.88 (2.72–22.84)8.27 (2.60–26.28)
  • a SPT, second primary tumor; IGF-I, insulin-like growth factor-I; IGFBP-3, IGF binding protein 3; OR, odds ratio; CI, confidence interval; 13 cRA, 13-cis-retinoic acid.

  • b Smoking-related SPTs include larynx, oral cavity, pharynx, lung, and bladder tumors.

  • c Multivariate logistic regression model was built to assess the predictive effects of IGF-I and IGFBP-3 for smoking-related SPT risk. The model simultaneously includes IGF-I, IGFBP-3, smoking status at registration, and treatment regimen.