Table 2.

Risk of adenomas; evidence of substantial effect in active arm

	Follow-up colonoscopy2 to 39 mo after beginning treatment (n = 267)		Follow-up colonoscopy33 to 36 mo after beginning treatment (n = 204)
	Placebo (n = 129)	DFMO/sulindac (n = 138)	Placebo (n = 97)	DFMO/sulindac (n = 107)
Detection of any adenoma
Cumulative incidence of adenomas detected at end of the treatment (%)	53 (41.1)	17 (12.3)	42 (43.3)	12 (11.2)
Risk ratio* (95% CI)		0.30 (0.18–0.49)		0.26 (0.15–0.46)
P		<0.001		<0.001
Detection of advanced adenomas†
Cumulative incidence of advance adenomas detected at end of the treatment (%)	11 (8.5)	1 (0.7)	9 (9.3)	1 (0.9)
Risk ratio* (95% CI)		0.085 (0.011–0.65)		0.10 (0.013–0.78)
P		0.001		0.004
Detection of advanced adenomas with size ≥ 1 cm
Cumulative incidence of advanced adenomas with size ≥ 1 cm detected at end of the treatment (%)	9 (7.0)	1 (0.7)	7 (7.2)	1 (0.9)
Risk ratio* (95% CI)		0.10 (0.013–0.81)		0.13 (0.016–1.03)
P		0.004		0.02
Detection of multiple adenomas (>1)
Patients with >1 adenoma, incidence (%)	17 (13.2)	1 (0.7)	15 (15.5)	1 (0.9)
Risk ratio* (95% CI)		0.055 (0.0074–0.41)		0.060 (0.0081–0.45)
P		<0.001		<0.001
Sensitivity analysis imputing adenoma for patients without an end-point determination‡
Cumulative incidence of adenomas detected at end of the treatment (%)	76/184 (41.3)	39/191 (20.4)
Risk ratio* (95% CI)		0.49 (0.36–0.69)
P		<0.001

^*Relative risk estimation by log-binomial regression. Likelihood ratio test P values are reported.

^†Advanced adenomas in the placebo group included tubulovillous (3), intramucosal carcinoma (2), size ≥1 cm (6), and one serrated adenoma with high-grade dysplasia; the one advanced adenoma in the treatment group was an adenoma >1 cm.

^‡Sensitivity analysis imputing adenoma for all patients without an end-of-study colonoscopy at the placebo rate of recurrence.