Table 4.

Detailed assessment of adverse events

	Placebo (n = 184)		DFMO/sulindac (n = 191)
	Aspirin (n = 69)	Non-aspirin (n = 115)	Aspirin (n = 77)	Non-aspirin (n = 114)
Cardiovascular
CAD, prior Hx*	0	1	3	0
CAD, new	0	1	1	0
MI, prior Hx	1	0	1	0
MI, new	0	0	3	1
CVA, prior Hx	0	0	0	0
CVA, new	0	1	1	1
CHF, prior Hx	1	0	1	0
Chest pain, new	0	4	0	4
Subtotal, prior Hx	2	1	5	0
Subtotal, new	0	6	5	6
Total	9 (4.9)		16 (8.4)
Risk ratio (95% CI)†	—		1.71 (0.78–3.78)
P	—		0.17
Hearing
Audiometric evaluations at end of study, no. patients with adverse events/no. patients in cohort (%)‡
Hearing loss at least 15 dB at ≥2 frequencies	24/123 (19.5)	35/136 (25.7)
Risk ratio (95% CI)†		1.32 (0.83–2.09)
P		0.23
Hearing loss at least 15 dB at ≥2 consecutive frequencies	12/123 (9.8)	25/136 (18.4)
Risk ratio (95% CI)†		1.88 (0.99–3.59)
P		0.05
Hearing loss at least 15 dB at ≥2 consecutive frequencies in the normal range (500–3,000 Hz)	7/123 (5.7)	14/136 (10.3)
Risk ratio (95% CI)†		1.81 (0.75–4.33)
P		0.17
Bilateral audiometric changes at 36 mo or at least 6 mo off-study
End of study, no. with bilateral changes (%)§	2/9 (22.2)	10/21 (47.6)
Risk ratio (95% CI)†	—	2.14 (0.58–7.88)
P	—	0.18
At 6 mo posttreatment3
Improved	1/2 (50.0)	3/10 (30.0)
Persistent	0/2 (0.0)	1/10 (10.0)
Unilateral hearing loss with development of bilateral hearing loss	1/2 (50.0)	2/10 (20.0)
Pending	0/2 (0.0)	4/10 (40.0)

^*Prior Hx, subjects had prior history.

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

^‡All self-reported hearing complaints were recorded. Hearing thresholds were assessed for 259 patients with follow-up audiograms measured from 18 to 36 mo after beginning of therapy.

^§Unilateral or bilateral change from baseline at 36 mo (n = 30; placebo, n = 9; DFMO/sulindac, n = 21).

^||Bilateral changes at 36 mo or at least 6 mo off-study (n = 12; placebo, n = 2; DFMO/sulindac, n = 10).