Table 10.

Protocol-wise (P1–P5) sample sizes (N) and bootstrap derived 95% confidence intervals (CI) required for a hypothetical therapeutic trial in Alzheimer’s disease (AD) or in mild cognitive impairment (MCI,) designed to detect a 25 percent change in the 24-month rate of hippocampal atrophy in comparison to a placebo group (with significance set at 0.05 and power at 0.8). Sample size calculations are reported as unadjusted or adjusted based on the mean 24-month protocol-specific rate of hippocampal atrophy observed in ADNI-1 normal elders. All rates of atrophy were entered as percentage change from baseline and normalized for each participant’s ADNI scan interval, which often was greater than 24-months.

Dx Group	P1		P2		P3		P4		P5
	n	95% CI	n	95% CI	n	95% CI	n	95% CI	n	95% CI
MCI Unadjusted	260	(202, 353)	339	254, 479)	225	(175, 306)	214	(165, 292)	217	(169, 293)
MCI Adjusted	891	(670, 1841)	1109	(834, 2661)	870	(594,1585)	902	(640, 1844)	893	(510, 1258)
AD Unadjusted	152	(100, 263)	207	(132, 381)	145	(97, 247)	135	(92, 229)	143	(95, 243)
AD Adjusted	287	(194, 601)	379	(250, 867)	297	(192, 585)	294	(198, 600)	302	(179, 534)