. 2011 Dec;55(12):5790–5797. doi: 10.1128/AAC.05044-11

Table 4.

Early responder results (lack of spread of lesion and afebrile) for the ITT population

Responder criterion (window)	No. (%)^a of early responders		OR^b	NI^c limit	% risk difference (lower CI)^d	P value
Responder criterion (window)	JNJ-Q2 (n = 83)	Linezolid (n = 78)	OR^b	NI^c limit	% risk difference (lower CI)^d	P value
Early responder, primary analysis (36-84 h)	62 (74.7)	62 (79.5)	0.76 (0.36, 1.59)	0.469	−4.8 (−19)	0.087
Responder, FDA definition (48-72 h)	51 (61.4)	45 (57.7)	1.14 (0.60, 2.15)	0.546	3.8 (−12.7)	0.024

Percentages are based on the number of patients in each treatment group.

Odds ratios and 95% Wald confidence intervals (CIs) for treatment group effect are from a logistic regression model, with treatment group and wound category as main effects and baseline infection site area and baseline temperature as covariates. An odds ratio of >1 indicates an increased odds of success in favor of JNJ-Q2 over linezolid.

NI limit, noninferiority limit. The lower confidence limit of the odds ratio of JNJ-Q2 over linezolid had to be greater than the NI limit to show evidence of noninferiority.

The threshold for the lower limit of the 95% confidence interval that was required to claim a difference was −15%.