Table 2 Pooled effects of overweight on disease-free survival and overall survival.

BMI comparison group		Sample size		Heterogeneity			Pooled effect
	Dependent variable	K^a	N	Q^b	p	I²	HR^c	95% CI	p	95% PI^d
BMI ≥ 25 vs BMI < 25	Disease-free survival	10	5,109	10.5	0.31	14.0	1.26	1.09–1.46	<0.01	N/A
	Adjusted for publication bias^e	15	—	—	—	—	1.13	0.97–1.32	Ns	—
BMI ≥ 25 vs BMI < 25	Overall survival	12	8,005	20.4	0.04	46.0	1.29	1.11–1.51	<0.01	0.83–2.34
	Adjusted for publication bias^e	18	—	—	—	—	1.08	0.92–1.27	Ns	—

^aK = number of studies.
^bQ-statistic: p-values < 0.1 and I² > 0.0 taken to suggest heterogeneity.
^cHR = hazard ratio (random effects), HR > 1.0 indicate overweight associated with poorer prognosis/increased mortality.
^d95% prediction interval calculated for heterogeneous results.
^eAs results (Egger’s test and funnel plot) suggested publication bias, missing studies were imputed and the pooled HR adjusted accordingly with the Duval and Tweedie Trim and Fill test³¹, K = number of published studies + number of imputed studies.
N= number of patients, Ns = Not significant, N/A = Not applicable.

Quick links

Search