Skip to main content
. 2016 May 31;6:26983. doi: 10.1038/srep26983

Table 3. Stratified pooled relative risks and 95% confidence intervals for highest versus lowest category of circulating beta-carotene and all-cause mortality.

Subgroups n Death Participants RR 95%CI I2 (%) Pa Pb
All studies   7 6137 25468 0.69 0.59, 0.80 37.10 0.145  
Duration of follow-up                 0.496
   < 10 years 4 565 6121 0.62 0.47, 0.82 0 0.494  
  > = 10 years 3 5572 19347 0.71 0.57, 0.88 67.20 0.047  
Sample size                 0.246
  <2,000 3 1038 2383 0.81 0.67, 0.97 0 0.600  
  > = 2,000 4 5099 23085 0.63 0.51, 0.78 46.60 0.132  
Population age at baseline                 0.246
  <50 4 5099 23085 0.63 0.51, 0.78 46.60 0.132  
  > = 50 3 1038 2383 0.81 0.67, 0.97 0 0.600  
Median level                 0.076
  <350 nmol/L 2 661 2439 0.54 0.42, 0.70 0 0.688  
  > = 350 nmol/L 5 5476 23029 0.75 0.66, 0.86 9.30 0.354  
Blood sample                 0.197
  Serum 5 5130 23239 0.63 0.52, 0.77 34.00 0.194  
  Plasma 2 1007 2229 0.82 0.68, 0.99 0 0.741  
Major confounders adjusted                  
History of disease                 0.560
  Yes 3 4546 17337 0.74 0.63, 0.88 0 0.703  
  No 4 1591 8131 0.64 0.49, 0.84 63.80 0.040  
Smoking and drinking                 0.962
  Yes 4 4759 21975 0.71 0.61, 0.83 1.40 0.385  
  No 3 1378 3493 0.66 0.46, 0.96 69.20 0.039  
BMI                 0.560
  Yes 3 4546 17337 0.74 0.63, 0.88 0 0.703  
  No 4 1591 8131 0.64 0.49, 0.84 63.80 0.040  
Physical activity                 0.572
  Yes 4 4692 19781 0.74 0.63, 0.86 0 0.839  
  No 3 1445 5687 0.62 0.44, 0.89 75.90 0.016  
Serum cholesterol level                 0.471
  Yes 3 4661 19627 0.74 0.63, 0.87 0 0.918  
  No 4 1476 5841 0.61 0.44, 0.85 65.90 0.032  
Vitamin supplementation use                 0.454
  Yes 2 4515 17183 0.75 0.64, 0.89 0 0.960  
  No 5 1622 8285 0.63 0.49, 0.82 54.60 0.066  

Abbreviation: RR, relative risk; CI, confidence interval; I 2, measure of heterogeneity; Pa, P value for heterogeneity within each group; Pb, P value for heterogeneity between subgroups in meta-regression.