meta-analysis use relative not absolute risk reduction, and so, by default or design, dramatically inflate the apparent value of the drugs. For example, in this meta-analysis of 464 164 people, 22 115 (4.76 per cent) suffered heart attacks and 17 890 (3.85 per cent) strokes during the average of four years that the trials lasted. If the drugs reduced heart attack risk by 25 per cent and stroke risk by 33 per cent, the percentage of subjects suffering a heart attack would be 4.05 per cent and a stroke
