Skip navigation

NNTs for lipid lowering


Bandolier was delighted to find that someone had got to grips with lipid lowering trials in primary and secondary prevention and pulled all the data together in a systematic review [1]. Even better was that results were presented as numbers-needed-to-treat (NNT).

Review

The review sought studies of primary, secondary, and tertiary prevention identified from electronic searching and reviews up to the end of 1995. Included studies were randomised and involved standard antidyslipidaemic therapy (diet, pharmaceuticals or surgery). All studies were single-blind, and most were double-blind.

Results

The main results are shown in the table for several different end-points, and in the graph for myocardial infarction (MI) or cerebrovascular (CV) death.
Number-needed-to-treat
Prevention studies Number of trials Number of subjects MI or CV death MI CV death All deaths
Primary prevention
All trials 7 29,683 69 (54 - 99) 78 (60 - 121) 347 (209 - 1981) 931 (212 - no benefit)
Secondary prevention
All trials 25 18,452 16 (13 - 19) 28 (23 - 36) 33 (26 - 50) 37 (26 - 67)
Newer studies 13 8,390 15 (13 - 19) 23 (18 - 32) 53 (41 - 84) 41 (31 - 66)
Diet only 7 2,407 15 (11 - 29) 42 (23 - no benefit) 25 (16 - 72) 29 (15 - no benefit)
Niacin 5 4,835 15 (11 - 30) 38 (21 - no benefit) 32 (16 - no benefit) 28 (16 - no benefit)
HMG-CoA reductase inhibitors 9 7,934 15 (13 - 19) 23 (18 - 31) 44 (34 - 67) 37 (28 - 58)


Primary prevention (no previous heart attack)


In seven primary prevention trials with 29,683 subjects, active treatment resulted in an average reduction of cholesterol of 13%, compared to an increase of 1% in the controls over an average duration of 4.9 years. This gave a NNT for death from heart attack or stroke of 69 (54 to 99). That is, 69 people have to have lipid lowering therapy for five years to prevent one of them dying from heart attack or stroke.

Secondary prevention


In 25 secondary or tertiary prevention trials with 18,452 subjects, active treatment resulted in an average reduction of cholesterol of 18%, compared to no change in the controls over an average duration of 4.9 years. This gave a NNT for death from heart attack or stroke of 16 (13 to 19). That is, 16 people have to have lipid lowering therapy for five years to prevent one of them dying from heart attack or stroke. These results were similar in the newer studies, those involving diet only, niacin, or coenzyme A reductase inhibitors.

Comment

This is an interesting paper, and worthwhile having in the filing cabinet. Some cautionary notes, though, since the figures in the tables do not always add up, and there is a conceptual flaw in the way in which negative NNTs (that is, those where control is better than treatment) are handled.

However, the way in which the benefits of treatment are handled cumulatively is illuminating. The early trials included some which did not have big effects (perhaps because they were small). The first trial in 1965 showed no benefit, and doesn't even appear on the graph because of that. If that was assumed to be the truth, then the benefits of lipid lowering would not have been discovered. As it was the NNTs settled down to about 15 by the time about 9,000 patients had been studied in 1988. Bandolier keeps seeing examples where the results of single, small trials can over-or under-estimate benefits of an intervention.

Reference:

  1. CM Rembold. Number-needed-to-treat analysis of the prevention of myocardial infarction and death by antidyslipidemic therapy. Journal of Family Practice 1996 42: 577-86.



previous or next story in this issue