Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Smoking and leg bypass grafts


Smoking is the single largest risk factor for development of peripheral arterial disease. Those with the worst disease necessitating bypass surgery have almost always been heavy smokers. The problem is to stop them smoking, not just because stopping smoking is a good thing of itself, but also because continued smoking is thought to affect the effectiveness of surgery. Meta-analysis of studies provides useful information about the effects of smoking, as well as a good example of how observational study design can affect results [1].


Authors searched for articles about peripheral artery disease from 1950 to 2004 in at least four electronic databases including Cochrane, together with reference lists and reviews. To be included studies had to provide information on primary, secondary, or cumulative patency rates of arterial reconstructive surgery in the lower extremities, and report rates for smokers and nonsmokers.

Patency could be determined by several methods, including pulse examination. Any form of determination of smoking was used. Any type of reconstructive surgery was allowed, with any graft material used in the surgery, with any length of follow up.


Information was available from 29 studies, four randomised trials (870 patients), 12 prospective (1722 patients) and 13 retrospective (2894 patients) observational studies. Follow up varied between six months and 10 years, but most studies were between three and five years.

The main results are in Table 1. Prospective studies (including one randomised trial, Figure 1) produced a large difference between nonsmokers and smokers, with smoking having a number needed to harm for reduced patency of 4 (3.4 to 5.2). Retrospective studies also had a greater patency rate for nonsmokers over smokers, but the deleterious consequences of smoking were less, with a number needed to harm of 9 (6 to 16).

Table 1: Effect of smoking on lower extremity bypass surgery grafts in smokers and nonsmokers, by study design

Number of
Patency in
Study type
Relative risk
(95% CI)
(95% CI)
1.4 (1.3 to 1.6)
4.1 (3.4 to 5.2)
1.2 (1.1 to 1.4)
8.8 (6.2 to 16)
1.4 (1.3 to 1.5)
6.3 (5.1 to 8.2)

Figure 1: Individual results on graft patency in smokers and nonsmokers in prospective studies

In prospective studies, how smoking was determined made no difference in patency rates, nor was there a difference between follow up times of less than two years and more than two years.


So smoking is not a good idea if you have lower extremity bypass surgery. Yet over half of the patients in these studies smoked, and at least some of those who said they didn't smoke probably lied about it.

The difference between prospective and retrospective studies here was large and statistically significant. There was no overlap of NNH confidence intervals, and a statistical test showed a high level of difference (p < 0.001).

Why was this? There was no obvious answer in the trials, but it may just be some interaction between the type of study, prospective or retrospective, and the propensity of patients to fib about smoking. There wasn't much difference between the two types of study for graft patency in smokers, but a much lower patency in 'nonsmokers' in retrospective studies. Of course, it may also be chance, but the numbers were quite large, with significant events, so this is an unlikely explanation.

So what can we tell patients about smoking and lower extremity bypass surgery? If they don't smoke, the chance of the surgery still working three to five years later is about 4 to 1 on. If they smoke, it's only evens, and they are likely to be in trouble.


  1. EM Willigendael et al. Smoking and the patency of lower extremity bypass grafts: a meta-analysis. Journal of Vascular Surgery 2005 42: 67-74.

previous or next story