Official: Smoking ban had no effect on England's heart attack rate

Data released today by the NHS shows that England's smoking ban had no discernable effect on the number of heart attack admissions recorded. In 2007/08, there were 66,933 admissions for acute myocardial infarction compared with 68,312 in 2006/07. This represents a 2% decline, slightly lower than the 2.8% drop in 06/07 and the 3.8% drop in 05/06.

Emergency admissions for acute myocardial infarction have declined in England every year since 1999, with the exception of 02/03 when there was a rise of 5.3%. When the unusual rise of 02/03 is excluded, the average annual decline in heart attack admissions is 2.3%. As such, the 2% drop in admissions since the smoking ban was enacted is entirely in line with the existing downward trend.

Several studies from the United States have shown sharp drops in heart attack admissions following the introduction of smokefree legislation. These include:

- a 40% drop in Helena, Montana

- a 41% drop in Pueblo, Colorado

- a 47% drop in Bowling Green, Ohio

One criticism of these studies is that they involve relatively small populations, thereby allowing minor variations in the number of actual cases to appear more profound when translated in percentage terms. Data from Florida, New York State, Oregon and California have failed to show any significant decline in heart attack admissions following the implementation of smoking bans.

The figures released by the NHS today provide data from every hospital in England (population: 51,000,000); this represents by far the largest "sample group" of any study to date. If smokefree legislation has a measurable impact on heart admissions, one would expect to see an unusually large drop in admissions in 2007/08.

The fact that the rate fell by just 2% - in line with the long-term trend - strongly suggests that the 'Helena hypothesis' is in error. This is supported by evidence from neighbouring Wales, where there was a fall of just 1% in the year following the smoking ban (compared to a fall of 6.3% the year before).

This data has several strengths. It comes from a health service that serves a population of 51 million. This compares very favourably with the modest populations of Helena (66,000), Pueblo (102,000) and Bowling Green (30,000). In addition, England has one of the toughest and most rigorously enforced smoking bans in the world. There are no exemptions for any workplace, bar, restaurant or venue. Even outdoor sports stadia and train platforms are included.

The "sample group" (the entire population of England) is many times larger than all the other smoking ban/heart attack studies combined. The size of the database, and the inclusion of data from every hospital in the country, rules out the possibility of these findings being due to statistical blips or local anomalies.

This data does, however, have one significant limitation. The English NHS shows admission figures for financial years but does not break down the information by month. Since the smoking ban began on July 1 2007, the 2007/08 figures include three months (April-June) which pre-dated the smoking ban. Even so, it would require a remarkable surge in admissions in these three months for the post-ban drop to significantly exceed 2% and it is, in practice, impossible for the true drop in admissions to be anywhere near the 40%, 20% or 10% claimed in some studies. The final figure will almost certainly fall between 1% and 3%.

A study from Scotland purported to show a 17% drop in acute coronary syndrome (ie. acute myocardial infarction plus unstable angina). Hospital data from Scotland showed this to be a huge exaggeration but the ISD Scotland data suffered from the minor limitation of not showing unstable angina and other angina separately. Fortunately, the NHS England data does make this distinction and so we can see that the decline in acute coronary syndrome admissions slowed after the smoking ban from 4.2% (2006/07) to 2.3% (2007/08).

Likewise, the decline in first-time admissions for acute myocardial infarction remained the same in both years (3.7%) and subsequent AMI admissions actually rose after the ban, by 5.7%. The rate of decline for unstable angina cases slowed after the ban, from 6.8% to 2.9%.

In short, there is no amount of statistical conjuring or selective analysis that can disguise the fact that these figures do not in any way support the claim that smoking bans are followed by sharp falls in heart attack admissions. The fact that several epidemiological studies have concluded otherwise only serves to underline the danger of taking selective evidence from small communities.

Published: 25.02.09

Christopher J. Snowdon is the author of Velvet Glove, Iron Fist: A History of Anti-Smoking

If you have any comments or corrections please send them to