Coronaries and controversy – the impact of smoke-free legislation on coronary heart disease

Friday 20 February 2009

A number of studies have been published recently showing a decline in admissions to hospital for heart attack following the implementation of smokefree laws in various countries. Many of these studies, including one on the impact of the Scottish smoking ban [1] have also been the subject of criticism by so-called “dissidents” who claim that the research is nothing more than ‘junk’ science. In a recently published review [2] ASH’s Director of Policy and Research, Martin Dockrell, looks behind the scenes to see what has prompted this criticism and to what extent it is justified. The analysis finds parallels with those who were in denial about the causes of AIDS long after the scientific debate was over.

The focus of the analysis was a study by Pell and colleagues of the impact of the Scottish smoke-free legislation which found that there was a 17% reduction in hospitalizations for acute coronary syndrome compared with a 4% reduction in England (which had not enacted smokefree legislation).

The ASH review acknowledges that many of the studies have differing strengths and that no single study can ever demonstrate causation. However, even before the Scottish study was published, the headline figures attracted criticism for being “over-hasty” and constituting “junk science”.

So who was behind the protest and was it justified? Dockrell’s review reveals that the strongest critics – Blastland and Dilnot [3] had in fact rejected the research before they had had the opportunity to examine it. Furthermore, ASH believes that all of their criticisms are unfounded and bear close resemblance to the views put forward by tobacco industry proponents.

Commenting on the dissidents, Martin Dockrell said:

“While careful analytical critiques of research are welcome, the recent unwarranted attacks on the authors of the smokefree research suggest an ulterior motive. We found an unholy alliance of conspiracy theorists, tobacco industry lobbyists and journalists impatient for a good story.”


Notes and links

1. Pell JP et al. Smokefree legislation and hospitalizations for acute coronary syndrome. NEJM 2008; 359: 482-91
2. Dockrell, M. Eye and heart at mortal war: coronaries and controversy in a smoke-free Scotland. Expert Review of Pharmacoeconomics & Outcomes Research 2009. Then click Ahead of print
3. Blastland, M. Just because. BBC News Magazine, Sept. 2008 ; Dilnot A, Blastland M. the worst junk stats of 2007. The Times 22 Dec. 2007.
4. Blastland’s criticisms of Pell et al. The following is a summary of Blastland’s arguments and ASH’s responses:

Researchers were slow to publish their data:
It is entirely normal and appropriate for researchers to present their initial findings at conferences for a discussion between colleagues in advance of peer reviewed publication. The Pell study appeared in one of the world’s most respected journals (New England Journal of Medicine) and there was no particular delay.

Findings do not match routine discharge data:
The authors never tried to disguise this difference and Pell was among the authors of a presentation of those routine data at the same conference but in this study they are measuring Acute Coronary Syndrome which is broader than Acute Myocardial Infarction (AMI), the definition used for routine discharge monitoring. The two are similar enough for broad comparisons but they have important differences. Pell’s measure for ACS was more precise and validated by lab tests every time. Many of the patients diagnosed with ACS did not get a discharge code of AMI.

The decline in heart attacks is part of a long term trend:
Pell also specifically acknowledges the long term trend, noting the exceptionally large drop in AMI admissions in 1999-2000 and refers to the 10 year average decline while Blastland cites only to the previous three years. Neither Pell nor Blastland offer any explanation for the exceptionally rapid decline in 1999-2000 a year with winter described as “the sunniest on record” with monthly average temperatures in Scotland from January to March 1.5-2oc warmer than normal.

The study does not prove of causation:
For a scientists to accept “causation” the effect has to be observed consistently across a number of studies, so in that sense no single study can ever demonstrate causation. However, all the published studies examining the effects of smokefree laws on heart attacks have found a similar result. There are also other tests for “causation” including a “plausible and consistent hypothesis” and a “dose/response” effect. Pell’s paper passes each of these tests.

Researchers have been unwilling to discuss their research with journalists:
Blastland complains that “attempts to obtain it [the data] or talk to the lead researcher have gone unanswered”. Perhaps the answer lies in Blastland’s own observations on what he calls “the flurry of excitable headlines”. While the authors may have declined to engage in the journalistic debate, Blastland and Dilnot have likewise yet to subject their analysis to peer review.