Changing conclusions on secondhand smoke in a sudden infant death syndrome review funded by the tobacco industry

Abstract

Background: Prenatal and postnatal exposure to tobacco smoke adversely affects maternal and child health. Secondhand smoke (SHS) has been linked causally with sudden infant death syndrome (SIDS) in major health reports. In 1992, the US Environmental Protection Agency (EPA) first noted an association between SHS and SIDS, and both prenatal exposure and postnatal SHS exposure were listed as independent risk factors for SIDS in a 1997 California EPA report (republished in 1999 by the National Cancer Institute) and a 2004 US Surgeon General report. The tobacco industry has used scientific consultants to attack the evidence that SHS causes disease, most often lung cancer. Little is known about the industry's strategies to contest the evidence on maternal and child health. In 2001, a review was published on SIDS that acknowledged funding from the Philip Morris (PM) tobacco company. Tobacco industry documents related to this review were examined to identify the company's influence on the content and conclusions of this review.

Methods: Tobacco industry documents include 40 million pages of internal memos and reports made available to the public as a result of litigation settlements against the tobacco industry in the United States. Between November 2003 and January 2004, we searched tobacco industry document Internet sites from the University of California Legacy Tobacco Documents Library and the Tobacco Documents Online website. Key terms included "SIDS" and names of key persons. Two authors conducted independent searches with similar key terms, reviewed the documents, and agreed on relevancy through consensus. Thirty documents were identified as relevant. Two drafts (an early version and a final version) of an industry-funded review article on SIDS were identified, and 2 authors independently compared these drafts with the final publication. Formal comments by PM executives made in response to the first draft were also reviewed. We used Science Citation Index in July 2004 to determine citation patterns for the referenced SIDS reviews.

Results: PM executives feared that SHS and maternal and child health issues would create a powerful and emotional impetus for smoke-free areas in the home, public areas, and the workplace. In response to the 1992 US EPA report on SHS, the Science and Technology Department of PM's Switzerland subsidiary, Fabriques de Tabac Reunies, searched for "independent" consultants to publish articles addressing SHS. The first industry-funded article was a literature review focusing on smoking and SIDS, conducted by consultant Peter Lee and co-author Allison Thornton, which stated that the association between parental smoking and SIDS could have been attributable to the failure to control fully for confounders. That first review has only been cited once, in the subsequent industry-funded review. In 1997, PM commissioned a consultant, Frank Sullivan, to write a review, with coauthor Susan Barlow, of all possible risk factors for SIDS. The first draft concluded that prenatal and postnatal smoking exposures are both independent risk factors for SIDS. After receiving comments and meeting with PM scientific executives, Sullivan changed his original conclusions on smoking and SIDS. The final draft was changed to emphasize the effects of prenatal maternal smoking and to conclude that postnatal SHS effects were "less well established." Changes in the draft to support this new conclusion included descriptions of Peter Lee's industry-funded review, a 1999 negative but underpowered study of SIDS risk and urinary cotinine levels, and criticisms of the conclusions of the National Cancer Institute report that SHS was causally associated with SIDS. In April 2001, the Sullivan review was published in the United Kingdom journal Paediatric and Perinatal Epidemiology, with a disclosure statement that acknowledged financial support from PM but did not acknowledge contributions from PM executives in the preparation of the review. By 2004, the Sullivan SIDS review had been cited at least 19 times in the medical literature.

Conclusions: PM executives responded to corporate concerns about the possible adverse effects of SHS on maternal and child health by commissioning consultants to write review articles for publication in the medical literature. PM executives successfully encouraged one author to change his original conclusion that SHS is an independent risk factor for SIDS to state that the role of SHS is "less well established." These statements are consistent with PM's corporate position that active smoking causes disease but only public health officials conclude the same for SHS. The author's disclosure of industry funding did not reveal the full extent of PM's involvement in shaping the content of the article. This analysis suggests that accepting tobacco industry funds can disrupt the integrity of the scientific process. The background of this SIDS review is relevant for institutions engaged in the debate about accepting or eschewing funding from the tobacco industry. Those who support acceptance of tobacco industry funds argue that academic authors retain the right to publish their work and maintain final approval of the written product, but this argument fails to recognize that the tobacco industry funds work to ensure that messages favorable to the industry are published and disseminated. Clinicians, parents, and public health officials are most vulnerable to the changed conclusions of the SIDS review. The national SIDS "Back to Sleep" campaign has been very successful in reducing SIDS rates. However, estimates of SIDS risk from SHS (odds ratios range from 1.4 to 5.1) have considerable overlap with estimates of risk from prone sleep positioning (odds ratios range from 1.7 to 12.9). With the Back to Sleep campaign well underway, efforts to address parental smoking behavior in both the prenatal and postnatal periods should be intensified. The tobacco industry's disinformation campaign on SHS and maternal and child health can be counteracted within clinicians' offices.