Page 119 - JSOM Summer 2018

An Ongoing Series

Epidemiologic Evidence and Possible Mechanisms for the
Association Between Cigarette Smoking and Injuries

Part 2: Is the Relationship Between Smoking and Injuries Causal?

Joseph J Knapik, ScD *; Sheryl A Bedno, MD 2
1

ABSTRACT
1
Part 1 of this series reviewed the epidemiologic evidence for association between cigarette smoking and injuries. However,
the association between cigarette smoking and injuries and an association between smoking and injuries does not nec-
possible biological and psychosocial mechanisms to account essarily imply that smoking directly causes the injury. There
for this relationship. In the present article, nine criteria are could be some intervening factor and it is important to know
explored to determine if smoking is a direct cause of injuries this from a clinical and a public health perspective. For exam-
(i.e., a causal relationship). There is substantial evidence that ple, in part 1, we cited literature demonstrating that smoking
1
individuals who smoked in the past have a higher subsequent was associated with risk taking. If it is risk-taking behavior
risk of injury. A recent meta-analysis found that smokers in the rather than smoking per se that is more directly related to the
military were 1.31 times more likely to be injured than non- greater risk of injury, smoking cessation by itself may not be
smokers and Servicemembers with low, medium, and high lev- effective in reducing injuries. In this case, it may be more effec-
els of smoking had 1.27, 1.37, and 1.71 times, respectively, the tive to attempt to modify aspects of the risk taking-behavior
risk of injury compared with nonsmokers. The association be- if the goal is to reduce injuries. This article, part 2 of this se-
tween smoking and injuries has been reported in at least 18 US ries, examines the possibility that there is a causal relationship
military studies and in 14 civilian studies in seven countries. between smoking and injuries; that is, that smoking directly
The biological plausibility of the association was discussed ex- causes injuries.
tensively in part 1 of this series. A possible alternative explana-
tion with sufficient data was that smokers may be risk takers Criteria for Determination of Causal Effects
and it is the risk-taking behavior that increases injury risk (not
smoking per se). Once an individual no longer smokes, a de- Is there sufficient evidence to assume that smoking causes in-
crease in injury risk has been reported for at least bone health juries? Criteria developed to determine whether an association
and wound healing. The effects of smoking do not appear to is causal are presented in Table 1. Each of these criteria is
2–4
be specific to one type of injury, possibly because of the nu- examined in this article to see if there are sufficient data to
merous compounds in tobacco smoke that could affect tissues infer that smoking can directly cause injuries, or at least some
and physiological processes, with evidence provided for bones, types of injuries.
tendons, and healing processes. The association was consis-
tent with other knowledge, with some evidence provided from Temporal Relationship
other types of medical problems and trends in smoking and The first criterion for a causal relationship is simply that the
injury-related mortality. In summary, the association between exposure should proceed the outcome. In this case, there is
smoking and injuries appears to meet many of the criteria for substantial evidence that individuals who smoked in the past
a causal relationship. (exposure) had a higher subsequent risk of injuries (outcome).
Those who did not smoke had a lower risk for many types of
Keywords: smoking; mortality; injury; epidemiology injuries. 5–22

Strength of Association
The second criterion is that the stronger the association, the
Introduction
more likely it is that smoking causes injuries. Studies that have
In part 1 of this series, we reviewed the possible biological and examined self-reported cigarette smoking and injuries in mili-
psychosocial mechanisms to account for the well-documented tary training have found that the relative risk of smoking has

*Correspondence to joseph.j.knapik.ctr@mail.mil
1 MAJ (Ret) Knapik served in the US military as a wheel vehicle mechanic, medic, Medical Service Corps officer, and Department of Defense civil-
ian employee. He is currently a senior epidemiologist/research physiologist with the Henry M. Jackson Foundation and an adjunct professor at
2
Uniformed Services University, Bethesda, MD; and Bond University, Robina, Australia. COL Bedno is chief, Department of Preventive Medicine,
at Womack Army Medical Center, Fort Bragg, NC, She is a preventive medicine and occupational medicine physician and researcher.
117