Both Duration and Pack-Years of Tobacco Smoking Should Be Used for Clinical Practice and Research

Tobacco smokers are more likely to develop cardiac and pulmonary diseases than never-smokers, and risk increases in a dose-dependent manner. Pack-years, the product of cigarettes per day in increments of packs (one measure of intensity) and duration (years smoked), is a cumulative exposure indicator of smoking burden, routinely used to estimate patients’ risk of developing smoking-related diseases for clinical practice, research, and healthcare policies. The use of pack-years is based on the assumption that both intensity and duration are equally important. Research suggests this is not the case; duration should be considered more important than intensity or the composite pack-years in determining risk of developing smoking-related diseases, such as lung cancer (1–3) and coronary artery disease (4), as well as affecting disease severity in chronic obstructive pulmonary disease (COPD) (5).

Evolution of Term Pack-Years to Define Tobacco Smoking History

In early clinical studies examining the dose–response relationship in smoking-related diseases, cigarettes per day and years smoked were used to define tobacco exposure. In a landmark study in 1950 examining the effect of tobacco smoking on the occurrence of lung cancer, Doll and Hill used smoking exposure metrics of cigarettes smoked in lifetime, age of starting smoking, and cigarettes per day (6). They arbitrarily chose cigarettes per day categories ranging from 1 to 4 cigarettes to >40 cigarettes/d for their analysis. All tobacco smoking measures were found to increase lung cancer risk, including cumulative cigarettes smoked in lifetime—a pack-years equivalent. About 95% of individuals diagnosed with lung cancer had smoked for >20 years, and 79% had initiated smoking at <20 years of age, indicating that exposure metrics that consider duration are highly predictive of disease risk (6). Pack-years were first reported in the medical literature in the mid-1950s, when Franklin and colleagues found a strong relationship between pack-years and lung function impairment in patients with COPD (7). Among major epidemiologic publications, it was not until the 1971 U.S. Surgeon General’s report on the health effects of smoking that pack-years were first mentioned (8). Numerous studies have since found a strong relationship between pack-years and the occurrence and severity of pulmonary and cardiac diseases but often did not undertake analyses comparing smoking duration with intensity on outcomes.

Key Dose–Response Studies in Lung Cancer

In one of the first studies to critically evaluate the dose–response relationship between tobacco smoking and the risk of lung cancer, Doll and Peto found the risk to be exponentially greater, between the fourth and fifth power, for years smoked, whereas cigarettes per day was to the second power (1). Pack-years or other cumulative exposure measures were not reported in this study. Nearly 2 decades later, Lubin and Caporaso reported persons smoking at a lower intensity over a long time period were more likely to develop lung cancer than those smoking at a higher intensity over a shorter time period (2). At higher smoking intensity, they found that the excess odds per additional pack-years of smoking were lower than for lower smoking intensities (excess odds ratio per pack-years: 0.293, 0.315, 0.247, and 0.203 for smoking intensities <20, 20–29, 30–39, and ≥40 cigarettes per day, respectively) (2). More recently, a case–control study found the odds ratio for lung cancer was greater with pack-years or duration than cigarettes per day (3). The odds of lung cancer increased more with every 10-U increase in exposure for duration (odds ranging from 1.23–28.94 for 20 to >50 yr of exposure) than for intensity (3.18–11.87 for 20 to >40 cigarettes per day) (3). Lung cancer mortality also appears to be more strongly affected by years smoked than smoking intensity. Among 746,485 smokers enrolled in the American Cancer Prevention Study II, the association between duration and lung cancer mortality (risk ratio for log-transformed years ranging from 3.5 to 16) was exponentially greater than for cigarettes per day (risk ratio for log transformed intensity ranging from 1.4 to 2.6) across strata of age and sex (9).

Key Dose–Response Study in Obstructive Lung Diseases

Obstructive lung disease also appears to be affected to a greater extent by smoking duration than intensity. In the COPDGene cohort of smokers (≥10 pack-year history), Bhatt and colleagues compared the dose–effect relationships of cigarettes per day, pack-years, and smoking duration with disease severity measures (5). There was a linear decrease in adjusted mean forced expiratory volume in 1 second/forced vital capacity with years smoked over all ranges of smoking intensity (β = −0.041 ± 0.004; P < 0.001) but a relatively flat slope for cigarettes per day across all ranges of smoking duration (β = −0.009 ± 0.009; P = 0.34). Computed tomography (CT) measures of emphysema and gas trapping were similarly more strongly associated with smoking duration than intensity (5).

Going beyond Using Pack-Years for COPD and Lung Cancer

First and foremost, patients and healthcare providers should recognize that it is the number of years smoked, and not how much tobacco one smokes daily, that is the most important measure associated with risk of adverse health outcomes in COPD and lung cancer. In other words, even small amounts of daily tobacco smoking over an extended time period have significant health risks.

Although there is sufficient evidence to show that COPD severity on the basis of clinical and radiographic measures is more dependent on years smoked than pack-years or cigarettes per day (5), it is unclear which smoking metric provides the best estimate for the risk of developing COPD. The implications of relying on pack-year history alone while not considering duration of smoking are substantial for lung cancer, in particular regarding population screening with low-dose CT. In the United States, the National Lung Screening Trial established the criteria for use of low-dose CT among current or former (quit within 15 yr) smokers, aged 55 to 74 years with at least a 30 pack-year smoking history (10), but no specific rationale was provided for this amount of tobacco smoking or cessation interval. In the NELSON study in Europe investigating lung cancer screening in the general population, inclusion criteria for smoking were based on a combination of years smoked and pack-years (11). High-risk subjects were defined as those between the age of 50 and 74 years, who had smoked at least 15 cigarettes/d for ≥25 years or 10 cigarettes/d for ≥30 years. Omission of low-intensity, long-term smokers from screening leads to disparities, particularly in black and Hispanic individuals who typically smoke a half-pack or less daily but often with longer durations (12). Compared with white individuals, black adults are more likely to be diagnosed with lung cancer and have the worst outcomes (hazard ratio for all-cause mortality, 1.35) but accrue greater survival benefit from lung cancer screening (hazard ratio for mortality, 0.61 vs. 0.86) (13).

There are some pragmatic issues also in support of reporting smoking duration: 1) patient recall is more reliable for years smoked than pack-years because use is variable over time, 2) widely used electronic medical record systems for patient care contain a field to document years of use, 3) tobacco years is a simplified approach to define exposure to smoking pipes, cigars, or hookah that are harder to quantify than cigarettes, and 4) duration provides a practical measure for population health surveys of tobacco use.

Conclusions

On the basis of current evidence, we believe it is time for a paradigm change in assessing and reporting patients’ smoking history to include both the composite of pack-years and the duration of smoking—whether it be for patient care or research. Low-intensity smokers and those who have decreased their daily consumption because “it is safer” should be informed that a few cigarettes each day can be as damaging to the lungs as a pack or more a day. Methodologic research is needed in epidemiologic and clinical studies of COPD and lung cancer to determine how to simultaneously incorporate measures of duration and intensity. In particular, further investigation is needed to determine if reliance on pack-years is appropriate to guide healthcare policy or for enrollment in clinical trials of COPD, or for other uses, all of which typically rely on a minimum of a 10 pack-year history.