Dear Editor,

Over the past few years, many studies have used cross-sectional data from the National Health Interview Survey (NHIS) [1, 2], the Behavioral Risk Factor Surveillance Survey (BRFSS) [3,4,5,6,7,8,9,10]), and the Population Assessment of Tobacco and Health Survey (PATH) [8, 11] to claim that e-cigarette use is associated with asthma [3, 6, 7] chronic obstructive pulmonary disease (COPD) [3, 6,7,8], emphysema, chronic bronchitis [7], stroke [4, 5], congestive heart disease [4] myocardial infarction (MI) [1, 2, 4], pre-diabetes [9, 10] and hypertension [11]. Although study authors acknowledge that smoking is a factor for these illnesses, they generally contend that their results for e-cigarettes are adequately adjusted for smoking. However, the applicability of NHIS and BRFSS is also challenged by the fact that they do not contain any information about when participants started vaping or smoking, and when they were diagnosed with the diseases.

The PATH data have information about when smokers and vapers started, and when they were diagnosed with diseases, but it has been ignored in previous reports [8, 11, 12]. We used the first wave of the PATH survey to provide these respective age distributions for diagnoses of COPD, emphysema, MI and stroke, and compared them with the age ranges of e-cigarette and smoking initiation.

Methods

We used data from the PATH Wave 1 Public-Use File (PUF), administered from September 2013 to December 2014. The PATH collected information from 32,320 civilian and non-institutionalized adults (ages 18 and older) in all 50 United States and the District of Columbia [13].

Tobacco use

We classified participants as never smokers if they had not smoked 100 cigarettes in their lifetime. Current smokers had smoked 100 + cigarettes and smoked every day or some days at Wave 1; former smokers had smoked at least 100 cigarettes in their lifetime and did not smoke at Wave 1. Both current and former smokers were asked their age when they first started smoking cigarettes fairly regularly.

Ever e-cigarette users were respondents who had ever used an e-cigarette even one or two times. They were asked about their age when they first used an e-cigarette.

Medical conditions

All participants were asked whether they had ever been diagnosed with multiple diseases. We chose four conditions that are strongly associated with smoking and for which claims have been made for e-cigarettes: COPD, emphysema, myocardial infarction and stroke. If participants answered affirmatively, a follow-up question asked them when they were first told about the diagnosis.

The PATH PUF provided the following respondents’ age range for both tobacco use and medical conditions: (1) less than 18 years old, (2) 18–24, (3) 25–34, (4) 35–44, (5) 45–54, and (6) 55 years old or older.

Notably, PATH variable names and questions can be found in supplemental Table 1.

Table 1 Cross-tabulation between age (years) first time started using an e-cigarette (Panel A) and first time started smoking cigarettes regularly (Panel B) with age (years) first told COPD, emphysema, MI and stroke
Full size table

Cross-tabulation analysis

Two-dimensional tables cross-tabulates participants’ age at first use of e-cigarettes and age at smoking cigarettes regularly with age when they were told they had individual diseases were obtained. For participants reporting the same age (group) for e-cigarette initiation and diagnosis, a second table cross-tabulating age of e-cigarette initiation with age of regular smoking was obtained

Results

Table 1, Panel A shows the cross-tabulation of age at first e-cigarette use (rows) and at first disease diagnoses (columns). The unshaded cells in the lower left half of the individual tables represent disease cases that cannot be associated with e-cigarettes, because the disease occurred prior to first use. The bold diagonals represent the disease cases that occurred in the same age group as the first e-cigarette exposure. In the remaining shaded cells are the cases that definitely occurred after first exposure. They represent 3.8% of COPD (21/548), 3.9% of emphysema (11/285), 6% of MI (12/200) and 3% of stroke (6/199).

Table 1, Panel B also presents the cross-tabulation results for smoking fairly regularly. In contrast to the findings for e-cigarettes, disease cases among smokers that definitely occurred after first exposure are 97% for COPD (881/908), 96% for emphysema (435/452), 98% for MI (432/439) and 93% for stroke (326/351).

In Panel A the largest numbers lie in the bold diagonals, in which participants’ diagnoses with these diseases occurred in the same age group as they first used e-cigarettes.

We investigated further by cross-tabulating between age when first using e-cigarettes and age at first smoking regularly among participants in the Table 1 Panel A diagonals. The shaded areas of Table 2 show that large majorities of participants with COPD (97%), emphysema (96%), MI (99%) and stroke (89%) who reported the same age group for first e-cigarette use and first diagnosis were smoking cigarettes regularly prior to using e-cigarettes. In fact, in most cases, it was decades earlier. Diagnosed cases in Table 2 who never smoked were very rare, only six COPD, five emphysema, one MI and seven stroke.

Table 2 Cross-tabulation between age first used an e-cigarette and age first started smoking cigarettes regularly among participants who reported the same ages for first used e-cigarettes and first told had diseases (i.e., diagonals in Table1)
Full size table

Discussion

Our simple cross-tabulation tables have revealed crucial information relevant to previous studies. Only 4% of COPD and emphysema cases, 6% of MIs and 11% of stroke had occurred in PATH participants who had used e-cigarettes prior to the diagnoses. In contrast, over 94% of all cases occurred after smoking cigarettes fairly regularly.

For those participants who were the same age range when they first used an e-cigarette and were first diagnosed, large majorities had smoked cigarettes regularly prior to e-cigarette initiation and only a few never smokers used e-cigarettes. This evidence indicates a potential reverse association between e-cigarette use and these diseases. In other words, having a diagnosis for respiratory and cardiovascular diseases leads smokers to use e-cigarettes. Even when the age of e-cigarette initiation and the age of diagnosis is the same in Table 2, Panel B, smoking remains a dominant factor with these diseases.

Studies based on cross-sectional data with no information on age of e-cigarette initiation and age of diagnosis invariably overestimate associations by including cases that were diagnosed before e-cigarette exposure [12, 14]. Although the authors of those studies did not make causal claims in the reports, university media releases [15, 16] and subsequent media articles invariably misled the public to believe that e-cigarette use increases risk for diseases.

Using PATH restricted data, which contains actual ages instead of age groups, would enhance the details of the results presented here, but it would not negate them. However, presentations of the findings from restricted data might be compromised by restrictions on cell counts involving small numbers.

The ages of e-cigarette initiation and first diagnosis were not addressed in previous studies using PATH data [8, 11], but those omissions resulted in retraction of another study claiming that vaping causes heart attacks [12]. Furthermore, our findings clearly apply to studies using cross-sectional data such as NHIS and BRFSS, which do not contain any information on exposure initiation or diagnosis. In the absence of that information, those studies exaggerate the associations and are simply unreliable.