TABLE 1.

Effect direction plot for biomarkers of harm, by comparison group.

Biomarker class	Biomarker	Group comparisons
		EC versus CC	EC versus EC + CC	Dual use (EC + CC) versus CC
Mercapturic acids	3‐HPMA (3‐hydroxypropylmercapturic acid)	↓↓ Cravo ↓↓ Morris	↔ McRobbie ↓ Goniewicz ↓↓ Morris ↔ Pulvers	↓↓ Morris
	SPMA (S‐phenylmercapturic acid)	↓↓ Cravo ↓↓ Morris	↓ Goniewicz ↓↓ Morris	↓↓ Morris
	HEMA (2‐hydroxyethylmercapturic acid)	↓↓ Morris	↑ Goniewicz ↔ Morris ↔ Pulvers	↓↓ Morris
	MHBMA (2‐hydroxy‐3‐buten‐1‐ylmercapturic acid)	↓↓ Morris	↓ Goniewicz ↓ Morris	↓↓ Morris
	HPMMA (3‐hydroxy‐1‐methyl propylmercapturic acid)		↓ Goniewicz ↔ Pulvers
	AAMA (N‐acetyl‐S‐(carbamoylethyl)‐l‐cysteine (synonym: 2‐carbamoylethylmercapturic acid))		↓ Goniewicz ↓↓ Pulvers
	CNEMA (2‐cyanoethylmercapturic acid)	↓↓ Morris	↓ Goniewicz ↓↓ Pulvers ↓↓ Morris	↓↓ Morris
	2‐HPMA (2‐hydroxypropylmercapturic acid)		↓ Goniewicz ↔ Pulvers
	3‐HMPMA (3‐hydroxy‐1‐methylpropyl‐mercapturic acid)	↓↓ Morris	↓↓ Morris%60	↓↓ Morris
	PMA (phenylmercapturic acid)		↓↓ Pulvers
	MMA (N‐nitrosodimethyamine)		↔ Pulvers
Nitrosamines	NNAL (4‐(methylnitrosamino)‐1‐(3‐pyridyl)‐1‐butanol)	↓↓ Cravo ↓↓ Morris	↓ Goniewicz ↔ Morris ↓↓ Pulvers	↓↓ Morris
Metabolites of polyaromatic hydrocarbons	1‐Hydroxyfluorene		↑ Goniewicz
	3‐, 4‐Hydroxyphenanthrenes		↑ Goniewicz
	2‐Hydroxyfluorene		↑ Goniewicz
	1‐Hydroxypyrene (1‐OHP)	↓↓ Morris	↑ Goniewicz ↓↓ Morris	↓↓ Morris
	3‐Hydroxyfluorene		↓ Goniewicz
	2‐Hydroxyphenanthrene		↑ Goniewicz
	1‐Hydroxyphenanthrene		↑ Goniewicz
	2‐Naphtol		↓ Goniewicz
Other known carcinogens	o‐tol (o‐toluidine)	↓↓ Morris	↓↓ Morris	↓↓ Morris
	1‐AN (1‐aminonaphthalene)	↓↓ Morris	↓↓ Morris	↓↓ Morris
	2‐AN (2‐aminonaphthalene)	↓↓ Morris	↓↓ Morris	↓↓ Morris
	NNN (N‐nitrosonornicotine)	↔ Morris	↔ Morris	↔ Morris
	3‐OH B[a]P (3‐hydroxybenxo[a]pyrene)	↓↓ Morris	↓↓ Morris	↓↓ Morris

Cravo et al. [8] reports data on a subset of participants who were confined for the first week of the study; all complied with study protocols during this period. Data are available in figures only; error bars do not overlap. In Gonieciwz et al. [9], all participants were given an electronic cigarette (EC) at baseline. After 2 weeks, statistically significant declines in 12 of 17 measured biomarkers of exposure to toxicants were observed. Authors also conducted a secondary analysis comparing changes at 2 weeks between those who switched completely to EC and those who used both EC and combustible cigarettes (CC). They give only absolute values (in their Supporting information, Table S3). In Morris et al. [6], all participants exclusively used EC at days 1–9 and then were randomized to three groups from days 10 to 14 (EC, CC and EC + CC); participants were confined for the duration of the study. Authors report between‐group differences and breakdown into two separate samples (studies 1 and 2). Study 1 has the larger sample size so is included in Table 1. Study 2 found consistent directions of effect for all markers. Pulvers et al. [13] assigned all participants to EC and reports median and interquartile ranges (IQRs) for a number of toxic exposures across three groups at 4 weeks: people who switched to EC for at least the first 2 weeks (n = 10); people who switched exclusively for the full 4 weeks (n = 6); and people who used both EC and CC throughout the entire 4‐week period (n = 21). The latter two groups are compared in Table 1. ↓↓ lower point estimate in EC group, with 95% confidence interval (CI)/error bars non‐overlapping; ↓ lower point estimate in EC group, with 95% CI/error bars/P‐value not provided; ↔ CI/error bars overlap; ↑ higher point estimate in EC group, with 95% CI/error bars/P‐value not provided)