Fanning the coals
Clearing the stem.
18+Reference resource. Nicotine is addictive. Tobacco use harms your health.
Clearing the stem.
Around 90% of the carbon monoxide and more than 95% of the benzo(a)pyrene in waterpipe smoke comes not from tobacco and not from the blend in the bowl, but from the burning charcoal. Almost everything else in this article follows from that single figure: replacing the base with a tobacco-free one mathematically cannot improve the picture — and, as the measurements show, it does not.
Monzer and colleagues (2008) separated the sources of toxicants in waterpipe smoke and obtained a distribution that is hard to argue with: "approximately 90% of the CO and 75–92% of the 4- and 5-membered ring PAH compounds originated in the charcoal. Greater than 95% of the benzo(a)pyrene in the smoke was attributable to the charcoal."
The check from the opposite direction came from a review by Shihadeh and colleagues (2015): benzo(a)pyrene "dropped from 170 to 9 ng/session when electrical heating was substituted for charcoal". An almost twentyfold fall — with the blend in the bowl unchanged.
From this follows a simple consequence worth holding in mind while reading any package. The contents of the bowl determine nicotine, flavour, and part of the aerosol. The charcoal determines carbon monoxide and carcinogenic PAHs. These are different sources, and by removing the tobacco you are working on the one that is not responsible for the bulk of the toxic load.
The key study is Shihadeh and colleagues (2012): double-blind, 31 smokers, with the products matched by flavour. That design detail matters: participants did not know what they were smoking, and therefore could not adjust their behaviour to expectations.
Yield per session:
| Measure | Tobacco preparation | Tobacco-free preparation | Cigarette |
|---|---|---|---|
| Nicotine, mg | 1.04 | < 0.01 | 0.73 |
| Carbon monoxide, mg | 155 | 159 | 12.0 |
| Tar, mg | 464 | 513 | 9.4 |
| Benzo(a)pyrene, ng | 51.8 | 66.1 | 6.6 |
| Formaldehyde, µg | 58.7 | 117.6 | 20.6 |
The authors' conclusion: "Nicotine yield was the only outcome that differed significantly between preparations. These findings contradict advertising messages that 'herbal' waterpipe products are a healthy alternative to tobacco products."
Note the shape of the result. One outcome differed significantly — nicotine. On the other four the tobacco-free preparation was no lower than the tobacco one, and for formaldehyde it was twice as high. The formaldehyde difference is not reported as statistically significant in that study, but no advantage for the tobacco-free product is visible in these numbers either.
An honest note on the limits of this study: acrolein was below the limit of detection for both products, so no comparison could be made for it at all, and benzene was not measured.
Soex, in the description of its Herbal line, claims "0% Nicotine" and "0% Tar". Measurements confirm the first number: less than 0.01 mg per session versus 1.04 mg for the tobacco blend — a difference of two orders of magnitude, and it is real.
Measurements do not confirm the second number at all. In the laboratory the tobacco-free product yielded 513 mg of tar per session — more than the tobacco one at 464 mg. And this is not about a single brand: "0% Tar" is a standard formulation for the category, and 513 mg is a measured yield for the category.
The gap has an explanation: "tar" in smoke is not carried over from the blend ready-made but formed by the thermal decomposition of the bowl contents and the charcoal. The absence of tar in the unburned product does not mean its absence in the smoke — these are two different substances in two different states. This is exactly why the WHO, in the second edition of its Advisory note (2015, p. 49), directs its objection not at the product but at the label: "Even 'tobacco free' or 'herbal' waterpipe alternatives contain large doses of toxicants, and the packaging should not be allowed to carry health or safety claims".
There is a second layer to this subject: even if the smoke were identical, smoker behaviour does not stay identical.
Cobb and colleagues (2012) compared sessions with a tobacco and a tobacco-free product and found a noticeable gap. Exhaled carbon monoxide: 30.9 ppm after the tobacco-free session versus 11.8 ppm after the tobacco one. Total smoke volume: 57 litres versus 31. Mean puff: 683 ml versus 423. The authors phrase it cautiously: "More smoke was inhaled during tobacco-free product use, potentially reflecting attempted regulation of nicotine intake."
The logic of the hypothesis runs like this: no nicotine arrives from the tobacco-free product, no sense of sufficiency follows, and the person smokes longer and deeper — and with the volume of smoke rises the dose of carbon monoxide, which has nothing to do with nicotine.
Honesty is required here, otherwise the thesis turns into a slogan. Shihadeh and colleagues (2012) did not see compensation. So this is a reproducible tendency — found by Cobb in 2012 and repeated in 2015 — but not a law that fires in every experiment. Cobb and colleagues (2015) added a detail: compensation is stronger in frequent smokers, and the largest rise in carboxyhaemoglobin was in that very group, on the placebo product.
The gap between subjective experience and biochemistry was shown most vividly by Blank and colleagues (2011).
Heart rate after a session rose by 8.6 beats per minute with the tobacco product versus 1.3 with the placebo — a large and expected difference, the pharmacological action of nicotine. Carboxyhaemoglobin, however, rose equally: 3.8% versus 3.9%. And participants rated craving relief, "buzz" and "satisfaction" the same for both products.
Two things follow. First: the subjective sense of "lightness" in a nicotine-free session is unrelated to how much carbon monoxide has bound to haemoglobin — the body reports nothing about that through sensation until high concentrations. Second: the satisfaction from the session was not delivered by nicotine. That participants could not tell the products apart on "buzz" means the ritual, the flavour and the mechanics of smoking supply a substantial part of the effect — which matters both for understanding the habit and for understanding why a nicotine-free version is not perceived as pointless.
Another slice is not toxicant yield but biological effect. Shihadeh and colleagues (2014) studied the effect of both kinds of smoke on lung cells. The result is stated in one line: "There were no significant differences across product in any measure."
The same group's summarising review (2015) pulls it together: "Except nicotine, smoke generated using tobacco-free preparations marketed for 'health conscious' users contains the same or greater doses of toxicants, with the same cellular effects as conventional products."
The WHO position (Advisory note on waterpipe tobacco smoking, second edition, 2015) is stated on page 32: "except for nicotine, smoke from tobacco-free waterpipe products has the same toxicant content and biological activity as that from tobacco-based products".
The same document gives the scale of the per-session load for a tobacco blend, 10 g of maassel on 1.5 quick-lighting charcoals: 2.94 mg of nicotine, 802 mg of tar, 145 mg of carbon monoxide.
The CDC puts it more briefly: "studies of tobacco-based shisha and 'herbal' shisha show that smoke from both preparations contains carbon monoxide and other toxic agents." It also gives the volume comparison that explains why a waterpipe session is a case of its own: roughly 90,000 ml of smoke in an hour versus 500–600 ml for a cigarette.
A separate line of data concerns not the smoker but the room — and it shows that the declared contents of a menu are a poor predictor of the contents of the air.
Hammal and colleagues (2015) measured heavy metals and PAHs in the unburned herbal product and found them at cigarette levels or above; one sample was particularly rich in lead, chromium, nickel and arsenic. For tar per session: 909 mg for the tobacco blend versus 2350 mg for one of the herbal samples. In cafés that on paper served only herbal products, PM2.5 and carbon monoxide were higher than in a casino where cigarette smoking was permitted, and nicotine was detected in one venue. A caveat: the full text of this work is paywalled; an abstract and a secondary account were used.
Zhou and colleagues (2015) surveyed New York hookah bars that declared herbal blends only: PM2.5 at 1179.9 µg/m³, carbon monoxide at 32 ppm, and nicotine detected in the air of every bar surveyed.
Chronic exposure is not the only storyline. A PubMed search on carbon monoxide poisoning in connection with waterpipe smoking returns 50 publications — no longer isolated case reports but an established category of presentations.
Abdul-Nabi and colleagues (2024): of 111 emergency-department presentations with carbon monoxide poisoning, 73.9% were waterpipe-related. In that group carboxyhaemoglobin was 19.7% versus 13.7% in the rest, and syncope occurred in 52.4% of cases versus 17.2%.
Inyang and colleagues (2020) estimated the scale in the United States: about 1371 presentations over 2011–2019, of which 66.8% were people aged 18–24.
An honest gap matters here: clinical reports do not distinguish the type of blend, and no case of poisoning specifically from a tobacco-free blend was found in the verified sources. But that gap is closed directly by the experimental data — carbon monoxide comes from the charcoal, the charcoal in a tobacco-free session is the same charcoal, and the CO yield in the laboratory was no lower.
One case where a nicotine-free option genuinely turned out cleaner was found: steam stones with electric heating (Schober and colleagues, 2017). "Compared to conventional waterpipes, the release of pollutants is lower. Nevertheless, smoking with Shiazo waterpipes is a source of health risks."
The reason for the improvement is the absence of charcoal, not the absence of nicotine. That is exactly what Monzer's figure predicts: remove the charcoal and you remove the main source of CO and PAHs. Even in that setup, however, formaldehyde, acetaldehyde and PAHs in the air rose, and an acrolein metabolite rose in the smokers' urine.
In Russia the absence of tobacco and nicotine does not take a product out of anti-tobacco legislation. Article 2, paragraph 1 of Law No. 15-FZ defines a waterpipe through the heating of tobacco or of products not containing tobacco leaf. Article 2, paragraph 3 classes a tobacco-free mixture for heating as nicotine-containing product, and Article 2, paragraph 4.2 (introduced by Federal Law No. 542-FZ of 28 December 2024) defines it as a type of nicotine-containing or nicotine-free article. Article 12, part 1, paragraph 14 prohibits the use of waterpipes on food-service premises; Article 16 prohibits advertising and the stimulation of sale and consumption; Article 20 prohibits sale to and consumption by minors.
The limits of knowledge deserve to be marked as clearly as the knowledge itself:
What produces most of the toxicants in waterpipe smoke?
The charcoal. About 90% of the carbon monoxide and more than 95% of the benzo(a)pyrene originate there, not in the blend in the bowl (Monzer et al., 2008).
Is there less carbon monoxide in the smoke of a tobacco-free blend?
No. In a double-blind comparison, 159 mg per session versus 155 mg for the tobacco blend; the difference is not significant.
Is "0% Tar" true for tobacco-free blends?
Possibly for the unburned product, but 513 mg of tar per session was measured in the smoke — more than for the tobacco blend (464 mg). Tar is formed during heating; it is not carried over ready-made from the blend.
Are nicotine-free blends smoked differently?
According to Cobb et al. (2012), yes: a larger smoke volume (57 l versus 31) and roughly threefold higher exhaled CO. But Shihadeh et al. (2012) did not see this effect, so it is a tendency, not a rule.
Does a nicotine-free session feel lighter?
Heart rate rises far less (+1.3 versus +8.6 bpm), while carboxyhaemoglobin rises equally (3.9% versus 3.8%). Subjective lightness and carbon monoxide dose are unrelated.
Are there cases where a nicotine-free option is cleaner?
One: steam stones with electric heating. The improvement came from the absence of charcoal, not the absence of nicotine, and the authors still call the method a source of health risks.
Related topics on the site: on which categories of nicotine-free blends actually exist — "'Protein Hookah' and Other Nicotine-Free Blends: What Is Behind the Trend"; on composition and brands — "Hookah Without Tobacco: Tea, Fruit, and Mineral Blends"; on nicotine doses — "How Much Nicotine Is in Hookah Smoke Compared to Cigarettes"; on indoor air — "Secondhand Hookah Smoke: What Non-Smokers Actually Inhale".