Article Text

Download PDFPDF

Thirdhand cigarette smoke in an experimental chamber: evidence of surface deposition of nicotine, nitrosamines and polycyclic aromatic hydrocarbons and de novo formation of NNK
  1. Suzaynn F Schick1,
  2. Kathryn F Farraro2,
  3. Charles Perrino3,
  4. Mohamad Sleiman4,
  5. Glenn van de Vossenberg5,
  6. Michael P Trinh6,
  7. S Katharine Hammond3,
  8. Bryan M Jenkins7,
  9. John Balmes1
  1. 1Department of Medicine, University of California, San Francisco, California, USA
  2. 2Department of Bioengineering, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
  3. 3Department of Public Health, University of California, Berkeley, Berkeley, California, USA
  4. 4Department of Environmental Energy Technologies, Lawrence Berkeley Laboratories, Berkeley, California, USA
  5. 5Department of Medicine, Radboud University, Nijmegen, The Netherlands
  6. 6International Medical Systems Inc, San Gabriel, California, USA
  7. 7Department of Biological and Agricultural Engineering, University of California, Davis, Davis, California, USA
  1. Correspondence to Dr Suzaynn F Schick, Department of Medicine, University of California, San Francisco, Box 0843, San Francisco, CA 94143-0843, USA; sschick{at}


Background A growing body of evidence shows that secondhand cigarette smoke undergoes numerous chemical changes after it is released into the air: it can adsorb to indoor surfaces, desorb back into the air and undergo chemical changes as it ages.

Objectives To test the effects of aging on the concentration of polycyclic aromatic hydrocarbons (PAHs), nicotine and tobacco-specific nitrosamines in cigarette smoke.

Methods We generated sidestream and mainstream cigarette smoke with a smoking machine, diluted it with conditioned filtered air, and passed it through a 6 m3 flow reactor with air exchange rates that matched normal residential air exchange rates. We tested the effects of 60 min aging on the concentration of 16 PAHs, nicotine, cotinine and tobacco-specific nitrosamines. We also measured sorption and deposition of nicotine, cotinine and tobacco-specific nitrosamines on materials placed within the flow reactor.

Results We observed mass losses of 62% for PAHs, 72%, for nicotine, 79% for N-nitrosonornicotine and 80% for 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK). Extraction of cotton cloth exposed to smoke yielded nicotine and NNK. The ratio of NNK:nicotine on the exposed cloth was 10-fold higher than that in aerosol samples.

Conclusions Our data suggest that the majority of the PAHs, nicotine, cotinine and tobacco-specific nitrosamines that are released during smoking in homes and public places deposit on room surfaces. These data give an estimate of the potential for accumulation of carcinogens in thirdhand cigarette smoke. Exposure to PAHs and tobacco-specific nitrosamines, through dermal absorption and inhalation of contaminated dust, may contribute to smoking-attributable morbidity and mortality.

  • Carcinogens
  • Environment
  • Secondhand smoke
  • Nicotine
  • Cotinine
View Full Text

Statistics from


For every eight smokers who die from smoking-related illnesses, one non-smoker dies from secondhand exposure.1 ,2 According to the 2006 report by the US surgeon general, 60% of non-smokers in the USA showed evidence of secondhand smoke (SHS) exposure.3 SHS contains more than 50 known carcinogens and has been declared a Group A carcinogen. SHS exposure results in 3400 deaths from lung cancer in the USA each year, and is also a known risk factor for nasal sinus cancer and breast cancer.2 ,4 Polycyclic aromatic hydrocarbons (PAHs) and tobacco-specific nitrosamines (TSNAs) are known human carcinogens that are present in SHS.5 ,6

Recent evidence has shown that many of the chemical components of SHS are not removed by ventilation and that they persist in indoor environments long after the cigarettes are extinguished. The term for this newly recognised set of chemical compounds deriving from secondhand smoke is ‘thirdhand smoke’ (THS).7 Research has shown that exposure to THS can increase urinary cotinine levels in non-smokers who are not exposed to SHS.8 ,9 Nicotine is the most abundant organic compound released during smoking. Previous research has shown that nicotine sorbs rapidly onto indoor surfaces8 ,10–19 and can react with ambient levels of ozone and nitrous acid to form the TSNAs 1-(N-methyl-N-nitrosamino)-1-(3-pyridinyl)-4-butanal (NNA), 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) and N-nitrosonornicotine (NNN), formaldehyde, n-methylformamide, nicotinaldehdye, cotinine and secondary organic aerosols.20–22

Evidence from animal exposure studies suggests that the chemical changes that occur when cigarette smoke ages and interacts with surfaces can increase the respiratory toxicity of SHS. Using data from studies performed by Philip Morris in the 1980s and 1990s, we found that animals exposed to smoke that had been aged for 30–90 min had a higher incidence of respiratory cell hyperplasia and metaplasia.23

We designed a system to age cigarette smoke to reproduce these normal chemical changes, for studies of the health effects of SHS exposure in human subjects. The system uses a smoking machine to burn cigarettes, dilutes the smoke with conditioned, filtered air and passes it through a large stainless steel flow cell or ‘aging chamber’ to mimic the normal ventilation patterns and surface interactions in homes and public places. The age of the smoke is controlled by the transit time through the system and was set to 60 min, reflecting residential air exchange rates.24 ,25 To test the effect of aging on smoke chemistry we:

  • Measured the concentrations of PAHs, nicotine, cotinine and TSNAs in fresh and aged cigarette smoke aerosol.

  • Tested the effects of mounting paper or cloth inside the stainless steel aging chamber on concentrations of nicotine, cotinine and TSNAs in fresh and aged cigarette smoke.

  • Analysed the cloth and paper exposed to smoke for nicotine, cotinine and TSNAs.


The smoke aging system

The smoke aging system is described in detail in Schick and Farraro et al.26 Briefly, smoke generated by an automatic smoking machine (Model TE-10z, Teague Enterprises, Woodland, California, USA) was diluted into conditioned, filtered air, and conducted through a 6 m3 stainless steel aging chamber. The aging chamber contained three vertically staggered baffles and two internal fans to promote mixing. See figure 1. Ozone from a Pen-Ray ozone generator (# 97-0068-01, UVP, Upland, California, USA) was injected into the chamber at 15 ppb to replace the ambient indoor levels of ozone that were removed by charcoal filtration of the dilution air.27 Smoke aerosol samples were collected from the ducting upstream of the aging chamber (Site 1), from the ducting downstream of the aging chamber (Site 2), from the ducting immediately before the human exposure chamber (Site 3) and from the exposure hood (Site 4).

Figure 1

Design of smoke aging system and location of sampling Sites.

Smoke from Site 1 was ∼3 min old. Smoke age at Sites 2, 3 and 4 was 60 min old. Average ventilation rates in homes in the USA are quite low, ranging from 0.37 to 0.88 air changes per hour (ACH).24 ,28 We chose 60 min aging time, or one ACH, to represent ventilation conditions when a home is occupied and ventilation systems or windows are in use.

Experimental conditions

Nicotine, cotinine and TSNA data are from a series of experiments conducted to test the effects of surface complexity on the loss of chemicals from the aerosol, and on changes in chemical composition during aging.26 Samples for these experiments were collected at Sites 1, 2 and 3. We compared the effects of three different conditions in the smoke aging chamber: stainless steel, 3 M chromatography paper (Whatman, a subsidiary of GE Inc) and 100% cotton terry cloth. For paper and cloth tests, ∼13 m2 of paper or cloth were attached to the baffles in the aging chamber. The baffles comprise ∼39% of the total internal surface area of the aging chamber. For each material, we conducted experiments with target particle concentrations at Site 3 of 100, 300, 600, 900 and 1800 μg/m3.26 Particle concentrations were determined gravimetrically. Each series of experiments took place over ∼10 days, with an average of 26 h of smoke generated per series. The chamber was ventilated until the particle concentration dropped to background levels after each experiment. We measured PAHs in smoke in a single experiment with the particle concentration at Site 3 at 500 μg/m3. The PAH experiments were conducted with cloth in the aging chamber. Samples were collected at Sites 1, 2 and 4.

Measurement of polycyclic aromatic hydrocarbons

PAH samples were collected using Chemcomb Model 3500 Speciation Sampling Cartridges (Thermo Electron, East Greenbush, New York, USA) with a PM10 inlet. Chemcomb cartridges were assembled with three honeycomb denuders upstream of three coated glass fibre filters. Denuders and filters were coated with finely ground XAD-4 resin beads and dried. Samples were collected at 10 L/min for 4 h. Flow rates on all pumps and mass flow controlled lines were calibrated with a Gilibrator soap bubble spirometer (#s 800268, D800286 and D800285, Sensidyne, Clearwater, Florida, USA). Extractions were done with dichloromethane. Reduced and filtered extracts were injected into a Hewlett Packard model 6890 gas chromatograph equipped with a 5972 Mass Selective detector. The capillary column used was a DB-17MS (30 m, 0.18 micron diameter, (50%-phenyl)-methylpolysiloxane. Background values derived from field blank Chemcomb samplers were subtracted from all values. The limit of detection was calculated as the mean of two field blank values plus 2 SDs. Values below the limit of detection were treated as zero values for this calculation. PAH analyses of the materials exposed to smoke in the aging chamber were not possible due to clogging of the column, likely by small particles suspended in the extract that were not removed by filtration.

Active sampling for nicotine, cotinine and TSNAs

Filter preparation

We used 47 mm teflon-coated glass fibre filters (#7221, Pall Emfab, Port Washington, New York, USA) which were dipped in an 11.25 mM solution of ascorbic acid in methanol and allowed to dry overnight. For sampling, two treated filters were preweighed and placed in series in 47 mm perfluoroalkoxy resin filter cassettes (#Part 401-21-47-30-21-2, Savillex, Eden Prairie, Minnesota, USA), and cassettes were wrapped in foil. After sampling and postweighing, filters were stored in foil packets at −20°C.

Sampling for sorption and deposition on materials in the smoke aging chamber

The test materials placed in the aging chamber were exposed to ∼26 h of smoke over the course of ∼10 days. After exposure, samples of the paper and cloth were cut and stored at −20°C.

Extraction and GC/MS/MS detection of nicotine, cotinine and TSNAs

Filters and 6 cm2 samples of the paper and cloth that were exposed to smoke in the aging chamber were extracted in high purity liquid chromatography (HPLC)-grade methanol following the protocol of Sleiman et al.29 Quinoline (32 μg/l) (Sigma–Aldrich, St Louis, Missouri, USA) was used as an internal standard. Samples were centrifuged and the supernatant was placed in amber glass Agilent vials and stored at −20°C until analysis. Gas chromatography/ mass spectrometry was performed on a Varian 4000-series gas chromatograph (Varian, Palo Alto, California, USA) with a Factor Four capillary column (30 m × 0.25 mm ID, DF=0.25. Full scan mass spectroscopy (MS) was performed for 0–12 min to quantify nicotine and cotinine, then tandem mass spectroscopy (MS/MS) was performed from 12–17.5 min to quantify TSNAs, and ionisation was turned off for the rest of the analysis time (17.5–25 min). Limits of detection for n-nitrosonornicotine (NNN), NNA, and 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) were 0.15, 0.12, and 0.27 ng mg−1, respectively.

Data analysis and terminology

The mass fraction (%) of material deposited on surfaces between paired sampling sites, Mdep, (deposition fraction) was found by computing:

Embedded Image 1

C1=analyte concentration at Site 1, pre-aging, and C2=analyte concentration at Site 2 or 3, post-aging. These calculations and basic descriptive statistical tests were performed using Microsoft Excel, 2010. Repeated measures analysis of variance (ANOVA) tests of mass loss were performed with surface and location of sampling site as factors, followed by pairwise comparisons using the Holm–Sidak test. Tests of correlation between nicotine and TSNA concentrations were performed using linear regression. ANOVA, Holm–Sidak and linear regression tests were done with SigmaPlot V.10.0 software.

Particles and gas-phase molecules adhere to surfaces through different physical processes. The accepted term for the adhesion of gas-phase molecules to surfaces is ‘sorption’. The accepted term for the adhesion of particles to surfaces is ‘deposition’, which includes both settling and impaction. In this paper, we used the term ‘deposition fraction’ to mean the fraction of a chemical lost from the aerosol between specific sampling sites, by any mechanism. The mechanisms of the species-specific mass loss summarised by this term may include sorption, settling and impaction, depending on the physical state of the chemical in question.


PAHs in smoke

The concentrations of acenaphthylene, acenaphthene, fluorene, anthracene, phenanthrene, fluoranthene, pyrene, benzo(α)anthracene, chrysene, benzo(β)fluoranthene, benzo(a)fluoranthene, benzo(a)pyrene and indeno(1,2,3-cd)pyrene decreased as the smoke aged between Site 1 (pre-aging) and Sites 2 and 3 (presubject). The percentage decrease per PAH ranged from 24% for acenaphthylene to over 90% for fluoranthene. The average mass loss for PAHs was 62%. The average mass loss for particle-phase PAHs was 56.5%, which is very close to the average total particulate material mass loss of 55%, observed in previous experiments with cloth in the aging chamber.26 From acenaphthene through pyrene, we observed an additional mass loss between Site 2 and Site 4. Concentrations of the two highest molecular weight PAHs, dibenz(a,h)anthracene (278 g/mol) and benzo-(g,h,i)perylene (276 g/mol), did not decrease with aging, although their concentrations were low initially (see figure 2).

Figure 2

Mass concentrations of total polycyclic aromatic hydrocarbons decrease with aging, as shown by the decreases in concentration between Site 1 (pre-aging) and Sites 2 and 3 (post-aging).

The PAHs we tested ranged in molecular weight from 128 to 278 g/mole, and the partitioning we observed between the vapour and particle phases followed the trend of their mass.30 In both concentrated and dilute smoke, the shift from the majority of the mass being in the vapour phase to the majority being in the particle phase occurred between pyrene (MW = 202.25) and benzo(α)athracene (MW=228.29). Aging did not affect vapour and particle phase partitioning significantly (see figure 3 for a representative distribution from Site 1).

Figure 3

Vapour and particle phase partitioning of polycylic aromatic hydrocarbons at Site 1 (pre-aging).

Aging decreases mass concentration of nicotine and TSNAs, but not cotinine.

As with the PAHs, we observed decreases in the concentrations of nicotine, NNN and NNK as the smoke aged between Site 1 and Sites 2 and 3. We did not detect any NNA in our samples. The mean deposition fraction for nicotine was 72% ± 13%; for NNK it was 80% ± SD 7%; for NNN it was 79% ± SD 7%; for cotinine it was −5% ± SD 39%. The mean deposition fraction for particulate matter in the same experiments was 49% ± SD 0.08% (reported in Schick and Farraro et al26). These graphs and calculations include only the 10 experiments for which there were values from all three sampling sites for nicotine, cotinine, NNN and NNK (see figures 4 and 5 for values for nicotine and the carcinogenic nitrosamine, NNK). Changes in cotinine values did not correlate with aerosol age and the ratio of cotinine to nicotine was highly variable, ranging from 0.02 to 2.0 (see online supplementary figure S1).

Figure 4

Mass concentrations of nicotine decrease with aging, as shown by the decreases in concentration between Site 1 (pre-aging) and Sites 2 and 3 (post-aging).

Figure 5

Mass concentrations of NNK decrease with aging, as shown by the decreases in concentration between Site 1 (pre-aging) and Sites 2 and 3 (post-aging).

Addition of terry cloth to aging chamber increases loss of nicotine, NNN and NNK, but has no effect on cotinine

Two-way ANOVA analysis of the effects of post-aging sampling location and materials mounted in the smoke aging chamber revealed statistically significant increases in mass deposition of particle-phase nicotine, NNN and NNK. Pairwise comparisons showed that, with terry cloth in the chamber, deposition of nicotine was 14% greater (p=0.014, F=7.55), deposition of NNN was 9% greater (p=0.002, F=12.91) and deposition of NNK was 8% greater (p=0.015, F=7.37) than when paper or stainless steel were present. There was no statistically significant relationship between cotinine mass deposition and materials in the aging chamber.

Relationship between concentrations of nicotine and TSNAs on filters

There was a positive and statistically significant correlation between the concentration of nicotine and NNN and between nicotine and NNK on filters at all sites (see online supplementary table S2). There was no statistically significant relationship between nicotine and cotinine on filters.

Nicotine, cotinine and NNK deposited on materials in the smoke aging chamber

We found nicotine and NNK, but not cotinine or NNN on the cloth samples. The two paper samples both had cotinine and NNK. One also had nicotine and NNN, the other did not (see online supplementary table S3). There was no significant correlation between nicotine and NNK concentrations on cloth. However, when we compared the ratio of NNK with nicotine for filters and materials exposed to smoke in the aging chamber, the ratio for the materials exposed to smoke was 10 times higher than that for the filters (average 0.0367 ± 0.0288 vs average=0.00377 ± 0.00205). The cloth was exposed to smoke for approximately 26 h, over the course of 10 days. The average mass of nicotine per square metre of cloth was 874 μg, and the average NNK mass was 17.9 μg.


Aging decreases mass concentrations of PAHs, nicotine, NNN and NNK in smoke

During 60 min of aging, the concentrations of all surveyed compounds except cotinine decreased. Average PAH concentration decreased 62%, nicotine decreased 72%, NNN decreased 79% and NNK decreased 80%. Potential explanations for this observation include air leakage into the system causing dilution, loss through sorption and deposition within the system and loss through chemical reactions. Air leakage into the system is unlikely because the system operated at a positive pressure of 0.25 inches water.26 Chemical reactions would have resulted in formation of significant quantities of nicotine-related compounds that would have been detected in the GC/MS/MS analyses. Sorption and deposition are the most likely explanations for the massive changes in chemical concentrations we observed between Site 1, where the smoke was approximately 3 min old, and Sites 2, 3 and 4, where it was 60 min old.

Relevance of smoke concentrations to real exposure scenarios

The range of smoke concentrations we tested (100, 300, 600, 900 and 1800 μg/m3 total particulate material) were chosen to represent exposure scenarios during which cigarettes were actively burning; 100 μg/m3 total particulate material represents a single cigarette in a room in a home, or one to two cigarettes in a larger space;3 ,31 ,32 300, 600 and 900 μg/m3 represent increasing numbers of cigarettes burning at one time, or the use of products with larger amounts of tobacco, like cigars;3 ,32–35 1800 μg/m3 represents a very extreme exposure, with multiple cigarettes burning in a small space. It is important to note that the majority of published values for nicotine concentrations in indoor spaces are time-averaged samples that include substantial periods of time when no cigarettes are burning.2 ,3 ,36 Median nicotine concentrations in the homes of smokers in the USA typically range from 0.1–6 μg/m3, but the sampling periods for these studies range from 14 h to several weeks.2 ,3 Our lowest nicotine values overlap this range: 3.4–6.5 μg/m3. All but the single highest nicotine concentration in our study (6.5–88 μg/m3) are within the range of values reported in homes with more than one smoker,37 and public places and workplaces where smoking is permitted.2 ,3 ,38–40

Aging did not decrease mass concentrations of cotinine

Cotinine was the only compound we tested that did not decrease in concentration as the smoke aged. There was no significant correlation between nicotine mass and cotinine mass, and no difference between data for stainless steel and cloth conditions. Cotinine is a stable oxidation product of nicotine and is almost always found wherever nicotine is present. Cotinine can form from nicotine sorbed to surfaces in the presence of ozone, and it is the highest concentration reaction product on glass and cotton surfaces.20 ,41 The concentration of ozone in the smoke aging chamber was 15 ppb. It is likely that nicotine reacted with ozone or other oxidant gases in the smoke to form cotinine. Cotinine is the single exception to the trend of significant mass loss, and the cotinine/nicotine ratio may indicate the extent of oxidation of nicotine.

Addition of terry cloth to aging chamber increased mass loss of nicotine, NNN and NNK from smoke

Substantial mass loss of nicotine, NNK and NNN occurred in the stainless steel chamber. (Average nicotine loss=64% ± 10%, average NNN loss=75% ± 7%, average NNK loss=75% ± 5%). However, when the chamber contained an additional 13 m2 of 100% cotton terry cloth, significantly more mass loss occurred (average nicotine loss=77% ± 10%, average NNN loss=83% ± 5%, average NNK loss=83% ± 7%). This may be further evidence for sorption and deposition as the cause for the changes in chemical concentration we observed. The terry cloth we used was 4.0 mm thick, and provided significantly greater surface area for sorption, deposition and chemical reactions than the flat, impermeable stainless steel. A study by Philip Morris, using a 30 m3 chamber with 0.75 ACH, reported deposition of 34% of total particulate matter when there were no furnishings, and deposition of 64% when the room was fully furnished.42 Van Loy et al12 performed a mass balance experiment by releasing pure, gas-phase nicotine into an unventilated stainless steel chamber. They recovered 80% of the nicotine and found that 99% of that nicotine sorbed to the walls of the chamber within 2 h. In a chamber with painted wallboard walls, the addition of carpet and other furnishings decreased the amount of nicotine and other semivolatile SHS compounds recovered from the air after smoking.15 Singer et al14 demonstrated that so much of the semivolatile compounds in SHS sorb and deposit on surfaces, that a person who lives in a room where someone chronically smokes 4 h a day can be exposed to more nicotine during the 20 h when no one is smoking than during the 4 h when the cigarettes are actually burning.

Nicotine and TSNA deposition fractions were higher than total smoke particulate matter deposition fraction

In a previous paper on the smoke aging system, we showed that the mean deposition fraction for total particulate matter was approximately 50%.26 Nicotine and TSNA deposition fractions were significantly higher.

High concentrations of nicotine and NNK on deposition samples

We found nicotine and NNK, but not cotinine or NNN, on the cloth samples exposed to smoke. Our paper samples gave inconsistent findings, but suggest that NNK is also found on paper exposed to cigarette smoke. Nicotine is found primarily in the vapour phase in SHS.18 ,43 ,44 These data suggest that sorption of nicotine occurred in our system, and that sorption and deposition may be the primary cause of the mass loss associated with aging of cigarette smoke. These data agree with the findings of numerous other studies.10–12 ,14 ,15 ,19 The 10-fold increase in the ratio of NNK to nicotine, between filter samples collected over 1–6 h, and materials exposed to ∼26 h smoke over ∼10 days, suggests that sorbed nicotine may react, perhaps by hydrolysis with water vapour and reaction with ozone,20 ,22 ,45 to form NNK. Another explanation for the apparent enrichment in NNK is that nicotine may desorb from the surfaces at a faster rate than NNK because the vapour pressure of nicotine is much higher. NNK is a strong carcinogen with reported cancer potency of 49 kg/mg.46 When given to laboratory animals, it primarily induces lung cancer, but nasal, pancreatic and liver cancers have also been observed.5

No NNA in filter or deposition samples

NNA is another oxidation product of nicotine, and Sleiman et al22 have shown that it forms from nicotine sorbed on paper surfaces, both under laboratory conditions and in samples from smokers’ vehicles. The analytical method used in our study was the same as used by Sleiman et al, but we detected no NNA in our samples. This may be because NNA has an aldehyde group that can react with ozone or other reactive gases. Because NNA is not found in fresh sidestream smoke or tobacco, it has been proposed as a marker of aged SHS and THS. Our data suggest that an elevated ratio of NNK to nicotine may be another potential marker of SHS and THS in the environment.


The 6 m3 smoke aging chamber is smaller than a real room and has a higher nominal surface to volume ratio (5.9 m−1). Typical surface to volume ratios in residences and offices are 2–3 m−1. This may have increased the deposition fractions of the compounds we studied. However, the chamber was made of stainless steel. When the baffles were covered with paper or cloth, only 39% of the total surface area was permeable or absorbent. Residential and public buildings are typically faced with permeable wallboard, and may contain carpet and furnishings with the ability to adsorb chemicals, and to absorb them into deeper layers. The ventilation rate (1 air change per hour) that we used, is higher than the ventilation rates observed in most homes (US average 0.71–0.76 ACH).25 ,47 Because the aging chamber had a higher ventilation rate and fewer highly complex sorbent surfaces, our data may, in fact, underestimate the deposition of particle-phase SHS compounds in a typical furnished room. The findings of Van Loy et al12 suggest that in a room with 0.5 ACH, all the nicotine released from a cigarette will sorb to surfaces.


Our data suggest that the majority of the PAHs, nicotine, cotinine and TSNAs that are released during smoking in homes and public places deposit on room surfaces. Our findings on PAHs, cotinine and TSNAs are novel and add to a body of work that demonstrates that the majority of the semivolatile organic compounds (tar) in SHS deposit on surfaces and furnishings when people smoke indoors. PAHs and TSNAs are highly carcinogenic, and it is important to understand where they are found in the environment and how we are exposed to them. We have not found published research on the concentrations of PAHs, nicotine or TSNAs extracted from furnishings exposed to cigarette smoke, but there is a growing literature on the presence of these compounds in house dust. Dust from the homes of smokers has higher levels of the same 16 PAHs tested in this study than dust from the homes of non-smokers48–50 A new study has also shown that nicotine and TSNAs are present in higher concentrations in the dust of smokers’ homes than non-smokers’ homes.51

Deaths from cancer comprise approximately ∼10% of the deaths attributed to SHS1 ,2 and ∼60% of the deaths attributed to active smoking.52 Smokers and non-smokers are both exposed to SHS and THS. Our data suggest that large quantities of PAHs and TSNAs deposit on surfaces where people smoke. These compounds may accumulate over time. Exposure to these THS compounds may contribute to the health effects attributable to smoking and SHS exposure.

What this paper adds

  • This paper demonstrates that ∼60% of the polycyclic aromatic hydrocarbons, ∼70% of the nicotine and ∼80% of the tobacco-specific nitrosamines in secondhand cigarette smoke stick to room surfaces and are not removed under normal ventilation conditions.

  • This paper also presents data that suggest that the nicotine that is stuck to room surfaces can react to form the lung carcinogen NNK, under normal ambient conditions.

  • Our data indicate that most of the carcinogens in secondhand smoke stay indoors when people smoke indoors. These carcinogens and nicotine are known as thirdhand smoke.


View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Contributors SFS and KFF are coauthors. SFS, KFF, BMJ, JB and SKH planned the experiments. KFF performed the nicotine and TSNA sample collection and extracted the samples. MS analysed these samples. GvdV and MPT collected the PAH samples and CP and MPT analysed them. SFS, KFF and SKH analysed the data. SFS wrote the paper.

  • Funding SFS, KFF and JB were supported by a grant from the Flight Attendants Medical Research Institute. MS was supported by the California Tobacco-Related Disease Research Program Grant 20KT-0051. Charles Perrino and SKH were supported by the Dr William Cahan, Distinguished Professor Award to SKH by Flight Attendants Medical Research Institute.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement One year after the publication date, unpublished data will be made available to reputable tobacco control researchers upon request.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.