Cigarettes and middle-ear disease and surgery

Exposure to cigarette smoking may affect the middle-ear function via a number of mechanisms. Cigarette smoke is proposed to impair the beating function of cilia and cause hyper-secretion of mucous by goblet cells within the Eustachian tube. This leads to Eustachian tube obstruction, dysfunction and trapping of mucus and bacteria within the middle ear with subsequent otitis media. Mast cell degradation, adenoidal tissue hyperplasia and impaired phagocytic function are further proposed to contribute to Eustachian tube dysfunction.^{Reference Becvarovski and Kartush8}

Pooled data from 7 homogeneous studies demonstrated that children exposed to second-hand smoke had a 48 per cent increased odds of developing recurrent acute otitis media (95 per cent confidence interval (CI), 1.08–2.04) and pooled data from 4 homogeneous studies demonstrated exposed children had a 38 per cent increased odds of developing otitis media with effusion (OME) (95 per cent CI, 1.23–1.55) compared with non-exposed children.^{Reference Strachan and Cook9}

A further large, industry funded systematic review of 58 trials by Thornton and Lee suggested the association between second-hand smoke exposure and recurrent acute otitis media (11 trials) and recurrent OME (23 trials) was present but weak owing to overestimation of the effect due to inconsistent adjustment for confounders and selection and information bias. Unfortunately, no meta-analyses of the data or calculation of heterogeneity was performed.^{Reference Thornton and Lee10}

In the most recent meta-analysis, pooled data from 14 studies demonstrated that children exposed to second-hand smoke because of post-natal maternal smoking had a 53 per cent increased odds of all cause middle-ear disease (acute otitis media, OME, recurrent otitis media and chronic otitis media) (95 per cent CI, 1.22–1.92). Furthermore, pooled data from 38 studies demonstrated children exposed to household second-hand smoke had 32 per cent increased odds of developing all cause middle-ear disease. However, considerable heterogeneity was seen in both analyses (I² = 73 per cent and I² = 73 per cent, respectively) owing largely to a variety of different observational study designs, methodological quality of studies and incomplete adjustment of confounders.^{Reference Jones, Hassanien, Cook, Britton and Leonardi-Bee11} Smoking cessation has been advocated as a key measure to address acute otitis media in low income environments, and NICE recommends parents of children with OME are advised that second-hand smoke exposure increases the risk of recurrent OME.^{12,Reference Brophy-Williams, Jarosz, Sommer, Leach and Morris13}

Cigarette smoking is associated with an increased risk of developing chronic suppurative otitis media (CSOM). In one trial from India (n = 4143), smokers had a 13 per cent increased odds of developing CSOM than non-smokers, and in another study, 68 per cent of patients with ear canal cholesteatoma were found in current or former smokers.^{Reference Gaur, Kasliwal and Gupta14,Reference Hertz and Siim15} In addition, smokers with CSOM have been found to have higher rates of active disease demonstrated by increased rates of otorrhoea and higher rates of aggressive cholesteatoma disease demonstrated in one study by a higher rate of further complications including exposed dura, exposed facial nerve and labyrinthine fistula.^{Reference Kaylie, Bennett, Davis and Jackson16} This may be attributable to antigenic substances within cigarette smoke that are proposed to induce an inflammatory response in middle-ear mucosa transforming it into a more secretory type.^{Reference Becvarovski and Kartush8}

Prognostic effects of cigarette smoking on middle-ear surgery have been investigated in several trials and have been subject to a prior review assessing outcomes up to six months (Table 2).^{Reference Becvarovski and Kartush8,Reference Kaylie, Bennett, Davis and Jackson16–Reference Kyrodimos, Stamatiou, Margaritis, Kikidis and Sismanis25} Cigarette smoking is proposed to lead to poorer outcomes because of defective middle-ear aeration secondary to the effects described above and also because of the vasoconstrictive effects of nicotine which have been demonstrated to adversely affect neovascularisation of newly grafted tissue.^{Reference Krueger and Rohrich26}

Table 2. Graft uptake and audiological outcomes between smokers and non-smokers following tympanoplasty

Trials have reported graft uptake results ranging from 40 to 100 per cent following tympanoplasty (Table 2). Trialists with increased success rates amongst smokers have attributed this to use of cartilage grafts rather than temporalis fascia owing to its metabolic properties and resistance to retraction, atelectasis and perforation.^{Reference Coelho, Peng, Thompson and Sismanis21,Reference Cox, Anderson, Russell and Dornhoffer24,Reference Kyrodimos, Stamatiou, Margaritis, Kikidis and Sismanis25} In one trial utilising both graft tissues, a 75 per cent failure was reported amongst smokers when temporalis fascia was used compared with similar success rates between smokers and non-smokers (88.9 per cent and 76.8 per cent, respectively) when cartilage grafts were used, further advocating the use of cartilage as routine in smokers.^{Reference Uguz, Onal, Kazikdas and Onal19}

Only one trial using temporalis fascia grafts reported similar outcomes between smokers and non-smokers and attributed this to the experience of the otologist, highlighting that alternative pre- and intra-operative factors may have a more profound effect on graft uptake rates.^{Reference Migirov, Lipschitz and Wolf22} Only one trial reported significantly increased odds of graft uptake failure in smokers despite using cartilage grafts (odds ratio, 8.16; 95 per cent CI, 1.74–36.89), but in this trial, an inlay technique was employed in contrast to the conventional underlay approach.^{Reference Lin, Wang, Weng and Lin20} Trialists have reported largely similar audiological outcomes between smokers and non-smokers following tympanoplasty with and without ossicular chain reconstruction (Table 2).^{Reference Visvanathan, Vallamkondu and Bhimrao17}

Smokers have been demonstrated to be at greater risk of post-operative complications. Becvarovski and Kartush found delayed perforation and atelectasis of the grafted tympanic membrane in 60 per cent of smokers (n = 9) compared with 20 per cent of non-smokers (n = 12) after 6 months following tympanoplasty.^{Reference Becvarovski and Kartush8} In a separate study, 12 per cent of smokers (n = 9) developed post-operative OME or tympanic membrane atelectasis warranting further secondary tympanostomy tube placement compared with 1.2 per cent of non-smokers (n = 2). Furthermore, this trial also found a higher rate of recurrence of cholesteatoma disease in smokers (n = 4 (5.3 per cent)) compared with non-smokers (n = 1 (0.6 per cent)), although this was not demonstrated in studies by Kaylie et al. and Becvarovski and Kartush.^{Reference Becvarovski and Kartush8,Reference Kaylie, Bennett, Davis and Jackson16,Reference Cox, Anderson, Russell and Dornhoffer24} Coelho and Peng found a greater proportion of smokers required revision surgery for persistent cholesteatoma, conductive hearing loss and tympanic membrane perforation compared with non-smokers (n = 6 (13.3 per cent) and n = 4 (4.7 per cent), respectively).^{Reference Coelho, Peng, Thompson and Sismanis21} In an assessment of 30-day post-operative outcomes following otological surgery, smokers have also been found to have increased odds of soft tissue complications including 89 per cent increased odds of superficial wound infections (95 per cent CI, 1.32–2.86) and almost 300 per cent increased odds of wound dehiscence (95 per cent CI, 1.26–11.60).^{Reference Kay-Rivest, Mascarella, Sewitch, Cloutier and Mijovic27}

Cigarettes and SNHL loss

Cigarette smoking is proposed to induce hearing impairment by inducing atherosclerosis within the internal auditory artery resulting in cochlear ischaemia and also through the direct ototoxicity of nicotine. The effect was demonstrated in early experimental studies in animals.^{Reference Maffei and Miani28} Similar trials in humans would be unethical; therefore, trialists have taken to observational study designs usually investigating the effect in conjunction with other atherogenic diseases. Pooled data in a meta-analysis of two homogeneous prospective cohort trials by Nomura et al. demonstrated cigarette smokers had almost double the risk (risk ratio, 1.97; 95 per cent CI, 1.44–2.70) of developing hearing loss compared with non-smokers.^{Reference Nomura, Nakao and Morimoto29}

Newer prospective trials and trials excluded from the pooled analysis have, however, demonstrated inconsistent results. Over a 5-year period (n = 705) Karlsmose et al. demonstrated almost equivalent risk of hearing impairment between smokers and non-smokers (relative risk, 0.97; 95 per cent CI, 0.74–1.26), but in the longest prospective trial to date, Cruickshanks et al. demonstrated that current smokers were 36 per cent more likely to develop hearing loss greater than 25 dB than non-smokers over a 15-year period (95 per cent CI, 1.05–1.77), possibly suggesting that longer follow-up periods are required to demonstrate the association between smoking and hearing as oxidative stresses and atherosclerosis within the internal artery progress and cochlear ischaemia ensues.^{Reference Karlsmose, Lauritzen, Engberg and Parving30,Reference Cruickshanks, Nondahl, Dalton, Fischer, Klein and Klein31} Despite this, Nakanishi et al. demonstrated a significantly increased risk (relative risk, 2.45 CI, 1.28–4.70) of high frequency hearing loss greater than 30 dB amongst smokers as early as 5 years. This was only seen at 4 kHz in men indicating that the basal turn of the cochlea may be affected first in smoking-related hearing loss and therefore identified earlier when reported independent of other frequencies.^{Reference Nakanishi, Okamoto, Nakamura, Suzuki and Tatara32} Certainly, trials investigating the character of early smoking-related hearing loss have found this to be more pronounced at higher frequencies.^{Reference Capra Marques de Oliveira and de Melo Tavares de Lima33–Reference Sumit, Das, Sharmin, Ahsan, Ohgami and Kato35}

The impact of cigarette smoke on SNHL has been investigated extensively through cross-sectional or ‘snapshot’ studies. The results have also been inconsistent but pointed towards increased prevalence of hearing impairment amongst smokers. Pooled data from 5 cross-sectional trials demonstrated cigarette smokers had a 30 per cent increased risk of having hearing loss compared to non-smokers (95 per cent CI 1.24–1.44). However, considerable heterogeneity was found between trials due to the variance in the definition of hearing loss.^{Reference Nomura, Nakao and Morimoto29}

More recently, a South Korean study (n = 12 935) demonstrated increased odds of bilateral high frequency hearing loss greater or equal to 25 dB (odds ratio 1.42; 95 per cent CI, 1.13–1.77) and bilateral hearing loss at speech frequencies (odds ratio 1.39; 95 per cent CI, 1.08–1.79) amongst smokers compared with non-smokers, indicating that cigarette smoking may be correlated not only with hearing loss but more importantly with clinically relevant hearing loss.^{Reference Chang, Ryou, Jun, Hwang, Song and Chae36} This is consistent with subjective data from prospective questionnaire-based studies that have demonstrated increased perceived hearing loss in smokers over a 5- and 10-year period compared with non-smokers.^{Reference Shargorodsky, Curhan, Eavey and Curhan37,Reference Burr, Lund, Sperling, Kristensen and Poulsen38}

In a recent large UK cross-sectional trial (n = 164 770) utilising the speech-in-noise test rather than pure tone audiometry, Dawes et al. demonstrated that current smokers had a 15 per cent increased odds of developing hearing loss compared with non-smokers (95 per cent CI, 1.09–1.21).^{Reference Dawes, Cruickshanks, Moore, Edmondson-Jones, McCormack and Fortnum39} In this study, a dose-related response was also demonstrated whereby long-term smokers had 30 per cent increased odds of hearing loss (95 per cent CI, 1.19–1.41) and medium-term smokers had 11 per cent increased odds of hearing loss (95 per cent CI, 1.03–1.19) compared with non-smokers. This was consistent with findings in the prospective trial by Nakanishi et al. where the risk of hearing loss significantly increased with increasing pack years.^{Reference Nakanishi, Okamoto, Nakamura, Suzuki and Tatara32} Studies have demonstrated that former smokers have a reduced risk of hearing loss compared with current smokers, highlighting the importance of smoking cessation for hearing preservation.^{Reference Nomura, Nakao and Morimoto29}

Exposure to second-hand smoke has been demonstrated to be harmful to cochlear physiology in children and adolescents exposed to second-hand smoke. This may occur as early as in utero as adolescents whose mothers smoked during pregnancy were demonstrated to have a 2.6-fold increased odds of unilateral low frequency hearing loss (95 per cent CI, 1.1–6.4) compared with children of non-smoking mothers.^{Reference Weitzman, Govil, Liu and Lalwani40} Children ‘heavily’ exposed to cigarette smoke were found to have 3 times the risk of having hearing loss to 20–25 dB (95 per cent CI, 18–8.3), and in a another study, adolescents exposed to second-hand smoke had an 82 per cent increased odds of hearing loss above 15 dB (95 per cent CI, 1.08–3.41) compared with those non-exposed. In this latter study, those with the highest exposure, measured objectively through serum cotinine levels, demonstrated increased odds of hearing loss (odds ratio, 2.72; 95 per cent CI, 1.46–5.06) further supporting the notion of a dose-related response relationship.^{Reference Talaat, Metwaly, Khafagy and Abdelraouf41,Reference Lalwani, Liu and Weitzman42}

There is also a suggestion that cigarette smoking has a synergistic effect on age-related hearing loss and noise-induced hearing loss.^{Reference Rigters, Metselaar, Wieringa, De Jong, Hofman and Goedegebure43–Reference Sung, Sim, Lee, Yoo, Lee and Kim46} In a ‘snapshot’ assessment of 52–99 year olds, Rigters et al. (n = 3315) demonstrated significantly increased prevalence of high frequency hearing loss in both men (β coefficient, 3.01) and women (β coefficient, 2.49) and low frequency hearing loss in women (β coefficient, 1.69) in smokers compared with non-smokers.^{Reference Rigters, Metselaar, Wieringa, De Jong, Hofman and Goedegebure43} Questionnaire-based studies have demonstrated a small but increased odds of developing hearing loss in smokers with a background of noise-induced hearing loss.^{Reference Ferrite, Santana and Marshall44–Reference Sung, Sim, Lee, Yoo, Lee and Kim46} In one trial, smokers with a background of noise-induced hearing loss had almost 4 times the prevalence of hearing loss (95 per cent CI, 2.81–5.52) compared with non-smokers or noise-exposed individuals. This was increased compared with individuals who only smoked cigarettes (adjusted prevalence ratio, 1.39; 95 per cent CI, 1.07–1.81) and those only exposed to noise (adjusted prevalence ratio, 2.66; 95 per cent, 1.86–3.82) highlighting the possible combined harmful effect on cochlear dysfunction.^{Reference Ferrite, Santana and Marshall44}

Electronic cigarettes

Electronic cigarettes are a less harmful alternative to traditional cigarettes because they do not contain the same number of carcinogenic components of cigarettes. Because of their relative modernity and lack of long-term safety data, their use is not advocated in favour of other nicotine replacement therapies, but they are not discouraged as a smoking cessation tool according to NICE.⁴⁷ Early studies have demonstrated respiratory irritation and reduced lung cell viability, which poses concern for middle-ear health due to the shared ciliated columnar epithelial lining.^{Reference Lerner, Sundar, Yao, Gerloff, Ossip and McIntosh48}

To date, one in vitro investigation of electronic cigarette liquid solution on human middle-ear epithelial cells has been conducted by Song et al. In this study, starved human middle-ear epithelial cells were incubated for 24 hours with electronic liquids of a variety of flavours from different manufacturers. Electronic liquids were found to induce cytotoxic effects on these cells and reduce their viability from 100 per cent to 32–62 per cent. Moreover, increased cytotoxicity was seen with menthol and fruit flavoured electronic liquid.^{Reference Song, Go, Mun, Lee, Im and Kim49} However, the clinical implications of this are currently unclear, and it is likely that trials investigating the clinical impact of electronic cigarettes on ear disease will face similar challenges as those faced by trialists investigating the effect of cigarette smoking, in particular at risk of confounding bias.

Electronic cigarettes contain nicotine and therefore may impact graft neovascularisation during tympanoplasty. Song et al. demonstrated increased human middle-ear epithelial cell cytotoxicity when nicotine was included, with cell viability reducing from 100 per cent to 25–43 per cent suggesting nicotine may have a role in middle-ear mucosal disease. This effect may be reduced with lower concentrations of nicotine containing electronic liquids but this was not tested in this study.^{Reference Song, Go, Mun, Lee, Im and Kim49} Recent animal studies have demonstrated nicotine to have a direct deleterious effect on cochlear outer hair cells, with reports of distorted shape, heterochromatic nuclei and vacuolated cytoplasms.^{Reference Abdel-Hafez, Elgayar, Husain and Thabet50} Despite this, there are currently no data for the sole clinical implications of nicotine in otology in humans. Propylene glycol is a Food and Drug Administration approved solvent that is commonly used in the chemical and food industry. Prior animal studies in rodents demonstrated administration of large concentrations in the middle-ear space induced inflammatory granulation tissue, subsequent Eustachian tube dysfunction and tympanic membrane retraction and cholesteatoma formation.^{Reference Huang, Shi and Yi51,Reference Vassalli, Harris, Gradini and Applebaum52} Song et al. found increased human middle-ear epithelial cell cytotoxicity with propylene glycol suggesting that concentrations in electronic liquids may also contribute to middle-ear mucosal disease.^{Reference Song, Go, Mun, Lee, Im and Kim49} Similarly, there are no in vivo clinical data in humans for propylene glycol.

Amongst the adolescent and paediatric population, there is concern over electronic cigarettes acting as a gateway drug to nicotine addiction and cigarette smoking.^{Reference East, Hitchman, Bakolis, Williams, Cheeseman and Arnott6} In the younger paediatric population, there is concern over accidental ingestion of the liquids. One case report found ingestion of electronic liquid solution led to bilateral SNHL with average thresholds of 63 dB for the right ear and 67 dB for the left ear.^{Reference D Demir and Topal53} This may be attributable to ethylene glycol, a constituent of anti-freeze solutions, which has been demonstrated to cause bilateral SNHL when ingested in extremely high concentrations in addition to other systemic effects.^{Reference Dezso, Ramsden, Saeed, Wessels and Mawman54,Reference Carr, Strachan, Quentin Summerfield, Kitterick, Tapper and Raine55} At present however, the effect of ethylene glycol in small doses as seen in electronic cigarettes on human ear cells is unknown.