Cannabis related side effects in otolaryngology: a scoping review

Background Cannabis has been rapidly legalized in North America; however, limited evidence exists around its side effects. Health Canada defines side effect as a harmful and unintended response to a health product. Given drug safety concerns, this study’s purpose was to review the unintended side effects of cannabis in otolaryngology. Methods The Preferred Reporting Items For Systematic Reviews and Meta-analysis extension for Scoping Reviews (PRISMA-ScR) protocol was used to conduct a scoping review of the MEDLINE, EMBASE, CINAHL, and CENTRAL databases. (PROSPERO: CRD42020153022). English studies in adults were included from inception to the end of 2019. In-vitro, animal, and studies with n < 5 were excluded. Primary outcome was defined as unintended side effects (defined as any Otolaryngology symptom or diagnosis) following cannabis use. Oxford Centre for Evidence-Based Medicine: Levels of Evidence and risk of bias using the Risk of Bias in randomized trials (RoB 2) and Risk of Bias in Non-Randomized Studies of Interventions (ROBINS-I) tools were assessed.. Two authors independently reviewed all studies; the senior author settled any discrepancies. Results Five hundred and twenty-one studies were screened; 48 studies were analysed. Subspecialties comprised: Head and Neck (32), Otology (8), Rhinology (5), Airway (5), Laryngology (1). Cannabis use was associated with unintended tinnitus, vertigo, hearing loss, infection, malignancy, sinusitis, allergic rhinitis, thyroid dysfunction, and dyspnea. About half (54.1%) of studies showed increased side effects, or no change in symptoms following cannabis use. Oxford Levels of Evidence was 2–4 with substantial heterogeneity. Risk of bias assessment with RoB2 was low to high and ROBINS-1 was moderate to critical. Conclusion This was the first comprehensive scoping review of unintended side effects of cannabis in Otolaryngology. The current literature is limited and lacks high-quality research Future randomized studies are needed to focus on therapeutic effects of cannabis in otolaryngology. Substantial work remains to guide clinicians to suggest safe, evidence-based choices for cannabis use. Graphic abstract


Introduction
Cannabis is one of the most commonly used drugs in North America. Nearly half of the American population has used cannabis at least once in their lifetime, with approximately 9% being current users [1]. Evolving societal perception has rapidly driven the legalization of cannabis, which has been structured to regulate it's production and sales, while promoting safe consumption [2]. Legalization has led to a perceived reduction in harm, which has been associated with an increased prevalence of cannabis consumption from 2002 to 2014 [2]. Evidence based research has followed this movement, yet a paucity of data concerning the side effects of this drug remains.

Methods
This scoping review was carried out according to a review protocol that has been published in the Prospective Register of Systematic Reviews (PROSPERO Registration number: CRD42020153022). A systematic review was the initial intention for this study, but the heterogeneity of the literature and lack of high-quality evidence precluded this, and therefore a scoping review was more appropriate. The reporting of this scoping review was conducted in accordance with the Preferred Reporting Items For Systematic Reviews and Metaanalysis extension for Scoping Reviews (PRISMA-ScR) statement [11]. The quality of the literature was assessed with the Oxford Centre for Evidence-Based Medicine: Levels of Evidence [12].

Data sources and search strategy
A literature search was performed in electronic databases, including MEDLINE, EMBASE, CINAHL, and CENTRAL databases of the Cochrane Library from inception through October 1, 2019. The search strategy was developed with assistance from a medical librarian. Search functions were designed to incorporate two subsections by [AND] Boolean operators. Subsections contained MeSH and field-designated search terms for otolaryngology related diseases and for cannabis. A cannabis search hedge was employed to identify both formal and informal terms for cannabis in the literature [13]. Additionally, reference lists from previously published reviews were screened for articles not identified in the initial search. Detailed search strategies are reported in "Appendix A".

Study selection
All articles identified via the literature search were exported to Covidence (Veritas Health Innovation Ltd., Melbourne, Australia), a systematic review management software. Study selection was independently undertaken by two authors (J.P.& J.H.) with discrepancies being resolved by consensus with the senior author (A.H.). Inclusion criteria included: (1) English language study, (2) adult population (≥ 18 years old), (3) sample size ≥ 5, (4) clinical study, (5) study subjects report use or were exposed to cannabis, and (6) report of otolaryngology related side effects (symptom or diagnosis) following cannabis use that were unintended. Of note, no pediatric studies were included in this scoping review. Most legislation for legal cannabis prohibits consumption in the pediatric population. The exclusion criteria included: (1) in-vitro or animal study and (2) inappropriate study or publication type (e.g. systematic review, literature review, or book chapter). Additionally, side effects involving the lower airways were excluded as this is generally considered outside of the scope of an Otolaryngologist, and primarily managed by Pulmonology Medicine. The areas of wound healing, analgesia and pain were excluded as they are non-specific to the head and neck. Lastly, the purpose of this scoping review was not to evaluate the therapeutic indications of cannabis in otolaryngology, but to evaluate the unintended otolaryngologic side effects of consuming cannabis recreationally or medically.

Main outcome
This study used Health Canada's definition of side effect [5]. The primary study concept was to examine unintended Otolaryngology related side effects following any level of cannabis exposure. Otolaryngology related side effects were defined as a pathologic diagnosis or symptom commonly treated by an Otolaryngologist.

Data extraction
Data extraction was conducted by a single author (J.P.) and reviewed by a second author (J.H.). Information extracted from each study included: title, first author, year of publication, study design and objectives, characteristics of study participants, intervention(s) and control(s), tobacco use, cannabis consumption and amount (e.g. joint year), primary study outcomes and main findings including otolaryngology related side effects after cannabis exposure.

Evaluation of risk of bias
Risk of bias assessment was conducted independently by two study authors (J.P.& J.H.) using the Risk of Bias in randomized trials (RoB 2) and Risk of Bias in Non-Randomized Studies of Interventions (ROBINS-I) tools [14,15]. The senior author (AH) settled any discrepancies.

Study selection
A study flow diagram is shown in Fig. 1. The comprehensive database search identified 614 articles and 12 additional articles were identified through the hand searching of reference lists. After removal of duplicates, 521 abstracts were screened. Full-text review of 117 articles excluded a further 69 articles. In total, 48 articles met the inclusion criteria and underwent data extraction and analysis.

Characteristics of included studies
Detailed characteristics of the 48 studies included are described in Table 1. Among the included studies there were: 17 case-control studies, 14 cross-sectional studies, 6 cohort studies, 6 case series, and 5 randomized controlled trials (RCTs). There was significant heterogeneity in study design, objective, and strength of the evidence reported. Quality of the literature ranged from level two to four, as shown in Table 1. The sample size of participants in the included studies ranged from 5 to 29,195. The amount of cannabis consumption was reported by the majority of studies (33). A total of 12 studies expressed cannabis consumption in joint-years, defined as the number of joints smoked per day, multiplied by the duration in years. This ranged from < 1 to 62.1 joint-years. Studies not expressing cannabis consumption in joint-years either failed to report the amount of cannabis use or instead reported frequency or weight of cannabis consumption. A large proportion of participants in the included studies also reported tobacco use in addition to cannabis consumption, ranging from 0 to 89.1%. All included articles reported on at least one otolaryngology related side effect following cannabis exposure.

Subspecialty synthesis
A wide variety of otolaryngology subspecialties were represented ( Fig. 2). Below is a synthesis of the results and highlights of themes, trends, and gaps categorized by subspecialty.
Overall, forming a firm conclusion on the effects of cannabis in H&N may be challenging, as many studies reported opposing findings. For example, Berthiller et al. described no association between cannabis use and the risk of H&N cancer [19]. However, Zhang et al. provided evidence for a strong dose-response pattern between cannabis use and the risk of H&N malignancy, while also reporting a synergistic effect of cannabis use and cigarette smoking on cancer risk [61]. This area was an unresolved knowledge gap in this scoping review. Level of evidence: Mostly 4.

Otology
The second most studied subspeciality area was otology, where eight studies evaluated cannabis' association with hearing loss (n = 6), vestibular dysfunction (n = 3) and tinnitus (n = 2) [22,31,38,44,51,52,57,58]. Please note that one study may have addressed several otologic side effects. Exposure to cannabis was also shown to promote changes in the auditory pathway and alter the function of outer hair cells, while chronic use of cannabis at high doses showed significant changes in vestibular function [22,52,57,58]. There were several excellent basic science otology studies, however, they were excluded because the goal was to keep a clinical focus for this review. Four of the eight otology studies were randomized controlled trials, so the level of evidence was highest for this subspecialty area. Level of Evidence: 2 and 4.

Rhinology
In rhinology, all five studies reported positive association between sinusitis (n = 4) and allergic rhinitis (n = 2) [31,33,53,56,63]. One rhinology study evaluated both diagnoses [56]. Cannabis smoking was associated with sinusitis in a large retrospective analysis of the US National Survey on Drug Use and Health database (n = 29,195), which was the largest sample size of all the studies in this scoping review [31]. Stokes et al. provided evidence to support cannabis as a clinically relevant aeroallergen on multi-test skin testing [53]. There was the least controversy in this subspecialty area. Level of Evidence: 3 and 4.

Airway
Five studies reported cannabis and its associations with upper airway issues, including sleep apnea (n = 3) and dyspnea (n = 2) [23,31,48,54,63]. Four studies described no change in risk following cannabis exposure and one study reported an increased risk. Primary evidence supports cannabinoids as a promising and safe treatment option for OSA [48]. One of the randomized controlled trials also concluded that there was no significant difference in adverse events between patients who were treated with cannabis and placebo [23]. A recent position statement by the American Academy of Sleep Medicine, however, recommended against medical cannabis for the treatment of OSA due to unreliable delivery methods and insufficient evidence of effectiveness, tolerability, and safety [64]. Level of Evidence: 2, 3, and mostly 4.

Laryngology
Only one study in laryngology explored the effect of cannabis on the voice [43]. This older study published in 1980 reported that marijuana users did not differ perceptually from non-smokers and cigarette smokers, and objective voice analysis was also similar. Darker vocal folds were seen on laryngoscopy. This subspecialty was the weakest with the largest knowledge gap. Level of evidence: 4.

Risk of bias assessment
Overall risk of bias for all RCTs was determined to be high, with some concerns in two studies, and low in the remaining three randomized studies ("Appendix B"). Bias in nonrandomized trials was moderate in 21 studies, serious in 12, and critical in 10 studies ("Appendix C"). The major sources of bias among serious and critical studies were due to selection bias, deviations from intended interventions, confounding and inappropriate methods to control for measured confounders.
The majority of nonrandomized studies demonstrated appropriate measurement of outcomes and reporting of outcome data.

Discussion
It is evident from this review that the Otolaryngology cannabis literature is limited but evolving. The included reports demonstrate that unintended side effects in otolaryngology cover a large spectrum of physiology and pathology. These included: tinnitus, vertigo, hearing loss, infection, malignancy, sinusitis, allergic rhinitis, thyroid dysfunction, and dyspnea. A number of challenges were observed in the included studies. Many studies contained small sample sizes with methodological errors, substantial heterogeneity of study types and varied outcome measures. A lack of standardization in the reporting of quantity, method of consumption, and length of time of cannabis use was common. Similar to cigarette smoking, several studies attempted to use pack-years, but this was inconsistent. Standard medical practice depends on patient self-reporting cannabis use, which leads to recall bias in the literature. Other studies inconsistently used measures to quantify the amount of cannabis in their experimental protocol (e.g. weight, puffs, joints). This did not always control for the type of cannabis, which has the potential to alter expected results substantially. The included studies were published over several decades, however an improvement in the quality of papers were seen over time. Five recent studies were RCTs, suggesting that as cannabis gains widespread acceptance, the medical community recognizes the need for robust evidence. Lastly, challenges arise when studying cannabis consumption with concomitant tobacco and alcohol use. The carcinogenic effects of cannabis, when smoked, is certainly another area of relevance. Although these factors were controlled for by some studies, they will remain confounders until their relationship can be explored entirely.
The medical use of cannabis is legal in 33 US states, and since 2012, 11 states have legalized recreational use of the drug [65]. Conservative estimates project legal cannabis sales to reach $30 billion by 2025, driven mainly by increased legalization of the drug across the country [66]. This swift legalization of cannabis and growth of related industry is likely to lead to the assumption among the general public that cannabis has an adequate safety profile, analogous to the perceived safety, and subsequent increase in consumption of alcohol following prohibition in the 1930's [67,68]. Anecdotally, cannabis has been used to treat a large variety of conditions ranging from acute pain and psychiatric diagnoses, to chronic inflammatory conditions [69]. However, only three indications have substantial evidence demonstrating cannabis as an effective treatment alternative. These include chronic pain in adults, oral antiemetic therapy in chemotherapy-induced nausea and vomiting, and an improvement in patient reported spasticity in multiple     sclerosis [70]. Outside of these few indications, there is limited and conflicting evidence to support the effectiveness of cannabis as a primary treatment for most medical conditions. This trend is similarly demonstrated in our review, with a disparity in consistency of the reported side effects.
As we learn more about cannabis, we will likely understand that there are many more interactions and side effects associated with varying amounts of use. Many of these effects will likely not be observed for years, when evidence on chronic use begin to emerge. Conversely, there is the possibility for positive uses that remain to be realized. Given this, it would be appropriate for all health practitioners to develop and incorporate a better understanding of this drug into their practice as the literature continues to evolve. There are obvious gaps in the knowledge of how cannabis affects otolaryngology, but this also suggests extensive opportunities for further research. Every subspecialty of Otolaryngology remains a potential area for further study. Presently, the areas of H&N oncology, otology and allergy demonstrate convincing effects of cannabis, making them easy study prospects. Furthermore, the quantification of cannabis use over time and understanding the potential risks may lead to harm reduction campaigns, notably with young and heavy users.
A small number of recent systematic reviews have studied cannabis in subspecialties within Otolaryngology, such as laryngology and H&N oncology. Meehan-Atrash et al. in [4], assessed the association between inhaled cannabis and voice disorders. Similar to our review, the group only identified a single clinical study specific to voice. De Carvalho et al. in [10] performed a meta-analysis to establish the relationship between marijuana and H&N cancer. This group reported no association between lifetime marijuana use and H&N malignancy. Although these researchers were able to reach a conclusion on H&N cancer risk, the findings of our review have dissuaded us from doing the same as we believe that much of the H&N oncology literature is divided and still in its infancy. Furthermore, de Carvalho et al. (2015) limited their meta-analysis to case-control studies. Considering the cannabis literature has evolved significantly since 2015, we believe that higher-quality studies, including longitudinal studies, are needed to reach a clear consensus. A narrative review on the evidence for the use of cannabis in otolaryngology was recently completed by Valentino and Mckinnon. However, they did not adhere to PRISMA guidelines and perform a scoping review of the literature [9]. They did not use the Oxford Centre for Evidence-Based Medicine Levels of Evidence or perform a risk of bias assessment to evaluate the quality of the studies.
The strengths of this scoping review included the rigorous PRISMA methodology, comprehensive peer reviewed search strategy, and the inclusion of a fair number of reports. This was an ideal methodology to study this topic, as scoping reviews are useful for evaluating emerging evidence. Broad areas can be examined to identify knowledge gaps and show how research is currently being conducted in this field, helping to define more precise questions. Recognizing the heterogeneity of the literature, the choice to use a scoping review methodology allowed for data synthesis, hypothesis generation and will serve as a precursor for systematic reviews and future study in this early field. Lastly, this review was the first to group unintended side effects to each subspecialty, and attempt to define a preliminary risk profile of cannabis (i.e. Head and neck malignancy). There were some limitations. This review was limited to English language studies only. Case studies with less than five patients were excluded to ensure that only higher-quality studies were included. This decision may have excluded very rare side effects. Furthermore, given the multiple methods of cannabis consumption, the severity and degree of side effects may vary. This will likely be overcome as the Cannabis literature becomes more robust. Basic science studies were excluded to maintain a clinical focus in this review. The study was limited to the adult population because most jurisdictions require a legal age limit to consume cannabis.

Conclusion
This scoping review was undertaken to better understand the safety profile of cannabis and its potential for unintended side effects related to Otolaryngology. Although the quality of evidence in the included reports was lacking, it has provided an overview of the available literature, potential pitfalls or challenges to study design, and was hypothesis-generating. This review indicated that cannabis use is associated with several side effects, including hearing loss, H&N malignancy, and HPV infection. However, much of the H&N literature remains divided on the actual risk of cannabis use. Significant opportunities exist for the otolaryngology community to better investigate the therapeutic effects of cannabis with highquality studies, particularly on the risk of long-term use and harm reduction. As the rapidly evolving cannabis market continues to grow, clinicians will be expected to guide and counsel patients considering medical or recreational use, and it is essential that they have access to high-quality objective evidence.