Dry mouth effects from drugs used for depression, anxiety, schizophrenia and bipolar mood disorder in adults: systematic review

Abstract

Background

Poor oral health is increasingly recognised as an important comorbidity in people with psychiatric illness. One risk factor is psychotropic-induced dry mouth.

Aims

To perform a systematic review of the severity of dry mouth due to psychotropic drugs in adults (CRD42021239725). Study quality was assessed using the Cochrane risk of bias tool.

Method

We searched the following databases: PubMed, EMBASE, PsycINFO, Cochrane Central Register of Controlled Trials, CINAHL and Web of Science. We included randomised controlled trials (RCTs) measuring the severity of drug-induced hyposalivation and xerostomia.

Results

Eighteen RCTs with 605 participants were included. Severity of drug-induced dry mouth was compared among eight drug classes and/or against placebo. All studies were published 20 to 40 years ago and included tricyclic antidepressants (TCAs), serotonin specific reuptake inhibitors (SSRIs) and other drug classes. Meta-analysis was not feasible owing to design heterogeneity. TCAs caused more severe dry mouth, both objectively and subjectively, than placebo or other drug classes. SSRIs were generally associated with less severe symptoms. However, there was no information on antipsychotics or more recently available antidepressants, and there was minimal information on mood stabilisers. Most studies were on healthy subjects, limiting the generalisability of findings. Only one study measured both objective and subjective dry mouth, which have different clinical implications.

Conclusions

Psychotropic-induced dry mouth is a poorly researched area, and well-designed RCTs of newer psychotropic drugs using standardised objective and subjective measures are indicated. Given the ongoing use of TCAs for treatment-resistant depression, prescribers need to remain vigilant for xerostomia.

Dry mouth can present as a sign (salivary gland hypofunction), symptom (xerostomia) or both.¹ Salivary gland hypofunction is diagnosed based on objective measures of decreased saliva production, whereas xerostomia is the subjective experience of oral dryness and is based on self-report.

In dentistry, dry mouth is an area of interest because of its negative effects on oral health, which include increased risks of dental caries, periodontal disease, tooth demineralisation, tooth sensitivity and oral candidosis.^1,2 It can also lead to dysgeusia, dysphagia and difficulties with speech and mastication.³

Psychotropic medications and xerostomia

Psychotropic medications are among the many causes of dry mouth.^1,3 The importance of the issue was highlighted in recent reviews of the oral health side-effects of psychotropics reported to drug companies such as the Monthly Index of Medical Specialties (eMIMs).⁴ Of the 57 identified drugs (23 antidepressants, 22 antipsychotics or mood stabilisers, and 12 anxiolytic or sedative medications), xerostomia was the most frequently reported side-effect (91%) of the 28 identified oral symptoms among all classes of medication.⁴

Xerostomia can occur through several mechanisms but is often secondary to the anticholinergic effect arising from deactivation of the M3 muscarinic receptor, leading to poorer oral health outcomes.^5–12 This is a significant problem given the range of agents and frequency of psychotropic prescription.^13,14

The current study

The existing literature on drug-induced dry mouth investigates the risk within a specific population or drug class but does not address severity. For instance, there are studies that investigate dry mouth in older people,^3,5,6 but these may overestimate the effects of drug-induced dry mouth in the general population, as salivary production reduces with age. This is an important gap to address because psychiatric illnesses often start in early adulthood, exposing patients to decades of dry mouth.¹⁵ Other work has assessed the risk of dry mouth in working-age adults but was restricted to a limited range of psychotropics and did not consider severity.^{7–10,16–20}

We therefore assessed the severity of both subjective and objective dry mouth secondary to psychotropic drugs in adults above 17 years old among eight drug classes and/or against placebo.

Methodology

We followed the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta Analyses and preregistered the study with PROSPERO (CRD42021239725). As this was a systematic review of the published literature, ethical approval and written informed consent were not needed.

Search strategy

A systematic search for studies was conducted using the following databases: PubMed, EMBASE, PsycINFO, Cochrane Central Register of Controlled Trials, CINAHL and Web of Science. We used medical subject headings, Emtree terms, titles and keywords related to dry mouth, antidepressants, antianxiety agents, antipsychotics, mood stabilisers and study design. The full search strategy can be found in Supplementary Table 1 available at https://doi.org/10.1192/bjo.2023.15. No publication date or language filters were applied to the searches.

Drugs used for each mental health condition were cross-checked with Martindale²¹ and Cockburn et al⁴ based on data from the Australian Medicines Handbook, Australian Therapeutic Guidelines, eMIMs and UpToDate databases. This provided a comprehensive list of drugs to include in the search terms, ensuring that all potentially relevant papers were searched.

Selection process

Duplications were eliminated using Endnote, and articles were independently evaluated by pairs of reviewers (C.X.W.T., M.T. or S.L.) based on title and abstract. All full-text articles were independently assessed by two reviewers to be ‘included’, ‘excluded’ or ‘maybe’. Any discrepancies were resolved by consulting a third reviewer.

Eligibility criteria

We included all randomised controlled trials (RCTs) in both in-patient and out-patient settings of drug-induced dry mouth, including xerostomia and hyposalivation, as primary outcome. Severity could be assessed subjectively by questionnaire or objectively through stimulated and unstimulated salivary flow rate, oral mucosal wetness or oral moisture meter. Studies could investigate multiple drugs, specific drugs or drug classes. Comparisons included placebo, usual treatment or another intervention.

We excluded observational studies (such as case–control, cohort and cross-sectional designs), as well as those on burning mouth syndrome (BMS).²² Although BMS is commonly associated with dry mouth, it is difficult to establish whether dry mouth is a result of drug use or BMS. Observational studies may not adequately control confounding between comparisons and hence were excluded.

Data collection process

Data extraction was conducted by four authors (C.X.W.T., M.T., S.L. and S.K.) with at least two independently collecting data from each study. A further author was available to resolve or check any differences between raters. This exceeded the original PROSPERO protocol, which states that one author's data extraction should only be checked by another for accuracy. Tables of included studies reported on the objective or subjective measures of dry mouth and the associated drugs. Drugs had head-to-head comparisons with other active agents, or indirect comparisons against placebo or both. Other information included each study's author and publication date, out-patient or in-patient setting, age, gender and diagnosis of subjects. We planned to meta-analyse any comparisons where there were more than two studies with sufficient uniformity of method and outcomes.

Risk of bias

Chosen papers were judged for quality using the Cochrane risk of bias tool.²³ Papers were scored as being of low, high or unclear risk of bias by two independent reviewers.

Results

The initial search identified 5199 references (Fig. 1). After duplicates had been removed, 1634 remained and, following screening at title, abstract and full-text levels, 18 studies were included (Fig. 1).^24–41 Where stated, all were in out-patient settings. Fourteen studies were conducted in healthy subjects and four in people with depression (Tables 1 and 2). The majority of the studies in healthy subjects were of single doses

Fig. 1 Flow diagram.

Table 1 Included studies measuring objective dry mouth

a. 2–4 h post-dose.

b. 5–6 h post-dose.

c. Both agents significantly worse than placebo, doxepin significantly worse than desipramine.

d. 10 h post-dose for whole-mouth results only.

e. Within the group, nomifensine and zimelidine inhibited salivation significantly less than imipramine and mianserin.

f. Up to 72 h post-dose.

g. After 5 weeks of therapy compared with placebo. However, amitriptyline was significantly worse than oxaprotiline.

TCA, tricyclic antidepressants; NS, non-significant. Data presented for highest dose of each agent. Studies marked in italics indicate where results were only presented in graphical form.

Table 2 Included studies measuring subjective dry mouth

VAS, Visual Analogue Scale.

Outcomes

Of the included studies, 14 focused on objective measures of hyposalivation (Table 1), three investigated subjective measures of xerostomia severity (Table 2) and one investigated both objective and subjective measures of severity (Tables 1 and 2).³⁶

Objective measures used one of two validated techniques that have been shown to give similar results (Table 1).^42–47 Two studies used the stimulated spit method.^30,41 This is where saliva is collected in a funnel following the sucking of a lozenge.⁴⁴ The others measured spontaneous saliva flow using cotton rolls. In this method, absorbent material is placed in the mouth for a fixed time (usually 2 min), and the increase in weight is compared at intervals pre- and post-dose.^45–47 Numerical values were therefore estimated using WebPlotDigitizer online software.⁴⁸ In terms of subjective data, two studies used visual analogue scales, and the other two used three- or four-point rating measures (Table 2).

Differences by drug class

Tables 1 and 3 summarise the results by drug class. Tricyclic antidepressants (TCAs) were the most commonly studied class and were included in all the studies, either as the agent of interest or the comparison drug (k = 18).^24–41 These were followed by selective serotonin reuptake inhibitors (SSRIs, k = 5)^{28,29,34,36,38} and tetracyclic antidepressants (TeCAs, k = 6).^{28,31,32,38,39,41} The remaining studies were on monoamine oxidase inhibitors (MAOIs, k = 2),^{28, 38} norepinephrine–dopamine reuptake inhibitors (NDRIs, k = 2),^28,38 serotonin antagonist and reuptake inhibitors (k = 2),^26,35 lithium citrate (k = 2)^28,38 and reboxetine, a norepinephrine reuptake inhibitor (NRI) (k = 1).³⁶

Table 3 Summary of comparisons

K, number of studies.

In general, TCAs caused more severe dry mouth on both objective and subjective measures in comparison with placebo and other drug classes. Amitriptyline was the most studied TCA, and all 12 studies found that it caused more severe dry mouth than placebo and all active agents on both objective and subjective measures. Comparing within the drug class, doxepin caused more severe dry mouth than desipramine.^24,37 When compared with placebo, there were mixed findings for desipramine, with it causing more severe dry mouth than placebo in two studies but the same as placebo in another.^24,37,40

In terms of TeCAs, both mianserin and oxaprotiline caused less severe dry mouth than amitriptyline on objective measures in two studies,^39,41 although mianserin had similar effects to other TCAs in two others.^28,38 It caused less severe dry mouth than nortriptyline subjectively in a further study.³² There were no comparisons within the drug class. When compared with placebo, TeCAs caused more severe dry mouth than placebo,^28,31,38 as well as lithium citrate and isocarboxazide.³⁸

There was less data on other drug classes. For SSRIs, fluvoxamine was the same as placebo in one study,²⁹ as was citalopram in another.³⁶ However, citalopram caused more severe dry mouth than placebo on subjective measures in a third study.³⁴ MAOI isocarboxazide, NRI reboxetine and mood stabiliser lithium citrate were the same as placebo,^28,36,38 whereas NDRI nomifensine caused significantly more severe dry mouth than placebo in one study,³⁸ but less severe symptoms than imipramine, nortriptyline and mianserin in two papers.^28,38

Risk of bias

Table 4 shows the risk of bias ratings. Eight out of 18 studies reported on methods to blind participants and assessors such as making placebo and intervention pills identical in appearance. A similar proportion had low attrition rates. However, the remaining domains were rated as unclear or at high risk of bias. Owing to the age of the studies (1972–2001), none used reporting guidelines such as CONSORT.

Table 4 Risk-of-bias judgment for each of six domains of bias for each included study

Discussion

To our knowledge, this is the first systematic review of both the objective and subjective severity of dry mouth associated with psychotropic medications, as previous reviews only assessed self-reported prevalence without considering severity.^3,7,9 We found that in both direct comparisons between drugs and indirect comparisons between drugs and placebo, newer psychotropic agents such as SSRIs caused less dry mouth objectively and subjectively. By contrast, TCAs (especially amitriptyline) were generally associated with a greater severity of dry mouth largely owing to their anticholinergic effects. This finding extends existing literature, including a previous meta-analysis that was restricted to the self-reported prevalence, not severity, of dry mouth in individuals receiving SSRIs.^9,49

Despite the scope of the review being extended to a range of psychotropic drugs, concerns included the relatively low number of included studies and their age; all were over 20 years old. In particular, there was no information on many commonly used psychotropics including fluoxetine, sertraline, duloxetine, venlafaxine and desvenlafaxine. This is despite xerostomia being identified as by far the most frequently reported symptom of 28 drug-company-reported oral side-effects among all classes of psychotropic medications.⁴ For instance, xerostomia has been reported as a common side-effect (>10%) of the following newer and commonly used antidepressants: citalopram, fluoxetine, fluvoxamine, paroxetine, sertraline, venlafaxine, desvenlafaxine and duloxetine.⁴ These were generally the same medications for which we could not find any literature on the severity of dry mouth when it does occur.

Of particular concern is that despite the potential oral health consequence of drug-induced dry mouth, many people do not access dental care. In a nationwide study of Australian service veterans, 40% (n = 50 679) were taking at least one medication that caused dry mouth, but fewer than half of them made a claim for dental services in the following year.⁵⁰

Limitations

This work had several limitations. Given that the studies were published between 20 and 40 years ago, most investigated medications that are less used today. This makes it difficult for practitioners to seek guidance from research to manage patients on newer psychotropic drugs. The age of the studies could therefore indicate a lack of awareness in studying the oral side-effects of psychotropic drugs, and that the impact of dry mouth on the oral health of patients with psychiatric illnesses may have been neglected in recent years.

In addition, meta-analyses of our included studies were not possible owing to the heterogeneity of agents, subjects, study designs and outcome measures. In addition, many papers only presented graphical data without providing raw numbers. Although we used online digitisation software to estimate numerical values, our results should be viewed with caution. Ten of the 18 included studies did not report on blinding methods.

There were comparatively few studies for many of the agents, and in some cases (e.g. desipramine and citalopram) the evidence was contradictory. There was even less information on mood stabilisers and none on antipsychotics, although these were included in our search strategy. Importantly, there was no information on the most commonly used antidepressants, including escitalopram, paroxetine, fluoxetine, sertraline, duloxetine, venlafaxine and desvenlafaxine. Furthermore, most studies were of single doses in limited numbers of healthy subjects instead of individuals with psychiatric illness, which limits the generalisability of findings to clinical practice. For instance, people with psychiatric illness may be taking several other medications or have physical comorbidities that can exacerbate dry mouth. All but two of the studies were also very small, which may mean that they were underpowered to detect significant findings. These factors may mean that our findings underestimate the burden in psychiatric populations.

Last, few studies measured both objective and subjective dry mouth, which have different clinical implications. Objective dry mouth directly increases the risk of oral diseases, whereas both subjective and objective dry mouth may potentially reduce compliance owing to patients’ inability to tolerate the adverse effects. Both objective and subjective measures are therefore important in providing guidance for practitioners in the management of patients’ treatment adherence and risk of oral diseases.

Implications

Box 1 summarises the key implications. First, more research on newer psychotropic drugs is needed, particularly antipsychotics and commonly used antidepressants such as fluoxetine, sertraline, duloxetine, venlafaxine and desvenlafaxine. Second, the presentation of results should be standardised, and they should not be restricted to graphs but rather include numerical data that could be pooled for meta-analysis. Third, in order to increase the generalisability of findings to clinical practice, future research should involve participants with psychiatric illnesses. Last, future research should include both objective and subjective measures of dry mouth.

Box 1 Study implications

• • Dry mouth is a common side-effect of many psychotropic medications.

• • There is less information on differences in the severity of symptoms between agents, both subjectively or objectively measured using salivary flow.

• • The present study was a systematic review of the literature on the severity of both subjective and objective dry mouth due to psychotropic drugs in adults.

• • Eighteen RCTs with 605 participants were included, and severity was compared among eight drug classes and/or against placebo.

• • All the studies were published 20 to 40 years ago and most investigated older drugs. As expected, TCAs caused more severe dry mouth, both objectively and subjectively, than placebo or newer agents.

• • No information was available for antipsychotic medications or many of the commonly used antidepressants, and minimal information was available on mood stabilisers. This limits the generalisability of the study findings.

• • The lack of research on the severity of psychotropic-induced dry mouth is surprising given its common occurrence; well-designed RCTs of newer psychotropic drugs are therefore indicated.

• • The extensive use of SSRIs has shifted the focus away from xerostomia, although this is an important side-effect given the continuing role of TCAs in the treatment of both chronic pain and treatment-resistant depression.

Conclusion

This systematic review suggests that newer psychotropic drugs are associated with less severe dry mouth. However, this is based on very limited evidence. Given that TCAs are still an important medication for chronic pain and treatment-refractory depression,⁵¹ both medical and dental practitioners should assess and manage the oral implications of dry mouth. This includes addressing other contributory factors (e.g. avoidance of caffeinated beverages, smoking cessation), advice on taking frequent sips of water throughout the day, and the use of oral lubricants, saliva substitutes or saliva stimulants, as well as management of any potential oral mucosal and dental complications. Finally, the lack of research on the severity of psychotropic-induced dry mouth is surprising given its common occurrence; well-designed RCTs of newer psychotropic drugs are therefore indicated.