Medications That Cause Dry Mouth As an Adverse Effect in Older People

A Systematic Review and Metaanalysis

Edwin C. K. Tan, PhD; Duangjai Lexomboon, PhD; Gunilla Sandborgh-Englund, PhD; Ylva Haasum, PhD; Kristina Johnell, PhD


J Am Geriatr Soc. 2018;66(1):76-84. 

In This Article


Protocol and Registration

The protocol for this review was registered with the International Prospective Register of Systematic Reviews (PROSPERO) (registration number CRD42016040033).

Search Strategy and Information Sources

A literature search was undertaken using Medline, Embase, Cochrane, Web of Science, and PubMed from 1990 to 2016. Medical Subject Headings (MeSH), Emtree subject headings, topics and key words related to dry mouth (xerostomia, asialia, hyposalivation, salivary dysfunction, salivary flow, salivary secretion, dry mouth), medication use (drugs, polypharmacy, specific drug classes), older people (aged, elder, senior, older), and study design (controlled trials, observational studies) were used. The full search strategy is included in the supplementary material (Supplementary Table S1). Searches were limited to English-language articles and excluded conference abstracts.

Eligibility Criteria

Studies were included in the review if they were original research published as a scientific peer-reviewed article from 1990 onwards; were intervention studies or observational studies (case-control studies, cohort studies, before-after studies and cross-sectional studies, conducted in humans); included persons aged 60 and older (if studies included subjects <60, data had to have been extractable for those ≥60; included a group exposed to drugs, including specific drugs and drug classes or number of drugs; included a comparator or control group that did not receive drugs or the drug of interest (studies that involved active comparators or persons who acted as their own controls were eligible); and reported on xerostomia, hyposalivation, or salivary gland hypofunction as adverse drug effects, measured in exposed and unexposed groups.

Study Selection

The titles and abstracts of studies were screened independently for relevance by two pairs of authors (ET and GS, DL and KJ). Full-text copies were obtained if a study appeared to meet the inclusion criteria or it was unclear whether it would meet the criteria. Two authors (ET and DL) independently reviewed the full texts to assess each study's suitability for inclusion. Disagreements or uncertainties about study inclusion were resolved by discussion in the presence of the two other authors (GS and KJ).

Data Extraction and Validity Assessment

Two authors (ET and DL) independently extracted data using a standardized data abstraction form. Data extracted included year of publication; study design; country; study setting; follow-up duration (for dry mouth outcome); study sample characteristics (age and relevant comorbidities); sample size; analytical sample size; year of data collection; definition of drug exposure, including name, dose, dose form, duration of use, and method used to determine drug exposure; definition of xerostomia or salivary gland hypofunction outcome; method used to determine presence of the outcome; and proportions in exposed and unexposed groups.

Methodological quality of studies was assessed using the Jadad score for randomized controlled trials (RCTs)[7] and an adapted version of the Newcastle-Ottawa Scale modified for cross-sectional studies.[8] For studies reporting pooled analyses, the original studies were assessed for quality. If at least one study did not meet a particular quality criterion, the pooled study was assigned a 0 for that quality criterion.


If 2 or more studies reported a similar drug exposure and primary outcome measure with appropriate extractable data, a metaanalysis was undertaken. Metaanalysis was performed using Review Manager version 5.3 (Nordic Cochrane Centre, Cochrane Collaboration, Copenhagen, Denmark). Random-effects models were used for pooling the data, and I 2 statistics were used to explore heterogeneity. The effect size for the metaanalysis was calculated as odds ratios (ORs) and risk differences (RDs). Significance was tested using Z-statistics.