A systematic review and meta-analysis of thiazide-induced hyponatraemia: time to reconsider electrolyte monitoring regimens after thiazide initiation?

AIMS: Hyponatraemia is one of the major adverse effects of thiazide and thiazide-like diuretics and the leading cause of drug-induced hyponatraemia requiring hospital admission. We sought to review and analyze all published cases of this important condition. METHODS: Ovid Medline, Embase, Web of Science and PubMed electronic databases were searched to identify all relevant articles published before October 2013. A proportions meta-analysis was undertaken. RESULTS: One hundred and two articles were identified of which 49 were single patient case reports. Meta-analysis showed that mean age was 75 (95% CI 73, 77) years, 79% were women (95% CI 74, 82) and mean body mass index was 25 (95% CI 20, 30) kg m(-2) . Presentation with thiazide-induced hyponatraemia occurred a mean of 19 (95% CI 8, 30) days after starting treatment, with mean trough serum sodium concentration of 116 (95% CI 113, 120) mm and serum potassium of 3.3 (95% CI 3.0, 3.5) mm. Mean urinary sodium concentration was 64 mm (95% CI 47, 81). The most frequently reported drugs were hydrochlorothiazide, indapamide and bendroflumethiazide. CONCLUSIONS: Patients with thiazide-induced hyponatraemia were characterized by advanced age, female gender, inappropriate saliuresis and mild hypokalaemia. Low BMI was not found to be a significant risk factor, despite previous suggestions. The time from thiazide initiation to presentation with hyponatraemia suggests that the recommended practice of performing a single investigation of serum biochemistry 7-14 days after thiazide initiation may be insufficient or suboptimal. Further larger and more systematic studies of thiazide-induced hyponatraemia are required.

On the alert: future priorities for alerts in clinical decision support for computerized physician order entry identified from a European workshop.

BACKGROUND: Clinical decision support (CDS) for electronic prescribing systems (computerized physician order entry) should help prescribers in the safe and rational use of medicines. However, the best ways to alert users to unsafe or irrational prescribing are uncertain. Specifically, CDS systems may generate too many alerts, producing unwelcome distractions for prescribers, or too few alerts running the risk of overlooking possible harms. Obtaining the right balance of alerting to adequately improve patient safety should be a priority. METHODS: A workshop funded through the European Regional Development Fund was convened by the University Hospitals Birmingham NHS Foundation Trust to assess current knowledge on alerts in CDS and to reach a consensus on a future research agenda on this topic. Leading European researchers in CDS and alerts in electronic prescribing systems were invited to the workshop. RESULTS: We identified important knowledge gaps and suggest research priorities including (1) the need to determine the optimal sensitivity and specificity of alerts; (2) whether adaptation to the environment or characteristics of the user may improve alerts; and (3) whether modifying the timing and number of alerts will lead to improvements. We have also discussed the challenges and benefits of using naturalistic or experimental studies in the evaluation of alerts and suggested appropriate outcome measures. CONCLUSIONS: We have identified critical problems in CDS, which should help to guide priorities in research to evaluate alerts. It is hoped that this will spark the next generation of novel research from which practical steps can be taken to implement changes to CDS systems that will ultimately reduce alert fatigue and improve the design of future systems.

A practical guide to monitoring for adverse drug reactions during antihypertensive drug therapy.

Monitoring of patients taking antihypertensive treatment can identify potential adverse drug reactions (ADRs). However, published guidelines give divergent or incomplete recommendations on monitoring for ADRs. Using a predetermined strategy, we undertook a systematic review to identify hypertension guidelines published from January 2001 to October 2011 with recommendations for monitoring for ADRs. We screened 88 abstracts and 187 web-based guidelines, and identified 19 published guidelines on monitoring the biochemical effects of antihypertensive drug therapy. We then produced a set of practical clinical guidelines, synthesized from those recommendations. Our recommendations are designed to provide efficient monitoring. They reduce the number of tests to a minimum consistent with safe practice and align monitoring schedules, so that creatinine, potassium and sodium concentrations are measured at the same times in all cases. The instructions for biochemical monitoring in current guidelines differ greatly, both in the extent of advice and in the detail provided. The current lack of consistent and workable instructions poses serious difficulties for practitioners. The recommendations distilled from this systematic review should help practitioners when they monitor therapy with antihypertensive drugs.

Developing consensus on hospital prescribing indicators of potential harms amenable to decision support.

AIMS: To develop a list of prescribing indicators specific for the hospital setting that would facilitate the prospective collection of high-severity and/or high-frequency prescribing errors, which are also amenable to electronic clinical decision support. METHODS: A two-stage consensus technique (electronic Delphi) was carried out with 20 experts across England. Participants were asked to score prescribing errors using a five-point Likert scale for their likelihood of occurrence and the severity of the most likely outcome. These were combined to produce risk scores, from which median scores were calculated for each indicator across the participants in the study. The degree of consensus between the participants was defined as the proportion that gave a risk score in the same category as the median. Indicators were included if a consensus of 80% or more was achieved. RESULTS: A total of 80 prescribing errors were identified by consensus as being high or extreme risk. The most common drug classes named within the indicators were antibiotics (n = 13), antidepressants (n = 8), nonsteroidal anti-inflammatory drugs (n = 6) and opioid analgesics (n = 6). The most frequent error type identified as high or extreme risk were those classified as clinical contraindications (n = 29 of 80). CONCLUSIONS: Eighty high-risk prescribing errors in the hospital setting have been identified by an expert panel. These indicators can serve as a standardized, validated tool for the collection of prescribing data in both paper-based and electronic prescribing processes. This can assess the impact of safety improvement initiatives, such as the implementation of electronic clinical decision support.

Improvement in the detection of adverse drug events by the use of electronic health and prescription records: an evaluation of two trigger tools.

PURPOSE: To test if two of the adverse event triggers proposed by the Institute of Healthcare Improvement can detect adverse drug events (ADEs) in a UK secondary care setting, using an electronic prescribing and health record system. METHODS: In order to identify triggers for over-anticoagulation and potential opioid overdose and we undertook a retrospective review of electronic medical and prescription records from 54,244 hospital admissions over a 1-year period, alongside a review of medical incident reports. Once prescription data were linked to triggers and duplicates were removed, case note review eliminated the false positive ADEs. Additionally, we tested the use of an electronic algorithm for the International Normalized Ratio (INR) ≥6 trigger. RESULTS: The INR ≥6 electronic trigger identified 46 potential ADEs and the naloxone electronic trigger identified 82 ADEs. Based on the available case note review, the INR ≥6 trigger had a positive predictive value (PPV) of 38 % (14/37) and the naloxone trigger had a PPV of 91 % (61/67). The electronic algorithm for the INR ≥6 trigger identified 12 ADEs, thus reducing the need of case note review. This was in comparison with one and two critical incidents reported in the trust medical incident reports system, which respectively related to over-coagulation with warfarin and over-sedation with opioid medication. CONCLUSIONS: We have integrated automated and manual methods of detecting ADEs using previously defined triggers. Incorporating electronic triggers in already established electronic health records with prescription and laboratory test data can improve the detection of ADEs, and potentially lead to methods to avert them.

Dose omissions in hospitalized patients in a UK hospital: an analysis of the relative contribution of adverse drug reactions.

BACKGROUND: The omission of charted (prescribed) doses for hospitalized patients is an important problem in the UK. Inappropriate drug omission can clearly lead to harm from lack of therapeutic effect. However, healthcare professionals administering medicines may decide that omission of a dose is appropriate in certain circumstances, e.g. when patients show signs of a possible adverse drug reaction (ADR). OBJECTIVE: The aim of this study was to characterize dose omissions to understand the factors that influence non-administration of therapy and to determine the proportion of doses that are appropriately omitted due to ADRs. METHODS: We used data from a bespoke hospital-wide electronic prescribing and administration system at University Hospitals Birmingham NHS Foundation Trust, Birmingham, UK. We extracted data on 6.01 million drug administrations during 2010 and then randomly selected four 7-day periods, concentrating on doses that were charted but not given. Omitted medicines were counted if either there was a charted 'non-administration' (i.e. an active acknowledgement of the omitted dose) or there was no charting of that dose (i.e. no record of either administration or omission). Paused medicines were not counted. When a dose was omitted, staff indicated the reasons for non-administration using codes ('hard coded') or free text in the electronic system. We used both to compare the contribution of different factors, including ADRs, to the total rates of dose omissions. RESULTS: In the four 7-day periods analysed, 60 763 (12.4%) of the 491 894 charted doses were omitted. The most common code was 'patient refused drug' (45.4%). Only 1.6% of doses were omitted for reasons of patient safety, of which 4 in 1000 omissions were coded as directly due to an ADR. CONCLUSIONS: Measures to improve the quality of care should seek to reduce dose omissions, but in some cases omission may be rational. Electronic medication administration records allow for detailed analysis of decisions made by healthcare professionals at the point of administration. While dose omissions related to ADRs are uncommon, they are important both for patient safety and for therapeutic decision making.

The potential of the internet.

The internet and the World Wide Web have changed the ways that we function. As technologies grow and adapt, there is a huge potential for the internet to affect drug research and development, as well as many other aspects of clinical pharmacology. We review some of the areas of interest to date and discuss some of the potential areas in which internet-based technology can be exploited. Information retrieval from the web by health-care professionals is common, and bringing evidence-based medicine to the bedside affects the care of patients. As a primary research tool the web can provide a vast array of information in generating new ideas or exploring previous research findings. This has facilitated systematic reviewing, for example. The content of the web has become a subject of research in its own right. The web is also widely used as a research facilitator, including enhancement of communication between collaborators, provision of online research tools (such as questionnaires, management of large scale multicentre trials, registration of clinical trials) and distribution of information. Problems include information overload, ignorance of early data that are not indexed in databases, difficulties in keeping web sites up to date and assessing the validity of information retrieved. Some web-based activities are viewed with suspicion, including analysis by pharmaceutical companies of drug information to facilitate direct-to-consumer advertising of novel pharmaceuticals. Use of these technologies will continue to expand in often unexpected ways. Clinical pharmacologists must embrace internet technology and include it as a key priority in their research agenda.

Biochemical monitoring of patients treated with antihypertensive therapy for adverse drug reactions: a systematic review.

Biochemical monitoring of patients treated with antihypertensive therapy is recommended in order to identify potential adverse reactions to treatment. We aimed to review the literature investigating the nature of biochemical monitoring in adults treated in primary care with antihypertensive drugs. Specifically, we wished to establish (i) the proportion of patients with biochemical baseline testing prior to the initiation of antihypertensive therapy; (ii) the proportion of patients with biochemical monitoring after initiation of antihypertensive therapy; (iii) the patient characteristics associated with biochemical monitoring; (iv) the frequency of biochemical monitoring after the initiation of antihypertensive therapy; and (v) the relationship, if any, between biochemical monitoring and adverse patient outcomes. We searched MEDLINE, EMBASE and Google Scholar from 1948 to 31 December 2010 using a combination of text words and search terms. Retrospective and prospective cohort studies, cross-sectional studies, randomized controlled trials or quasi-randomized controlled trials, and audits of current clinical practice were included. Clinical trials, case reports and case series were excluded. Studies were included if they provided data on the proportion of patients treated with antihypertensive therapy in primary care who had any biochemical monitoring before or after the initiation of therapy. In total, 15 studies were included in our review, which used a wide variety of definitions of monitoring prior to and after the initiation of antihypertensive therapy. From 17% to 81% of patients treated with antihypertensive drugs had a baseline biochemical test and from 20% to 79% had any follow-up monitoring. In only 7 of the 12 studies that examined follow-up monitoring did more than half of the patients have any monitoring. Overall, this systematic review provides evidence that monitoring as recommended by published guidelines is not commonly undertaken. Only two studies were identified that examined patients with both baseline testing and follow-up monitoring. Omission of one or the other limits the ability to analyse the effect of treatment on electrolyte concentrations or renal function. There is limited research on the patient factors associated with monitoring and further work is required to determine the impact of monitoring on adverse patient outcomes. Important barriers to effective monitoring exist and this review emphasizes that these have not yet been overcome.

Metal ions: supporting actors in the playbook of small ribozymes.

Since the 1980s, several small RNA motifs capable of chemical catalysis have been discovered. These small ribozymes, composed of between approximately 40 and 200 nucleotides, have been found to play vital roles in the replication of subviral and viral pathogens, as well as in gene regulation in prokaryotes, and have recently been discovered in noncoding eukaryotic RNAs. All of the known natural small ribozymes - the hairpin, hammerhead, hepatitis delta virus, Varkud satellite, and glmS ribozymes--catalyze the same self-cleavage reaction as RNase A, resulting in two products, one bearing a 2'-3' cyclic phosphate and the other a 5'-hydroxyl group. Although originally thought to be obligate metalloenzymes like the group I and II self-splicing introns, the small ribozymes are now known to support catalysis in a wide variety of cations that appear to be only indirectly involved in catalysis. Nevertheless, under physiologic conditions, metal ions are essential for the proper folding and function of the small ribozymes, the most effective of these being magnesium. Metal ions contribute to catalysis in the small ribozymes primarily by stabilizing the catalytically active conformation, but in some cases also by activating RNA functional groups for catalysis, directly participating in catalytic acid-base chemistry, and perhaps by neutralizing the developing negative charge of the transition state. Although interactions between the small ribozymes and cations are relatively nonspecific, ribozyme activity is quite sensitive to the types and concentrations of metal ions present in solution, suggesting a close evolutionary relationship between cellular metal ion homeostasis and cation requirements of catalytic RNAs, and perhaps RNA in general.

Republished paper: Where errors occur in the preparation and administration of intravenous medicines: a systematic review and Bayesian analysis.

OBJECTIVE: To investigate the overall probability of error in preparing and administering intravenous medicines; to identify at which stage of the process an error is most likely to occur; and to determine the impact of error correction on the error probability. Design Systematic review and random-effects Bayesian conditional independence modelling. METHODS: Medline and EMBASE were searched for studies on intravenous medicines. The error rates of each stage were extracted. These, expert estimates, and error rates from generic tasks, were used in a Bayesian conditional independence model to find error 'hot-spots.' The main outcome measure was the probability of at least one error occurring during intravenous therapy. RESULTS: Nine published studies were identified for inclusion in the systematic review and meta-analysis. The overall probability of making at least one error in intravenous therapy was 0.73 (95% credible interval (CrI) 0.54 to 0.90). If error-checking was introduced at each stage of the process, the overall rate fell to 0.22 (95% CrI 0.14 to 0.31). Errors were most likely in the reconstitution step. Removing the reconstitution step by providing preprepared injections would reduce the overall error rate to 0.17 (95% CrI 0.09 to 0.27). CONCLUSIONS: Intravenous therapy is complex and error-prone. Error-checking at each stage could reduce the error probability. The use of preprepared injections may help by eliminating errors in the reconstitution of drug and diluent. However, it will be important to ensure that benefits are not outweighed by practical disadvantages such as an increase in selection errors.

Oversight: a retrospective study of biochemical monitoring in patients beginning antihypertensive drug treatment in primary care.

WHAT IS ALREADY KNOWN ABOUT THIS SUBJECT: Guidelines recommend biochemical monitoring of patients treated with antihypertensive agents, although there is little primary evidence for these recommendations. WHAT THIS STUDY ADDS: Patients treated for hypertension in general practice often have no biochemical tests before, or in the 6 months after, starting drug treatment. AIMS: Guidelines on the management of hypertension have recommended baseline testing of serum electrolyte and creatinine concentrations before treatment since the 1990s. We wished to examine the extent of laboratory monitoring in patients with newly diagnosed hypertension and newly treated with antihypertensive drugs. METHODS: We carried out a retrospective analysis of 74,096 patients in the General Practice Research Database (GPRD) aged 18 years and older with newly diagnosed hypertension and prescribed a single antihypertensive agent. We determined the number of patients with a laboratory test for serum electrolyte and creatinine (or urea) concentrations prior to the first antihypertensive drug prescription and in the 6 months after and patient factors associated with subsequent monitoring. RESULTS: Thirty-four thousand nine hundred and forty-seven patients (47%) had at least one biochemical test in the 12 months prior to beginning antihypertensive treatment, and 26,946 (36%) had at least one biochemical monitoring test in the 6 months after beginning antihypertensive treatment. Thirteen thousand five hundred and four (18%) had both baseline and monitoring tests. Baseline tests were normal in 11,671 patients (86%), of whom 10,213 (88%) had normal tests at first monitoring. Monitoring was significantly more likely in patients treated with ACE inhibitors than thiazides (adjusted OR 1.90; 95% CI 1.80, 2.00), older patients (adjusted OR 1.23; 95% CI 1.11, 1.36) [individuals aged 80-89 years compared with <40 years], and patients with diabetes mellitus (adjusted OR 2.03; 95% CI 1.91, 2.16). CONCLUSION: Biochemical testing at baseline and monitoring after starting treatment is often omitted in newly diagnosed hypertensive patients. When both are tested, one in eight normal results becomes abnormal.

Correlates of spontaneous reporting of adverse drug reactions within primary care: the paradox of low prescribers who are high reporters.

AIM(S): To examine Primary Care Trust (PCT) demographics influencing general practitioner (GP) involvement in pharmacovigilance. METHODS: PCT adverse drug reaction (ADR) reports to the Yellow Card scheme between April 2004 and March 2006 were obtained for the UK West Midlands region. Reports were analysed by all drugs, and most commonly reported drugs ('top drugs'). PCT data, adjusted for population size, were aggregated. Prescribing statistics and other characteristics were obtained for each PCT, and associations between these characteristics and ADR reporting rates were examined. RESULTS: During 2004-06, 1175 reports were received from PCTs. Two hundred and eighty (24%) of these reports were for 14 'top drugs'. The mean rate of reporting for PCTs was 213 reports per million population. A total of 153 million items were prescribed during 2004-06, of which 33% were 'top drugs'. Reports for all drugs and 'top drugs' were inversely correlated with the number of prescriptions issued per thousand population (r(s)=-0.413, 95% CI -0.673, -0.062, P < 0.05, and r=-0.420, 95% CI -0.678, -0.071, P < 0.05, respectively). Reporting was significantly negatively correlated with the percentages of male GPs within a PCT, GPs over 55 years of age, single-handed GPs within a PCT, the average list size of a GP within a PCT, the overall deprivation scores and average QOF total points. ADR reports did not correlate significantly with the proportion of the population over 65 years old. CONCLUSIONS: Some PCT characteristics appear to be associated with low levels of ADR reporting. The association of low prescribing areas with high ADR reporting rates replicates previous findings.

Laboratory monitoring and adverse patient outcomes with antihypertensive therapy in primary care.

PURPOSE: The monitoring of serum electrolyte and creatinine concentrations in patients treated with antihypertensive therapy is recommended. We wished to examine the relationship between laboratory monitoring and adverse patient outcomes. METHODS: We carried out a retrospective cohort study using the General Practice Research Database (GPRD). Patients aged 18 years or older with newly diagnosed hypertension and prescribed a single antihypertensive agent were included. Monitoring was defined as any laboratory test for serum electrolyte and creatinine (or urea) concentrations within 6 months of starting treatment. RESULTS: We identified 74 096 patients who were newly diagnosed with hypertension and prescribed a single antihypertensive agent. Twenty six thousand nine hundred forty six (36.4%) patients had any biochemical laboratory measurement within 6 months. Three hundred ten patients (0.4%) died, 1451 (2%) were admitted to hospital at least once and 29 749 (40.2%) discontinued their first course of antihypertensive treatment within 6 months. Patients were more likely to be admitted to hospital if their biochemistry had been monitored after beginning treatment (adjusted hazard ratio (HR) 1.37; 95%CI 1.21-1.55). They were also marginally more likely to discontinue treatment (adjusted HR 1.04; 95%CI 1.02-1.07). They were not significantly more likely to die (adjusted HR 1.21; 95%CI 0.87-1.67). CONCLUSIONS: Biochemical testing at baseline and monitoring after starting treatment is often omitted in newly diagnosed hypertensive patients. Those patients who are monitored are more likely to be admitted to hospital and to discontinue initial antihypertensive therapy, but not to die. Many biochemical adverse drug reactions are found only by laboratory monitoring.

Where errors occur in the preparation and administration of intravenous medicines: a systematic review and Bayesian analysis.

OBJECTIVE: To investigate the overall probability of error in preparing and administering intravenous medicines; to identify at which stage of the process an error is most likely to occur; and to determine the impact of error correction on the error probability. DESIGN: Systematic review and random-effects Bayesian conditional independence modelling. METHODS: Medline and EMBASE were searched for studies on intravenous medicines. The error rates of each stage were extracted. These, expert estimates, and error rates from generic tasks, were used in a Bayesian conditional independence model to find error 'hot-spots.' The main outcome measure was the probability of at least one error occurring during intravenous therapy. RESULTS: Nine published studies were identified for inclusion in the systematic review and meta-analysis. The overall probability of making at least one error in intravenous therapy was 0.73 (95% credible interval (CrI) 0.54 to 0.90). If error-checking was introduced at each stage of the process, the overall rate fell to 0.22 (95% CrI 0.14 to 0.31). Errors were most likely in the reconstitution step. Removing the reconstitution step by providing preprepared injections would reduce the overall error rate to 0.17 (95% CrI 0.09 to 0.27). CONCLUSIONS: Intravenous therapy is complex and error-prone. Error-checking at each stage could reduce the error probability. The use of preprepared injections may help by eliminating errors in the reconstitution of drug and diluent. However, it will be important to ensure that benefits are not outweighed by practical disadvantages such as an increase in selection errors.

The pathophysiology of medication errors: how and where they arise.

1. Errors arise when an action is intended but not performed; errors that arise from poor planning or inadequate knowledge are characterized as mistakes; those that arise from imperfect execution of well-formulated plans are called slips when an erroneous act is committed and lapses when a correct act is omitted. 2. Some tasks are intrinsically prone to error. Examples are tasks that are unfamiliar to the operator or performed under pressure. Tasks that require the calculation of a dosage or dilution are especially susceptible to error. 3. The tasks of prescribing, preparation, and administration of medicines are complex, and are carried out within a complex system; errors can occur at each of many steps and the error rate for the overall process is therefore high. 4. The error rate increases when health-care professionals are inexperienced, inattentive, rushed, distracted, fatigued, or depressed; orthopaedic surgeons and nurses may be more likely than other health-care professionals to make medication errors. 5. Medication error rates in hospital are higher in paediatric departments and intensive care units than elsewhere. 6. Rates of medication errors may be higher in very young or very old patients. 7. Intravenous antibiotics are the drugs most commonly involved in medication errors in hospital; antiplatelet agents, diuretics, and non-steroidal anti-inflammatory drugs are most likely to account for 'preventable admissions'. 8. Computers effectively reduce the rates of easily counted errors. It is not clear whether they can save lives lost through rare but dangerous errors in the medication process.

Should doctors who make clinical errors be charged with manslaughter? A survey of medical professionals and members of the public.

If a doctor is grossly negligent and the patient dies as a result, the doctor can be charged with manslaughter. We have investigated the difference in opinion between medical professionals and the public on whether doctors should face criminal charges following different fatal medical errors. We conducted a survey of 40 medical professionals and 40 members of public, using a set of questions about negligence and manslaughter relating to four real-life cases of doctors charged with manslaughter where eventual outcomes were known. Medical professionals and the public agreed that lessons could be learnt from all four cases and that an independent review of each case should be carried out. However, across all cases, the public were more likely to respond that the doctor should be charged with manslaughter (OR = 2.1; 95% CI = 1.3-3.2). The public and, to a lesser extent, medical professionals still hold individuals responsible following a death due to medical error. This has implications for those who advocate a systems-based approach for assessing the root causes of medical errors, where there is a limited focus on individual accountability.

Ethnic differences in the risks of adverse reactions to drugs used in the treatment of psychoses and depression: a systematic review and meta-analysis.

BACKGROUND: Factors such as age, sex and disease state alter a patient's susceptibility to adverse drug reactions (ADRs). Ethnicity may also alter the risk of an ADR. OBJECTIVE: To review the evidence for ethnic differences in susceptibility to adverse reactions to drugs used to treat psychoses and depression. DATA SOURCES: We searched MEDLINE (from 1951), EMBASE (from 1974), and PsycINFO (from 1950) to March 2006. STUDY SELECTION: Studies were included if there was a mention of ethnicity, ethnic or racial groups and a description of a procedure to investigate ADRs specifically or a description of ADRs that were a result of drugs in therapeutic use. Studies selected by any two reviewers were retained if they referred to drugs used in the treatment of psychoses and related disorders or antidepressant drugs. Of 124 studies describing ADRs to antipsychotics or antidepressants, 51 reported data from different ethnic groups. DATA EXTRACTION: Data were extracted independently from those studies selected for inclusion by two reviewers, using a standard data extraction form. Studies were assessed for bias in order to determine the quality of the study. DATA SYNTHESIS: In a pooled analysis of patients treated with antipsychotics, the relative risk (RR) of tardive dyskinesia in Black compared with White patients was 1.03 (95% CI 0.85, 1.24); the RR of extrapyramidal symptoms in East Asian compared with non-East Asian patients was 1.38 (95% CI 1.11, 1.72); the RR of hyperglycaemia in Black compared with non-Black patients was 1.55 (95% CI 0.95, 2.53); and the RR of diabetes mellitus in non-White compared with White patients was 1.35 (95% CI 0.95, 1.92). It was impossible to perform pooled analysis of data from studies investigating antidepressants due to insufficient data. CONCLUSIONS: We found limited evidence of ethnic differences in the risk of ADRs. The clinical implications of these results remain unclear because of confounding factors. Further progress will require improved recruitment of patients from different ethnic groups and an established consensus on how to define ethnicity.