Drug-Induced Orthostatic Hypotension in UK Primary Care: A Self-Controlled Case Series Study

Abstract

Background

Drug-induced orthostatic hypotension (OH) is common and increasing worldwide with rising polypharmacy. Real-world evidence is crucial for safe prescribing. We examined the risk of OH associated with commonly prescribed drugs in UK primary care among adults aged ≥60 years.

Methods

This was a self-controlled case series using routinely collected UK primary care data from the IQVIA Medical Research Database (IMRD). Individuals with an OH diagnosis and drug exposure between 1 January 2000 and 31 December 2018 were included. Drug classes potentially causing OH were coded by British National Formulary sub-chapters. Incidence rate ratios (IRRs) of OH were compared across a 90-day pre-exposure period and three post-prescription risk periods (day 1–28; day 29–56; day 57+) against periods outside these risk periods (reference period).

Results

Of 2,597,077 eligible individuals, 41,005 (1.6%) had a first-recorded OH diagnosis; 51% were women. Increased OH risk was associated with initiation of all cardiovascular system (CVS), central nervous system (CNS), and urinary retention drugs, and some anti-diabetic drugs. Across most classes, OH risk peaked sharply after initiation (day 1–28) and then declined. Drugs causing OH via sympathetic inhibition, diuresis, or mixed mechanisms had the highest risk (eg, uroselective alpha-blockers: IRR 4.94, 95% CI 4.38–5.58) during day 1–28. Vasodilation-causing drugs had lower risk (eg, calcium-channel blockers: IRR 1.88, 95% CI 1.58–2.23). An increased OH risk was observed before drug initiation when the indication was an independent OH cause (eg, loop diuretics for heart failure: IRR 1.74, 95% CI 1.59–1.91).

Conclusion

This study provides insights into OH risk across 17 drug classes in older adults, considering timing, drug system, and mechanism. Consistent trends show that OH risk peaked in the first 4 weeks post-initiation and reduced thereafter. Drugs causing OH through sympathetic inhibition, diuresis, or mixed mechanisms carried the greatest risk. Elevated OH risk in the pre-exposure period may suggest underlying indications that independently cause OH. These findings have important implications for primary care.

Introduction

Orthostatic hypotension (OH) is defined as a reduction in systolic blood pressure (BP) of ≥20 mmHg or diastolic BP of ≥10 mmHg within 3 minutes of assuming an upright posture,¹ and affects around one-third of older adults.^2,3 Drug-induced OH is the most common form of OH.⁴ The resulting cerebral and cardiac hypoperfusion is associated with an increased risk of falls and fractures,^5,6 myocardial infarction (MI),^7–9 stroke,^2,7 atrial fibrillation,⁷ dementia,^5,6 and mortality.^5,7,8 Over 250 medications are associated with OH.¹⁰ The main mechanisms by which drugs cause OH are sympathetic inhibition, vasodilation, and diuresis,^1,4,11 disrupting the coordinated autonomic and cardiovascular response which normally regulates postural changes in the body. Older adults are additionally at increased risk of OH owing to the higher prevalence of cardiovascular and neurological disorders, which are independently associated with OH.¹⁰ Better recognition and management of drug-induced OH could improve several outcomes for older adults. As polypharmacy rises worldwide in an ageing population,¹² drug-induced OH is of increasing concern.¹³

To date, there are limited studies on drug-induced OH. Our previous systematic review and meta-analysis of randomised controlled trials (RCTs) found that a broad range of drugs causing sympathetic inhibition (eg, alpha-blockers, beta-blockers, antidepressants, antipsychotics) were most likely to cause OH.¹¹ However, these results were limited since RCTs focus on efficacy rather than adverse effects¹⁴ and often exclude older adults.¹⁵ A further narrative review identified a similar range of cardiovascular and psychoactive drugs as potential culprits.⁴ However, the included studies were mostly small cross-sectional studies with limitations. Previous studies have been unable to examine temporal trends in drug-induced OH and have been limited by sample size.

The objective of this study was to examine the risk of drug-induced OH associated with all potential medication classes in a large real-world representative population of adults aged ≥60 years in UK primary care.

Materials and Methods

These self-controlled case series (SCCS) analyses followed the Strengthening the Reporting of Observational in Epidemiology (STROBE) reporting guideline. This study was approved by the Scientific Research Committee (SRC) (SRC reference number: 20SRC055).

Data Source

We used UK general practice data from anonymised routine electronic healthcare records in IQVIA medical research data (IMRD), a database which includes over 18 million patients from over 750 general practices in UK primary care.¹⁶ This version of IMRD incorporates data from The Health Improvement Network (THIN), a Cegedim database. Reference made to THIN is intended to be descriptive of the data asset licensed by IQVIA. IMRD is broadly representative of age, sex, practice size, geographical distribution, and sociodemographic characteristics.¹⁷

In the UK, access to healthcare is free at the point of use and individuals typically register with a general practitioner (GP) in their local area. Approximately 98% of the UK population is registered with a GP¹⁸ and over 90% of healthcare contacts are in this setting.¹⁹ GPs are usually the first point of call for a problem, and they also manage chronic conditions (eg, diabetes, hypertension).²⁰

During routine healthcare, GPs electronically record patient information such as symptoms, diagnoses, and prescriptions. Prescription data are coded automatically when entered and are essentially complete in IMRD (although excluding a small proportion of drugs purchased over the counter or prescribed in hospital).

We included data from adults aged ≥60 years between 1 January 2000 and 31 December 2018 (the latest date for which data were available) registered with a GP practice contributing data to IMRD at acceptable quality and mortality reporting levels.^21,22

The Self-Controlled Case Series (SCCS)

The SCCS method is used to study the association between a time-varying exposure and an outcome event.²³ It examines when exposure occurred in individuals experiencing the outcome. Thus, it is an alternative method to cohort and case–control studies. This method was originally developed to investigate associations between vaccination and acute adverse events,^23,24 and has since been used widely in pharmacoepidemiology,²³ primarily examining the acute adverse effects of drugs.^25–27

The SCCS analysis uses only cases and examines an observation period within individuals. Therefore, the risk of the outcome is compared during periods of drug exposure (exposure risk period) compared to periods outside these risk windows (reference period) to determine whether the risk of the event is greater during drug exposure. The SCCS only requires cases, and individuals act as their own control; it compares risk within individuals rather than between individuals. This eliminates fixed confounding (eg, sex, ethnicity, social deprivation, and genetic factors).^28,29 It offers a method of assessing adverse drug events in real-world complex older adults with polypharmacy and comorbidity, who would otherwise have been excluded from RCTs.

Exposure

We identified medications with the potential to cause OH from our previous systematic review¹¹ and a recent narrative review on drug-induced OH.⁴ We identified drug classes according to British National Formulary (BNF) sub-chapters based on therapeutic action. The BNF is commonly used by prescribers in the UK³⁰ and drug codes are readily available in our data source, IMRD. We excluded drug classes that were not relevant to primary care (ie, those initiated in specialist settings or prescribed in <100 people) and excluded drugs that were taken for intermittent medical conditions, “pro re nata” (PRN). Drug code lists were developed using established methods.³¹ A prescription was considered eligible if it was the first prescribed prescription for the patient and it occurred during the study observation period. A prescription episode was deemed to be continuous if there was a gap of ≤90 days between consecutive prescriptions. This method has been used previously.³² The end date of drug exposure was defined as the last prescription date of the continuous episode plus 30 days, to account for the final prescription duration.²⁹

Outcome

The outcome was defined as the first recorded OH diagnosis that occurred during the observation period. Incident OH was identified through a specific code list with high specificity, which has been described previously.³³ Our previous study³³ examined the incidence of recorded cases of OH in IMRD, which was lower than expected compared to studies in screened older community populations.³ However, the SCCS analysis does not require complete recording of the outcome, as we are examining periods within individual cases.

Study Population and Follow-Up

Our study population included patients for whom the observation period for each individual started on the latest of 1 January 2000, their registration date, the date of recording of data at acceptable quality and mortality reporting levels, or the year in which they had their 60th birthday plus a washout period of 12 months of continuous registration in IMRD. The washout period ensured that any historic prescriptions or diagnoses that were transferred from previous medical records were not included as an incident prescription.

Eligible patients were those with both the exposure of interest (prescription) and the outcome of interest (incident OH) during the study period (1 January 2000 to 31 December 2018). The end of the observation period for each individual was the earliest of 31 December 2018, the last data collection date for the practice, transfer out of the practice, death, or their 100th birthday.

Statistical Analysis

The SCCS makes use of conditional Poisson regression to calculate the incident rate ratios (IRRs) comparing the risk of OH during four exposure risk periods (90-days pre-prescription, day 1–28, day 29–56, and day 57+ post-prescription) with periods outside these risk windows (reference period) (Figure 1). These risk periods were agreed based on previous studies³⁴ and defined in conjunction with an advisory group including GPs and clinical pharmacologists to reflect pragmatic prescribing practices. We adjusted for age in 10-year age bands: 60–69, 70–79, 80–89, and 90–99 years.

Figure 1.

Illustration of self-controlled case series study design. The figure illustrates a single individual with drug exposure during their observation period, demonstrating the different exposure risk periods.

We included a 90-day pre-exposure risk period to address a key SCCS assumption that occurrence of an event (OH) should not appreciably affect subsequent exposures (prescriptions). In this case, an OH diagnosis may temporarily reduce the likelihood of a clinician prescribing a medication and thus exaggerate the relative rate of OH events occurring during exposed versus unexposed periods.²⁴

Results

We found 2,597,077 people aged ≥60 years who were eligible between 1 January 2000 and 31 December 2018. Among these, 41,005 (1.6%) had a first recorded OH diagnosis during the study period: 19,979 (49%) men and 21,026 (51%) women.

We examined drugs from four BNF chapters based on physiological systems (cardiovascular system [CVS], central nervous system [CNS], urinary tract disorders, and endocrine system), among which we analysed 17 drug classes, including diuretics, beta-adrenoreceptor-blocking drugs, antihypertensives, nitrates, calcium-channel blockers (CCBs), hypnotics, oral antipsychotics, antidepressants, urinary retention drugs, erectile dysfunction drugs, antidiabetic drugs, and neuropathic pain agents. The results per analysis are presented in Table 1.

Table 1.

Number of People Included per Analysis

British National Formulary Section	No. of People	90-Day Pre-Exposure Period IRR (95% CI)	Day 1–28 IRR (95% CI)	Day 29–56 IRR (95% CI)	Day 57 to End of Prescribing Episode IRR (95% CI)
Cardiovascular System
Alpha-adrenoreceptor-blocking drugs	5912	1.63 (1.40–1.90)	4.88 (4.18–5.72)	2.40 (1.91–3.00)	2.43 (2.24–2.64)
Beta-adrenoreceptor-blocking drugs	10,051	1.86 (1.57–2.21)	3.70 (3.00–4.57)	2.26 (1.71–2.97)	1.55 (1.42–1.70)
Potassium-sparing diuretics and aldosterone antagonists	4445	1.96 (1.71–2.25)	3.56 (2.97–4.27)	3.38 (2.73–4.05)	2.07 (1.87–2.29)
Thiazides and related diuretics	7937	1.73 (1.51–1.96)	3.21 (2.72–3.79)	1.91 (1.54–2.38)	1.64 (1.52–1.77)
Loop diuretics	12,123	1.74 (1.59–1.91)	2.80 (2.47–3.18)	2.01 (1.73–2.35)	1.73 (1.61–1.84)
Renin–angiotensin system drugs	14,917	1.82 (1.66–2.00)	3.18 (2.81–3.61)	2.48 (2.15–2.86)	1.94 (1.85–2.04)
Nitrates	6437	1.54 (1.33–1.77)	2.03 (1.64–2.49)	1.79 (1.42–2.25)	1.92 (1.68–2.19)
Calcium-channel blockers		1.84 (1.67–2.03)	1.88 (1.58–2.23)	1.86 (1.56–2.22)	1.30 (1.22–1.39)
Central Nervous System
Selective serotonin reuptake inhibitors	8899	2.10 (1.89–2.29)	4.22 (3.76–4.74)	2.52 (2.15–2,94)	1.62 (1.50–1.77)
Hypnotics	6531	1.80 (1.60–2.02)	1.99 (1.64–2.41)	1.46 (1.12–1.86)	1.34 (1.14–1.57)
Tricyclic and related antidepressant drugs	8313	1.11 (0.97–1.27)	2.12 (1.79–2.50)	1.39 (1.11–1.73)	1.27 (1.13–1.43)
Antipsychotics	2772	2.05 (1.76–2.39)	1.79 (1.35–2.37)	1.17 (0.81–1.71)	1.27 (1.00–1.49)
Other antidepressant drugs	4656	1.99 (1.75–2.26)	2.17 (1.76–2.68)	1.23 (0.92–1.64)	0.91 (0.81 −1.03)
Genitourinary Disorders
Drugs used for urinary retention	9456	1.48 (1.30–1.67)	4.94 (4.38–5.58)	2.95 (2.52–3.46)	2.36 (2.21–2.52)
Endocrine System
Sulfonylureas	2209	1.50 (1.19–1.89)	1.81 (1.25–2.62)	1.57 (1.04–2.35)	1.16 (1.03–1.32)
Biguanides	3663	1.17 (0.95–1.45)	1.86 (1.39–2.50)	1.93 (1.44–2.59)	1.19 (1.08–1.31)
Other antidiabetic drugs	1956	1.16 (0.88–1.53)	1.05 (0.63–1.76)	1.45 (0.93–2.26)	1.12 (0.97–1.38)

Notes: Data are shown as incidence rate ratios (IRRs) and 95% confidence intervals (CI) for orthostatic hypotension associated with drug classes. Pronounced effects are observed during day 1–28 of drug exposure. Statistically significant results are highlighted in bold.

Cardiovascular System Drugs

Among CVS drugs, we examined diuretics, beta-adrenoreceptor blockers, alpha-adrenoreceptor blockers, renin–angiotensin system drugs, and CCBs, including 5500–15,000 people in the analyses. All CVS drug exposures were associated with OH. We observed the highest increased risk of OH during the day 1–28 risk period, which steadily reduced from day 29+ in all classes. The highest risk of OH was observed with drugs causing sympathetic inhibition in the day 1–28 risk period (eg, alpha-blockers: IRR 4.88, 95% CI 4.18–5.72), followed by drugs causing diuresis (eg, potassium-sparing diuretics: IRR 3.56, 95% CI 2.97–4.27), then drugs causing vasodilation (eg, CCBs: IRR 1.88, 95% CI 1.58–2.23). For all classes we observed an increased risk of OH in the 90-day period leading up to drug treatment, ranging from IRR 1.54 (95% CI 1.33–1.77) in those treated with nitrates to IRR 1.96 (95% CI 1.71–2.25) in those treated with potassium-sparing diuretics and aldosterone antagonists (Table 1).

Central Nervous System Drugs

Among CNS drugs, we examined antidepressants, hypnotics, and oral antipsychotics, including 2000–9000 people in the analyses. All CNS drug exposures were associated with OH. We observed the highest increased risk of OH during the day 1–28 risk period, which steadily reduced from day 29+ in all classes except for oral antipsychotics. Overall, the risk of OH was highest with selective serotonin reuptake inhibitor (SSRI) antidepressants during days 1–28 (IRR 4.22, 95% CI 3.76–4.74). The risk of OH was similar for all other drug classes, approximately a 1.5–2-fold increased risk of OH in the day 1–28 risk period. For most drugs we also observed a substantial elevated risk in the 90 days leading up to treatment (Table 1). Further details on antidepressants can be found in our previous publication.³⁴

Genitourinary Disorders

We examined medications used for urinary retention (eg, alfuzosin, tamsulosin), including over 9400 people in the analysis. We observed an increased risk of OH in all exposure risk periods; the risk of OH was highest during the day 1–28 period (IRR 4.94, 95% CI 4.38–5.58). Again, we observed an elevated risk (IRR 1.48, 95% CI 1.30–1.67) in the 90 days leading up to initiation of treatment (Table 1).

Endocrine System

Among endocrine system drugs, we examined oral antidiabetic drug classes: sulfonylureas (eg, gliclazide), biguanides (eg, metformin), and “other antidiabetic drugs” (which included SGLT-2 inhibitors, DPP-4 inhibitors, and pioglitazones). We found an increased risk of OH with sulfonylureas and biguanides, but no increased risk with the “other antidiabetic drugs” class. For sulfonylureas, the risk was greatest in the day 1–28 risk period (IRR 1.81, 95% CI 1.25–2.62), whereas for biguanides (eg, metformin), the risk of OH was elevated in both the first and second risk periods, between days 1 and 56 (Table 1).

Overall Trends in Timing of Drug-Induced OH

Across all drug classes – CVS, CNS, genitourinary disorders, and the endocrine system – in drugs associated with an increased risk of OH, we observed a trend where the risk peaked during the day 1–28 risk period and steadily reduced in the following risk periods, day 29–56 and day 57+. This was consistent across drug classes, regardless of the underlying mechanism of action (Figures 2 and 3).

Figure 2.

Graphical representation of incidence rate ratios (IRRs) for orthostatic hypotension (OH) associated with drug classes according to time, to illustrate trends in timing of drug-induced OH. Risk of OH peaks during day 1–28 for most drug classes.

Figure 3.

Graphical representation of incidence rate ratios (IRRs) for orthostatic hypotension (OH) associated with endocrine and genitourinary drug classes according to time, to illustrate trends in timing of drug-induced OH. Risk of OH peaks in day 1–28 for most drug classes.

Overall Trends by Drug Mechanism

We divided the drug classes within each system according to the primary mechanism of action underlying OH (Table 1). The mechanisms of action include diuretics, sympathetic inhibitors, vasodilators,^1,11 and mixed (where there are multiple coexistent mechanisms contributing to OH or where the primary mechanism is unclear).⁴ Overall trends indicate that drugs causing OH through sympathetic inhibition, diuresis, or mixed mechanisms are associated with the highest risk, demonstrating a 2–5-fold increased risk immediately after initiation of treatment (day 1–28 period). In contrast, we found that drugs causing OH through vasodilation either were not associated with OH or a had moderately increased risk, up to 2-fold, in the day 1–28 risk period. The mechanisms related to oral antidiabetics and OH are unclear,³⁵ other than for SGLT-2 inhibitors, which probably cause OH through diuresis.³⁵ SGLT-2 inhibitors were among the “other antidiabetic drugs” class in this study, so could not be differentiated.

Overall Trends in Pre-Exposure Period

There was a consistent trend towards an increased risk of OH in the 90-day pre-exposure risk period for most drug classes (Figures 2 and 3). This suggests that the main indication for the prescription may be an independent cause of OH. All CVS drugs had an approximately doubled risk of OH in the period immediately before the drug initiation (90-day pre-exposure period (eg, beta-adrenoreceptor-blocking drugs: IRR 1.86, 95% CI 1.57–2.21). Similarly, among CNS drugs, SSRI antidepressants, “other antidepressant drugs”, antipsychotics, and hypnotics were associated with a doubled risk before drug initiation (eg, SSRIs: IRR 2.10, 95% CI 1.89–2.29). Among the oral anti-diabetic drugs, sulfonylureas were associated with an increased risk of OH in the 90-day pre-exposure risk period (IRR 1.50, 95% CI 1.19–1.89).

Discussion

Summary

Our SCCS study spanning 17 widely prescribed drug classes covering the CVS, CNS, genitourinary system, and endocrine system, including 1500–15,000 people aged ≥60 years per analysis (a real-world population that might have otherwise been excluded from RCTs) provides important insights into drug-induced OH. We found an increased risk of OH associated with acute exposure to all CVS and CNS drugs, drugs used for urinary retention, and some oral antidiabetics. We found important overarching trends regardless of drug class. This includes a striking risk in drugs causing OH via sympathetic inhibition, diuresis, or mixed mechanisms, with a 4–5-fold increased risk of OH for alpha-blockers (eg, prazosin), SSRI antidepressants (eg, sertraline), and drugs for urinary retention (eg, tamsulosin) in the early day 1–28 risk period. In contrast, drugs causing OH through vasodilation (eg, CCBs) were associated with a moderate risk of OH. We found consistent trends in timing whereby the risk of OH peaked soon after initiation (day 1–28) and subsequently reduced with time across all drug classes. In addition, we found a consistent increased risk of OH in the 90-day pre-exposure risk period for all drugs, which suggested that the underlying condition/illness was an independent cause of OH.

Strengths and Limitations

This is the first study to provide a data-driven overview of all potential drug classes associated with OH. We used routine data from a real-world older population including over 41,000 people aged ≥60 years, including some people who would otherwise have been excluded from RCT studies. Important overarching principles on drug-induced OH have emerged from this study across all drug classes that will help to guide prescribing and risk management of drug-induced OH. The main strengths of this study include the large sample sizes in each drug class for the analysis, and the SCCS analysis, which offered an invaluable method to closely examine the risk of OH during specific periods of exposure, in an older population, and accounting for all time-invariant confounding.

However, there were limitations. We examined drugs by class and did not have sufficiently large samples to investigate drugs individually within classes or to compare doses and half-lives. We acknowledge that the issue of a prescription in primary care may not always indicate medication “use”, as there may be a proportion of patients who were prescribed the drug but did not take the medication. Thus, our estimates may slightly underestimate the effect. On the other hand, there may have been a small number of prescriptions prescribed in hospital (eg, psychiatry clinics) that would not have been captured in primary care records, although this would be unusual for the majority of our included drugs. Furthermore, we only examined cases of recorded OH in primary care. There will be a proportion of people with “masked” or asymptomatic OH, who will not have been included in this study. While the SCCS method accounts for fixed confounding, we could not control for the development of new conditions, or changes in the severity of the underlying condition, that were independently associated with OH. This was mitigated somewhat by the use of short risk periods. Furthermore, there may have been some comorbidities or concomitant medications independently associated with OH (eg, Parkinson’s disease) that we were not able explore in the analysis, which may have led to residual confounding. We acknowledge that some drug classes also had smaller cohorts (eg, oral antipsychotics) and therefore limited statistical power and wider confidence intervals. Furthermore, direct comparison of IRRs across subclasses with different sample sizes has its limitations.

There may be some potential reporting bias related to patients being more likely to report and attribute symptoms of OH to a drug when it is initiated, rather than if symptoms emerge later with drug exposure. However, we found that the acute trends observed were consistent across most of the 19 drug classes. We also used narrow windows of drug exposure to increase our confidence that the effects observed in this study were related to drug initiation.

Comparison with Other Studies

Our finding that the risk of drug-induced OH peaks in the first 28 days of exposure and reduces thereafter, regardless of drug class, is novel. Few studies have examined the timing of drug-induced OH. This is reminiscent of the “first-dose phenomenon” first documented for prazosin (an alpha-adrenoreceptor blocker) by Bendall et al in 1975.³⁶ This phenomenon was characterised by severe symptomatic OH that occurred within 90 minutes after a first dose. It has not been widely documented for other drug classes; however, additional treatment with other antihypertensive agents (such as beta-blockers) results in an amplified effect.³⁷ A more recent SCCS study on tamsulosin examining temporal risk windows for hypotension found that the first 8 weeks after initiation was the highest risk period.³⁸ The manifestation of drug-induced OH may emerge more rapidly in older people owing to differences in pharmacokinetics.³⁹ Metabolic changes related to ageing (eg, changes in liver mass and renal function) could affect the metabolism and elimination of the drug, resulting in higher circulating concentrations of active substance.^39,40 Pharmacodynamic sensitivity may also increase in older people.³⁹ The only exceptions in our study, where the risk of OH did not peak during day 1–28, were oral antipsychotics and “other antidiabetics”. Regarding antipsychotics, some may have been initiated in secondary care, resulting in a delay in recording in the GP records, which may explain why the risk of OH was greatest during the 90-day pre-exposure period. The “other antidiabetics” comprised a heterogeneous group including SGLT-2 inhibitors, DPP-4 inhibitors, and pioglitazones (which could not be differentiated), which may limit our interpretation of this analysis.

Our study found that drugs causing OH primarily via sympathetic inhibition, diuresis, or mixed mechanisms were associated with a greater risk of OH than drugs causing vasodilation, which aligns with the literature.^1,11 We found that alpha-blockers (non-selective and uroselective) and SSRI antidepressants were associated with the greatest risk of drug-induced OH, among all classes. While most studies have identified alpha-blockers as high risk,^4,11,38 the evidence for SSRIs is mixed. We have explored SSRIs in more detail in a separate study.³⁴ Our previous systematic review of RCTs found that SSRIs were not associated with OH.¹¹ However, this evidence was based on trials in a younger population (mean age 57 vs 71 years in this study). More consistent with our results, Briggs et al reported a doubled risk of OH with SSRIs (odds ratio 2.11, 95% CI 1.25–3.57) in a cross-sectional study.⁴¹ The causal pathway, however, is unclear as SSRIs have multiple mechanisms of action associated with OH, including sympathetic inhibition and vasodilation. These mechanisms could also be potentiated in an older cohort.⁴²

Finally, we found an increased background risk of OH during the 90-day pre-exposure period for all drugs where the condition being treated was an independent risk factor for OH itself. These included drugs prescribed for cardiac conditions, hypertension,^43–45 neurodegenerative disorders causing dysautonomia (eg, urinary retention), neuropsychiatric conditions (eg, depression, anxiety),⁴ and diabetes.^43,46 Tricyclic antidepressants (TCAs) were not associated with an increased OH risk at baseline in our study. In UK general practice, TCAs are now more commonly prescribed in older adults for insomnia and pain,⁴⁷ rather than for depression – conditions that have not been associated with OH in the literature.⁴³ Biguanides (metformin) was also not associated with an increased risk of OH at baseline. Although diabetes is a risk factor for OH, metformin is commonly used as first-line therapy,⁴⁸ early in the trajectory of diabetes, when autonomic dysfunction and OH may not have developed.

Conclusions

Drug-induced OH is complex to manage in older adults and often overlooked. Our study has identified important overarching principles to help manage drug-induced OH in practice, regardless of drug class. Older adults are particularly vulnerable to drug-induced OH for several reasons, including changes in physiology due to ageing, a high prevalence of neurodegenerative conditions (with dysautonomia), CVD, and polypharmacy. Clinicians should consider preventive advice (eg, increased hydration, reduced alcohol consumption, fall-prevention measures) and/or monitoring for symptoms (eg, dizziness on standing) or measurement of postural BP during the first month of initiating a drug, where the risk of drug-induced OH is greatest. For most drug classes, this risk decreases after day 28. It is also important to consider the mechanism of action; drugs causing sympathetic inhibition should be prescribed with caution in older people at high risk of OH, and vasodilators may be considered in preference. Finally, it is important to consider the patient’s baseline risk of OH and drug indication. For people at high risk owing to comorbid conditions (eg, hypertension, diabetes) or polypharmacy (on multiple medications that may lead to increased OH burden), clinicians should consider measuring baseline postural BP prior to starting a new drug. Our overview table (Table 1) provides a summary across a range of widely prescribed drug classes that can be used as a tool to guide decision making. Future research should explore potential synergistic effects in people with polypharmacy and the development of a dedicated management tool for drug-induced OH.

Funding Statement

This work was funded by the Vivensa Foundation. CB was funded by a Research Training Fellowship (RTF1906\131).

Ethics Approval

IMRD data collection, to provide anonymous patient data to the researchers, was approved by the National Health Service South-East Multicentre Research Ethics Committee in 2003. This study was approved by the Scientific Research Committee (SRC) on 24 August 2020 (SRC approval reference number: 20SRC055).

Disclosure

The authors report no conflicts of interest in this work.