Master of Clinical Pharmacy Candidate
Jeffery Hughes BPharm Grad Dip Pharm MPharm PhD*
Head, School of Pharmacy
Richard Parsons BSc MSc PhD
Ya Ping Lee BPharm PhD
School of Pharmacy, Curtin University of Technology, Perth, Australia and Curtin Health Innovation and Research Institute, Perth, Australia
Received date: 4 October 2009; Accepted date: 25 November 2009
AimThe aim of this study was to examine the risk of cardiovascular diseases among users of both inhaled (ipratropium bromide or tiotropium bromide) and oral (oxybutynin and propantheline, solifenacin, tolterodine) anticholinergics. MethodA retrospective study was undertaken on data obtained from the Food and Drug Administration (FDA) from subjects who had received either an inhaled or oral form of an anticholinergic drug and experienced some side effect during the period from 1988 to 2009. The recorded data included: patient’s age, sex, list of drugs and side effects. Side effect rates for the anticholinergic drugs were compared using univariate (Chi-square) and multivariate (logistic regression) methods. ResultsThe files from the FDA held data for 36 491 different subjects, of whom 2610 (7.15%) experienced a cardiovascular or neurovascular side effect. Subjects were classified as taking the oral (45%) or inhaled (55%) class of the drug, with only 109 subjects (0.3%) taking drugs in both forms. Side effect rates differed between anticholinergic drugs. Stroke and hypertension were significantly more common for subjects taking oral anticholinergic drug compared with tiotropium,while other reported vascular side effects (cardiac ischaemia or arrhythmiascardiac failure, cardiac arrest) tended to be more commonly associated with the use of inhaled anticholingerics. These differences persisted after adjustment for age and gender.Conclusion This observational study of recorded side effects showed that, except for stroke and hypertension, patients who were treated with an inhaled anticholinergic drug appeared to be at higher risk of developing neurovascular or cardiovascular side effects, than those treated with an oral drug.However, physicians should also be aware that oral anticholinergic drugs may have similar adverse impacts on health. Further studies on the association between anticholinergic drugs and cardiovascular and neurovascular side effects are recommended.
anticholinergics, cardiovascular disease, neurovascular, side effects, stroke, tiotropium
How this fits in with quality in primary care
What do we know?
Inhaled anticholinergics, namely tiotropium and ipratropium have been implicated in causing adverse neuro- and cardiovascular events, although the evidence is not consistent. Further, the presence of COPD may also be a contributing factor to such events.
What does this paper add?
Based on adverse drug reactions reported to the Food and Drug Administration (FDA) in the USA, anticholingeric agents appear to differ in their propensity to cause adverse neuro- and cardiovascular effects. Inhaled anticholinergics appear to pose a greater risk overall for vascular (neuro- and cardio-) adverse events compared to oral agents. However, for some vascular adverse effects, in particular stroke and hypertension, oral agents may pose a greater risk. These findings should be borne in mind when prescribing such agents to at risk patients.
Inhaled anticholinergic bronchodilators are widely used in chronic obstructive pulmonary disease (COPD). Ipratropium bromide is a short-acting bronchodilator and its duration of action does not exceed six hours. In contrast, tiotropium bromide is a long-acting bronchodilator whose effect lasts for more than 24 hours. It exerts its effect by blocking M3 muscarinic receptors. About eight million patients with COPD worldwide have been prescribed tiotropium to manage their symptoms since its approval in 2002. It is an effective drug, producing improvement in lung function, dyspnoea and exercise tolerance, as well as reducing both respiratory mortality and exacerbations.
A 4-year UPLIFT trial of tiotropium in patients withCOPDdemonstrated a reduction in cardiovascular side effects associated with tiotropium (such as myocardial infarction and congestive heart failure) compared with a placebo group. Similar results in the published pooled safety analysis of tiotropium demonstrated that cardiovascular adverse events did not occur more frequently among patients receiving tiotropium.
However, other studies demonstrated that patients on anticholinergic drugs were at risk of having cardiovascular and/or neurovascular side effects. Singh et al undertook a systematic review to determine whether inhaled anticholinergics increased the risk of major cardiovascular events. Their review involved 17 randomised controlled trials (RCTs) in which patients had more than 30 days follow-up, involving 14 783 patients of average age 49–68 years. The majority of patients using anticholinergics in the trials were males (56.7 – 99%). Inhaled anticholinergics increased the risk of the composite end point, myocardial infarction (risk ratio 1.53, 95% CI 1.05–2.23) and cardiovascular death (risk ratio 1.80, 95% CI 1.17–2.77), but not stroke or death from all causes. This increased risk was seen in long-term (follow up from 48 weeks to five years) but not short-term (between six and 26 weeks) use. Ipratropium had a risk ratio for major cardiovascular events of 1.70 (95% CI 1.19–2.42) while tiotropium had a risk ratio of 1.43 (95% CI 0.95–2.16).
Oral anticholinergic drugs such as tolterodine are commonly used in the treatment of overactive bladder (OAB) syndrome, which is characterised by urgency (with or without urge incontinence), usually with frequency and nocturia.[5,6] Antimuscarinic drugs prevent involuntary bladder contraction by blocking muscarinic receptors M2 and M3 which are found in the human bladder. The heart contains a high density of M2 muscarinic receptors as well, thus tachycardia can be a negative result of blocking these receptors, due to an imbalance between the sympathetic and parasympathetic effects on the heart.
Anticholinergic drugs are considered to be safe and well tolerated agents despite untoward side effects such as constipation, blurred vision and dry mouth, which result from the blockage of muscarinic receptors.[6,7] Tachycardia was not considered to be a serious side effect of anticholinergic drugs in OAB studies.However, even a small elevation in heart rate can be problematic, as the prevalence of OAB increases with advancing age, with about 17% of the general population who are 40 years or older havingOABsyndrome.Anincreased heart rate in older patients elevates their risk of developing cardiovascular comorbidities such as heart failure.[5,7,8]
This study was undertaken to compare the neurovascular and cardiovascular side effects which have been reported in patients treatedwith either inhaled or oral anticholinergic drugs.
A literature review was performed by searching Medline using terms such as ‘inhaled anticholinergics’, ‘oral anticholinergics’ and ‘cardiovascular disease’, without date limitations. Cardiovascular diseases were categorised in this report as cardiac ischaemia, hypertension, arrhythmia, cardiac failure, cardiac arrest, ECG abnormality and unspecified cardiac events, while neurovascular disease in this study indicated stroke. A request was made to the FDA for all the case reports of side effects associated with the use of anticholinergics between 1988 and 2009. Analyses of this database are presented in this report.
All patients were treated with an oral or inhaled form of an anticholinergic drug. The FDA file listed all side effects which these patients were recorded to have experienced between 1988 and 2009. There was no information in the file concerning subjects who were taking these drugs and had no side effects. Hence the present study aimed to compare the profile of side effects (cardiovascular, neurovascular or other) of the drugs, classified as oral or inhaled, and also to consider them separately.
The drug which each subject (patient) was taking at the time of the side effect was recorded. If a subject was taking more than one anticholinergic drug, they were allocated to a separate ‘mixed’ group. Tables were constructedshowing thedrug takenby the type of side effect, and a Chi-square statistic was obtained to identify if there were any differences in side effect profile for the different drugs. A similar table was constructed after classifying the drugs into oral or inhaled categories. A logistic regression model was developed to identify if any differences between rates of vascular side effects (cardiovascular and neurovascular taken together) persisted after adjustment for age and gender. Finally, a similar logistic regression model was applied only to those subjects having a vascular side effect, to identify any differences between the rates of cardiac and neurological side effects between drugs. Results were presented as percentages of subjects experiencing the side effect (univariate analyses) and odds ratios (OR) and their 95% confidence intervals (multivariate analyses). In all analyses, a P-value 0.05 was considered to indicate a significant association.
The FDA file listed side effects in some detail. These were classified into categories as follows:
• stroke carotid arteriosclerosis, carotid artery disease, carotid artery occlusion, carotid artery stenosis, cerebral infarction, cerebral ischaemia, cerebrovascular accident, transient ischaemic attack
• cardiac ischaemia acute myocardial infarction, myocardial infarction, myocardial ischaemia, angina pectoris, acute coronary syndrome
• arrhythmia arrhythmia, bradycardia, cardiac pacemaker, palpitation, atrialfibrillation, atrial tachycardia, bradyarrhythmia, cardiac flutter, atrioventricular block, sinus arrhythmia, sinus bradycardia, ventricular arrhythmia, ventricular extrasystole, ventricular fibrillation, tachyarrhythmia, tachycardia, ventricular tachycardia, cardioversion
• cardiac failure cardiac failure, ventricular dysfunction, ventricular failure
• cardiac arrest cardio-respiratory arrest, cardiogenic shock
• ECG abnormalities electrocardiogram Q waves, electrocardiogram QRS, electrocardiogram QT, electrocardiogram T waves
• cardiac death sudden cardiac death, accidental death
• unspecified cardiac events cardiac disorder, ventricular hypertrophy.
All side effects listed above other than stroke (‘neurovascular’) were classified as ‘cardiovascular’ side effects for the purpose of analyses.
The FDA provided data on a total 36 491 subjects, 11 296 males with a mean age of 69.8 14.5 years and 21 839 females with a mean age of 67.5 15.6 years, while the gender of 3356 patients was unknown. In 61% of cases the age was missing from the record. However, these cases were divided approximately equally between genders. Table 1 shows the numbers of people taking each anticholinergic drug and experiencing each type of neurovascular and cardiovascular side effect. The column headed ‘Combined’ shows subjectswhowere takingmore than one anticholinergic drug. Of the 1199 subjects in this group, 1192 were taking two drugs, and seven subjectswere taking three.
With the categorisation of cardiovascular and neurovascular side effects as stated in the method, records were classified as having either one, both or neither of these side effects. When tabulated against the type of drug taken, differences in the profile of side effects emerge (Table 2).
Side effect profiles of the different drugs were significantly different. A small number of subjects experienced both neurological and cardiovascular side effects, very few records in the file were associated with propantheline, and the majority of side effects were of a non-cardiovascular and non-neurological type (approximately 93% overall). However, there were differences in the proportions of subjects experiencing the two types of vascular side effect. Table 3 suggests that cardiovascular side effects may be more common for subjects taking ipratropium (11.4%) than other drugs (generally approximately 6%) and neurovascular side effects may be more common for oxybutynin (2.8%) than other drugs (approximately 1% or less).
Note that the ‘Combined’ row in Table 2 includes subjects who took any combination of drugs, whereas in Table 3 it includes only people who took drugs by different routes, so that people who took two oral drugs appear in the ‘Oral’ row in Table 3 but in the ‘Combined’ row of Table 2. The highly significantChisquare statistic for Table 3 indicates a difference in side effect profile between the oral and inhaled drugs.
In order to account for any differences in side effect rate attributable to differences in the age or gender of the subjects, a logistic regression analysis was undertaken, with age, gender and type of drug as independent variables. The outcome variable for the regression analysis was the occurrence of a vascular side effect (cardio- or neurovascular) versus another side effect. The results of this analysis are shown in Table 4. Because the number of vascular side effects amongst people taking anticholinergic drugs by both oral and inhaled routes was only 5/109, this small group was excluded from the regression modelling procedure. This means that the regression is based on 36 382 records rather than the full database of 36 491 records (Table 4).
The analysis shows that the oral drugs had a significantly lower association with vascular side effect than the inhaled drugs, after adjustment for age and gender. Females were less likely than males to experience a vascular side effect. The younger age groups carried a similar risk of vascular event to the older age group (over 78 years), but the large group of records with missing ages corresponded to a much higher risk group. There are a number of possible explanations for this, but the main point is that by including these subjects, their increased risk of vascular events can be taken into account when examining the odds ratio associated with the route of administration. The alternative would be to exclude these subjects from analysis, but this would weaken the overall results as they form such a large group (n = 22 213).
In order to investigate any differences between individual drugs, a second logistic regression model was applied to the data (Table 5). Subjects taking any combination of drugs were excluded from this analysis, leaving a total of 35 267 records.
The data in Table 5 indicate that, after adjustment for age and gender, there remain some significant differences between the various anticholinergic drugs in the chances of a vascular side effect occurring. With the tiotropium group as reference, ipratropium and oxybutynin had significantly greater proportions of reported vascular side effects, while solifenacin use appeared to be associated with a significantly lower risk of these side effects. This analysis demonstrated a significant difference in the risk of a vascular side effect between the two inhaled drugs (tiotropium and ipratropium), with tiotropium appearing to have a much lower risk. The chance of a vascular side effect with the tolterodine group appeared to be not significantly different from that of the tiotropium group.
The analyses above treat all vascular side effects as a single group. In order to identify if the drugs vary in their association with particular side effects, a third series of logistic regression models was applied to the records relating to vascular side effects only. In this way, the differences already noted above were removed, and the analyses could focus on any differences in proportions of cases in each drug group associated with the specific end points. Odds ratios for age and gender are not presented as they are not the primary focus for this research, but the odds ratios for the drugs presented in Table 6 have been adjusted for these variables.
The results show that the use of inhaled anticholinergics significantly increases the risk of developing cardiovascular side effects when compared to the use of oral anticholinergic agents as a group, when adjustment has been made for age and gender (adjusted OR 0.79; 95% CI 0.73–0.86, P < 0.0001). These findings were similar to a systemic review of 17 clinical trials involving 14 783 patients which suggested that inhaled anticholinergics increase the risk of mortality and/or cardiovascular events.[4,9] In comparison to tiotropium, ipratropium appeared more likely to be associated with neuro- and cardiovascular events (AOR 1.78; 95% CI 1.58–2.00). However, when looking at individual adverse effects ipratropium was less likely to be associated with stroke (AOR 0.62; 95% CI 0.40–0.96, P 0.05), but more likely to be associatedwith cardiac arrhythmias (AOR 1.31, 95% CI 1.04–1.64, P 0.05), cardiac arrest (AOR 3.70; 95% CI 2.38–5.75, P<0.0001), ECG abnormalities (AOR 3.37; 95% CI 1.65–7.72) and other unspecified cardiac events (AOR 1.80; 95% CI 1.21–2.68, P = 0.004).
These results differ from those of Singh et al from which suggested that ipratropium was overall less likely to cause major cardiovascular events than tiotropium, however, this may be due to the broader array of cardiovascular events included as end points in this study. A large clinical trial did demonstrate a significant increase in cardiovascular mortality and morbidity in patients with ipratropium compared with a placebo. It also showed a higher incidence of supraventricular tachycardia amongst ipratropium users, which was attributed to the drug’s vagolytic effects.
There is a suggestion that the risk–benefit ratio of inhaled anticholinergic treatment should be assessed for patients according to their cardiovascular tendency. Patients who are at a higher risk of developing cardiovascular diseases and who have mild to moderate symptoms of COPD should not use anticholinergics while, those without any cardiac risks but with severe COPD symptoms are strongly recommended to use these agents.
Oral anticholinergic drugs are widely used to manage OAB in elderly patients who are prone to urge incontinence. Such drugs act in OAB by blocking cholinergic M2 and M3 receptors in bladder smooth muscle. M2 receptors are present in the heart as well and are responsible for slowing heart rate, thus blocking these receptors may lead to tachycardia. OAB patients tend to be elderly and any impact on their heart rate will increase their risk of cardiovascular comorbidities. Because of this, it was decided to undertake a comparison of risk of cardiovascular and neurovascular side effects with inhaled anticholinergic (tiotropium) and oral anticholinergics.
Solifenacin succinate is a more recently developed antimuscarinic drug.[8,12] Solifenacin has a higher selectivity to M3 than to M2 receptors compared with other antimuscarinic drugs such as tolterodine. Solifenacin has shown efficacy in improving symptoms of OAB, and is reported to be well tolerated without any serious cardiovascular side effects.
The analysis comparing solifenacin to tiotropium demonstrated that solifenacin was less likely to be associated with neuro- and cardiovascular events (AOR 0.65; 95% CI 0.54–0.77). However, in the case of each specific side effect, solifenacin was associated with significantly greater risk of hypertension (AOR 2.59; 95% CI 1.53–4.36, P = 0.0004) and ECG abnormality (AOR 6.26; 95% CI 2.39–16.36, P = 0.0002), when compared to tiotropium.
There is a case report of an 81-year-old female who had been using solifenacin 5 mg once daily for about three weeks and developed QT prolongation and torsade de pointes (TDP). Whilst ECG abnormalities were reported with solifenacin, the FDA files lacked critical information such as past medical history, genetic factors and electrolyte status. Without such information one cannot be sure whether solifenacin was indeed the sole cause of these side effects or if there were other factors involved.
The FDA data failed to reveal any significant association between the use of solifenacin and stroke, cardiac ischaemia, arrhythmias, cardiac failure, cardiac arrest or unspecified cardiac events. No heart rate elevation was reported in an open-label, post-marketing surveillance study in which the cardiac safety of solifenacin in 4450 patients with OAB was examined during a 12-week course of treatment. One of the possible explanations for this result is that receptors which are exposed to long-term muscarinic receptor antagonists can be modified, in a similar way to tissue denervation, which causes increased regulation of the receptors. Therefore, the heart becomes less responsive to the heart rate elevating effects of solifenacin. Also, this study demonstrated that solifenacin has no effect on blood pressure during a 12-week course, but no explanation of this resultwas provided. In contrast, hypertension was significantly associated with solifenacin according to the FDA results (P = 0.0004).
Tolterodine is a competitive, non-selective M2/M3 receptor blocker.[5,14] Due to its bladder selectivity, tolterodine has been a well tolerated agent compared with other antimuscarinic drugs. Further, based on randomised, double-blind studies, age does not appear to be a factor in determining the safety profile of tolterodine, as no laboratory (clinical chemistry and haematological parameters) or electrocardiographic changes were reported among older patients who were on tolterodine.[14,16,17]
In comparison to tiotropium, tolterodine appears to have a similar propensity to be associated with both cardiovascular and neurovascular events (AOR 0.94; 95% CI 0.85–1.04, P = 0.2428). However, the analysis of individual side effects demonstrated that tolterodine was more likely to be associated with stroke (AOR 2.13; 95% CI 1.58–2.87, P < 0.001) and hypertension (AOR 4.40; 95% CI 3.21–6.03, P < 0.0001), but less likely to be associated with cardiac ischaemia (AOR 0.54; 95% CI 0.40–0.72, P < 0.0001), arrhythmia (AOR 0.68; 95% CI 0.55–0.84, P = 0.0002) and cardiac failure (AOR 0.51; 95% CI 0.37–0.70, P < 0.0001).
Ventricular arrhythmia, atrial fibrillation, cardiac failure, palpitations, bradycardia, collapse, transient ischaemic attack and hypertension are included in post-marketing reports of neurovascular and cardiovascular side effects of tolterodine. These are reported infrequently and discussed in the manufacturer’s product information but supporting evidence of causality is lacking.
The results of a study in the UK, using a technique of prescription event monitoring, showed that tolterodine at the recommended daily dose is a well tolerated agent in general practice and is associated with infrequent cardiac side effects. It showed that only 0.3% of patients using tolterodine had cardiovascular side effects such as tachycardia or palpitations, atrial fibrillation and chest pain. In contrast, in another study involving 162 healthy participants who were aged 50 or over, tolterodine users experienced a significant increase in heart rate compared with users of darfinacin (a highly selective M3 receptor blocker) and a placebo. Tachycardia is a serious event as it may increase the risk of patients developing congestive heart failure and arrhythmias.
Smoking, hypertension, hyperlipidaemia, low physical activity and obesity are all risk factors which have an important role in causing stroke. The FDA data provided did not have sufficient information for patients on tolterodine to determine whether it was solely the drug which was responsible for patients’ cerebrovascular events (stroke) or whether they were due to other risk factors. Furthermore, tolterodine is metabolised completely by the liver and excreted renally. Thus, patients with hepatic and/or renal impairment will have a higher serum concentration and a longer elimination half-life.[16,19] The FDA data did not contain information on patients’ renal or liver function. Renal and/or hepatic dysfunction could influence patients’ risk of tolterodine related side effects.
Oxybutynin is an anticholinergic drug with some selectivity for M1 and M3 over M2 receptors. A study of 21 patients taking oxybutynin to treat their OAB symptoms suggested that oxybutynin does not cause QTc prolongation or tachycardia at recommended doses even if the patients are at high risk of developing cardiovascular diseases. This may relate to the selectivity of oxybutynin for M3 compared to M2 receptors.
However, higher doses of oxybutynin may cause QT prolongation. A 34-year-old female who ingested an overdose of nearly 100 mg of oxybutynin experienced hypertension and sinus tachycardia with frequent ventricular ectopics and bigeminy.
Based on the FDA data oxybutynin is more likely to be associated with cardiovascular and neurovascular events compared with tiotropium (AOR 1.38; 95% CI 1.16–1.64). This was true in the case of stroke (AOR 3.52; 95% CI 2.34–5.28, P < 0.001), hypertension (AOR 2.99; 95% CI 1.85–4.82, P < 0.0001), cardiac ischaemia (AOR 0.23; 95% CI 0.11–0.46, P < 0.0001) and arrhythmia (AOR 0.52; 95% CI 0.36–0.73, P = 0.0002).
About half of the hypertensive patients were on antihypertensive medications to control their blood pressure, which indicates that they suffered from hypertension prior to the commencement of oxybutynin. It is important to acknowledge that hypertension is a significant risk factor for cardio- and neurovascular events, and that patients’ overall cardiovascular risk needs to be taken into consideration when interpreting the results of this study.
Propantheline bromide is an antimuscarinic agent which has been used in the past in duodenal ulcer treatment to inhibit acid secretion. It acts by blocking both M1 and M2 receptors. M1 receptors are located on the postganglionic neurons in the stomach wall and M2 receptors are located on the parietal cell membranes. It is also used in the treatment of OAB, but its use has been reduced with the introduction of newer agents such as oxybutynin. Analysis of the FDA data failed to demonstrate any significant increase in the risk of adverse cardiovascular or neurovascular events, possibly because of the small number of patients in each side effects group.
In a placebo controlled, double-blind study of 10 duodenal ulcer patients, a significant increase in heart rate (P< 0.05) was seen in those who were treated with propantheline. This was the only evidence found of an association between oral propantheline and risk of cardiovascular events.
This was an observational study of subjects taking an anticholinergic drug and experiencing side effects which were voluntarily reported to the FDA. The analyses aimed to compare the side effects profiles of different drugs. However, the analyses could not compare the absolute rate of side effects as there were no data on the number of subjects who were taking these drugs but did not experience any side effects. Because the data assembled by the FDA are reported on a voluntary basis, its completeness is somewhat uncertain. If under-reporting of side effects is different for the different drugs, this could lead to bias.
Some important information was not available from this database, including: past medical history, and details of patients’ weight, smoking status, cholesterol levels and blood pressure, all of which have a negative influence on the heart and brain. There were no data on the patients’ liver and renal function which might influence the drug safety profile. The age of 22 242 patients (61%) was missing fromthe FDA data. This group of patients was treated as a separate group in the analysis so that some adjustment for their unknown age could be made. With age missing for so many subjects, an alternative approach would have been to exclude any adjustment for age in the analyses. This approach was tried and showed that none of the conclusions in relation to the drugs was changed, although the odds ratios and their confidence intervals varied slightly. This stability in the results with or without adjustment for age gives confidence that the conclusions are valid. Accurate ages for these patients might have made the comparisons between drugs more precise, as both advancing age and male gender influence the risk of developing cardiovascular and cerebrovascular disease.
Inhaled anticholinergic drugs were associated with neurovascular and cardiovascular side effects. In comparing tiotropium with oral anticholinergics, these adverse effects appeared to be less common with tolterodine and solifenacin but were reported more frequently with oxybutynin. Patients who are using the oral anticholinergics should be aware that they may suffer such side effects. We suggest that anticholinergic drugs should be used with caution by patients who are at increased risk of cardiovascular disease, and that their use should be avoided in patients who are suffering severe cardiovascular disease.
A large randomised controlled trial would be required in order to confirm the trends found in this observational study and to determine the true nature of the risk of neuro- and cardiovascular adverse events amongst users of anticholinergic agents, both inhaled and oral.
Not commissioned; externally peer reviewed.