Summary
Background
Digoxin is highly potent and efficacious for treatment of heart failure (HF) and/or atrial fibrillation (AF) yet compliance is often poor.
Aims
To examine prevalence rates of non-compliance with digoxin; variations between clinical settings, types of non-compliance and methods of detection; and potential factors influencing non-compliance with digoxin.
Methods
This was a systematic review and meta-analysis of prospective observational studies of non-compliance with digoxin in patients with HF and/or AF, published in English. The studies were identified through these bibliographic databases: MEDLINE, EMBASE, CINAHL, IPA and Cochrane CENTRAL. Subgroup analysis examined the influence of clinical settings, types of non-compliance and methods of detection.
Results
Ten studies met the inclusion criteria, comprising 1841 patients with HF and/or AF. The corresponding prevalence rates of non-compliance for outpatients, after hospital discharge and inpatients were 43.1% (interquartile range [IQR] 29–48%), 25% (95% confidence interval [CI] 12–37%) and 4.5%, respectively. In patients with HF and AF co-morbidities, the prevalence rate of non-compliance with digoxin was 38.7% (IQR 27–46%); the corresponding prevalence rates of overdosing and underdosing were 33.04% (IQR 22–49%) and 33.8% (95% CI 25–42%), respectively. Rates varied depending on the methods of detecting non-compliance. Regularity of prescribed dose, diuretic use, coronary artery bypass, implantable cardioverter-defibrillator, number of office visits and pill boxes demonstrated strong associations with non-compliance with digoxin.
Conclusions
Non-compliance with digoxin is prevalent among patients with HF and/or AF. A better understanding of the factors influencing compliance and improved intervention strategies are necessary to increase digoxin compliance.
Résumé
Contexte
La digoxine est très puissante et efficace contre l’insuffisance cardiaque (IC) et/ou la fibrillation auriculaire (FA), cependant l’observance du traitement est souvent médiocre.
Objectifs
Examiner la prévalence de la non-observance au traitement par digoxine et ses variations en fonction des données cliniques, des types de non-observance et de leur mode de détection ; en identifier d’éventuels facteurs prédictifs.
Méthodes
Synthèse et méta-analyse de dix études observationnelles prospectives. Patients atteints d’IC et/ou de FA traités par digoxine. Nous avons identifié par un examen systématique des bases de données bibliographiques (MEDLINE, EMBASE, CINAHL, IPA et à Cochrane CENTRAL), les études prospectives, publiées en anglais, sur la non-observance à la digoxine chez les patients présentant une IC et/ou une FA. Nous avons réalisé des analyses de sous-groupes pour examiner l’influence des conditions de traitement, des types de non-observance et de méthodes de détection sur la non-observance.
Résultats
L’analyse statistique a été réalisée avec STATA v10. Dix études regroupant un total de 1841 patients répondaient aux critères d’inclusion. La prévalence de la non-observance chez les patients ambulatoires, après hospitalisation, et en cours d’hospitalisation est respectivement de 43,07 % (Intervalle InterQuantile [IIQ] 29,02–48,24 %) ; 25,10 % (IC 95 % 12,20–37,90 %) ; et 4,51 %. Dans le groupe des patients ayant une complication liée à l’IC ou à la FA, la prévalence de la non-observance à la digoxine est de 38,74 % (IIQ 27,59–46,93 %) ; et les prévalences correspondant à celles des sur- et des sous-dosages sont respectivement de 33,04 % (IIQ 22,15–49,42 %) et 33,8 % (95 % CI 25,40–42,20 %). Ces taux varient en fonction des types de non-observance et de leurs méthodes de détection. L’utilisation de doses stables de digoxine, la coprescription de diurétiques, les antécédents de pontage coronaire, d’implantation de défibrillateur, la fréquence du suivi médical et l’utilisation d’un pilulier sont fortement corrélés à la non-compliance.
Conclusions
La prévalence de la non-compliance au traitement par digoxine des patients en IC et/ou FA est importante. Une meilleure compréhension des facteurs qui l’influencent et le développement de stratégies d’intervention efficaces sont nécessaires pour améliorer l’observance des prescriptions de digoxine.
Introduction
Non-compliance with medication regimens is considered as one of the major problems in patients with heart failure (HF) and/or atrial fibrillation (AF); it can lead to intoxication or missed therapeutic effects . Digoxin is the cardiac glycoside most commonly used to treat these patients and has many positive factors: reduced morbidity, reduction of mortality at low dose and low cost with a favourable cost-effectiveness ratio . Currently, digoxin is recommended for the treatment of the following conditions: HF in sinus rhythm with reduced left ventricular ejection fraction and symptoms of HF, despite the use of diuretics, beta-blockers, angiotensin-converting enzyme inhibitors/angiotensin receptor blockers or aldosterone receptor antagonists ; AF with HF or left ventricular dysfunction . Although information on compliance with this drug is regarded as one of the important factors for the effective use of digoxin in these groups of patients , such information is rare. Previous systematic reviews have examined non-compliance with medications in HF patients, but limited attention has been given to the issues of study design, clinical settings, types of non-compliance and methods for detecting non-compliance . This systematic review and meta-analysis addresses these questions by discussing the prevalence rates of non-compliance with digoxin in patients with HF and/or AF and describes potential factors influencing non-compliance with digoxin.
Methods
Data source and study selection
The inclusion criteria for our review were as follows: prospective observational studies that provided sufficient data to calculate the prevalence of non-compliance with digoxin; studies conducted in patients with congestive HF (CHF) and/or AF; studies published in English.
Outcome measures
This review focused on prevalence rates of non-compliance with digoxin as the primary outcome measure and the effect size of the associations between potential factors and non-compliance with digoxin as the secondary outcome measure.
Search strategy
The following bibliographic databases were systematically searched from their inception date to June 2012: MEDLINE, Excerpta Medica Database (EMBASE), Cumulative Index to Nursing and Allied Health Literature (CINAHL), International Pharmaceutical Abstracts (IPA) and Cochrane CENTRAL. The key words ‘digoxin’; ‘(non) compliance’; ‘(non) adherence’; ‘(congestive) heart failure’ and ‘atrial fibrillation’ were used in combination with MeSH terms and EMTREE. The literature retrieval was supplemented by hand searching the reference lists of all identified articles.
Data extraction and manipulation
Data were extracted on study design and prevalence rate of non-compliance with digoxin was recorded on a data extraction form. Other information included methods for detecting non-compliance, study population, study setting, the use of digoxin (i.e. digoxin dose) and the effect size of the associations between influencing factors and non-compliance with digoxin. One study provided the effect size from both univariate and multivariate analyses . The direct association between a potential factor and non-compliance from univariate analysis rather than multivariate analysis was used to determine the effect size of the association between an influencing factor and non-compliance with digoxin. One study reported prevalence rates at different points of time after discharge; we used the prevalence rate detected at the time closest to those in other included studies . In addition, one study reported outcome of non-compliance as a percentage of compliance but did not define the benchmark for classification as compliant ; patients who recorded a compliance percentage greater or equal to 80% were categorized as compliant . Cochrane’s risk of bias and the component approach were adopted for assessing study quality .
Data analysis
Prevalence rates of non-compliance with digoxin were calculated for each study; these were derived by dividing the number of patients who were non-compliant by the total number of patients using digoxin. The effect size ( r ) was calculated from exact P values when r was not reported in the original studies . The effect size, in behavioural science research, can be considered small when r ≤ 0.10, medium when r = 0.25 and large when r ≥ 0.40 .
A statistical test of heterogeneity was performed using the Cochran-Mantel-Haenszel method. The degree of between-study heterogeneity was assessed using χ 2 and I 2 tests to determine whether it would be appropriate to compute a meta-analytical summary estimate . The summary weighted mean differences and 95% confidence intervals (CI) were calculated based on a random-effects model using the DerSimonian-Laird method . Confidence intervals for I 2 were also obtained to convey uncertainty in I 2 . The results across these studies were summarized using the median and interquartile range (IQR) if there was significant heterogeneity between studies. The included studies were divided into subgroups to explore possible reasons for heterogeneity: the settings in which non-compliance had occurred (outpatient, after hospital discharge, or inpatient); the types of non-compliance (overdose or underdose); the methods for detecting non-compliance (pill count, interview, measurement of serum digoxin concentration [SDC] or combination method [interview confirmed by SDC]). If one study reported more than one method for detecting non-compliance for each type of non-compliance, we used data from objective measures (i.e. SDC measurement or combination method) for the analysis. STATA statistical software was used to perform all statistical analyses (version 10.0; StataCorp, College Station, TX, USA).
Results
Study and patient characteristics
Ten studies met the inclusion criteria, comprising a total of 1841 patients with CHF and/or AF ( Fig. 1 ) and covering the period 1973–2010. Eight studies reported prevalence rates of non-compliance with digoxin in outpatients , three studies reported prevalence rates after hospital discharge and one study reported prevalence rates in inpatients . A summary of the study characteristics is presented in Table 1 . The studies were undertaken in North America , Europe , Asia and Africa . Six studies were undertaken before the era of angiotensin-converting enzyme inhibitors/angiotensin receptor blockers and beta-blockers . All studies were conducted in older adults or the elderly. Four studies were conducted in patients with co-morbidities of CHF with AF, including a total of 877 patients . The mean daily digoxin maintenance dose, the number of daily digoxin tablets and the total number of tablets of all medications reported in the included studies are also shown in Table 1 .
| Study (year) [Ref.] | Country | Study population | Types of non-compliance | Method of detecting non-compliance |
|---|---|---|---|---|
| Weintraub (1973) | USA | Outpatients with CHF and/or arrhythmia | Underdosing/overdosing | Interview; interview + SDC |
| Johnston (1978) | Ireland | Outpatients | Underdosing/overdosing | SDC; pill count |
| Johnston (1978) | Ireland | Patients discharged from hospital b | Underdosing/ | SDC |
| Overdosing | SDC; pill count | |||
| Interview + SDC | ||||
| Taggart (1981) | Ireland | Outpatient | Overdosing | SDC; pill count |
| Wiseman (1991) | South Africa | Outpatients with HF | Underdosing | Interview + SDC |
| Kruse (1992) | Germany | Elderly patients with HF and/or AF discharged from hospital | Underdosing | Pill count i |
| Miura (2000) | Japan | Outpatients with CHF and/or AF | Overdosing | SDC |
| Patients discharged from hospital | Overdosing | SDC | ||
| Inpatients with CHF and/or AF | Overdosing | SDC | ||
| Miura (2001) | Japan | Outpatients with CHF and supraventricular tachycardia | Overdosing | SDC |
| Modares-Mosadegh (2001) | Iran | Outpatients with CHF | Underdosing/overdosing | SDC |
| Muzzarelli (2010) | Switzerland | Symptomatic CHF (NYHA > class II) | Underdosing | Interview; SDC; interview + SDC |
| Study (year) [Ref.] | No. of participants (No. of men) | Mean digoxin dose (SD) (mg/d) | Mean No. of daily digoxin tablets (SD) | Mean No. of daily tablets (SD) | ||||
|---|---|---|---|---|---|---|---|---|
| C | NC | C | NC | C | NC | C | NC | |
| Weintraub (1973) | 75 (NA) | 37 (NA) | 0.25 (NA) | 0.24 (NA) | NA | NA | NA | NA |
| Johnston (1978) | 27 (12) | 23 (7) | 0.22 (0.12) | 0.22 (0.12) | 1.3 (0.5) | 1.5 (0.7) | 4.3 a (3.3) | 5.3 a (3.5) |
| Johnston (1978) | 25 c (14) | 18 c,g (5) | NA | NA | 1.26 (0.55) | 1.22 (0.56) | 6.6 (3.7) | 6.6 (3.9) |
| 22 d (NA) | 8 d (NA) | NA | NA | NA | NA | NA | NA | |
| 14 e (NA) | 6 e (NA) | NA | NA | NA | NA | NA | NA | |
| Taggart (1981) | 25 (NA) | 23 (NA) | 0.25 (NA) | 0.25 (NA) | 1,2,4 h (NA) | 1,2,4 h (NA) | NA | NA |
| Wiseman (1991) | 82 (NA) | 55 (NA) | NA | NA | NA | NA | NA | NA |
| Kruse (1992) | 5 j (2) | 4 j (2) | 0.25 (NA) | 0.25 (NA) | 1 (NA) | 1 (NA) | NA | NA |
| Miura (2000) | 253 (NA) | 72 (NA) | 0.21 (0.08) | 0.24 (0.09) | NA | NA | NA | NA |
| 45 f (NA) | 10 f (NA) | NA | NA | NA | NA | NA | NA | |
| 496 (NA) | 22 (NA) | NA | NA | NA | NA | NA | NA | |
| Miura (2001) | 218 (17) | 213 (16) | NA | NA | NA | NA | NA | NA |
| Modares-Mosadegh (2001) | 63 (32) | 56 (27) | 0.19 (NA) | 0.21 (NA) | 0.5 or 1 tablets/day or 5 or 6 tablets/week | 0.5 or 1 tablets/day or 5 or 6 tablets/week | 3.64 (1.14) | 3.78 (1.36) |
| Muzzarelli (2010) | 30 k (24) | 10 k (9) | NA | NA | NA | NA | 13 (5) | 8 (4) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
