Adherence to medication and drug monitoring in apparent treatment-resistant hypertension

Abstract

Poor drug adherence is one of the main reasons for the failure to achieve treatment targets in hypertensive patients. In patients who receive pharmacological treatment, assessment of drug adherence is of the utmost importance. The aim of this review is to present an update of the methods available to reveal and monitor non-adherence in patients with apparent treatment-resistant hypertension. Methods for monitoring adherence are divided into indirect and direct methods. The indirect methods are mainly based on self-reported adherence and can easily be manipulated by the patient. Directly observed therapy and therapeutic drug monitoring are examples of direct methods. There are limitations and advantages to all of the methods, and because of the patient’s ability to manipulate the outcome of indirect methods, direct methods should be preferred. Therapeutic drug monitoring and directly observed therapy with subsequent ambulatory blood pressure measurement are considered to be reliable methods and should be used more in the routine assessment of patients with apparent treatment-resistant hypertension.

Keywords:

• Blood pressure-lowering treatment
• drug adherence
• resistant hypertension

Introduction

Hypertension is one of the major risk factors for cardiovascular disease and mortality and is present in about 30% of the adult population.[1] Lifestyle changes alone or in combination with pharmacological treatment are often enough to achieve blood pressure (BP) control.[2] Some patients suffer from secondary hypertension and will not achieve target BP until successful treatment of the underlying condition, while the remaining patients with persistently elevated BP are often defined as resistant to pharmacological treatment. Treatment-resistant hypertension (TRH) is defined as office BP > 140/90 mmHg, despite treatment with three or more antihypertensive drugs, including at least one diuretic, in the highest and/or maximally tolerated doses.[3] The prevalence of TRH is assumed to be 3–30%[4] depending on the population and the definition used. Still, the prevalence of true TRH, excluding patients with poor drug adherence, white-coat hypertension and spurious and secondary hypertension, is not known. The variation in prevalence is mainly due to an absence of large, prospective studies aiming to investigate the prevalence of true TRH using forced titration of antihypertensive drugs.[5]

Poor drug adherence, although known for decades,[6,7] is once again actualized with the development and introduction of novel methods for treating TRH [4,8–15]. An example of these new methods is renal denervation, using catheter-based ablation of the sympathetic nerves running in the adventitia of the renal arteries. In the Oslo-RDN study, using directly observed therapy (DOT), and in several studies using therapeutic drug monitoring (TDM), 19–86% of the subjects were classified as non-adherent to pharmacological treatment, including both partial and complete non-adherence.[8–15] TDM has been proven not only to be efficient in unveiling poor drug adherence among hypertensive patients, but also to contribute to increased compliance, at least in clinical trials.[11] The results of the above-mentioned studies [8–15] emphasize the necessity of providing simple methods to investigate and improve poor drug adherence among hypertensive patients who fail to reach target BP.

Methods in the monitoring of drug adherence

There are multiple methods available for assessing drug adherence (Table 1). The methods can be divided into indirect and direct methods, and then further into subgroups of each category.[16,17]

Table 1. Available methods in drug monitoring.

Indirect methods		Direct methods
Self-report	Pill count	Witnessed intake	Therapeutic drug monitoring
Questionnaire	Manual	Directly observed therapy	Blood sample analysis
Diary	Electronic pill box		Urine sample analysis
Interview	Prescription registries

Indirect methods

Several scales and questionnaires are used in the evaluation of patients’ compliance to treatment.[18] The most commonly used questionnaire is the Medication Adherence Questionnaire by Morisky et al.,[18] although there is no consensus on the questionnaire of choice. Medical diaries and patient interviews are other opportunities for investigating drug adherence. When using medical diaries, patients report which medication they are using and when the drug was consumed. The diaries are later reviewed by their physician. The method of interviewing is based on a consultation where the physician retrieves information from the patient regarding their administration of medication.

Besides interviews, counting pills is the most commonly used method in the evaluation of drug adherence.[16] This is an easy method based on manually counting the remaining pills brought to the consultation by the patient.

Methods using electronic registration have been available since 1988.[17] The first such method was developed as an alternative to DOT in the treatment of tuberculosis.[19] In its simplest form, an example is the electronic pill box, which upon opening registers the time of administration of the medication. This technology is constantly evolving. Recently, a device has been developed that only registers when the medication has been ingested. The patient ingests a sensor along with the prescribed medications, and a wearable patch registers the ingestion and synchronizes with a mobile app to simplify everyday monitoring.[20]

An optional method is to use prescription registries, either by prescription fill rates or by assessing renewal rate, offering the physician the opportunity to monitor adherence. In addition, prescription registries can be used to investigate the prevalence of TRH.[21] They can also be used to investigate prescribed medications in a population, with the aim of evaluating whether the physician prescribes antihypertensives according to guidelines.

Direct methods

DOT is used to ensure intake of medication before assessing the treatment effect. In the assessment of hypertensive patients, the method is based on the patient taking the prescribed medications from the original packaging in the correct doses, while being observed by a physician or trained nurse, before ambulatory BP measurement. The patient is often observed for some time to prevent them from spitting out the medications or vomiting, and for safety reasons, as a non-adherent patient may experience severe hypotension when all the medications are taken at once.

TDM is another direct method and it is based on the detection and quantification of antihypertensive medications. The analyses can be conducted with respect to the active antihypertensive drugs, or more easily by detecting metabolites of the drugs in blood or urine samples. Measurements are based on precipitation and solid-phase extraction with subsequent liquid chromatography–mass spectrometry.[22] TDM is widely used in clinical practice regarding a range of different medications, e.g. to monitor serum concentrations of antiepileptics or antidepressants. The analyses for antihypertensives are mainly used in research or forensic medicine, despite being available for clinical use in the USA [11] and several European countries.

Discussion

Patients failing to reach target BP are a great challenge in antihypertensive treatment.[4] A large proportion (19–86%) of these patients do not take their medication as prescribed [8–15] (Table 2). The importance of unveiling poor drug adherence when treating hypertensive patients has been known for decades and was emphasized in the literature as early as 1988.[7,23] Poor drug adherence, and the subsequent incorrect diagnosis of TRH, may have major implications on the patient as well as on society in general. Uncontrolled hypertension will lead to reduced quality of life and possibly fatal consequences, as a result of increased general morbidity. Apparent TRH will also present an economic burden to the healthcare system, as a consequence of extensive and unnecessary medical follow-up, demanding resources from other healthcare services. Currently, there is no consensus regarding the optimal investigation and evaluation of patients with poor drug adherence. There are several methods available, and knowledge about the limitations and advantages of the methods is of utmost importance to the physician (Table 3).

Table 2. Overview of the percentage classified as non-adherent in different studies.

Reference	No. of patients	Methods used	Total percentages classified as non-adherent^a	Population and characteristics
Ceral et al., 2011 (12)	84	TDM, blood sample	65.5%	Population: patients with difficult-to-control hypertension
				Inclusion: BP > 150/95 mmHg, and a minimum of three antihypertensives from different classes
Jung et al., 2013 (10)	76	TDM, urine sample	53.0%	Population: patients with uncontrolled hypertension
				Inclusion: BP > 140/90 mmHg, and a minimum of three antihypertensives from different classes
Strauch et al., 2013 (9)	163	TDM, blood sample	47.0%	Population: outpatients with resistant essential hypertension
				Inclusion: BP > 140/90 mmHg and a minimum of three antihypertensives including a diuretic
Strauch et al., 2013 (9)	176	TDM, blood sample	19.0%	Population: hospitalized patients with resistant essential hypertension
				Inclusion: BP > 140/90 mmHg and a minimum of three antihypertensives including a diuretic
Fadl Elmula et al., 2013, 2014 (13,14)	83	DOT + 24 h ABPM	29.3%	Population: patients with apparent treatment-resistant hypertension evaluated for possible renal denervation
				Inclusion: SBP > 140 mmHg and daytime SABPM > 135 mmHg, and a minimum of three antihypertensives including a diuretic
Brinker et al., 2014 (11)	56	TDM, blood sample	54.0%	Population: patients with treatment-resistant hypertension
				Inclusion: BP > 140/90 mmHg and a minimum of three antihypertensives including a diuretic
Tomaszewski et al., 2014 (15)	208	TDM, urine sample	25.0%	Population: new referrals, follow-up patients and referrals to possible renal denervation with hypertension
				Inclusion: BP > 140/90 mmHg
Florczak et al., 2015 (8)	36	TDM, blood sample	86.1%	Population: patients with primary resistant hypertension
				Inclusion: SBP > 140 mmHg, a minimum of four antihypertensives, and at least one sign of non-adherence^b
Hameed et al., 2015 (24)	50	DOT + 24 h ABPM	50.0%	Population: patients with uncontrolled hypertension evaluated for non-adherence
				Inclusion: reduction of SABP > 5 mmHg using DOT and 24 h ABPM, compared with previous non-DOT 24 h ABPM, and a minimum of three antihypertensives

^aThe total percentage classified as non-adherent includes both partial and complete non-adherence; ^bsigns of non-adherence include tachycardia while using an adequate dose of β-blocker or low potassium plasma levels when taking spironolactone.

TDM, therapeutic drug monitoring; DOT, directly observed therapy; ABPM, ambulatory blood pressure measurement; BP, blood pressure; SBP, systolic blood pressure; SABPM, systolic ambulatory blood pressure measurement; SABP, systolic ambulatory blood pressure.

Table 3. Assessment of the different methods for investigating adherence.

Methods	Advantages	Limitations
Self-report
Questionnaire	Easy to administer; low cost; identify causes of non-adherence	Manipulation by the patient (patients pretend to overestimate their medication intake)
Diary	Easy to administer; low cost	Manipulation by the patient (patients pretend to overestimate their medication intake)
Interview	Easy to administer; low cost	Manipulation by the patient (patients pretend to overestimate their medication intake)
Pill count
Manual	Easy to administer; low cost	Manipulation by the patient (patients does not return untaken medications)
Electronic pill box	Easy to administer; possibility to monitor drug habits (exact time, etc.)	Manipulation by the patient (opening of the box is enough for registration, does not secure actual intake)
Prescription registries	Easy to administer; low cost	Manipulation by the patient (renewal or pick-up is enough for registration, does not secure actual intake)
Witnessed intake
Directly observed therapy	Easy to administer; secure actual intake; important part of the evaluation of patients with apparent treatment-resistant hypertension	Potentially expensive in a long-term perspective; potential “white coat-adherence”
Therapeutic drug monitoring
Blood sample analysis; urine sample analysis	Secure actual intake; detection of altered metabolism of pharmacological treatment	Expensive in a long-term perspective if frequently used; potential “white coat-adherence”; unknown time and dose of taken drug

The patient’s ability to manipulate the results is a major limitation to several of the methods used in the evaluation of drug adherence,[17] and it is well known that patients tend to overestimate their adherence.[25] For this reason, methods based on self-reported adherence and drug consumption are often used in combination with other means to assess drug adherence.[25] Nonetheless, questionnaires have proven useful in separating non-adherent patients from adherent patients who are truly non-responders to treatment.[26] An advantage of the questionnaires, with the possibility of this method being qualitative, is the opportunity to reveal reasons for non-adherence, and this should therefore be considered useful in the follow-up of non-adherent patients.

The above-mentioned opportunity of the patient to manipulate the results is also the most important limitation of all methods regarding the quantification of consumed drugs, i.e. pill counting. Discarding medications, abstaining from bringing the medications to consultation or not consuming them when opening the electronic pill box, as well as filling the prescription at the pharmacy without taking medications, are all examples of apparent compliance, which will severely overestimate drug adherence. It has been shown that 28.3% of prescriptions are never filled at the pharmacy, and regarding antihypertensives, the percentage is 19.5%.[27] One might speculate on whether examples of discarding medications are more relevant in countries with low healthcare fees, for instance as in Western European countries. Low prescription fill rates may be more common in countries where the patients pay for the medications themselves. Nonetheless, pill counting is a simple and inexpensive method for monitoring compliance, given an adequate rate of consultations. The main advantage of the electronic pill box is the possibility of monitoring the exact time of consumption of the drug, which could promote taking medication at correct hours. Currently, the method giving the most reliable estimate of compliance is registration after digestion of the medication. The use of prescription renewal rates and electronic pill boxes is more likely to be relevant in the long-term follow-up of compliance, especially with the introduction and development of electronic prescription registries.

The only method ensuring the administration of correct medication in correct dosage and at the correct hours is DOT, which, in combination with subsequent ambulatory BP measurement, is suitable for confirming the effect of pharmacological treatment. DOT was originally developed for the treatment of patients with tuberculosis,[28] and the differences in the treatment of hypertension and tuberculosis are not to be trivialized. Unlike the limited duration of tuberculosis treatment, antihypertensive treatment is often permanent. There are also ethical aspects regarding the use of DOT in treatment. DOT can easily be justified as part of the treatment of tuberculosis, owing to the fact that it is a highly contagious disease, often with fatal outcomes if left untreated. The same arguments cannot be used for treating hypertension. However, our experiences from the Oslo-RDN study indicate an important role for DOT in the examination and evaluation of apparent TRH. Being able to ensure the effects of prescribed medication is likely to make it easier for the doctor and patient to discuss further management of an elevated BP. The potential BP-lowering effect of DOT in the treatment of hypertension, both in short- and long-term perspectives, requires more research. Still, in our opinion, DOT in combination with ambulatory BP measurement should be part of the routine evaluation of poor drug adherence in patients with apparent TRH.

A final option in the investigation of adherence is TDM. There are two major benefits to using TDM. By detecting drug substances or metabolites in blood or urine samples, adherence can easily be investigated. In addition, by evaluating serum levels of the active substance, assessment of whether the prescribed dosage generates an adequate serum concentration can be made. Thus, the use of TDM can contribute towards revealing altered drug metabolism, e.g. due to CYP polymorphisms. This can be confirmed with further pharmacogenetic investigations. As a consequence of individual pharmacogenetic polymorphisms, the reduced effect of consumed medication may lead to insufficient reduction of BP, while a serum concentration of the medication above therapeutic levels is more likely to induce side-effects, possibly contributing to reduced drug adherence. Combining TDM and pharmacogenetic investigations can aid in optimizing the antihypertensive treatment on an individual level. Thus, the patient is more likely to achieve target BP.

Multiple studies assessing adherence to pharmacological treatment have discussed the use of TDM. Although investigating a moderate number of patients, Brinker et al. showed that 54% of patients diagnosed with TRH did not have detectable values of antihypertensive medication in blood samples or had serum concentrations of antihypertensive medication below reference value.[11] The patients were unaware of the blood sample analysis, and before being confronted with the results of the analysis all the participants denied poor drug adherence. After being confronted with the results, and then given counselling to overcome poor drug adherence, BP was subsequently reduced by 46 ± 10/26 ± 14 mmHg without any changes in prescribed medication. In the drug-adherent group, BP was reduced by 12 ± 17/7 ± 7 mmHg. The proportion of non-adherent patients considered to have TRH or difficult-to-control hypertension found in this study is in line with other recent studies using TDM.[8–10,12,15]

Despite the promising results of TDM in research, there are several aspects that need to be considered before implementation in clinical practice. There is a well-known phenomenon in which patients improve their behaviour during observation, e.g. patients start to take their prescribed medication when they are part of a clinical trial. This is referred to as the “Hawthorne effect”, and often the result is overestimation of the impact of an intervention.[29] An advantage of TDM in monitoring drug adherence is the reduced possibility of manipulation of the results by the patient. A limitation of the method is the lack of opportunity to ensure administration of the medication at the correct hours, since the full effect of treatment depends on regular administration.[17]

As with the use of DOT, ethical challenges apply to TDM as well. In several studies using TDM, the patients were unaware of the blood or urine sample measurements, and in routine clinical practice this may be an unacceptable approach. One option for the physician in the treatment of patients who fail to reach target BP is to obtain consent to perform TDM in future consultations. By not performing TDM at every consultation, an element of surprise is maintained, and the results of the TDM measurements in a long-term perspective are more reliable. There are two other important aspects to consider: first, whether the patients who consent to TDM are already compliant before its introduction; and secondly, how regularly TDM should be performed during routine follow-up.

Another limitation concerning all kinds of intermittent testing, including DOT and TDM, is the phenomenon known as “white-coat adherence”. The patient’s drug adherence often improves before and after a consultation where BP measurements are performed, leaving the impression of an adherent patient. Nevertheless, using DOT and TDM provides further evidence that the medications have the desired antihypertensive effect, and the above-mentioned limitation should not be used to exclude DOT and TDM in the initial investigation of apparent TRH.

Pharmacological non-adherence is a complex challenge and the aim of this review has, so far, been to discuss methods to reveal and monitor poor drug adherence. When treating non-adherent patients, there are also other important aspects to remember, including the patients’ expectations for treatment, reasons for non-adherence and possible endeavours to improve drug adherence. The patient’s autonomy has to be respected, and the choice to accept pharmacological treatment for hypertension is the patient’s own. Despite the major implication of hypertension on future risk of cardiovascular disease, the condition is, in most cases, not characterized by any symptoms at all. The role of the doctor as an educator and health informant is of the paramount importance, and the significance of discussing different aspects of hypertension, including the long-term effects and treatment of hypertension, with the patient, cannot be emphasized enough. Reasons for non-adherence may be multiple and complex, e.g. intolerable side-effects, reduced faith in conventional medicine or misunderstandings related to the dosage and duration of pharmacological treatment. In multicultural societies, challenges related to communication may be enhanced. It is important that the physician acknowledges the presence of non-adherence, and assessments of the patient’s drug adherence must be performed in the initial phase of treatment. In an ideal world, the patient would give honest answers to questions regarding drug consumption. However, some patients do not, and this is the reason why the above-mentioned methods are used and new methods are being developed. At the end of the day, a patient’s health is his or her own responsibility, and the duty of the healthcare system is of limited extent.

Even if there is an effect of being monitored,[30] using one or more of the methods for investigating poor drug adherence does not guarantee the improvement of adherence. The current research in this field is scarce, and properly designed, prospective studies are necessary to determine methods to use in unveiling and monitoring non-adherence. Together with the effective methods discussed above, the relationship between patient and physician should be the foundation for optimal treatment. In addition, education of patients has proven to increase drug adherence,[11] at least in clinical trials, and patients monitoring their own BP with home measurements have been shown to have improved drug adherence.[31] Furthermore, prescribing long-acting, multidrug single-pill medication will simplify treatment and improve compliance.[17]

In conclusion, poor drug adherence is one of the main reasons for failing to achieve target BP. Several methods are available for revealing and monitoring poor drug adherence. TDM and DOT with subsequent ambulatory BP measurement should be used in the routine assessment of patients with apparent TRH.

Acknowledgements

The authors gratefully acknowledge the assistance and expert advice of Ulla Hjørnholm and Vibeke Kjær of the Section for Cardiovascular and Renal Research, and they thank Professor Sverre E. Kjeldsen of the Department of Cardiology, Oslo University Hospital, Ullevaal, for commenting on this paper.

Declaration of interest

ACL has received lecture honoraria from Hemo Sapiens and Merck, Sharpe & Dome. FEMFE has received lecture honoraria from Hemo Sapiens, Medtronic and Amgen. AH has received lecture honoraria from Merck, Sharpe & Dome and Amgen. PAE, SH and JEM declare no conflicts of interest.