Impact of single-pill combination therapy on adherence, blood pressure control, and clinical outcomes: a rapid evidence assessment of recent literature.

OBJECTIVE: The 2018 European Society of Cardiology/European Society of Hypertension Guidelines for the management of arterial hypertension raised the need for evidence to support the use of single-pill combination (SPC) therapy in preference to free-dosed therapy for hypertension. This systematic rapid evidence assessment sought to determine if initiating SPC therapy improves adherence, blood pressure (BP) control and/or cardiovascular outcomes vs. initiation of free-dose combination therapy.
METHODS: Rapid evidence assessment conducted in MEDLINE, EMBASE, and Cochrane Library (1 January 2013-11 January 2019) to identify studies investigating SPC therapy for adults with hypertension. Information on adherence/persistence, BP lowering/goal attainment, and cardiovascular outcomes/events were extracted via two-phase screening process. Studies not focusing on adherence, persistence, or compliance with SPC therapy were excluded. Methodological quality was assessed using appropriate scales.
RESULTS: Of 863 citations, 752 failed to meet inclusion or were duplicates. Twenty-nine studies remained following full-text screening. Just four studies (14%) were randomized controlled studies; 25 (86%) were observational. A range of SPC therapies were studied, with calcium channel blocker/angiotensin receptor blocker combinations most common (11/29 studies). Adherence and persistence were generally higher with SPC vs. free-dose combination therapy; 15 studies (54%) directly compared adherence and four (14%) compared persistence. Patients achieving BP targets ranged from 25 to 89%. Despite all studies investigating patients with hypertension only 16 (55%) reported change in BP. Few studies reported on cardiovascular outcomes. Methodological reporting was often suboptimal.
CONCLUSION: Adherence and/or persistence were generally higher in patients taking antihypertensives as SPC vs. free-dose combination; however, methodological reporting was suboptimal to facilitate comparison. Specifically designed, well reported studies are required to determine if the increased adherence/persistence seen in patients on SPC regimen leads to improved BP control and/or cardiovascular outcomes.

INTRODUCTION

The Global Burden of Disease Study (2017) suggests that high blood pressure (BP) was, and has been for at least the past decade, the world's leading risk factor for attributable disability-adjusted life-years (DALYs), accounting for ∼10.4 million deaths, and attributed∼218 million DALYs [1]. Furthermore, the global prevalence of elevated BP of at least 140 mmHg continues to rise [2,3]. This prevalence of arterial hypertension (AH) is relatively consistent, irrespective of income status [4,5], and identification and management are suboptimal around the world [5-8]. Hypertension also becomes progressively more common with advancing age, with a prevalence higher than 60% in people aged 60 years and older [2,4]. Therefore, as the population ages, it is predicted that the impact of AH will rise even further, if management is not optimized.

According to most international guidelines, including the European Society of Cardiology (ESC)/European Society of Hypertension (ESH) 2018 guidelines for the management of AH in adults, aged at least 18 years, AH is defined as a SBP at least 140 mmHg and/or a DBP at least 90 mmHg [9]. However, in 2017, joint societies in the USA, including the American Heart Association (AHA)/American College of Cardiology (ACC) lowered the level at which ‘hypertension should be defined, to at least 130/80 mmHg [10]. This consequently raised the prevalence of patients with hypertension according to this updated definition [2]. The ESC/ESH noted during the release of their guidelines in 2018 that the majority of patients do not achieve BP less than 140/90 mmHg, and therefore their focus was to improve goal attainment, rather than to lower the definition of BP beyond the traditional AH definition of BP at least 140/90 mmHg [9]. ‘Normal’ BP in patients aged less than 65 years according to both guidelines is, however, defined at a similar level, either as BP 120–129/80–89 mmHg [9] or as BP less than 120/80 mmHg [10]. Studies have demonstrated that attainment of ‘normal’ BP reduces the risk of future cardiovascular events [9-11]. However, goal attainment remains suboptimal throughout the world, even using BP goal of less than 140/90 mmHg [2,5-8].

Five classes of antihypertensive drugs are first-line therapies for the management of AH: angiotensin-converting enzyme inhibitors (ACEis), angiotensin 1 receptor (AT1) blockers (ARBs), calcium channel blockers (CCBs), and diuretics, with beta-blockers being recommended for patients with compelling indications [9,10]. These can be given either individually as a monotherapy; in combination as multiple pills; as a single-pill combination (SPC) [or alternatively termed fixed-dose combination (FDC)] agent that combine two or more drugs in one pill; or as a combination of individual and combined therapies.

Despite some differences in definitions, both the 2018 ESC/ESH and the 2017 ACC/AHA guidelines recommend that most patients with BP at least 140/90 mmHg (or >20/10 mmHg above target) initiate two antihypertensive agents as initial therapy (if tolerated) to reduce BP towards goal [9,10]. ESC/ESH guidelines have recommended the use of initial SPC therapy for patients requiring more than one antihypertensive therapy since 2013 [12], and in 2018 the guidelines were updated to include recommendations for initial SPC therapy for most patients with uncomplicated AH, with the exception of very old (≥80 years) or frailer patients and/or patients with low-risk grade 1 hypertension [9]. The 2017 ACC/AHA guidelines also recommend initiation of two antihypertensive therapies for patients with BP more than 20/10 mmHg above target, but this can be given either as two pills in a free combination or as an SPC therapy [10]. SPC therapy is recognized to significantly improve patient adherence [13,14] and through improved adherence, some evidence suggests SPC therapy can improve BP-lowering and cardiovascular outcomes [14]; conversely, some evidence suggests no difference in BP effects, as when patients are adherent with SPC therapy and free-dose combination therapy, BP reduction is similar [13,15]. Despite some clinical evidence for the use of SPCs in the treatment of AH, a lack of quality evidence and a lack of definitive conclusions have been reported, regarding whether SPC or free-dose combination therapy leads to better efficacy [15]. In addition, there are other considerations around prescribing SPC therapies, including risk of treatment duplication, and the ability to individually titrate different components of the single pill with the incidence of any adverse event [16]. As such, further clarification of the benefits is needed for healthcare professionals. Through a systematic review of the literature, we sought to investigate the impact of initiating patients with AH onto SPC therapy vs. a free combination of therapeutic agents (herein termed free-dose combination therapy) on medication adherence, BP control and cardiovascular outcomes. This comparison was identified as an unmet need during discussion of the ESC/ESH 2018 guidelines [9]. Through a rigorous analysis, we also sought to highlight gaps in the literature to inform future design and development of studies in this area, and to support international guideline recommendations.

METHODS

Study design and data source

A systematic rapid evidence assessment (REA) was conducted within MEDLINE, EMBASE and the Cochrane Library Database between 1 January 2013 and 11 January 2019 (Appendix I). This time-span covers the treatment period since release of the 2013 ESC/ESH guidelines [12], where use of SPC therapy for patients with markedly high baseline BP or at high cardiovascular risk who required more than one antihypertensive agent was made. We sought to reflect changing treatment practices following these recommendations, and to bring literature up-to-date following release of guidelines in late 2017, and 2018 [9,10]. An REA enables a systematic assessment of peer-reviewed literature, in a streamlined and efficient fashion, while adhering to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses protocol guidelines [17]. The REA differs from a traditional systematic literature review (SLR) as it does not include grey or non-English language sources, and provided us with a more focused methodological approach to answer the specific question raised by the ESC/ESH [9].

Search strategy: inclusion and exclusion criteria

The full search strategies for each of the databases are given in Appendix I (Table S1). The inclusion criteria confined the search to studies in adults with AH, where at least one treatment arm was given as SPC therapy (or termed FDC therapy, bi or dual or triple SPC/FDC therapy). Terms specifically included in the search strategy for adherence included: adherence, compliance, persistence; and for cardiovascular outcomes included: venous thromboembolic events, peripheral arterial disease, myocardial infarction, heart failure (ischemic or congestive), chronic kidney disease, transient ischemic attack or stroke, aortic aneurysms, atrial fibrillation, atherogenic index, palpitations/tachycardia, chest pain/angina, percutaneous coronary intervention (PCI), or peripheral vascular intervention (Table S1). SPC therapies considered in the search were based on commonly approved therapeutic agents, according to the US Food and Drug Administration, and including a combination of any first-line antihypertensive agents (i.e. a CCB, ARB, ACEi, diuretic and/or beta-blocker). A full list of specific therapeutic agents added to the search strategy is given in Table S2.

A rigorous, two-stage screening process was carried out: Stage 1 involved screening titles and abstracts of identified publications, and retaining relevant literature for full-text review. The second stage involved full-text screening and retaining literature that met full search criteria. Studies were retained for full text-review that included information on adherence and/or persistence and/or compliance measures. Information on BP measurement and control and/or cardiovascular outcomes (or cardiovascular events) and/or mortality or morbidity (including cardiovascular or non-cardiovascular-related death) were extracted alongside information on adherence and/or persistent measures.

Studies that did not focus on adherence, persistence, or compliance as an assessment, or that did not focus on combination/SPC treatment arms of interest were excluded. Studies that focused on pregnant patients or other nonarterial causes of hypertension were also excluded, as were studies that only discussed monotherapy or that focused on antihypertension combinations with statins (i.e. only one antihypertensive agent), anti-angina agents, or that focused on behavioral outcomes only (including patient education/reminder system/nutrition/exercise/herbal etc.). Studies that focused on clinical outcomes other than BP control, BP goal attainment, or cardiovascular outcomes were not eligible for inclusion, such as primary assessment of blood measures (e.g. lipid profiles or inflammatory markers), drug survival rates, retention rates, discontinuation, or dropout rates. Studies that were preclinical, Phase 1 studies, case reports, or studies of less than 30 patients; a letter to the editor, opinion piece, or review or meta-analysis, were excluded from this REA.

Data extraction

Full publications of retained records were examined in detail for information relating to study design, participants and population, intervention, treatment, and adherence, BP and cardiovascular outcomes, according to Population, Interventions, Comparator, Outcomes, and Study designs (PICOS) methodology. Information on adverse events was captured from retrieved studies, including information on serious adverse events (SAEs). Observational data were summarized using descriptive statistics.

Methodological reporting

Methodological quality assessment was conducted for all randomized controlled trials (RCTs) according to the National Institute of Health and Care Excellence (NICE) Single Technology Appraisal (STA) recommendation [18]. NICE STA involves scoring studies as ‘low’, ‘unclear’, or ‘high’ across 11 domains [18]. Observational studies were assessed using the Newcastle-Ottawa Quality Assessment Scale [19]. The Newcastle-Ottawa scale involves rating elements of study selection, comparability and exposure. A study can be awarded a maximum of one star for each numbered item within the Selection and Outcome categories. A maximum of two stars can be given for Comparability [19]. A ‘good’ quality score required 3 or 4 stars in selection, 1 or 2 stars in comparability, and 2 or 3 stars in outcomes. A ‘fair’ quality score required 2 stars in selection, 1 or 2 stars in comparability, and 2 or 3 stars in outcomes. A ‘poor’ quality score reflected 0 or 1 star(s) in selection, or 0 stars in comparability, or 0 or 1 star(s) in outcomes.

RESULTS

Study selection

Of the 863 citations identified, 756 remained after duplicates were removed. A further 645 articles failed to meet inclusion criteria, based on title and abstract screening. Overall, 111 publications were screened in full, and 82 were excluded due to reasons outlined in Fig. 1. As a result, 29 articles were considered suitable for inclusion [14,20-47].

FIGURE 1

Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram.

Study populations and antihypertensive therapy prescribed

All 29 studies reported on adherence and/or persistence in patients taking SPC therapy for hypertension. Four studies were RCTs (14%) [35,37,38,46]; the remainder (n = 25; 86%) were observational studies [14,20-34,36,39-45,47] (Table S3). Sample sizes varied considerably, from n = 75 to n = 79 958 for SPC therapy and n = 73 to n = 383 269 for free-dose combination therapy. Thirteen studies (45%) were carried out in Europe, seven in Asia (24%), with other studies carried out across different continents (Table S3). The mean age across studies ranged from 46.8 to 71.5 years. Follow-up also varied widely, from 0.5 to 60 months, depending on study. Within the studies, reporting of baseline demographics varied. Of note, only 17 studies (59%) reported baseline BP, although all focused on management of patients with ‘hypertension’ as defined by their inclusion criteria.

A range of SPC therapy combinations were reported in the identified publications (Fig. 2). CCB/ARB combinations were the most common type of SPC therapy used by 11 treatment arms, followed by CCB/ACEi SPC, by 10 treatment arms. Amlodipine (amlo) besylate was the most commonly identified, specific antihypertensive agent in the SPC therapies, regardless of drug class. In the studies identified, amlo was used in treatment combination with all classes of antihypertensive, that is, in dual-SPC with an ARB, ACEi, beta-blocker or diuretic, and in 3-SPC therapy in combination with ARB/hydrochlorothiazide (HCTZ) and ACEi/HCTZ. The most commonly used ACEi was perindopril (per), and valsartan (val) was the most commonly used ARB in an SPC therapy, although some observational studies did not specify the type of ARB or ACEi within the SPC therapy. No combination agent included amiloride or spironolactone.

FIGURE 2

Single-pill combination therapies utilized in studies identified by the rapid evidence assessment of recent literature.

Adherence and persistence

Twenty-eight studies reported data on adherence. One study reported data on persistence only (Table 1). All four RCTs reported adherence, although generally reported adherence for the whole study population rather than for each treatment arm separately, and specific comparison of adherence assessments was only reported by one of the RCT (Webster et al. [46]). Proportion of days covered (PDC) was the most commonly used measure of adherence, by nine of 28 studies (32%) (Table 1). Generally a definition of ‘adherent’ was PDC at least 80% in these studies (Table 1). Other measures for adherence included tablet count (n = 3), patient self-reported adherence (n = 4), medication possession ratio (MPR) (n = 2) and Morisky Medication Adherence Scale (n = 2). Medication discontinuation (>60 days), ingestion, physician-rated and Hill-Bone Compliance were used to measure adherence in one study each; four studies did not report the method used to measure adherence, but noted that adherence was assessed.

TABLE 1

Studies reporting assessments of adherence or persistence with single-pill combination therapy or with free-dose combination therapy

1st Author (year) [REF], study type	Sample size	Treatment (SPC/free)	Treatment details	Follow-up (months)	Measure of adherence /persistence	Adherent/persistent,a n (%)	P value
(a) Measures of adherenceb
Bramlage (2018) [22] – Obs	10 938	SPC	ram/amlo	12	MPR ≥ 80%	5699 (52.1)	<0.001
	60 525	Free	ram + amlo	12	MPR ≥ 80%	19 913 (32.9)
	1413	SPC	cand/amlo	12	MPR ≥ 80%	1195 (84.6)	<0.001
	9082	Free	cand + amlo	12	MPR ≥ 80%	5286 (58.2)
Webster (2018) [46] – RCT	349 (321, adherence calculation)	SPC	tel/amlo/chlo	6	Ingestion ≥ 4 of last 7 days	305 (95.0)	0.82
	351 (336, adherence calculation)	Free	Not specified	6	Ingestion ≥ 4 of last 7 days	318 (94.6)
Degli Esposti (2018) [24] – Obs	3597	SPC	per/amlo	12	PDC ≥ 80%	2597 (72.2)	NR
	20 423	Free	Not specified	12	PDC ≥ 80%	12 969 (63.5)
Lauffenburger (2017) [31] – Obs	79 958	SPC	Not specified	12	PDC ≥ 80%	40 505 (51.3)	NR
	383 269	Mono	Not specified	12	PDC ≥ 80%	161 356 (42.1)
Levi (2016) [32] – Obs	4522	SPC	olm/amlo	6	PDC ≥ 80%	1684 (55.1)	<0.001
	2090	Free	olm + amlo	6	PDC ≥ 80%	291 (15.9)
Machnicki (2015) [34] – Obs	1884	SPC	amlo/val/HCTZ	12	PDC ≥ 80%	1040 (55.2)	<0.0001
	1884	Free	amlo + val + HCTZ	12	PDC ≥ 80%	629 (33.4)
Hsu (2015) [27] – Obs	5725	SPC	ARB/thiazide diuretic	24	MPR ≥ 80%	1747 (30.5)	NR
	1623	Free	ARB + thiazide diuretic	24	MPR ≥ 80%	371 (22.9)
Degli Esposti (2014) [25] – Obs	239	SPC	olm/amlo	6	PDC ≥ 80%	188 (78.7)	NR
	20 769	Free	Not specified	6	PDC ≥ 80%	13 084 (63)
Xie (2014) [47] – Obs	8516	SPC (triple-pill)	olm/amlo/HCTZ or val/amlo/HCTZ	12	PDC ≥ 80%	4710 (55.3)	<0.0001
	7842	Mix (2 pills)	olm/amlo + HCTZ or val/amlo + HCTZ, or olm/HCTZ + amlo, or val/HCTZ + amlo	12	PDC ≥ 80%	3171 (40.4)
	1107	Free (3 pills)	olm + alm + HCTZ or val + alm + HCTZ	12	PDC ≥ 80%	361 (32.6)
Panjabi (2013) [39] – Obs	4005	Mix (2 pills)	amlo/HCTZ + ARB	34.2	PDC ≥ 80%	1033 (25.8)	<0.001
	1335	Free (3 pills)	amlo + HCTZ + ARB	34.6	PDC ≥ 80%	184 (13.8)
	6082	Mix (2 pills)	amlo/HCTZ + ACEi	33.6	PDC ≥ 80%	1448 (23.8)	<0.001
	3041	Free (3 pills)	amlo + HCTZ + ACEi	33.3	PDC ≥ 80%	487 (16.0)
	609	Mix (2 pills)	amlo/HCTZ + BB	34.7	PDC ≥ 80%	173 (28.4)	<0.001
	1218	Free (3 pills)	amlo + HCTZ + BB	35.6	PDC ≥ 80%	166 (13.6)
(b) Measures of persistencec
Bramlage (2018) [22] – Obs	10 938	SPC	ram/amlo	12	Prescription fill	7186 (65.7)	<0.001
	60 525	Free	ram + amlo	12	Prescription fill	29 415 (48.6)
	1413	SPC	cand/amlo	12	Prescription fill	784 (55.5)	<0.001
	9082	Free	cand + amlo	12	Prescription fill	3914 (43.1)
Lauffenburger (2017) [31] – Obs	79 958	SPC	Not specified	12	Prescription refill	40 938 (51.2)	NR
	383 269	Monotherapy	Not specified	12	Prescription refill	164 806 (43)
Simons (2017) [42] – Obs	9340 (700 for persistence calculation)	SPC	amlo/per	42	Prescription refill	350 (50)	NR
	3093 (78 for persistence calculation)	Free	amlo + per	42	Prescription refill	21 (27)
Hsu (2015) [27] – Obs	5725	SPC	ARB/thiazide diuretic	24	Prescription refill	1494 (26.1)	<0.0001
	1623	Free	ARB + thiazide diuretic	24	Prescription refill	317 (19.5)
Machnicki (2015) [34] – Obs	1884	SPC	amlo/val/HCTZ	12	30-day Tx gap	882 (46.8)	<0.0001
	1884	Free	amlo + val + HCTZ	12	30-day Tx gap	445 (23.6)
Xie (2014) [47] – Obs	8516	SPC (triple-pill)	olm/amlo/HCTZ or val/amlo/HCTZ	12	Prescription refill	7541 (88.55)	<0.0001
	7842	Mix (2 pills)	olm/amlo + HCTZ or val/amlo + HCTZ, or olm/HCTZ + amlo, or val/HCTZ + amlo	12	Prescription refill	6363 (81.14)
	1107	Free (3 pills)	olm + alm + HCTZ or val + alm + HCTZ	12	Prescription refill	869 (78.5)

ACEi, angiotensin-converting enzyme inhibitor; amlo, amlodipine; ARB, angiotensin 1 receptor (AT1); BB, beta-blocker; cand, candesartan; CCB, calcium channel blocker; chlo, chlorthalidone; HCTZ, hydrochlorothiazide; medox, medoximil; MPR, medication possession ratio; NR, not reported; Obs, observational; olm, olmesartan; PDC, proportion of days covered; per, perindopril; ram, ramipril; RCT, randomized-controlled trial; SPC, single-pill combination; tel, telmisartan; val, valsartan.

aN (%) adherent or persistent based on study criteria.

bNedogoda et al (2017) (RCT) reported ‘very good’ adherence at 97.6 ± 6.4% (undefined measure) across the whole study population but did not differentiate between SPC and free-combination therapy [38]. Tung et al. (2015) reported mean PDC and is not included in the table [44]. Verma et al. (2017) (Obs) reported median PDC and is not included in the table [14]. Kumagia et al. (2013) (Obs) adherence was self-reported ingestion and is not included in the table [30]. Simonyi et al. (2016) (Obs) compared SPC of two different drug combinations and is not included in the table (no coadministered therapy comparator) [43].

cTung et al. (2015) (Obs) reported data as mean persistence days and is not included in table [44].

For the 15 studies (54%) that directly compared adherence with SPC therapy vs. free-dose combination therapy, adherence was generally higher with SPC therapy than with free-dose combination therapy (Table 1). Only six of 15 studies included statistical comparison of treatment arms. Of these six studies, only Webster et al.[46] did not see a difference in adherence between SPC and free-dose combination therapy (P = 0.82), although this study did use ingestion as a measure of adherence, whereas the other studies used PDC or MPR.

Seven studies (24%) discussed persistence assessments, of which only four studies directly compared SPC vs. free-dose combination therapy (Table 1). All four of these studies demonstrated a significant improvement in persistence with SPC vs. free-dose combination therapy (Table 1).

Change in blood pressure and blood pressure control

Only 14 studies (48%) reported the number of patients reaching their target BP, although none reported the time to achieve BP control (Fig. 3). BP was consistently assessed as less than 140/90 mmHg in studies that defined BP targets. The percentage of patients who achieved target BP ranged from 25% (reported at 1 month) [37] to 89% (reported at 4 months) [45]. All four RCTs reported on BP goal attainment following randomization to treatment [21,37,38,46]. However, of these, just two studies provided statistical comparison of treatment arms, both of which demonstrated a significant improvement in BP goal attainment with SPC vs. free-dose combination therapy (Fig. 3).

FIGURE 3

Studies reporting patients achieving blood pressure goal on single-pill combination therapy vs. free-dose combination therapy.

Despite all studies investigating antihypertensive treatments, only half of the studies [16/29 studies (55%)] clearly reported change in BP during the study. All of these 16 studies, which included all four RCTs, noted BP reduction following antihypertensive treatment, although the time-to-follow-up varied. The majority of these 16 studies reported a significant change in BP from baseline. Three studies (19%) compared change in BP between SPC and free-dose combination regimens [22,38,46]. Of these three studies, Webster et al.[46] reported a significant difference in BP with SPC vs. free-dose combination therapy (reduction in SBP –29.3 vs. –20.6 mmHg; P < 0.001), whereas Bramlage et al.[22] reported no significant difference in change in BP with SPC vs. free-dose combination (change in SBP –9.1 vs. –8.7 mmHg; P = 0.964) and Nedogoda et al.[38] reported change in BP was similar between triple SPC [per/indapamide (ind)/amlo] and two-pill therapy (dual SPC per/ind + amlo; change in SBP –21.5 vs. –20.0 mmHg).

Cardiovascular morbidity and mortality

There were no RCTs that investigated cardiovascular outcomes (Table 2); however, some studies provided information regarding cardiovascular outcomes with SPC therapy. For example, Simons et al.[42] specifically discussed cardiovascular mortality over 48-month follow-up, and the adjusted hazard ratio for risk of death over 48 months for single-pill vs. two-pill therapy (any CCB/ACEi therapy vs. amlo + per therapy) was 1.83 [95% confidence interval (CI) 1.55–2.16]. In addition, hazard ratio for discontinuation with SPC vs. two-pill therapy was 1.86 (95% CI 1.74–1.99) [42]. The authors noted that the mortality outcome may be an overestimate due to residual confounding [42]. Verma et al.[14] reported on the hazard ratio of death for patients taking any SPC vs. free-dose combination therapy in Canada, and found (in an intention-to-treat analysis) a significantly lower rate of composite primary outcome, respectively (3.4 events/100 persons vs. 3.9 events/100 persons; hazard ratio = 0.89; 95% CI, 0.81–0.97; P < 0.01) [14]. However, for an on-treatment analysis (only patients who remained adherent with medication), there was no significant difference between treatment arms, demonstrating the importance of adherence in the treated population (hazard ratio = 1.06; 95% CI, 0.86–1.31; P = 0.60) [14]. Tung et al.[44] reported event rates with SPC of amlo/val vs. free-dose combination therapy (any CCB + any ARB) after a median follow-up of 15.2 months in a data set from Taiwan. This study also reported a reduction in major adverse cardiovascular events (hazard ratio = 0.83; 95% CI, 0.73–0.94; P < 0.001), and a reduction in the rates of heart failure, PCI, and malignant dysrhythmia with SPC vs. free-dose combination therapy [44].

TABLE 2

Reported cardiovascular-related outcomes in studies comparing single-pill combination and free-dose combination regimens

			n/N [%]
1st Author (year) [REF], study type	Treatment (SPC/free)	Treatment details	Hypotension	Oedema	Syncope	Angina	MI	HF	Stroke
Verma (2018) [14] – Obs	SPC	NR	NR	NR	NR	NR	46/8258 (0.6)	19/8261 (0.2)	39/8243 (0.5)
	Free	NR	NR	NR	NR	NR	34/6322 (0.5)	11/6330 (0.2)	26/6320 (0.4)
Webster (2018) [46] – RCT	SPC	tel/amlo/chlo	2 (0.6)	6 (1.7)	18 (5.2)	11 (3.2)	NR	NR	NR
	Free	NR	2 (0.6)	5 (1.4)	10 (2.8)	10 (2.8)	NR	NR	NR
Nedogoda (2017) [38] – RCT	SPC	per/ind/amlo	NR	NR	NR	NR	NR	NR	NR
	Free (2 pills)	(per/ind) + amlo	1 (1.3)	NR	NR	NR	NR	NR	NR
Tung (2015) [44] – Obs	SPC	amlo/val	NR	NR	NR	NR	19 (0.58)	70 (2.12)	197 (5.97)
	Free	ARB + CCB	NR	NR	NR	NR	122 (0.92)	431 (3.26)	864 (6.54)
Panjabi (2013) [39] – Obs	Mix (2 pills)	amlo/HCTZ + ARB	NR	NR	NR	NR	55 (1.37)	242 (6.04)	66 (1.65)
	Free (3 pills)	amlo + HCTZ + ARB	NR	NR	NR	NR	17 (1.27)	90 (6.74)	27 (2.02)
	Mix (2 pills)	amlo/HCTZ + ACEi	NR	NR	NR	NR	68 (1.12)	282 (4.64)	96 (1.58)
	Free (3 pills)	amlo + HCTZ + ACEi	NR	NR	NR	NR	40 (1.32)	147 (4.83)	56 (1.84)
	Mix (2 pills)	amlo/HCTZ + BB	NR	NR	NR	NR	30 (0.49)	33 (5.42)	11 (1.81)
	Free (3 pills)	amlo + HCTZ + BB	NR	NR	NR	NR	20 (1.64)	101 (8.29)	26 (2.13)

ACEi, angiotensin-converting enzyme inhibitor; amlo, amlodipine; ARB, angiotensin 1 receptor (AT1) blocker; BB, beta blockers; BP, blood pressure; CCB, calcium channel blocker; chlo, chlorthalidone; HCTZ, hydrochlorothiazide; HF, heart failure; ind, indapamide; MI, myocardial infarction; per, perindopril; tel, telmisartan; val, valsartan.

Summary of adverse events

In general, study medication was well tolerated in the 15 studies (52%) that reported on drug-related adverse events. Depending on the study, adverse events were captured by clinical report or from information included in clinical records. Although of note, none of the studies were designed to be safety studies, and furthermore many of these studies were observational and allowed concomitant medications as required. Therefore, adverse event observations should be considered in light of these points. Of the 15 studies that included information on adverse events, eight (53%) reported SAEs, of which five studies (33%) specifically compared treatment with SPC vs. free-dose combination therapy. In these studies, the proportion of patients reporting SAEs on SPC therapy was generally 1% or less (Table S4). Only one study reported SAEs in free vs. SPC therapy, where SAEs were similar (7.7% with SPC vs. 6.0% with free). No deaths were reported to be directly related to study treatment. Discontinuation from study medication due to adverse events was not widely reported.

The most commonly reported adverse events in the 15 studies that reported on individual adverse events were oedema (10 publications) and hypotension (seven publications).

Methodological reporting practices

Four studies were RCTs and assessed according to the NICE STA checklist. Of these four studies, most had low or unclear ratings across the methodological assessment scales (Table S5A). Twenty-five observational studies were assessed according to the Newcastle-Ottawa Quality assessment [19]. Aspects of methodological reporting varied, with a median score of 8 seen across the 25 studies (Table S5B); the maximum score possible on the Newcastle-Ottawa scale was 9 which was scored by nine studies [14,27,29,31,32,34,42,44,47].

DISCUSSION

This systematic REA found that adherence and persistence are generally higher in patients initiating antihypertensive mediation as SPC therapy rather than taking their anti-hypertensive therapy as multiple tablets, in free-dose combinations. However, we also found a paucity in data comparing BP lowering, BP control and/or cardiovascular outcomes in patients taking SPC therapy vs. free-dose combination therapy, with just four of 29 studies (14%) conducted as RCTs and, of these RCTs, only one reported a direct comparison of adherence for SPC vs. free-dose combination therapy. Within the studies identified, clear comparison of BP lowering or goal attainment for the SPC vs. free-dose combination therapy was presented in a minority of studies, and links to improved adherence were unclear. As such, more high-quality evidence is required to determine if the increases in adherence seen in patients on SPC therapy translate into better BP control and ultimately improve long-term cardiovascular outcomes.

Since the development of SPC therapies in the hypertension arena and their recommendation for clinical use, debate has been rife on the benefits and drawbacks of prescribing therapies as a single pill to patients requiring more than one antihypertensive therapy, rather than in free-dose combinations that may be more easily and individually titrated, especially should an adverse event occur [16,48,49]. Data suggest that the majority of patients with AH will require more than one antihypertensive therapy to achieve BP goals [9,10], with complementary combinations of therapeutics providing, in some cases, improved tolerability alongside additional BP lowering towards therapeutic targets [50,51]. However, evidence is limited beyond the BP-lowering benefits of SPC therapies, and longer-term benefits on morbidity and/or mortality outcomes is largely assumed from efficacy data from RCTs involving the parent drugs. Recent guidelines from the ESC/ESH recommend that the majority of patients with AH initiate two antihypertensive agents and preferably as an SPC, to reduce BP quickly towards target [9]. Individual evidence suggests more frequent BP control after 1 year for patients initiating SPC therapy compared with monotherapy, in addition to a steeper dose-BP response curve [52]. However, in presenting their guidelines, the ESC/ESH committees sought further evidence on the impact of single pill vs. multiple-drug treatment strategies on adherence to treatment, BP control and clinical outcomes [9]. In addition to investigating these points, our study sought to highlight the gaps in literature to guide future study design and development to fully answer the unmet need highlighted by the ESC/ESH in 2018.

By assessing the existing literature, we highlight a number of areas where improvements could be made to study reporting, design, or conduct (Table 3). First, specifically designed and powered RCTs are lacking, as just 10% supported evidence for SPC therapy, with just three RCTs providing statistical evidence for adherence and/or efficacy. In addition, well-designed, longer-term observational studies that support the efficacy observations from the RCTs are also required (Table 3). The majority of studies conducted in the past 5 years were observational (86%), which are inherently limited by the data they are using. As such, well-designed studies around international recommendations are still needed. Our assessment of methodological reporting highlighted that many studies were considered ‘low’ on basic aspects of reporting RCT design (on the NICE STA methodological scoring). Just nine of the 25 observational studies (36%) scored full ‘points’ on the Newcastle-Ottawa Quality assessment, indicating better reporting of fundamental aspects of observational study methodology, but still only in the minority of studies. Gaps in evidence will only be answered by ensuring full reporting of these, and other, fundamental points of methodology (Table 3).

TABLE 3

Gap in evidence for assessing efficacy of single-pill combination therapy vs. free-dose combination therapy

Topic	Evidence gap	Recommendations
Study design	RCTs [only 14% (4/29 in past 5 years)] Long-term prospective studies Outcome studies	Well-designed studies providing a direct comparison between adherence and BP control and/or CV outcomes for studies comparing SPC and free-dose combination therapies Longer-term follow-up to analyze efficacy Consistent recording and reporting of BP
Length of follow-up	Variation in follow-up length and time points made comparisons between studies difficult Few studies reported results beyond six months to one year of follow-up	Multiple and standardized time points should be included in future studies Longer follow-up periods will provide useful information on the long-term adherence benefits of SPC therapy
Assessment of efficacy	Approx. half of studies reported BP measurement or patients reaching BP goals Time to reach BP control not reported How BP was measured was rarely reported	Standard inclusion of BP efficacy measurement(s) Clear definition of time to reach BP control Follow standardized techniques for accurate BP measurement according to up-to-date guidelines Employ ABPM/HPBM/app technology (or similar) to improve assessment of BP out-of-office
Adherence/persistence measure	No standardized reporting of adherence or persistence measurements	Clear and consistent reporting measures for adherence (PDC and MPR are most commonly used) to facilitate comparison between studies Standardized and recommended reporting methodology for persistence Use of validated scales to confirm adherence in specific patient populations
Adverse events	AEs were rarely reported in studies comparing patients receiving SPC and free-combination therapies Few studies reported the grade of AE	Specific studies to determine if increased adherence on SPC therapy is associated with a different safety profile, rather than physicians having to extrapolate information from the safety profile of individual components of the SPC
Patient/physician preferences	Few studies reported patient or physician preferences for SPC or free-combination therapies, or types of antihypertensive medications being prescribed	Evaluate patient preferences for medication type and other factors affecting adherence to aid understanding of preferences around SPC therapy Educate physicians’ treatment decisions as increased importance is being placed on patient preferences across all aspects of health care

ABPM, ambulatory blood pressure monitoring; AE, adverse event; BP, blood pressure; CV, cardiovascular; HBPM, home blood pressure monitoring; MPR, medication possession ratio; PDC, proportion of days covered; RCT, randomized controlled trial; SPC, single-pill combination.

One of the key driving forces behind the benefits for using SPC therapies is evidence that higher patient pill burden negatively affects adherence [47,53,54], and in turn lower adherence leads to poorer efficacy outcomes [55-57]. Although patients with hypertension and other cardiovascular risk factors may have comorbidities and be taking other medications [9,10], reducing pill count from three or two separate pills to one SPC would still impact their overall pill burden [53,58]. In line with our conclusion, previous literature analyses have demonstrated higher adherence for patients treated with SPC vs. free-dose combination therapy [15,59]. However, we found that direct comparisons were in the minority of studies; indeed from the four RCTs identified, only one directly compared adherence with SPC and comparator therapy [46]. Similarly, only four studies (14%) included a direct comparison of persistence between SPC and free-dose combination therapy, none of which were RCTs [22,27,34,47]. Adherence is typically defined as an MPR or PDC above 80% [60], and 11 studies that used PDC or MPR in the current analysis did use this definition within their reporting. However, many studies simply included the statement that ‘adherence was good...’, or similar, limiting accurate representation of medication-taking behavior.

Another limitation to qualitative assessment of adherence with SPC vs. free-dose combination therapy was the use of many different types of assessment for adherence and persistence, each with their own intricacies. There are limitations associated with any self-reported measures [60], and scales should be validated in the patient population of interest [61-64], to ensure they can be relied upon [65]. Of note, monitoring ingestion by returned tablets was the only study that did not demonstrate a difference in adherence [46]. Regardless, more standardized assessment would facilitate comparison among studies, and a better fundamental understanding of medication-taking behavior. Therefore, we suggest there are evidence gaps in methodological design regarding the assessment of adherence/persistence, with a need to standardize definitions of ‘adherent/persistent’ (Table 3). There are many aspects that affect adherence beyond those directly related to the treatment, including patient-related and physician-related factors or socioeconomic reasons. Patient preference was assessed in only two studies [23,41], and no study reported on physician preference. As such, there is a lack of information available on all aspects of adherence, which is needed to fully understand why patients become non-adherent to their medications (Table 3). International guidelines may consider promoting methods to improve adherence, such as those issued by the WHO based around the five dimensions of adherence [66].

Assessment of effectiveness in observational studies depends on the data available within the healthcare records; however, to confirm effectiveness of an antihypertensive, and that patients are taking their medication, BP needs to be routinely and accurately measured. For example, it is very surprising and a major shortcoming that only half of studies clearly reported BP reduction over time, and only approximately half of studies reported information on BP goal attainment. Guidelines clearly outline target BP for patients of different ages, risk levels and with different comorbidities [9,10]. These goals are developed through rigorous assessment of clinical evidence and academic debate, and although United States and European Union guidelines differ in their definition of ‘hypertension’, ‘normal’ BP is similarly defined (BP < 120 mmHg or BP 120–129 mmHg for the United States and European Union, respectively [9,10]). The importance of consistency and accuracy in measuring BP is something that has been commented on for decades [67-69] but still remains a major drawback when comparing clinical studies or real-world data [67]. To assess efficacy or effectiveness of any antihypertensive therapy, BP should be accurately recorded as standard medical practice after therapy initiation. Just three studies (10%) included ambulatory BP monitoring (ABPM) [26,35,37] and methodology of BP monitoring was generally insufficiently reported. Many studies have debated the need for more standardized and careful methodology for BP monitoring [67-69] and/or the use of ABPM/home BP monitoring, where costs and equipment allow [9,10]. Some investigations are ongoing in this area to confirm if out-of-office BP can improve cardiovascular outcomes [70-72], in an attempt to more accurately monitor BP away from the in-office setting. This includes investigating how cell phone apps, such as those developed by the ESC/ESH, may assist with BP monitoring [73,74]; but recent studies highlight that only the minority of existing apps meet quality and accuracy standards [75]. With the acceleration of technological interventions in recent times, improvements in monitoring apps is inevitable. But, improving study reporting and consistency in BP monitoring will aid interpretation of BP recording, and how the number reflects ‘real’ BP (Table 3).

In addition, there are gaps in the literature for studies assessing BP goal attainment with SPC therapy vs. free-dosed combination therapy, ideally in an RCT alongside other efficacy outcomes and/or retrospective observational studies that allow assessment of information on cardiovascular outcomes (Table 3). Moreover, studies are needed that not only demonstrate the continued efficacy of SPC therapies, but that allow loss of efficacy to be investigated. For example, highlighting the need to up-titrate by dose, or to add additional medications (e.g., two-pill SPC to three-pill SPC), and/or to address loss of adherence. Few studies provided statistical evidence for both improved adherence and improved efficacy or effectiveness of SPC therapy vs. free-dose combination therapy. In the longer term, evidence is needed to support a meta-analysis or similar, to determine if SPC therapies lower cardiovascular mortality, as indicated by some individual observational studies. Of interest, very recently the WHO added, for the very first time, a SPC comprising two antihypertensive medications to the WHO Essential Medicines List [76], thereby acknowledging that the use of SPC is the emerging best practice for safe, effective, rapid, and convenient hypertension control worldwide [77].

Our study should be considered in light of its limitations. We conducted a systematic rapid-evidence assessment to focus our methodology within databases known to publish high-quality evidence, but this methodology is less exhaustive than a full SLR. It is possible that by conducting an SLR, information published outside of EMBASE, MEDLINE, and COCHRANE may have been captured, and the number of studies included changed as a result. Our assessments are naturally limited by the information available in the publications. Therefore, further information on BP measurement, for example, may not have been included due to word constraints or similar. The use of concomitant medications in all patient groups should be considered when efficacy observations are compared. The aim of each study would have varied, depending on if any specific patient type was being investigated. Higher risk patients would have been managed differently, and the selection of medications would have been made in light of any comorbidities or existing medications. For example, no SPC included the combination of a thiazide diuretic + potassium-sparing diuretic, which can be prescribed in SPC in specific patient groups [78]. Finally, we summarize information published from RCT trials (14%) alongside observational studies of healthcare records, and the difference in these approaches should be considered when interpreting our observations. These observational studies reflect real-world clinical practice, and by selecting both types of study conducted in the past 5 years, we have been able not only to support adherence observations for selecting SPC therapy, but also to acknowledge the gaps in literature, by study type.

In conclusion, adherence and persistence were generally higher for patients taking antihypertensive medication as a SPC rather than as free-dose combination therapy, although a number of different ways of measuring adherence and persistence limited the ability to directly compare studies. In addition, few studies compared BP control and/or fewer still reported cardiovascular outcomes in patients prescribed SPC vs. free-dose combination therapy. Most surprisingly, only half of studies clearly reported change in BP during the study, and fewer still included detail on the methodology used to measure BP, limiting our ability to effectively compare the BP-lowering outcomes of these hypertension management approaches. As a result, future studies need to work on filling these gaps in the literature (Table 3), with specifically designed and reported studies aiming to determine if an increase in adherence with SPC regimen leads to better efficacy outcomes. Furthermore, better methodological reporting and recommendations from international bodies on standardizing assessments would also aid fuller comparison of medication-taking behavior and allow the benefits of SPC therapy to be more fully explored.

ACKNOWLEDGEMENTS

Medical writing support was provided by Karen Burrows of Engage Scientific Solutions, and was funded by Pfizer. Assistance with data extraction was carried out by Catherine Rolland and Jon Carthy of CURO Payer Evidence (part of the Envision Pharma group) and was funded by Pfizer Inc.

This study was sponsored by Pfizer. CURO Payer Evidence (part of the Envision Pharma group) were paid consultants to Pfizer in relation to conducting the rapid evidence assessment.

Conflicts of interest

K.T. has received research grants or honoraria for advisory boards and lectures from Medtronic, Servier, Bayer, Menarini, Novartis, AstraZeneca, Boehringer Ingelheim, Pfizer, Chiesi, Sanofi, Amgen, ELPEN, Recordatti and Winmedica. He was a member of the Task Force of the 2018 ESC/ESH Hypertension Guidelines. R.K. has received honoraria for consultancy, lectures, and support for research from Bayer Pharma, Berlin-Chemi Menarini, Daiichi Sankyo, Ferrer, Sanofi and Servier. He was a member of the Task Force of the 2018 ESC/ESH Hypertension Guidelines. G.S., J.V.V., and T.H. are full-time employees of Upjohn a division of Pfizer Inc.