Impact of switching to polypill based therapy by baseline potency of medication: post-hoc analysis of the SPACE Collaboration dataset

Ruth Webster1, PhD, MIPH(hons), MBBS(hons), BMedSc(hons)

Chris Bullen2, PhD, MBChB

Anushka Patel1, MBBS, SM, PHD

Vanessa Selak2, PhD, MBChB

Sandrine Stepien1,

Simon Thom3, MBBS, MD

Anthony Rodgers1, PhD, MBChB

1The George Institute for Global Health, University of New South Wales, Sydney2National Institute for Health Innovation, University of Auckland, Auckland3Imperial College, London

Corresponding Author:Dr Ruth WebsterThe George Institute for Global HealthPO Box M201,Missenden RdCamperdown NSW 2050AustraliaPh: +61 2 8052 4557Fax: +61 2 8052 4502E: rwebster@georgeinstitute.org.au

Abstract word count: 328

Manuscript word count: 2469

FundingThe authors have received grants from several research charities and national funding agencies for research on cardiovascular polypills, and from Dr Reddys Ltd for co-ordination of the SPACE program (www.spacecollaboration.org). The polypills used in the SPACE trials were manufactured and supplied by Dr Reddy’s Ltd free of charge. Some authors received funding from Dr Reddy’s Laboratories Ltd to attend investigator meetings related to the polypill (VS, AC, AP, NR, AR, ST, AW, RW).

1

Abstract

Background: Fixed dose combinations of cardiovascular therapy (‘polypills’) have now been launched in several dozen countries. There is considerable clinical interest in the effects of switching to polypill-based care from typical current treatment regimens, especially if polypills contain components at sub-maximal dosage.

Methods:

The SPACE Collaboration includes three trials of polypill based care vs usual care in patients with established CVD or at high calculated risk. Individual patient data for 3140 trial participants were combined. Patients were categorized according to the potency of the statin and the number of BP lowering medications they were taking at baseline. Effects on adherence to anti-platelet medication, systolic blood pressure (SBP) and LDL cholesterol stratified by baseline potency of medication were determined using fixed effects models.

Results:

Randomisation to the polypill group was associated with improved SBP at 12 months, but this improvement varied according to baseline BP regimen: -3.3, -5.9, -2.5 and +1 mmHg for patients taking 0, 1, 2 and 3+ BP lowering medications at baseline. For changes in LDL cholesterol at 12 months, significant improvements in LDL cholesterol were seen for those taking no statin (-0.21 mmol/l; 95% CI: -0.34 to -0.07), less potent statin (-0.16 mmol/L; 95% CI: -0.29 to -0.04) and equipotent statins (-0.14 mmol/L; 95% CI -0.26 to -0.02) at baseline.

Conclusion: The adherence benefits of polypills tend to offset the loss of potency from use of individual components with lower dose potency, and to facilitate improvements in multiple risk factors.

Keywords: polypill, combination therapy, cardiovascular prevention, meta-analysis

2

Introduction

A large amount of variation exists in preventive medication regimens prescribed to those at

high risk of CVD. Availability of individualized therapy has failed to achieve adequate

coverage of recommended, proven CVD preventive medications globally. In high income

countries only approximately 50% of patients take the recommended combination of statin,

blood pressure lowering and anti-platelet therapy(1), with coverage much lower in lower

income countries(2).

Fixed dose combination (FDC) cardiovascular ‘polypills’ (combination medications

containing blood pressure lowering and statin medication with or without aspirin) improve

adherence(3-5), but this benefit could be negated if polypill components were less potent than

alternative component medicines in the same class. Specifically, clinical concern has been

expressed about the lesser potency of simvastatin that has been included in some polypills,

compared to atorvastatin and rosuvastatin and whether this may be more important than any

adherence advantages from taking the polypill.(6) Perceived lack of dose titration options of

the polypill components has also led to concerns over potential for risk factor control to be

diminished compared to conventional approaches to individualized therapy.(6)

The SPACE Collaboration is an international group of academic investigators who have

conducted three similar, large, randomized trials of polypill based care compared to usual

care in patients with established cardiovascular disease or at high calculated risk.(3)

Individual patient data meta-analysis (IPD) demonstrated improvements in the use of

recommended combination therapies for patients randomized to polypill based care, with

consequent improvements in systolic blood pressure (SBP) and low density lipoprotein

cholesterol (LDL-C).(3) Patient enrolment in all SPACE Collaboration trials necessitated that

3

all components of the polypill (i.e. anti-platelet, cholesterol lowering, and combination BP

lowering therapy) were clinically indicated. However not all patients were currently

prescribed all of these components, since there were no limitations on which medications or

dosage the patient should be taking prior to enrolment. Additionally, physicians were allowed

to add medications on top of the polypill in order to achieve risk factor targets. The extent to

which this occurred has not previously been reported. Therefore, it is possible that patients

who were receiving more than two BP lowering medications at baseline may have had their

therapy down-titrated by randomization to the polypill with possible adverse impacts on their

BP control, despite improvements in adherence.

In this paper we report post-hoc subgroup analyses of the SPACE Collaboration dataset to:

1. Describe patterns of treatment intensity at 12 months follow-up in both randomized

groups compared to baseline, and

2. Investigate any differential effect of randomisation to the polypill arm in those who

were on a more potent statin at baseline or on ≥ 3 BP lowering medications at

baseline.

Methods

Details of the SPACE Collaboration IPD analysis methods have been previously reported in

detail.(7) In brief, three trials were collaboratively planned, conducted and analysed, based on

the same protocol, with minor regional adaptations: UMPIRE,(8) with 1004 participants from

Europe (United Kingdom, Ireland, and The Netherlands), and 1000 participants from India;

Kanyini-GAP (N=623) conducted in Australia(9) and IMPACT (N=513) conducted in New

Zealand.(10) The three trials utilized a prospective, randomised open-label, blinded endpoint

4

(PROBE) design. A prospectively planned IPD meta-analysis was also registered with the

Australian New Zealand Clinical Trial Registry: ACTRN12612000980831.

Patient population

All patients were required to have indications for each category of antiplatelet, cholesterol

lowering and, BP lowering therapy, based on either established cardiovascular disease or

absolute risk greater than 15% over 5 years using a Framingham based risk calculator.(11)

Study medication

Usual care comprised cardiovascular preventive medications (including antiplatelet,

cholesterol lowering and BP lowering medicines as separate medicines) as prescribed at the

treating doctor’s discretion. No efforts were made to influence treatment provided as usual

care and no changes were required following enrolment into the study.

If patients were randomised to the polypill, 2 polypill versions were available for prescription

containing 75mg aspirin, 40mg simvastatin, 10mg lisinopril and either 12.5mg

hydrochlorothiazide or 50mg atenolol. Additional anti-platelet, cholesterol lowering and BP

lowering medication were allowed to be prescribed as required to meet BP and cholesterol

targets which were left to the prescribers’ clinical opinion. If the patient withdrew from

taking the polypill for any reason the treating doctor was advised to commence individual

agents, however no restriction was placed on what medications could be prescribed to replace

the polypill (e.g. any statin in any dose could be prescribed).

Medication use was documented at baseline and any changes that occurred were noted at

each trial visit.

Categorization of intensity of treatment

5

Doses of individually prescribed statins were categorized as being less potent, equipotent or

more potent than 40mg simvastatin as contained in the polypill according to published

definitions of equivalence (Table 1).(12)

Blood pressure lowering medications were identified as being prescribed as monotherapy,

dual therapy, triple therapy or as part of ≥ four concurrent therapies.

Numbers and intensity of statins and BP lowering medications were determined based on all

medications that were taken regardless of whether as part of a polypill or as individual

medications.

All medication data was collected via patient self-report. Adherence to medication was

defined as taking the medication at least 4 out of the previous 7 days.

Statistical methods

Analyses were performed on the combined dataset using one-stage meta-analyses.(13) For

BP and cholesterol, analyses consisted of a linear mixed model with the month-12 value as

the outcome, the baseline value and the treatment arm as fixed effects, and a random trial

intercept and fixed trial-by-treatment interaction. For adherence and other dichotomous

outcomes, a log-binomial regression with a fixed treatment effect, a random trial intercept

and fixed trial-by-treatment interaction was used. No adjustments were made for post-hoc

subgroup analysis.

Results

A total of 3140 patients were randomized across the three SPACE Collaboration trials.

Patients were predominantly male with a mean age of 62 years of age (Table 2). At baseline

mean SBP was 139 mmHg and mean LDL-C 2.4 mmol/L. About three quarters of

6

participants had established cardiovascular disease. Ninety-one percent of participants were

prescribed at least 1 BP lowering medication, 84% prescribed a statin and 87% prescribed an

anti-platelet medication at baseline. Seventy-four percent were prescribed all three

medication types.

Change in antiplatelet therapy from baseline to 12 months

Of the participants prescribed anti-platelet therapy at baseline, just over half (58.4%) were

prescribed aspirin alone and 30% prescribed another form of anti-platelet therapy. At 12

months, a greater proportion of participants in both arms were prescribed some form of anti-

platelet agent (92% in the usual care arm vs 96% in the polypill arm) with aspirin remaining

the most common agent.

Change in number of BP lowering agents from baseline to 12 months

Table 3 shows changes in number of BP lowering medications from baseline to 12 months.

Cells are colour-coded to demonstrate those who were on a lesser number of BP lowering

medications by 12 months compared to baseline, those who stayed at the same number of

medications and those who increased the number of BP lowering medications taken. In the

usual care arm, 171 patients had no data at 12 months and 145 patients had no data in the

intervention arm. At baseline overall, 67% of participants were prescribed at least 2 drugs. At

12 months, 83% of patients in the usual care arm continued prescribed the same number of

BP lowering drugs that they were prescribed at baseline with 5% decreasing the number

taken and 12% increasing the number taken. In the polypill group overall, 52% took the same

number of BP lowering medications at 12 months with 9% prescribed less and 39%

prescribed more. The largest increase in number of BP lowering medications taken occurred

in those prescribed 0 or 1 BP lowering drug at baseline (consistent with being randomized to

the polypill containing 2 BP lowering medications).

7

Changes in potency of statins prescribed from baseline to 12 months

Table 3 shows changes in intensity of statin prescription from baseline to 12 months.

Participants in the usual care arm predominantly remained at the same potency of statin with

a small number decreasing (4% of participants) and increasing (8% of participants) their

intensity of statin dosing. In the polypill arm, 22% of participants had a reduction in the

potency of their statin dose, 32% remained the same and 37% either started or had an

increase in the potency of their statin dose.

Primary outcomes of adherence, SBP and LDL-C stratified by potency of medication taken at

baseline

Figure 1 shows polypill treatment effects on adherence to aspirin therapy, SBP and LDL-C at

12 months follow-up, stratified by potency of anti-platelet, BP lowering and statin therapy at

baseline.

Randomisation to the polypill group was associated with improved adherence to anti-platelet

therapy for all categories of baseline anti-platelet use. At 12 months, randomisation to the

polypill group was associated with overall improved SBP at 12 months, but this effect

progressively declined with an increasing number of BP lowering medications used at

baseline: : -3.3, -5.9, -2.5 and +1 mmHg for patients prescribed 0, 1, 2 and ≥3 BP lowering

medications at baseline. For those prescribed 3 or more medications at baseline, the 95%

confidence intervals ranged from -2.1 to 6.5 mmHg.

For changes in LDL cholesterol at 12 months, randomisation to polypill was associated with

significant improvements in LDL cholesterol among those prescribed no statin (-0.21 mmol/l;

95% CI: -0.34 to -0.07), less potent statin (-0.16 mmol/L; 95% CI: -0.29 to -0.04) and

equipotent statins (-0.14 mmol/L; 95% CI -0.26 to -0.02) at baseline, while there was no clear

8

difference among those prescribed more potent statins at baseline (0.07 mmol/L; 95% CI -

0.05 to 0.18).

When the effect of the polypill on BP and LDL-C was examined by subgroups defined by

different levels of baseline medications that would not be expected to directly impact the

outcome, no heterogeneity was seen. Thus, even if a neutral or negative impact on a

particular risk factor was seen due to difference in potency, beneficial effects of the polypill

on the alternate risk factor was preserved.

Figure 1 shows the treatment effect on adherence to aspirin, SBP and LDL cholesterol based

on overall potency of combined treatment regimen at baseline. There were relatively few

patients prescribed more intensive treatment across all component medications so confidence

intervals were wide, but there were no findings consistent with potential benefit or harmful

effect of switching to a polypill (RR 1.0, 95%CI 0.9 to 1.1 for aspirin adherence; 0.94 mmHg,

95%CI: -3.6 to 5.48 for SBP; 0.049 mmol/L, 95%CI: -0.136 to 0.233 for LDL). Patients who

were prescribed all three treatment modalities (but not all at highest intensity) showed a 30%

(95% CI 1.3 to 1.4) improvement in adherence to aspirin, 3.33 (95%CI: -4.84 to -1.83)

mmHg improvement in SBP and 0.12 (-0.183 to -0.063) mmol/L improvement in LDL

cholesterol. The improvements in those prescribed less than optimal treatment were more

marked with up to 500% improvement in adherence to aspirin, 4.6mmHg improvement in

SBP and 0.15 mmol/l improvement in LDL cholesterol (although LDL cholesterol difference

non-significant due to wide confidence intervals).

Discussion

Our analysis demonstrates that overall randomization to polypill based therapy was

associated with greater self-reported adherence to optimal combinations of therapy, lower BP

9

levels and lower LDL-C levels. When stratified by baseline antiplatelet use, no heterogeneity

was observed in the effects on BP and LDL-C. When stratified by baseline potency of statins,

there was heterogeneity observed in the effect on LDL-C with higher potency at baseline

associated with less LDL-C lowering effect. A similar result was seen for BP lowering

medications.

However, rather than concentrating solely on the effect of individual drugs on risk factor

levels, our analyses indicate that treatment with a polypill impacts multiple risk factors

together because it enforces use of multiple risk reducing medications simultaneously.

Therefore, use of a polypill may improve BP levels and concurrently address lipids levels and

risk of thrombosis.

The clinical implications of these findings are two-fold: first, switching to polypill-based

care will tend to give most benefits when patients are undertreated or not treated for all three

treatment modalities with up to 500% improvement in adherence to aspirin, 4.6mmHg

improvement in SBP and 0.15 mmol/L improvement in LDL cholesterol (although difference

in LDL cholesterol non significant). Second, when a patient is receiving intensive treatment

for one modality (e.g. potent statin therapy), switching to a polypill with less potency (such as

a moderate potency statin) may not worsen control of the factor targeted by the medication

(e.g. LDL-cholesterol) but may have benefits on other modalities (i.e. BP control and aspirin

adherence). In other words, improvements in overall polypill adherence appeared to balance

out potential reductions in efficacy from the use of a less potent component medication.

Unless a patient is already prescribed antiplatelet medication, high potency statin and at least

3 BP lowering medications, the benefits of polypill based therapy on global patient

10

cardiovascular risk appear to be clear. The literature has clearly shown that the proportion of

patients at high risk who fall into this ‘optimal treatment’ group is very low.(1,2)

There is debate about the need for polypill studies with cardiovascular event endpoint data

given the already proven efficacy of these medicines. However, some suggest a large trial of

polypill-based care vs usual care with long-term follow-up is needed.(14) Additional

evidence on longer-term adherence, and in diverse patient populations and health care

settings would be helpful. Finally, further research is needed on the effects of using polypills

with the most potent components (such as rosuvastatin and atorvastatin) and/or multiple dose

versions , now that these statins are off-patent.

Strengths of this study include the large number of patients and use of individual patient data

for meta-analysis. All three included studies were conducted so as to mimic routine practice

as much as possible to provide insights into how polypills would be utilized in clinical

practice.

Limitations include lack of power in certain patient subgroups (including those receiving all

three treatments at optimal doses at baseline) due to low numbers. The patients in the SPACE

Collaboration trials were an unusually well treated group, possibly due to selection bias for

trial participants. Three quarters of patients were on all three component medications

(regardless of dose) at baseline compared to more conservative estimates of around 50% in

high income settings and as low as 3% in low income settings.(1,2)

We conclude that the conventional paradigm of treating risk factors separately and titrating

medications individually may present barriers to optimal medication use and adherence.

11

Conclusions

Use of a polypill based regimen improves risk factor control in the majority of patients

regardless of potency of baseline individualized medication regimen; the benefits are largest

in those who are undertreated or poorly adherent which comprises the vast majority of

patients globally. Use of a polypill may assist in facilitating risk reduction across multiple

risk factors with less cost and inconvenience to patients.

References:

1. Webster RJ, Heeley EL, Peiris DP et al. Gaps in cardiovascular disease risk management in Australian general practice. Medical Journal of Australia 2009;191:324-9.

2. Yusuf S, Islam S, Chow CK et al. Use of secondary prevention drugs for cardiovascular disease in the community in high-income, middle-income, and low-income countries (the PURE Study): a prospective epidemiological survey. Lancet 2011;378:1231-43.

3. Webster R, Patel A, Selak V et al. Effectiveness of fixed dose combination medication ('polypills') compared with usual care in patients with cardiovascular disease or at high risk: A prospective, individual patient data meta-analysis of 3140 patients in six countries. International journal of cardiology 2016;205:147-56.

4. Castellano JM, Sanz G, Penalvo JL et al. A polypill strategy to improve adherence: results from FOCUS (Fixed-dose Combination Drug for Secondary Cardiovascular Prevention) Project. Journal of the American College of Cardiology 2014.

5. Bramlage P, Sims H, Minguet J, Ferrero C. The polypill: An effective approach to increasing adherence and reducing cardiovascular event risk. European journal of preventive cardiology 2017;24:297-310.

6. Webster R, Castellano JM, Onuma OK. Putting polypills into practice: challenges and lessons learned. Lancet 2017;389:1066-1074.

7. Webster R, Patel A, Billot L et al. Prospective meta-analysis of trials comparing fixed dose combination based care with usual care in individuals at high cardiovascular risk: the SPACE Collaboration. International journal of cardiology 2013;170:30-5.

8. Thom S, Poulter N, Field J et al. Effects of a fixed-dose combination strategy on adherence and risk factors in patients with or at high risk of CVD: the UMPIRE randomized clinical trial. JAMA : the journal of the American Medical Association 2013;310:918-29.

9. Patel A, Cass A, Peiris D et al. A pragmatic randomized trial of a polypill-based strategy to improve use of indicated preventive treatments in people at high cardiovascular disease risk. European journal of preventive cardiology 2014.

12

10. Selak V, Elley CR, Bullen C et al. Effect of fixed dose combination treatment on adherence and risk factor control among patients at high risk of cardiovascular disease: randomised controlled trial in primary care. BMJ 2014;348.

11. Anderson KM, Odell PM, Wilson PW, Kannel WB. Cardiovascular disease risk profiles. American heart journal 1991;121:293-8.

12. Helfand M, Carson S, Kelley C. Drug Class Review on HMG-CoA Reductase Inhibitors (Statins): Final Report. Portland (OR), 2006.

13. Debray TPA, Moons KGM, Abo-Zaid GMA, Koffijberg H, Riley RD. Individual Participant Data Meta-Analysis for a Binary Outcome: One-Stage or Two-Stage? PLoS ONE 2013;8:e60650.

14. Piepoli MF, Hoes AW, Agewall S et al. 2016 European Guidelines on cardiovascular disease prevention in clinical practice: The Sixth Joint Task Force of the European Society of Cardiology and Other Societies on Cardiovascular Disease Prevention in Clinical Practice (constituted by representatives of 10 societies and by invited experts)Developed with the special contribution of the European Association for Cardiovascular Prevention & Rehabilitation (EACPR). European heart journal 2016;37:2315-81.

13

Table 1: Defined statin equivalence.(12)

Statin Less potent dosages Equipotent dosages More potent

dosages

Simvastatin <40mg 40mg (=dose in the

polypill)

>40mg

Pravastatin <80mg 80mg Nil

Fluvastatin ≤80mg Nil Nil

Atorvastatin <20mg 20mg >20mg

Rosuvastatin <5mg 5mg >5mg

Lovastatin <80mg 80mg Nil

Pitavastatin <4mg 4mg >4mg

14

Table 2: Baseline Characteristics.

Baseline CharacteristicPolypill

N = 1569

Usual Care

N = 1571

Age, years (SD) 62.3 (10.6) 62.0 (10.9)

Female, n (%) 398 (25.4%) 381 (24.3%)

Office BP, mmHg (SD)

Systolic BP 139.2 (20.8) 139.8 (21.0)

Diastolic BP 79.0 (12.1) 79.5 (11.9)

Lipid fractions, mmol/L (SD)

Total cholesterol 4.2 (1.0) 4.3 (1.3)

HDL cholesterol 1.1 (0.3) 1.1 (0.3)

LDL cholesterol derived 2.4 (0.9) 2.4 (0.9)

Triglycerides 1.6 (1.1) 1.6 (1.0)

Current smoker, n (%) 312 (19.9%) 322 (20.5%)

Creatinine, µmol/L (SD) 88.1 (30.3) 88.6 (25.0)

No history of symptomatic

cardiovascular disease, n (%)377 (24%) 367 (23%)

History of coronary heart disease, n

(%)1021 (65.1%) 1025 (65.3%)

History of cerebrovascular disease, n

(%)216 (13.8%) 231 (14.7%)

History of peripheral vascular disease,

n (%)92 (5.9%) 70 (4.5%)

Diabetes mellitus, n (%) 581 (37.0%) 542 (34.5%)

15

Table 3: Change in treatment intensity of BP lowering therapy and statins taken at 12 months. Participants remain in the same row, i.e. data at 12 months shows outcomes for patients in each category at baseline

Baseline 12 Months

Usual care (N=1400) Polypill (N=1424)

Usual care Polypill

No BP

lowering 1 2 >=3

No BP

lowering 1 2 >=3

Number of observations 1571 1569 67 329 560 444 49 86 793 496

BP lowering agents at baseline

No BP lowering agents 105

(49%)

110

(51%)

51

(55%)

19

(21%)

16

(31%)

6

(7%)

21

(21%)

10

(10%)

58

(59%)

10

(10%)

1 BP lowering agent 392

(47%)

441

(53%)

9

(3%)

278

(80%)

42

(12%)

18

(5%)

10

(3%)

53

(13%)

275

(69%)

59

(15%)

2 BP lowering agents 634

(51%)

607

(49%)

7

(1%)

25

(4%)

474

(83%)

64

(11%)

12

(2%)

15

(3%)

383

(69%)

144

(26%)

≥3 BP lowering agents 440

(52%)

411

(48%)0

7

(2%)

28

(7%)

356

(91%)

6

(2%)

8

(2%)

77

(21%)

283

(76%)

No statin Less Equi More No statin Less Equi More

N 1508 1519 171 409 394 385 86 49 1127 123

No statin 240 (49%) 252 (51%)133

(62%)

30

(14%)

26

(12%)

24

(11%)

30

(13%)

3

(1%)

182

(81%)

11

(5%)

Less potent 420 (50%) 422 (50%)13

(3%)

365

(92%)

8

(2%)

9

(2%)

23

(6%)

38

(10%)

329

(83%)

5

(1%)

Equipotent 437 (52%) 411 (48%)21

(5%)

8

(2%)

355

(88%)

21

(5%)

19

(5%)

4

(1%)

339

(88%)

24

(6%)

16

Baseline 12 Months

Usual care (N=1400) Polypill (N=1424)

Usual care Polypill

No BP

lowering 1 2 >=3

No BP

lowering 1 2 >=3

More potent 411 (49%) 434 (51%)4

(1%)

6

(2%)

5

(1%)

331

(96%)

14

(4%)

4

(1%)

277

(73%)

83

(22%)

Changed to lower number of BP lowering

agents/lower potency of statin

Stayed at the same number of BP lowering

agents/same potency of statin

Changed to higher number of BP lowering

agents/higher potency of statin

17

18

Figure 1: Adherence to anti-platelet medication, systolic blood pressure and LDL cholesterol at 12 months follow-up stratified by baseline potency

of individual classes of medication and overall treatment intensity.

Intensive treatment defined as prescribed anti-platelet medication, >2 blood pressure lowering medications and a more potent statin.

All 3 modalities defined as all other patients prescribed anti-platelet, blood pressure lowering and statin medication, at any dose.

2 out of 3 modalities defined as all other patients prescribed at least 2 out of 3 of anti-platelet, blood pressure lowering and statin medication, at any

dose.

Less than 2 modalities includes all other patients prescribed none, or only 1 of antiplatelet, blood pressure lowering and statin medication, at any dose.

19