Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
Confidential: For Review OnlyThe effectiveness of first-line antihypertensive drugs
according to age and ethnicity: a cohort study in UK primary care
Journal: BMJ
Manuscript ID BMJ-2019-053556
Article Type: Research
BMJ Journal: BMJ
Date Submitted by the Author: 11-Nov-2019
Complete List of Authors: Sinnott, Sarah-Jo; London School of Hygiene and Tropical Medicine, Non-communicable disease epidemiologyDouglas, Ian; London School of Hygiene and Tropical Medicine, Non-communicable disease epidemiologySmeeth, Liam; London School of Hygiene and Tropical Medicine, Non-communicable disease epidemiologyWilliamson, Elizabeth; London School of Hygiene and Tropical Medicine, Medical StatisticsTomlinson, Laurie; LSHTM, Non-communicable disease epidemiology
Keywords: hypertension, blood pressure, calcium channel blocker, ace-inhibitor, angiotensin receptor blocker, effectiveness, age, ethnicity
https://mc.manuscriptcentral.com/bmj
BMJ
Confidential: For Review Only
1
Title
The effectiveness of first-line antihypertensive drugs according to age and ethnicity: a
cohort study in UK primary care
Author names
Sarah-Jo Sinnott PhD1, Ian J Douglas PhD1, Liam Smeeth PhD1, Elizabeth Williamson PhD2,
Laurie A Tomlinson PhD1
1 Department of Non-Communicable Disease Epidemiology, London School of Hygiene and
Tropical Medicine, Keppel Street, London, WC1E 7HT, United Kingdom2 Department of Medical Statistics, London School of Hygiene and Tropical Medicine, Keppel
Street, London, WC1E 7HT, United Kingdom
Correspondence:
Sarah-Jo Sinnott
Department of Non-Communicable Disease Epidemiology, London School of Hygiene and
Tropical Medicine, Keppel Street, London, WC1E 7HT, United Kingdom
Tables: 2
Figures: 5
Page 1 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
2
Abstract
Objective: To study whether age-based (</≥55yrs) and ethnicity-based (black/non-black) treatment
recommendations in United Kingdom (UK) clinical guidelines for hypertension translate to better
reductions in blood pressure (BP) in current routine clinical care.
Design: Observational cohort study
Setting: UK Primary Care
Participants: New-users of angiotensin converting enzyme inhibitors/angiotensin receptor blockers
(ACEI/ARB), calcium channel blockers (CCB) and thiazide diuretics.
Main outcome measure: Change in systolic blood pressure in new-users of ACEI/ARB versus CCB,
stratified by age and gender. In secondary analyses we compared new-users of CCB versus diuretics.
We used a negative outcome (Herpes Zoster) to detect residual confounding and a positive outcome
(expected drug effects) to show that the study identified expected associations.
Results: We identified 84,440 new-users of ACEI/ARB, 67,274 of CCB and 22,020 of thiazides with a
median of 4 (IQR 2-6) BP measurements during one year follow-up. For non-black, non-diabetic
people <55 years initiating CCB, systolic BP fell by 1.69mmHg (99%-CI -2.52,0.86) more than
ACEI/ARB users at 12 weeks and by 0.40mmHg (99%-CI -0.98,0.18) in those ≥55 years. Across six
finer age categories, we found that CCB were more effective than ACEI/ARB in reducing systolic BP
only in those ≥75 years. Comparing initiators of CCB versus ACEI/ARB, among black people systolic
BP reduction was 2.15mmHg (99%-CI -6.17,1.87) greater with CCB, while for non-black people the
reduction was 0.98mmHg (99%-CI -1.49,-0.47) greater for CCB.
Conclusions: We found similar patterns of effectiveness for CCB versus ACEI/ARB in those aged <55
and ≥55 years. CCB produced numerically greater BP reductions than ACEI/ARB among black people
compared to non-black people, but the confidence intervals for reductions achieved were
overlapping for the two strata. Our results suggest that the current UK algorithmic approach to first
line antihypertensive treatment does not lead to greater BP reductions. Specific indications could be
considered in treatment recommendations.
Page 2 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
3
Summary box
What is already known on this topic
Current NICE guidance for first-line antihypertensive choice is based on age and ethnicity,
with the understanding that people of black African or African-Caribbean family ethnicity
(hereafter referred to as ‘black people’ to reflect diversity) and older people have lower
levels of renin and a salt-sensitive hypertension phenotype. Calcium channel blockers (CCB)
are thus recommended in non-diabetic people aged ≥55 years and black people. Angiotensin
converting enzyme inhibitors/angiotensin receptor blockers (ACEI/ARB) are recommended in
people aged <55 years.
Since the current guidance was developed the medical literature has evolved incrementally,
including older populations in trials and with meta-analysis indicating that the effectiveness
of first-line antihypertensive drugs does not differ by age (</>65 years). Additionally, the use
of categorisations of ethnicity to guide clinical decision making has been questioned.
It is not known if current recommended drug choices lead to greater reductions in BP in
routine clinical care.
What this study adds
This propensity score matched cohort study shows that initiation of either CCB or ACEI/ARB
leads to similar blood pressure (BP) reductions in people aged <55 years and those aged ≥55
years.
In analyses using finer age strata, we found that CCB lead to greater reductions in BP only in
those aged ≥75 years.
Our results for ethnicity indicated numerically greater reductions in BP for black people
initiating CCB versus ACEI/ARB than seen in non-black people but the confidence intervals
overlapped between the two groups.
Page 3 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
4
Introduction
High blood pressure (BP), or hypertension, affects more than one in four adults globally and is a
major modifiable risk factor for morbidity and mortality.1 Guideline based approaches to
pharmacotherapy for hypertension have been adopted internationally to simplify clinical practice
and improve BP control.2-4 Central to the recommendations of these guidelines is the understanding
that the effect of antihypertensive drugs differs among specific populations.
In the United Kingdom (UK), National Institute for Health and Care Excellence (NICE) guidelines
recommend angiotensin-converting enzyme inhibitors and angiotensin receptor blockers (ACEI/ARB)
as first-line treatment for hypertension in people aged <55 years, and calcium channel blockers
(CCBs) for those aged ≥55 years, replacing CCB with thiazide diuretics for those with drug
intolerance.2 The presence of an age-based recommendation is unique amongst major international
guidelines for the treatment of hypertension,3,4 and is based on differences in the activity of the
renin-angiotensin system with age.5-7 Since this threshold was introduced in the first iteration of
NICE hypertension guidance in 2004, the evidence base for hypertension treatment in older age,
including the use and safety of ACEI/ARB drugs in older populations has evolved.8,9
Secondly, in NICE guidelines CCB or thiazides are recommended as first-line treatment for black
people of African or Caribbean family origin (hereafter referred to as ‘black people’ to reflect
diversity), and the same drugs are recommended, after consideration of comorbidities, in
international guidelines.2-4 The pathophysiology of hypertension in this population has been thought
to differ importantly from people of Caucasian heritage; lower levels of renin result in a reduced
antihypertensive response to drugs that block the renin-angiotensin system such as ACEI/ARB.10
However, treatment recommendations based on historical categorisations of ethnicity have recently
been criticised given that ethnicity can be considered a social construct rather than a biological one,
and as the proportion of people with mixed ethnic heritage increases.11
Contemporary routine care is characterised by an increasingly older, more ethnically diverse and
multi-morbid population. For people initiating antihypertensives, it is not known whether current
age- and ethnicity-based treatment recommendations translate to greater BP reductions in these
settings.
The Quality and Outcomes Framework in the UK ensures that BP is regularly measured and recorded
in patients’ electronic health records (EHR) in primary care.12 This, along with complete information
on drugs prescribed, means that these anonymised data are a rich and high quality resource for
examining antihypertensive drug effectiveness. Therefore, framing our questions around the current
Page 4 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
5
NICE algorithm for drug treatment of hypertension, we sought to determine whether initiation of
CCB compared to ACEI/ARB led to differences in BP reduction across age and ethnicity strata.
Page 5 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
6
Methods
Data
We conducted a cohort study, using the Clinical Practice Research Database (CPRD-GOLD) linked to
Hospital Episodes Statistics (HES). CPRD-GOLD is a nationally representative repository of
prospectively collected, anonymised EHRs from primary care in the UK, that has been extensively
validated.13 It holds data on demographic information, health related behaviours, test results
including BP measurements, diagnoses, and prescriptions for more than 11 million people.13 The HES
database records all hospital admissions for patients covered by the National Health Service (NHS)
who receive treatment from either English NHS trusts or independent providers.14 Almost 60% of
general practices included in the CPRD-GOLD are linked to HES.13 We used linked data in this study
to improve completeness of ethnicity recording.15
Cohort entry
We identified new-users of antihypertensive drugs (CCB, ACEI/ARB and thiazides), defined as people
who had a prescription for one of these drugs in the study period 1st January 2007 to 31st December
2017, with no prior use of any antihypertensive in the preceding year. People entered the cohort
(index date) on the date of the first prescription. People were eligible for inclusion from the latest of:
study start date (1st January 2007); one year after GP registration (to allow time for recording of
covariates in the GP record); the date the GP practices’ data recording practices were considered of
adequate standard to be included in CPRD-GOLD; or the persons’ 18th birthday. People were
included until the earliest of: end of the study period (31st December 2017), last data collection from
the GP practice, date of leaving the GP practice, or death.
Exclusion
We sought to determine the effects of drugs on BP solely in those treated for hypertension.
Therefore we excluded people who initiated any of the study drugs without recorded BP
measurements in the year prior to cohort entry, and those whose BP was at target or lower
(<140/90mmHg, according to current NICE guidelines).2 We also excluded people who initiated more
than one antihypertensive drug on the index date.
Outcomes
For our main analysis we investigated the change in systolic BP from baseline at 12 weeks, 26 weeks
and 52 weeks. In a secondary analysis, we repeated this for diastolic BP.
To investigate the reliability of our data and methods in this context we included positive control
analyses; outcomes that we expected to differ according to drug prescribed. These outcomes were
Page 6 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
7
incidence of ankle swelling, gout and angioedema. We also included a negative control outcome;
Herpes Zoster. Our rationale was that none of the antihypertensive drugs studied should be
associated with an altered risk of Herpes Zoster – thus, if an association was found between any
drug group and Herpes Zoster this would suggest unmeasured confounding in our analyses.16
Covariates
Baseline systolic and diastolic BP values were taken from CPRD, using the closest measurement to,
or on, the index date. Using a priori knowledge about factors that could influence treatment choice
or response we defined multiple covariates: age, sex, smoking, alcohol use, body mass index
diabetes, myocardial infarction, stroke, heart failure, arrhythmia, peripheral vascular disease, cancer,
depression and chronic obstructive pulmonary disease. Chronic comorbidities were defined using
records from information recorded in both CPRD and HES using all available data. We determined
whether each person had ever previously been prescribed statins, anti-platelet agents, proton pump
inhibitors, insulin and loop diuretics. To capture polypharmacy, we measured medication usage from
multiple British National Formulary chapters in the year prior to index date. To capture health
service use in the year prior to index date, we determined how often a person visited the GP
surgery. To account for socioeconomic status we used quintiles of person-level deprivation based on
the 2010 English Index Multiple Deprivation scores. We calculated baseline estimated Glomerular
Filtration Rate (eGFR) from the most recent creatinine value recorded in CPRD within one year prior
to index date using the CKD-EPI equation, categorised as follows: no CKD (eGFR≥60); stage 3a (eGFR
45-59); stage 3b (eGFR 30-44); stage 4 (eGFR 15-29); stage 5 (eGFR<15).
To retain power in our main analysis, we imputed missing data under the missing at random
assumption. Data for renal function were approximately 30% missing; our previous work has
demonstrated that people without recorded renal function have similar health outcomes to those
with normal renal function after adjusting for covariates such as age, diabetes and vascular
disease.17 In the imputation model we included all explanatory variables listed above, including the
outcome variable (first systolic and diastolic BP measurement during follow-up). We conducted
diagnostics using the midiagplots function in Stata.18 We combined treatment effects across five
imputed datasets to obtain one overall estimate in each of our analyses.19
Analysis
Reflecting current UK NICE guidance2, we sought to answer two specific research questions in
relation to BP reduction: 1) in a non-diabetic, non-black population does the effectiveness of CCB
versus ACEI/ARB differ in those aged </≥55 years? 2) In a non-diabetic population, does the
effectiveness of CCB versus ACEI/ARB differ by black or non-black ethnicity? We also compared
Page 7 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
8
thiazides versus CCB to give insight into the comparative effectiveness of drugs that are
recommended as alternative choices.2
Therefore, among new-users of ACEI/ARB or CCB, within each of four strata defined by age (</≥55
years) and ethnicity (black/non-black), we estimated the propensity to be prescribed CCB versus
ACEI/ARB using a logistic regression model including all the covariates listed above, except diastolic
BP. Then, within each stratum we matched one CCB user to one ACEI/ARB user on their propensity
score within a caliper of 1%. We used linear mixed models to model change in systolic BP from
baseline at 12, 26 and 52 weeks with a random intercept for each person and fixed effects for
exposure drug, using splines over time to allow the effects of exposure to vary flexibly over time.
Such models allow for correlations within persons for longitudinal BP measurements and also
accommodate the unbalanced nature of BP recordings in the data.20 People without a BP
measurement during one year follow up were dropped from the BP analysis (n=10,840). The same
analysis was used to compare thiazides versus CCB.
We used Cox proportional hazards models for the control outcomes within each of the propensity
score matched strata defined above.
To reduce the risk of type I error with multiple testing we defined a priori that we would report 99%
confidence intervals (CI).
Subgroup and sensitivity analyses
To look more closely at the effectiveness of CCB versus ACEI/ARB and TZD versus CCB in reducing
systolic BP according to age in a non-black and non-diabetic population we used fine age categories
for stratum specific analyses: <55, 55-<60, 60-<65, 65-<70, 70-<75 and ≥75 years.
We explored the effectiveness of thiazide-like diuretics (indapamide and chlortalidone) versus CCB
and also the effectiveness of older thiazide agents versus CCB.
As our main analysis represents an ‘intention-to-treat’ approach, we also carried out an ‘on-
treatment’ analysis whereby we censored each persons’ follow-up BP data if they switched to or
added another antihypertensive or discontinued their drug, defined as no repeat prescription within
90 days of prior prescription. We carried out a complete case analysis to explore the impact of our
multiple imputation approach. Finally, as a post-hoc addition we repeated our main analysis with
95% CI.
Patient involvement
No patients were involved in setting the research question or the outcome measures, nor were they
involved in developing plans for design or implementation of the study. No patients were asked to
Page 8 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
9
advise on interpretation or writing up of results. We are not able to disseminate the results of the
research directly to study participants because the data used were anonymised.
Page 9 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
10
Results
From more than 1 million users of antihypertensive drugs between 2007-2017, we identified 84,440
new users of ACEI/ARB, 67,274 new users of CCB and 22,040 new users of thiazides (Figure 1).
People taking thiazides were more likely to be female and were older than other antihypertensive
users, while people taking ACEI/ARB group had a higher prevalence of myocardial infarction, heart
failure and diabetes. For each drug comparison, we matched people on propensity score and
achieved good balance on all covariates within age and ethnicity strata (Appendix 1 and 2). Within
the matched stratum of people with black ethnicity for each drug comparison, there was a 1mmHg
difference in systolic BP at baseline. There was a median of 4 (IQR 2-6) BP measurements available
during one year follow up for each people.
*Insert Figure 1*
*Insert Table 1*
Change in systolic BP for CCB vs ACEI/ARB by age
In those aged <55 years, systolic BP changed from 162.8mmHg to 140.7mmHg at 12 weeks in new
users of CCB in comparison to a change from 162.7mmHg to 142.2mmHg in new users of ACEI/ARB.
In relative terms, CCB reduced systolic BP by 1.69mmHg (99% CI -2.52, -0.86) more than ACEI/ARB in
those aged <55 years (Table 2). In those aged ≥55 years, systolic BP changed from 163.8mmHg to
143.2mmHg at 12 weeks in new users of CCB in comparison to a change from 164.6mmHg to
144.6mmHg in new users of ACEI/ARB (Figure 2), a relative difference of -0.40mmHg (99% CI -0.98,
0.18) in favour of CCB (Table 2). Over one year follow-up, there was no difference in effectiveness
between CCB and ACEI/ARB for systolic BP in either age group (Table 2 and Figure 2).
*Insert Table 2*
*Insert Figure 2*
Change in systolic BP for CCB vs ACEI/ARB by ethnicity
Page 10 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
11
Among black people, systolic BP changed from 156.0mmHg to 140.1mmHg at 12 weeks in new users
of CCB in comparison to a change from 158.4mmHg to 144.1mmHg in new users of ACEI/ARB. In
relative terms, CCB reduced systolic BP by 2.15mmHg (99% CI -6.17, 1.87) more than ACEI/ARB in
black people (Table 2). Among non-black people, systolic BP changed from 163.6mmHg to
143.5mmHg at 12 weeks in new users of CCB in comparison to a change from 163.7mmHg to
143.8mmHg in new users of ACEI/ARB (Figure 2), a relative difference of 0.98mmHg (99% CI -1.49, -
0.47) greater reduction among CCB initiators. Over one year follow-up, there was no difference in
effectiveness between CCB and ACEI/ARB for either non-black or black people (Table 2 and Figure
2).
We found results of similar magnitude and direction for diastolic BP (Appendices 3- 4)
Change in systolic BP for thiazides versus CCB
For those aged >55 years, CCB were more effective at lowering systolic BP compared to thiazide
diuretics at 12 and 26 weeks. However, the reductions observed were similar to those observed in
those aged <55 years, indicated by overlapping confidence intervals. Similarly, overlapping
confidence intervals for the reductions in each of the ethnicity strata indicated no evidence for a
difference in effectiveness for thiazides versus CCB in black and non-black people (Table 2 and
Figure 3).
*Insert Figure 3*
We observed similar results for diastolic BP (Appendices 3-4).
Control analyses
The incidence of drug specific side effects was as expected for each drug comparison; there was
increased incidence of ankle swelling for new users of CCB in comparison to new users of ACEI/ARB
(HR 1.77, 99% 1.61, 1.94) and reduced ankle swelling for new users of thiazides in comparison to
new users of CCB (HR 0.69, 99% CI 0.61, 0.78) (Figure 4). There was no evidence that the incidence
of Herpes Zoster differed between drug groups in either drug comparison (Figure 4). Stratum
specific negative and positive control analyses are in Appendix 6.
*Insert Figure 4*
Figure 3: Achieved systolic blood pressures for thiazides and CCB in each stratum
Page 11 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
12
Sensitivity analyses
Across all age categories, among non-black, non-diabetic people, we found that the CCB were more
effective than ACEI/ARB at reducing systolic BP only in those aged ≥75 years (12 week change -
3.0mmHg, 99% -4.05, -1.93, 26 weeks change -1.90mmHg, 99% CI -3.03, -0.77 and 52 weeks change -
1.86mmHg, 99% CI -3.08, -0.64) (Figure 5). In the thiazide versus CCB comparison, CCB were more
effective than thiazides in all age categories at 12 weeks, but the difference between the drugs
became negligible with increasing time (Figure 5).
*Insert Figure 5*
In the whole propensity score matched cohort, we found that thiazide-like drugs (indapamide and
chlortalidone) were of similar effectiveness to CCB in terms of lowering systolic BP, and that CCB
were more effective than older thiazide agents (Appendix 7).
Our ‘on treatment’ analysis provided results that were broadly similar to the main analysis. The
numerical reduction in BP was greater among black people commencing CCB than ACEI/ARB (-
3.65mmHg, 99% CI -7.99, 0.70) than we observed in non-black people (-0.35mmHg, 99% CI -0.94,
0.24), although the 99% confidence intervals were overlapping for both strata indicating no strong
evidence for a difference, similar to the main analysis (Appendix 8).
The complete case analysis found similar results to the main analysis (Appendix 9).
Page 12 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
13
Discussion
In this propensity-score matched cohort study of more than 150,000 hypertensive people in routine
care, we found no evidence to support a difference in the effectiveness of CCB versus ACEI/ARB
across those aged above and below 55 years of age. We did, however, find that CCB were more
effective than ACEI/ARB at reducing systolic BP in people older than 75 years. Our results indicated a
numerically greater reduction in systolic BP for CCB versus ACEI/ARB among black people compared
to non-black people but the confidence intervals for the reductions overlapped between the two
strata indicating no evidence of difference.
It is possible that people initiating antihypertensives other than first-line options suggested by
current guidance were prescribed them for other indications. However, we endeavoured to reduce
confounding by indication by: including only people with hypertension; undertaking our analyses
only in non-diabetics; and using propensity-score matching on multiple factors, inclusive of cardiac
failure. This latter approach achieved highly balanced cohorts for each drug comparison within each
stratum. Furthermore, we used multiple strategies to help detect bias in our matched cohorts; our
negative outcome analysis demonstrated no evidence of selection bias or residual confounding by
worse health status given that there was no increased risk of Herpes Zoster in any drug group.16
Additionally, our positive outcome analyses demonstrated that our data and methods were sensitive
to drug effects that we would anticipate to be causally associated with specific antihypertensive
drugs.
Our results related to ethnicity have wide margins of uncertainty which may have occurred for
several reasons. First, we compared changes in BP between black and non-black people, consistent
with current UK guidance.2 Much of the original clinical research that informs this guidance
compared people of African-Caribbean family history with people classified as Caucasian. If people
of non-black, non-Caucasian ethnicity have intermediate responses to the antihypertensive drug
classes, our inclusion of them in the non-black group may have contributed to the overlapping
confidence intervals we observed for non-black and black people. Secondly, we had small numbers
of black people initiating ACEI/ARB as a first line antihypertensive, suggesting that current NICE
guidance is closely adhered to. Third, the uncertainty we observed may be truly reflective of wide
variation in the activity of the renin-angiotensin system across all ethnicities.
Our outcomes relied on clinic measurements of BP, rather than BP measured using ambulatory
devices, which is recommended by NICE for the confirmation of hypertension. However, clinic
measurements of BP are used in primary care to monitor hypertension and to guide treatment
decisions. Therefore, the outcomes we have used in this study reflect the provision of care as it
Page 13 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
14
occurs routinely. Furthermore, we do not expect any differential misclassification of the measured
blood pressure by variation in measurement depending on the drug prescribed.
NICE guidelines recommend ACEI/ARB as first-line treatment for hypertension for those aged <55
years.2 This age threshold, currently absent in international guidelines, is based on the idea that
hypertension in younger people is more commonly characterised by high renin levels.5,6 The age cut-
off contained in NICE guidelines is supported by studies limited by small sample size (<60 people),5,6
short follow up periods (<6 weeks),5,6 restricted to men7 and geographical location,5 and unpublished
data.21 To support drug recommendations in those aged ≥55 years the ASCOT, VALUE and ALLHAT
trials are cited although none demonstrate clear effect modification by age (usually <60/>60 or
<65/>65 age categorisation) for any included clinical outcome.22-24 Our age-stratified results are in
line with a meta-analysis of 31 trials including more than 190,000 patients that demonstrated no
difference in BP reduction or cardiovascular events for any antihypertensive drug class between
young (<65yrs) or older (>65 years) patients.25 Our result of no difference in effectiveness between
CCB and ACEI/ARB in those aged ≥55years also agrees with the results of a randomised study that
examined effectiveness of first line BP-lowering agents across the <55/≥55years age categorisation.26
Since the first inclusion of the </≥55 years age threshold in NICE guidance (CG18) in 2004, the
literature has evolved to capture older populations with subgroup analyses based on older ages
(</>75years) or subgroups defined by frailty.8,9 In our study, greater effectiveness of CCB over
ACEI/ARB was seen only in those aged ≥75 years. This may reflect a higher prevalence of isolated
systolic hypertension in this subgroup (57% in ≥75 years versus 40% in ≥55 years in our data) for
which CCB may be more effective.27,28
International guidelines for hypertension are similar to NICE guidelines in that ethnicity is a major
consideration in choosing a first-line antihypertensive drug, although do not use the algorithmic
approach of UK guidance. Ethnicity-based recommendations are supported by strong evidence from
the ALLHAT randomized clinical trial which demonstrated thiazides or CCBs are more effective at
lowering BP compared to ACEI/ARB in black people.24 We have suggested above a number of
reasons why our results may differ from established clinical evidence. Notwithstanding, the value of
making clinical decisions based on the colour of an individual’s skin is increasingly questioned given
the growing proportion of people who classify themselves as multi-ethnic or of mixed heritage.11
Furthermore, our understanding of gene-environment interactions in the development of
hypertension is now improving.29 Thus, the value of using categorisations of ethnicity as a proxy
predictor for drug response is uncertain in current society, though is likely to be informed by ongoing
research.30
Page 14 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
15
In conclusion, we found that CCB, relative to ACEI/ARB, achieved reductions in systolic BP of similar
magnitude in those aged above and below 55 years. However, in those aged ≥75 years, CCB was
more effective than ACEI/ARB in reducing BP. This suggests that people aged between ≥55 and 75
years who are currently offered CCB, but have indications for renin-angiotensin blockade such as
proteinuria, may be missing the therapeutic benefits of ACEI/ARB. We found no evidence that CCB
was more effective than ACEI/ARB among black people compared to other ethnic groups, although
power was limited and there was uncertainty in the estimate. Our results suggest that the algorithm
for choice of pharmacotherapy in UK NICE guidance could be simplified. Moving towards a choice of
any of the three major antihypertensive drug classes with suggested compelling indications for their
use would align the UK with international guidance, in particular with regard to age.
Page 15 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
16
Author’s contributions
S-JS, LAT, ID, EW and LS conceived and devised the study, S-JS and EW analysed the data. All authors
contributed to the interpretation of the data. S-JS drafted the article and all authors (IJD, EW, LS,
LAT) reviewed and edited the manuscript, and approved the version to be published. The
corresponding author attests that all listed authors meet authorship criteria and that no others
meeting the criteria have been omitted. S-JS is guarantor and accepts full responsibility for the
conduct of the study, had access to the data, and controlled the decision to publish.
Funding
S-JS is funded by a Wellcome Trust Sir Henry Wellcome fellowship (107340/Z/15/Z). LAT is funded by
a Wellcome Trust Intermediate Clinical Fellowship (101143/Z/13/Z). The funders had no role in the
design or conduct of this research.
Competing interests:
IJD is funded by, holds stock in and has consulted for GSK. LS has received grants from GSK and has
received grants from the Wellcome trust, the MRC, the NIHR, the British Heart foundation and
Diabetes UK outside of the submitted work and is a Trustee of the British Heart Foundation. S-JS, LAT
and EW have no relevant conflicts of interest to disclose.
Ethical approval
The study protocol was approved by the London School of Hygiene and Tropical Medicine ethics
committee (No. 16251) and the Independent Scientific Advisory Committee for the Medicines and
Healthcare products Regulatory Agency (No. 18_312R). The protocol was made available to
reviewers as supplementary material.
Transparency declaration
The manuscript’s guarantor (SJS) affirms that the manuscript is an honest, accurate, and transparent
account of the study being reported; that no important aspects of the study have been omitted; and
that any discrepancies from the study as planned have been explained.
Data sharing
No additional data available.
Page 16 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
17
License
“I Sarah-Jo Sinnott the Corresponding Author of this article contained within the original manuscript which includes any diagrams & photographs within and any related or stand alone film submitted (the Contribution”) has the right to grant on behalf of all authors and does grant on behalf of all authors, a licence to the BMJ Publishing Group Ltd and its licencees, to permit this Contribution (if accepted) to be published in the BMJ and any other BMJ Group products and to exploit all subsidiary rights, as set out in our licence set out at: http://www.bmj.com/about-bmj/resources-authors/forms-policies-and-checklists/copyright-open-access-and-permission-reuse.”
Please tick one or more boxes as appropriate:
. I am one author signing on behalf of all co-owners of the Contribution.
Page 17 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
18
1. Olsen MH, Angell SY, Asma S, et al. A call to action and a lifecourse strategy to address the global burden of raised blood pressure on current and future generations: the Lancet Commission on hypertension. The Lancet 2016; 388(10060): 2665-712.2. National Institute for Health and Care Excellence. Hypertension in adults: diagnosis and management.[NG136]. Available from https://www.nice.org.uk/guidance/ng136/evidence. 2019.3. Whelton PK, Carey RM, Aronow WS, et al. 2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA guideline for the prevention, detection, evaluation, and management of high blood pressure in adults: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. Journal of the American College of Cardiology 2018; 71(19): e127-e248.4. Williams B, Mancia G, Spiering W, et al. 2018 ESC/ESH Guidelines for the management of arterial hypertension: The Task Force for the management of arterial hypertension of the European Society of Cardiology and the European Society of HypertensionThe Task Force for the management of arterial hypertension of the European Society of Cardiology and the European Society of Hypertension. Journal of hypertension 2018; 36(10): 1953-2041.5. Dickerson JC, Hingorani AD, Ashby MJ, Palmer CR, Brown MJ. Optimisation of antihypertensive treatment by crossover rotation of four major classes. The Lancet 1999; 353(9169): 2008-13.6. Deary AJ, Schumann AL, Murfet H, Haydock SF, Foo RS-Y, Brown MJ. Double-blind, placebo-controlled crossover comparison of five classes of antihypertensive drugs. Journal of hypertension 2002; 20(4): 771-7.7. Materson BJ, Reda DJ, Cushman WC, et al. Single-Drug Therapy for Hypertension in Men--A Comparison of Six Antihypertensive Agents with Placebo. New England Journal of Medicine 1993; 328(13): 914-21.8. Beckett NS, Peters R, Fletcher AE, et al. Treatment of hypertension in patients 80 years of age or older. New England Journal of Medicine 2008; 358(18): 1887-98.9. SPRINT Research Group. A randomized trial of intensive versus standard blood-pressure control. New England Journal of Medicine 2015; 373(22): 2103-16.10. Sagnella G. Why is plasma renin activity lower in populations of African origin? Journal of human hypertension 2001; 15(1): 17.11. Bonham VL, Green ED, Pérez-Stable EJ. Examining how race, ethnicity, and ancestry data are used in biomedical research. Jama 2018; 320(15): 1533-4.12. Hippisley-Cox J, Vinogradova Y, Coupland C. Time Series Analysis for selected clinical indicators from the Quality and Outcomes Framework 2001-2006. Leeds: The Information Centre for Health and Social Care 2007.13. Herrett E, Gallagher AM, Bhaskaran K, et al. Data Resource Profile: Clinical Practice Research Datalink (CPRD). Int J Epidemiol 2015; 44(3): 827-36.14. Herbert A, Wijlaars L, Zylbersztejn A, Cromwell D, Hardelid P. Data Resource Profile: Hospital Episode Statistics Admitted Patient Care (HES APC). International Journal of Epidemiology 2017; 46(4): 1093-i.15. Mathur R, Bhaskaran K, Chaturvedi N, et al. Completeness and usability of ethnicity data in UK-based primary care and hospital databases. Journal of public health 2013; 36(4): 684-92.16. Lipsitch M, Tchetgen ET, Cohen T. Negative controls: a tool for detecting confounding and bias in observational studies. Epidemiology (Cambridge, Mass) 2010; 21(3): 383.17. Schmidt M, Mansfield KE, Bhaskaran K, et al. Serum creatinine elevation after renin-angiotensin system blockade and long term cardiorenal risks: cohort study. BMJ 2017; 356: j791.18. Eddings W, Marchenko Y. Diagnostics for multiple imputation in Stata. Stata Journal 2012; 12(3): 353.19. Leyrat C, Seaman SR, White IR, et al. Propensity score analysis with partially observed covariates: How should multiple imputation be used? Stat Methods Med Res 2017: 962280217713032.
Page 18 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
19
20. Fitzmaurice G.M, Laird N.M, J.H. W. Applied Longitudinal Analysis (Second Edition): Chapter 8 Linear Mixed Effects Models.: Wiley; 2011.21. ASCOT. From reference number 55 in Full NICE guideline for Hypertension "ASCOT Steering Committee. Age-Stratified Analysis of Blood Pressure Responses. Personal communication: 06". 2006.22. Dahlöf B, Sever PS, Poulter NR, et al. Prevention of cardiovascular events with an antihypertensive regimen of amlodipine adding perindopril as required versus atenolol adding bendroflumethiazide as required, in the Anglo-Scandinavian Cardiac Outcomes Trial-Blood Pressure Lowering Arm (ASCOT-BPLA): a multicentre randomised controlled trial. The Lancet 2005; 366(9489): 895-906.23. Zanchetti A, Julius S, Kjeldsen S, et al. Outcomes in subgroups of hypertensive patients treated with regimens based on valsartan and amlodipine: An analysis of findings from the VALUE trial. Journal of hypertension 2006; 24(11): 2163-8.24. Wright J, Bakris G, Greene T. Major outcomes in high-risk hypertensive patients randomized to angiotensin-converting enzyme inhibitor or calcium channel blocker vs. diuretic. the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). ACC Current Journal Review 2003: 37.25. Blood Pressure Lowering Treatment Trialist's Collaboration. Effects of different regimens to lower blood pressure on major cardiovascular events in older and younger adults: meta-analysis of randomised trials. BMJ 2008; 336(7653): 1121-3.26. Mahmud A, Feely J. Choice of first antihypertensive: simple as ABCD? American journal of hypertension 2007; 20(8): 923-7.27. Staessen JA, Fagard R, Thijs L, et al. Randomised double-blind comparison of placebo and active treatment for older patients with isolated systolic hypertension. The Systolic Hypertension in Europe (Syst-Eur) Trial Investigators. Lancet 1997; 350(9080): 757-64.28. Hansson L, Lindholm LH, Ekbom T, et al. Randomised trial of old and new antihypertensive drugs in elderly patients: cardiovascular mortality and morbidity the Swedish Trial in Old Patients with Hypertension-2 study. Lancet 1999; 354(9192): 1751-6.29. Jones DW, Hall JE. Racial and ethnic differences in blood pressure: biology and sociology. Am Heart Assoc; 2006.30. Mukhtar O, Cheriyan J, Cockcroft JR, et al. A randomized controlled crossover trial evaluating differential responses to antihypertensive drugs (used as mono-or dual therapy) on the basis of ethnicity: The comparIsoN oF Optimal Hypertension RegiMens; part of the Ancestry Informative Markers in HYpertension program—AIM-HY INFORM trial. American heart journal 2018; 204: 102-8.
Page 19 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
20
Table 1: Descriptive characteristics for new-users of ACEI/ARB, CCB and thiazides at baseline.ACEI/ARB
87,440n (%)
CCB 67,274n (%)
Thiazides 22,040n (%)
Females 37158 (42.5) 33506 (49.8) 14277 (64.8)
Age (years)<50 29079 (33.3) 9204 (13.7) 2372 (10.8)50-54 15178 (17.4) 5377 ( 8.0) 1341 ( 6.1)55-59 11190 (12.8) 9522 (14.2) 2302 (10.4)60-64 10215 (11.7) 11686 (17.4) 3394 (15.4)65-69 7795 ( 8.9) 10746 (16.0) 3205 (14.5)70-74 5763 ( 6.6) 8421 (12.5) 3139 (14.2)75+ 8220 ( 9.4) 12318 (18.3) 6287 (28.5)
Systolic BP mean (sd) 160.2 (16.6) 165.8 (18.1) 164.3 (17.2)Diastolic BP mean (sd) 94.8 (11.4) 93.8 (12.1) 91.5 (11.2)
EthnicityWhite 73247 (83.8) 55477 (82.5) 19,253 (87.4)South Asian 2949 (3.4) 1,684 (2.5) 372 (1.7)Black 914 (1.1) 3,096 (4.6) 389 (1.8)Other/mixed/unknown 3,214 (3.7) 2394 (3.6) 707 (3.2)Missing 7116 (8.1) 4623 (6.9) 1319 (6.0))
Index Multiple Deprivation1 (least deprived) 20,174 (23.1) 15,540 (23.1) 40,491 (22.9)2 20,481 (23.4) 15,822 (23.5) 41,637 (23.6)3 17,616 (20.2) 13,457 (20.0) 35,634 (20.2)4 16,301 (18.6) 12,795 (19.0) 33,233 (18.8)5 (most deprived) 12,810 (14.7) 9,631 (14.3) 25,655 (14.5)Missing 58 (0.1) 29 (0.04) 17 (0.1)
Body Mass IndexUnderweight <18.5 646 ( 0.7) 863 ( 1.3) 361 ( 1.6)Healthy weight 18.5-24.9 17317 (19.8) 17609 (26.2) 6152 (27.9)Overweight 25-29.9 31198 (35.7) 25329 (37.7) 7755 (35.2)Obesity ≥30 35119 (40.2) 20498 (30.5) 6841 (31.0)Missing 3160 ( 3.6) 2975 ( 4.4) 931 ( 4.2)
SmokingNon-smoker 29956 (34.3) 23302 (34.6) 7984 (36.2)Current smoker 18183 (20.8) 12632 (18.8) 3830 (17.4)Ex-smoker 39206 (44.8) 31168 (46.3) 10171 (46.1)Missing 95 ( 0.1) 172 ( 0.3) 55 ( 0.2)
AlcoholNon-drinker 7929 ( 9.1) 7060 (10.5) 2663 (12.1)Current drinker 67932 (77.7) 50627 (75.3) 16236 (73.7)Ex-drinker 7332 ( 8.4) 6197 ( 9.2) 2093 ( 9.5)Missing 4247 ( 4.9) 3390 ( 5.0) 1048 ( 4.8)
ComorbiditiesMyocardial Infarction 2489 ( 2.8) 449 ( 0.7) 158 ( 0.7)Stroke 2671 ( 3.1) 2199 ( 3.3) 785 ( 3.6)Heart Failure 1122 ( 1.3) 294 ( 0.4) 114 ( 0.5)Peripheral Vascular Disease 2427 ( 2.8) 2872 ( 4.3) 896 ( 4.1)Diabetes 15095 (17.3) 4289 ( 6.4) 975 ( 4.4)Depression 6677 ( 7.6) 4756 ( 7.1) 1547 ( 7.0)
Page 20 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
21
COPD 3141 ( 3.6) 3748 ( 5.6) 1250 ( 5.7)Cancer 6455 ( 7.4) 7822 (11.6) 2729 (12.4)Herpes Zoster 4872 ( 5.6) 5067 ( 7.5) 1939 ( 8.8)Gout 4365 ( 5.0) 3207 ( 4.8) 645 ( 2.9)Angioedema 61 ( 0.1) 67 ( 0.1) 5 ( 0.0)
CKDNo CKD 57963 (66.3) 37990 (56.5) 11824 (53.6)Stage 3a 6132 ( 7.0) 4726 ( 7.0) 2126 ( 9.6)Stage 3b 1541 ( 1.8) 999 ( 1.5) 393 ( 1.8)Stage 4 218 ( 0.2) 251 ( 0.4) 44 ( 0.2)Stage 5 28 ( 0.0) 96 ( 0.1) ~Missing 21558 (24.7) 23212 (34.5) 7651 (34.7)
DrugsAntiplatelet agents 11868 (13.6) 9590 (14.3) 3699 (16.8)Statins 19769 (22.6) 13740 (20.4) 4232 (19.2)Proton Pump Inhibitors 29775 (34.1) 26374 (39.2) 8139 (36.9)Insulin 2352 ( 2.7) 422 ( 0.6) 84 ( 0.4)Loop diuretics 3109 ( 3.6) 2813 ( 4.2) 1607 ( 7.3)NSAIDs 56108 (64.2) 44292 (65.8) 14663 (66.5)
Primary Care Consultations<5 consultations 22829 (26.1) 17330 (25.8) 5563 (25.2)5-9 consultations 28577 (32.7) 20855 (31.0) 7237 (32.8)10-14 consultations 16106 (18.4) 12243 (18.2) 4121 (18.7)15-19 consultations 8091 ( 9.3) 6591 ( 9.8) 2081 ( 9.4)≥20 consultations 11837 (13.5) 10255 (15.2) 3038 (13.8)Missing 0 ( 0.0) 0 ( 0.0) 0 ( 0.0)
Page 21 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
22
Table 2: Difference in systolic blood pressure since drug initiation for CCB versus ACEI/ARB and thiazide versus CCB.
Difference between CCB and ACEI/ARB for change in systolic BP(mmHg) since baseline (99% CI)
n 12 weeks 26 weeks 52 weeks
Age <55* 20,964 -1.69 (-2.52, -0.86) -0.48 (-1.47, 0.51) -0.53 (-1.96, 0.91)Age≥55* 58,396 -0.40 (-0.98, 0.18) 0.63 (0.06, 1.21) 0.85 (-0.11, 1.81)
Non-black** 73,726 -0.98 (-1.49, -0.47) 0.11 (-0.42, 0.64) 0.08 (-0.81, 0.96)
Black** 894 -2.15 (-6.17, 1.87) 0.55 (-3.39, 4.49) 2.28 (-5.36, 9.92)
Difference between thiazide and CCB for change in systolic BP(mmHg) since baseline (99% CI)
Age <55* 5,724 1.51 (-0.03, 3.06) 0.07 (-1.70, 1.83) 0.34 (-2.39, 3.07)Age≥55* 32,464 2.16 (1.35, 2.96) 1.43 (0.56, 2.31) 0.17 (-0.95, 1.30)
Non-black** 35,876 2.10 (1.37, 2.82) 1.19 (0.40, 2.0) 0.32 (-0.74, 1.38)
Black** 648 -0.28 (-5.69, 5.14) -1.75 (-6.98, 3.48) 2.01 (-5.46, 9.47)
*In a non-black, non-diabetic population** In a non-diabetic populationAll analyses are 1:1 PS matched within each stratum. Each n consists of half CCB and half ACEI/ARB (or half thiazide and half CCB). For the CCB vs ACEI/ARB comparison a negative result means CCB produced larger reductions in systolic BP and a positive results indicates that ACEI/ARB produced larger reductions in systolic BP. For the thiazide vs CCB comparison a negative result means thiazides produced larger reductions in systolic BP and a positive results indicates that CCB produced larger reductions in systolic BP. Changes in systolic BP with 95% CI are in Appendix 5.
Page 22 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
All antihypertensive drug users in CPRD eligible
for HES linkage and eligible for study inclusion,
2007 – 2017
N= 1,034,036
Prevalent users
ACE/ARB = 615,740 CCB = 454,499
Thiazide = 310,552
New-users
ACE/ARB = 120,722 CCB = 91,432
Thiazide = 31,473
New-users
ACE/ARB = 120,722 CCB = 91,432
Thiazide = 31,473
ACE/ARB = 116,601 CCB = 87,934
Thiazide = 29,059
Remove people who initiate more than
one agent on index date:
ACEI/ARB + CCB = 2,727 ACEI/ARB + thiazide = 1,667
CCB + thiazide = 1,027 ACE + CCB + thiazide = 0
Single drug new-users
ACE/ARB = 114,463 CCB = 85,346
Thiazide = 29,053
Remove second instances of new-use:
ACEI/ARB= 2,138 CCB = 2,588 Thiazide= 6
Number of new users with BP readings in year prior to cohort entry
ACEI/ARB =106,948
CCB = 77,816 Thiazide = 25,956
Number of new users with high BP readings in
year prior to cohort entry
ACEI/ARB =87,440 CCB = 67,274
Thiazide = 22,040
Page 23 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
Page 24 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
Page 25 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
Page 26 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
Page 27 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
1
List of Appendices
Appendix 1: Descriptive characteristics and propensity score distributions before and after
matching, CCB vs ACE-I/ARB .......................................................................................................... 2
Appendix 2: Descriptive characteristics and propensity score distributions before and after
matching, thiazides vs CCB............................................................................................................. 5
Appendix 3: Diastolic blood pressure results .................................................................................. 8
Appendix 4: Diagrams of diastolic blood pressure results ............................................................... 9
Appendix 5: 95% CI for changes in systolic blood pressure ........................................................... 11
Appendix 6: Negative and positive outcomes in age-based and ethnicity-based strata ................. 12
Appendix 7: Comparison of thiazide-like diuretics with calcium channel blockers ........................ 14
Appendix 8: On treatment analysis .............................................................................................. 15
Appendix 9: Complete case analysis ............................................................................................. 16
Appendix 10: RECORD-PE checklist .............................................................................................. 17
Page 28 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
2
Appendix 1: Descriptive characteristics and propensity score distributions before and after matching, CCB vs ACE-I/ARB
Appendix 1 Table : Descriptive characteristics for initiators of calcium channel blockers (CCB) versus initiators of ACE-inhibitors/angiotensin receptor blockers (ACE-I/ARB) in matched stratum specific cohorts
Aged <55yrs Aged ≥55yrs Non-Black Black
ACE-I/ARB CCB ACE-I/ARB CCB ACE-I/ARB CCB ACE-I/ARB CCB
10482 10482 29198 29198 36863 36863 447 447
Females 4620 (44.1) 4675 (44.6) 13768 (47.2) 13549 (46.4) 17249 (46.8) 17380 (47.1) 251 (56.2) 250 (55.9)
Age (years)
<50 6256 (59.7) 6278 (59.9) 0 ( 0.0) 0 ( 0.0) 5723 (15.5) 5723 (15.5) 258 (57.7) 258 (57.7)
50-54 4226 (40.3) 4204 (40.1) 0 ( 0.0) 0 ( 0.0) 3801 (10.3) 3801 (10.3) 75 (16.8) 75 (16.8)
55-59 0 ( 0.0) 0 ( 0.0) 7403 (25.4) 7403 (25.4) 6669 (18.1) 6669 (18.1) 42 ( 9.4) 42 ( 9.4)
60-64 0 ( 0.0) 0 ( 0.0) 7229 (24.8) 7229 (24.8) 6591 (17.9) 6591 (17.9) 22 ( 4.9) 22 ( 4.9)
65-69 0 ( 0.0) 0 ( 0.0) 5417 (18.6) 5417 (18.6) 5055 (13.7) 5055 (13.7) 20 ( 4.5) 20 ( 4.5)
70-74 0 ( 0.0) 0 ( 0.0) 3818 (13.1) 3818 (13.1) 3646 ( 9.9) 3646 ( 9.9) 19 ( 4.3) 19 ( 4.3)
75+ 0 ( 0.0) 0 ( 0.0) 5331 (18.3) 5331 (18.3) 5378 (14.6) 5378 (14.6) 11 ( 2.5) 11 ( 2.5)
Ethnicity**
White 8627 (82.3) 8247 (78.7) 25797 (88.4) 25593 (87.7) 34686 (94.1) 34172 (92.7) 0 ( 0.0) 0 ( 0.0)
South asian 415 ( 4.0) 588 ( 5.6) 442 ( 1.5) 515 ( 1.8) 827 ( 2.2) 1077 ( 2.9) 0 ( 0.0) 0 ( 0.0)
Black 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 447 (100.0) 447 (100.0)
Other/mixed/unknown 421 ( 4.0) 673 ( 6.4) 893 ( 3.1) 925 ( 3.2) 1350 ( 3.7) 1614 ( 4.4) 0 ( 0.0) 0 ( 0.0)
Missing 1019 (9.72) 974 (9.29) 2066 (7.08) 2165 (7.41) 0 (0.00) 0 (0.00) 0 (0.00) 0 (0.00)
BMI**
Underweight <18.5 71 ( 0.7) 67 ( 0.6) 348 ( 1.2) 323 ( 1.1) 386 ( 1.0) 378 ( 1.0) 0 ( 0.0) 2 ( 0.4)
Healthy weight 18.5-24.9 1893 (18.1) 1857 (17.7) 7863 (26.9) 7758 (26.6) 8969 (24.3) 8975 (24.3) 77 (17.2) 81 (18.1)
Overweight 25-29.9 3402 (32.5) 3438 (32.8) 11505 (39.4) 11653 (39.9) 13976 (37.9) 13917 (37.8) 178 (39.8) 179 (40.0)
Obesity >=30 4718 (45.0) 4663 (44.5) 8260 (28.3) 8267 (28.3) 12119 (32.9) 12093 (32.8) 178 (39.8) 178 (39.8)
Missing 398 ( 3.8) 457 ( 4.4) 1222 ( 4.2) 1197 ( 4.1) 1413 ( 3.8) 1500 ( 4.1) 14 ( 3.1) 7 ( 1.6)
SMOKING**
Non-smoker 3874 (37.0) 3835 (36.6) 9774 (33.5) 9785 (33.5) 12379 (33.6) 12557 (34.1) 250 (55.9) 246 (55.0)
Current smoker 2740 (26.1) 2756 (26.3) 4931 (16.9) 4823 (16.5) 7122 (19.3) 7087 (19.2) 64 (14.3) 58 (13.0)
Ex-smoker 3859 (36.8) 3878 (37.0) 14458 (49.5) 14511 (49.7) 17328 (47.0) 17136 (46.5) 133 (29.8) 143 (32.0)
Missing 9 ( 0.1) 13 ( 0.1) 35 ( 0.1) 79 ( 0.3) 34 ( 0.1) 83 ( 0.2) 0 ( 0.0) 0 ( 0.0)
ALCOHOL**
Non-drinker 1055 (10.1) 1028 ( 9.8) 2499 ( 8.6) 2475 ( 8.5) 3202 ( 8.7) 3129 ( 8.5) 124 (27.7) 120 (26.8)
Current drinker 8094 (77.2) 8086 (77.1) 22763 (78.0) 23019 (78.8) 28856 (78.3) 28885 (78.4) 264 (59.1) 262 (58.6)
Ex-drinker 822 ( 7.8) 799 ( 7.6) 2391 ( 8.2) 2362 ( 8.1) 2978 ( 8.1) 3099 ( 8.4) 38 ( 8.5) 48 (10.7)
Missing 511 ( 4.9) 569 ( 5.4) 1545 ( 5.3) 1342 ( 4.6) 1827 ( 5.0) 1750 ( 4.7) 21 ( 4.7) 17 ( 3.8)
Comorbidities
Myocardial Infarction 14 ( 0.1) 16 ( 0.2) 326 ( 1.1) 335 ( 1.1) 371 ( 1.0) 347 ( 0.9) 1 ( 0.2) 0 ( 0.0)
Stroke 157 ( 1.5) 159 ( 1.5) 1225 ( 4.2) 1325 ( 4.5) 1438 ( 3.9) 1461 ( 4.0) 6 ( 1.3) 8 ( 1.8)
Heart Failure 11 ( 0.1) 11 ( 0.1) 205 ( 0.7) 215 ( 0.7) 267 ( 0.7) 225 ( 0.6) 2 ( 0.4) 3 ( 0.7)
Peripheral Vascular Disease 119 ( 1.1) 121 ( 1.2) 1159 ( 4.0) 1161 ( 4.0) 1279 ( 3.5) 1208 ( 3.3) 2 ( 0.4) 6 ( 1.3)
Diabetes 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0)
depression 1022 ( 9.8) 994 ( 9.5) 1746 ( 6.0) 1684 ( 5.8) 2658 ( 7.2) 2674 ( 7.3) 23 ( 5.1) 34 ( 7.6)
COPD 168 ( 1.6) 170 ( 1.6) 1645 ( 5.6) 1594 ( 5.5) 1812 ( 4.9) 1729 ( 4.7) 5 ( 1.1) 8 ( 1.8)
Cancer 380 ( 3.6) 376 ( 3.6) 3581 (12.3) 3578 (12.3) 3889 (10.5) 3819 (10.4) 14 ( 3.1) 21 ( 4.7)
Herpes Zoster 391 ( 3.7) 359 ( 3.4) 2312 ( 7.9) 2345 ( 8.0) 2553 ( 6.9) 2575 ( 7.0) 10 ( 2.2) 10 ( 2.2)
Gout 407 ( 3.9) 409 ( 3.9) 1555 ( 5.3) 1587 ( 5.4) 1864 ( 5.1) 1882 ( 5.1) 11 ( 2.5) 9 ( 2.0)
Angioedema 16 ( 0.2) 17 ( 0.2) 15 ( 0.1) 11 ( 0.0) 25 ( 0.1) 27 ( 0.1) 0 ( 0.0) 1 ( 0.2)
CKD**
no CKD 7041 (67.2) 5619 (53.6) 17340 (59.4) 15650 (53.6) 22626 (61.4) 19675 (53.4) 271 (60.6) 238 (53.2)
stage 3a 156 ( 1.5) 139 ( 1.3) 3556 (12.2) 3166 (10.8) 3550 ( 9.6) 3260 ( 8.8) 30 ( 6.7) 31 ( 6.9)
Page 29 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
3
stage 3b 32 ( 0.3) 35 ( 0.3) 784 ( 2.7) 681 ( 2.3) 852 ( 2.3) 724 ( 2.0) 4 ( 0.9) 5 ( 1.1)
stage 4 15 ( 0.1) 9 ( 0.1) 121 ( 0.4) 93 ( 0.3) 131 ( 0.4) 112 ( 0.3) 2 ( 0.4) 2 ( 0.4)
stage 5 5 ( 0.0) 14 ( 0.1) 13 ( 0.0) 33 ( 0.1) 13 ( 0.0) 45 ( 0.1) 1 ( 0.2) 1 ( 0.2)
Missing 3233 (30.8) 4666 (44.5) 7384 (25.3) 9575 (32.8) 9691 (26.3) 13047 (35.4) 139 (31.1) 170 (38.0)
Drugs
Antiplatelet agents 516 ( 4.9) 496 ( 4.7) 4625 (15.8) 4731 (16.2) 4996 (13.6) 5084 (13.8) 26 ( 5.8) 29 ( 6.5)
Statins 742 ( 7.1) 751 ( 7.2) 6456 (22.1) 6542 (22.4) 7038 (19.1) 6967 (18.9) 31 ( 6.9) 34 ( 7.6)
Proton Pump Inhibitors 3270 (31.2) 3213 (30.7) 11057 (37.9) 10853 (37.2) 13830 (37.5) 13754 (37.3) 149 (33.3) 156 (34.9)
Insulin 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0)
Loop diuretics 216 ( 2.1) 194 ( 1.9) 1331 ( 4.6) 1332 ( 4.6) 1481 ( 4.0) 1584 ( 4.3) 9 ( 2.0) 12 ( 2.7)
NSAIDs 6478 (61.8) 6537 (62.4) 19290 (66.1) 19380 (66.4) 24392 (66.2) 24547 (66.6) 286 (64.0) 303 (67.8)
Primary Care Consultations**
<5 consultations 3351 (32.0) 3370 (32.2) 7734 (26.5) 7794 (26.7) 9633 (26.1) 9639 (26.1) 128 (28.6) 130 (29.1)
5-9 consultations 3045 (29.0) 3055 (29.1) 9925 (34.0) 10039 (34.4) 12300 (33.4) 12284 (33.3) 139 (31.1) 117 (26.2)
10-14 consultations 1687 (16.1) 1654 (15.8) 5268 (18.0) 5269 (18.0) 6736 (18.3) 6834 (18.5) 75 (16.8) 87 (19.5)
15-19 consultations 800 ( 7.6) 823 ( 7.9) 2598 ( 8.9) 2521 ( 8.6) 3357 ( 9.1) 3308 ( 9.0) 49 (11.0) 57 (12.8)
>=20 consultations 1599 (15.3) 1580 (15.1) 3673 (12.6) 3575 (12.2) 4837 (13.1) 4798 (13.0) 56 (12.5) 56 (12.5)
Missing 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0)
Systolic BP mean (sd) 162.7 (17.2) 162.8 (19.2) 164.6 (16.8) 163.8 (16.2) 163.7 (17.0) 163.6 (17.1) 158.4 (15.4) 156.0 (14.5)
Diastolic BP mean (sd) 100.4 (10.0) 101.0 (11.3) 92.4 (11.2) 92.5 (11.2) 94.2 (11.5) 94.5 (11.9) 96.7 (10.2) 96.9 ( 9.9)
Page 30 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
4
Before Matching After Matching
<55years
≥55years
Non-Black
Black
02
46
8
kde
nsity p
sco
re_im
p
0 .2 .4 .6 .8propensity score
_ACE/ARB _CCB
02
46
8
kde
nsity p
sco
re_im
p
0 .2 .4 .6 .8propensity score
_ACE/ARB _CCB
01
23
4
kde
nsity p
sco
re_im
p
0 .2 .4 .6 .8 1propensity score
_ACE/ARB _CCB
01
23
4
kde
nsity p
sco
re_im
p
0 .2 .4 .6 .8 1propensity score
_ACE/ARB _CCB
01
23
kde
nsity p
sco
re_im
p
0 .2 .4 .6 .8 1propensity score
_ACE/ARB _CCB
0.5
11.5
22.5
kde
nsity p
sco
re_im
p
0 .2 .4 .6 .8propensity score
_ACE/ARB _CCB
02
46
8
kde
nsity p
sco
re_im
p
0 .2 .4 .6 .8 1propensity score
_ACE/ARB _CCB
01
23
45
kde
nsity p
sco
re_im
p
.4 .6 .8 1propensity score
_ACE/ARB _CCB
Page 31 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
5
Appendix 2: Descriptive characteristics and propensity score distributions before and
after matching, thiazides vs CCB
Appendix 2 Table: Descriptive characteristics for thiazides versus calcium channel blockers (CCB) in matched stratum specific cohorts
Aged <55yrs Aged ≥55yrs Non-Black Black
CCB Thiazides CCB Thiazides CCB Thiazides CCB Thiazides
2862 2862 16232 16232 17938 17938 324 324
Females 1881 (65.7) 1902 (66.5) 10392 (64.0) 10414 (64.2) 11708 (65.3) 11636 (64.9) 222 (68.5) 223 (68.8)
Age (years)
<50 1751 (61.2) 1778 (62.1) 0 ( 0.0) 0 ( 0.0) 1616 ( 9.0) 1616 ( 9.0) 161 (49.7) 161 (49.7)
50-54 1111 (38.8) 1084 (37.9) 0 ( 0.0) 0 ( 0.0) 985 ( 5.5) 985 ( 5.5) 54 (16.7) 54 (16.7)
55-59 0 ( 0.0) 0 ( 0.0) 2048 (12.6) 2048 (12.6) 1861 (10.4) 1861 (10.4) 32 ( 9.9) 32 ( 9.9)
60-64 0 ( 0.0) 0 ( 0.0) 3035 (18.7) 3035 (18.7) 2795 (15.6) 2795 (15.6) 19 ( 5.9) 19 ( 5.9)
65-69 0 ( 0.0) 0 ( 0.0) 2879 (17.7) 2879 (17.7) 2708 (15.1) 2708 (15.1) 26 ( 8.0) 26 ( 8.0)
70-74 0 ( 0.0) 0 ( 0.0) 2819 (17.4) 2819 (17.4) 2676 (14.9) 2676 (14.9) 14 ( 4.3) 14 ( 4.3)
75+ 0 ( 0.0) 0 ( 0.0) 5451 (33.6) 5451 (33.6) 5297 (29.5) 5297 (29.5) 18 ( 5.6) 18 ( 5.6)
Ethnicity**
White 2304 (80.5) 2322 (81.1) 14478 (89.2) 14689 (90.5) 16833 (93.8) 17023 (94.9) 0 ( 0.0) 0 ( 0.0)
South asian 142 ( 5.0) 126 ( 4.4) 280 ( 1.7) 178 ( 1.1) 444 ( 2.5) 303 ( 1.7) 0 ( 0.0) 0 ( 0.0)
Black 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 324 (100.0) 324 (100.0)
Other/mixed/unknown 187 ( 6.5) 149 ( 5.2) 459 ( 2.8) 463 ( 2.9) 661 ( 3.7) 612 ( 3.4) 0 ( 0.0) 0 ( 0.0)
Missing 229 (8.00) 265 (9.26) 1015 (6.25) 902 (5.56) 0 (0.00) 0 (0.00) 0 (0.00) 0 (0.00)
BMI**
Underweight <18.5 16 ( 0.6) 14 ( 0.5) 319 ( 2.0) 302 ( 1.9) 309 ( 1.7) 308 ( 1.7) ~ ~
Healthy weight 18.5-24.9 454 (15.9) 462 (16.1) 4983 (30.7) 5000 (30.8) 5109 (28.5) 5157 (28.7) 75 (23.1) 54 (16.7)
Overweight 25-29.9 842 (29.4) 856 (29.9) 5936 (36.6) 6034 (37.2) 6472 (36.1) 6473 (36.1) 80 (24.7) 92 (28.4)
Obesity >=30 1431 (50.0) 1449 (50.6) 4251 (26.2) 4219 (26.0) 5267 (29.4) 5300 (29.5) 160 (49.4) 170 (52.5)
Missing 119 ( 4.2) 81 ( 2.8) 743 ( 4.6) 677 ( 4.2) 781 ( 4.4) 700 ( 3.9) 7 ( 2.2) 7 ( 2.2)
SMOKING**
Non-smoker 1074 (37.5) 1086 (37.9) 5642 (34.8) 5755 (35.5) 6255 (34.9) 6354 (35.4) 183 (56.5) 185 (57.1)
Current smoker 700 (24.5) 723 (25.3) 2548 (15.7) 2543 (15.7) 3048 (17.0) 3073 (17.1) 46 (14.2) 50 (15.4)
Ex-smoker 1085 (37.9) 1050 (36.7) 7993 (49.2) 7895 (48.6) 8575 (47.8) 8470 (47.2) 95 (29.3) 89 (27.5)
Missing 3 ( 0.1) 3 ( 0.1) 49 ( 0.3) 39 ( 0.2) 60 ( 0.3) 41 ( 0.2) 0 ( 0.0) 0 ( 0.0)
ALCOHOL**
Non-drinker 268 ( 9.4) 304 (10.6) 1771 (10.9) 1863 (11.5) 2033 (11.3) 2072 (11.6) 111 (34.3) 102 (31.5)
Current drinker 2221 (77.6) 2212 (77.3) 12029 (74.1) 12090 (74.5) 13250 (73.9) 13377 (74.6) 178 (54.9) 177 (54.6)
Ex-drinker 226 ( 7.9) 228 ( 8.0) 1588 ( 9.8) 1551 ( 9.6) 1729 ( 9.6) 1719 ( 9.6) 22 ( 6.8) 32 ( 9.9)
Missing 147 ( 5.1) 118 ( 4.1) 844 ( 5.2) 728 ( 4.5) 926 ( 5.2) 770 ( 4.3) 13 ( 4.0) 13 ( 4.0)
Comorbidities
Myocardial Infarction ~ ~ 121 ( 0.7) 130 ( 0.8) 133 ( 0.7) 130 ( 0.7) ~ ~
Stroke 30 ( 1.0) 27 ( 0.9) 642 ( 4.0) 637 ( 3.9) 646 ( 3.6) 660 ( 3.7) ~ ~
Heart Failure ~ ~ 83 ( 0.5) 90 ( 0.6) 104 ( 0.6) 92 ( 0.5) ~ ~
Peripheral Vascular Disease 26 ( 0.9) 25 ( 0.9) 707 ( 4.4) 711 ( 4.4) 768 ( 4.3) 727 ( 4.1) ~ ~
Diabetes 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0)
Depression 294 (10.3) 301 (10.5) 1027 ( 6.3) 986 ( 6.1) 1213 ( 6.8) 1251 ( 7.0) 7 ( 2.2) 13 ( 4.0)
COPD 53 ( 1.9) 45 ( 1.6) 1011 ( 6.2) 1013 ( 6.2) 1067 ( 5.9) 1040 ( 5.8) 6 ( 1.9) ~
Cancer 102 ( 3.6) 105 ( 3.7) 2242 (13.8) 2267 (14.0) 2365 (13.2) 2330 (13.0) 9 ( 2.8) 13 ( 4.0)
Herpes Zoster 124 ( 4.3) 115 ( 4.0) 1618 (10.0) 1595 ( 9.8) 1659 ( 9.2) 1623 ( 9.0) 8 ( 2.5) 8 ( 2.5)
Gout 51 ( 1.8) 54 ( 1.9) 495 ( 3.0) 497 ( 3.1) 549 ( 3.1) 529 ( 2.9) 6 ( 1.9) 5 ( 1.5)
Angioedema ~ ~ ~ ~ ~ ~ ~ ~
CKD**
no CKD 1586 (55.4) 1626 (56.8) 8828 (54.4) 8623 (53.1) 9731 (54.2) 9601 (53.5) 170 (52.5) 166 (51.2)
stage 3a 31 ( 1.1) 36 ( 1.3) 1653 (10.2) 1850 (11.4) 1634 ( 9.1) 1811 (10.1) 14 ( 4.3) 17 ( 5.2)
stage 3b ~ ~ 333 ( 2.1) 319 ( 2.0) 311 ( 1.7) 319 ( 1.8) ~ ~
stage 4 0 ( 0.0) 0 ( 0.0) 38 ( 0.2) 33 ( 0.2) 39 ( 0.2) 33 ( 0.2) ~ ~
Page 32 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
6
stage 5 0 ( 0.0) ~ 7 ( 0.0) ~ 12 ( 0.1) ~ ~ ~
Missing 1244 (43.5) 1198 (41.9) 5373 (33.1) 5406 (33.3) 6211 (34.6) 6172 (34.4) 138 (42.6) 136 (42.0)
Drugs
Antiplatelet agents 140 ( 4.9) 132 ( 4.6) 3001 (18.5) 2944 (18.1) 3002 (16.7) 3026 (16.9) 19 ( 5.9) 21 ( 6.5)
Statins 173 ( 6.0) 168 ( 5.9) 3296 (20.3) 3319 (20.4) 3364 (18.8) 3368 (18.8) 27 ( 8.3) 26 ( 8.0)
Proton Pump Inhibitors 846 (29.6) 845 (29.5) 6183 (38.1) 6190 (38.1) 6932 (38.6) 6838 (38.1) 83 (25.6) 79 (24.4)
Insulin 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) ~ ~
Loop diuretics 132 ( 4.6) 120 ( 4.2) 1167 ( 7.2) 1131 ( 7.0) 1257 ( 7.0) 1234 ( 6.9) 7 ( 2.2) 11 ( 3.4)
NSAIDs 1799 (62.9) 1820 (63.6) 11103 (68.4) 10872 (67.0) 12221 (68.1) 12092 (67.4) 195 (60.2) 212 (65.4)
Primary Care Consultations**
<5 consultations 779 (27.2) 799 (27.9) 4069 (25.1) 4068 (25.1) 4357 (24.3) 4380 (24.4) 113 (34.9) 103 (31.8)
5-9 consultations 1008 (35.2) 1003 (35.0) 5494 (33.8) 5438 (33.5) 6064 (33.8) 6037 (33.7) 95 (29.3) 96 (29.6)
10-14 consultations 477 (16.7) 462 (16.1) 3000 (18.5) 3076 (19.0) 3414 (19.0) 3404 (19.0) 61 (18.8) 62 (19.1)
15-19 consultations 239 ( 8.4) 215 ( 7.5) 1577 ( 9.7) 1563 ( 9.6) 1729 ( 9.6) 1741 ( 9.7) 17 ( 5.2) 20 ( 6.2)
>=20 consultations 359 (12.5) 383 (13.4) 2092 (12.9) 2087 (12.9) 2374 (13.2) 2376 (13.2) 38 (11.7) 43 (13.3)
Missing 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0) 0 ( 0.0)
Systolic BP 156.9 (17.3) 157.4 (16.4) 166.7 (17.5) 166.7 (16.7) 165.3 (17.7) 165.2 (16.9) 158.5 (17.4) 159.6 (16.2)
Diastolic BP 99.0 (10.6) 97.5 ( 9.8) 91.0 (11.6) 90.8 (11.0) 92.2 (11.8) 91.6 (11.1) 96.9 (10.2) 97.7 (10.1)
Page 33 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
7
Before Matching After Matching
<55years
≥55years
Non-Black
Black
Page 34 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
8
Appendix 3: Diastolic blood pressure results
Appendix 3 Table: Difference in diastolic blood pressure since index date for CCB versus ACE-I/ARB and thiazide versus CCB.
Difference in Diastolic BP between CCB and ACE-I/ARB (99% CI)
n 12 weeks 26 weeks 52 weeks
Age <55* 20,964 -0.98 (-1.39, -0.56) -0.28 (-0.79, 0.22) 0.11 (-0.76, 0.97)
Age>55* 58,396 -1.06 (-1.39, -0.72) -0.14 (-0.47, 0.18) 0.01 (-0.47, 0.49)
Non-black** 73,726 -1.06 (-1.27, -0.85) -0.20 (-0.43, 0.02) -0.02 (-0.44, 0.40)
Black** 894 -2.46 (-4.11, -0.82) -0.96 (-3.40, 1.49) 0.99 (-2.48, 4.46)
Difference in Diastolic BP between thiazide and CCB (99% CI)
Age <55* 5,724 1.29 (0.40, 2.18) 0.62 (-0.22, 1.45) 0.83 (-0.75, 2.42)
Age>55* 32,464 1.80 (1.44, 2.15) 1.36 (1.00, 1.71) 0.80 (0.15, 1.45)
Non-black** 35,876 1.75 (1.41, 2.08) 1.24 (0.93, 1.56) 0.84 (0.29, 1.40)
Black** 648 0.82 (-1.11, 2.74) -0.26 (-2.86, 2.34) 1.92 (-3.04, 6.89)
*In a non-black, non-diabetic population ** In a non-diabetic population All analyses are 1:1 PS matched within each stratum. Each n consists of half CCB and half ACEI/ARB (or half thiazide and half CCB). For the CCB vs ACEI/ARB comparison a negative result means CCB produced larger reductions in systolic BP and a positive results indicates that ACEI/ARB produced larger reductions in systolic BP. For the thiazide vs CCB comparison a negative result means thiazides produced larger reductions in systolic BP and a positive results indicates that CCB produced larger reductions in systolic BP.
Page 35 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
9
Appendix 4: Diagrams of diastolic blood pressure results
Appendix 4 Figure 1: Achieved diastolic blood pressures for calcium channel blocker (CCB) and ACE-
inhibitor/angiotensin receptor blocker (ACE-I/ARB) in each stratum
Calcium Channel Blockers vs ACE-inhibitors/Angiotensin Receptor Blockers: Achieved diastolic BP since baseline in a non-black, non-diabetic population
Calcium Channel Blockers vs ACE-inhibitors/Angiotensin Receptor Blockers: Achieved diastolic BP since baseline in a non-diabetic population
Page 36 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
10
Appendix 4 Figure 2: Achieved diastolic blood pressures for thiazide and calcium channel blocker (CCB)
in each stratum
Thiazide vs Calcium Channel Blockers: Achieved diastolic BP since baseline in a non-black, non-diabetic population
Thiazide vs Calcium Channel Blockers: Achieved diastolic BP since baseline in a non-diabetic population
Page 37 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
11
Appendix 5: 95% CI for changes in systolic blood pressure
Appendix 5 Table: Difference in systolic blood pressure since drug initiation for calcium channel blockers (CCB) versus ACE-inhibitors/angiotensin receptor blockers
(ACEI/ARB) and thiazide versus CCB, 95% CI.
Difference between CCB and ACEI/ARB for change in systolic BP(mmHg) since baseline (95% CI)
n 12 weeks 26 weeks 52 weeks
Age <55* 20,964 -1.69 (-2.32, -1.06) -0.48 (-1.24, 0.27) -0.53 (-1.62, 0.57)
Age>55* 58,396 -0.40 (-0.84, 0.04) 0.63 (0.20, 1.07) 0.85 (0.12, 1.58)
Non-black** 73,726 -0.98 (-1.37, -0.59) 0.11 (-0.30, 0.51) 0.08 (-0.60, 0.75)
Black** 894 -2.15 (-5.21, 0.91) 0.55 (-2.45, 3.55) 2.28 (-3.53, 8.09)
Difference between thiazide and CCB for change in systolic BP(mmHg) since baseline (95% CI)
Age <55* 5,724 1.51 (0.34, 2.69) 0.07 (-1.28, 1.41) 0.34 (-1.74, 2.42)
Age>55* 32,464 2.16 (1.55, 2.77) 1.43 (0.77, 2.10) 0.17 (-0.68, 1.03)
Non-black** 35,876 2.10 (1.55, 2.65) 1.19 (0.58, 1.80) 0.32 (-0.49, 1.13)
Black** 648 -0.28 (-4.40, 3.84) -1.75 (-5.73, 2.23) 2.01 (-3.68, 7.69)
*In a non-black, non-diabetic population ** In a non-diabetic population All analyses are 1:1 PS matched within each stratum. Each n consists of half CCB and half ACEI/ARB (or half thiazide and half CCB). For the CCB vs ACEI/ARB comparison a negative result means CCB produced larger reductions in systolic BP and a positive results indicates that ACEI/ARB produced larger reductions in systolic BP. For the thiazide vs CCB comparison a negative result means thiazides produced larger reductions in systolic BP and a positive results indicates that CCB produced larger reductions in systolic BP.
Page 38 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
12
Appendix 6 Figure 1: Negative and positive outcomes for calcium channel blockers (CCB) versus ACE-inhibitors/angiotensin receptor blockers (ACEI/ARB) in age-based and ethnicity-based strata. ~ is used when cell sizes are <5
Appendix 6: Negative and positive outcomes in age-based and ethnicity-based strata
Page 39 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
13
Appendix 6 Figure 2: Negative and positive outcomes for thiazide versus calcium channel blockers (CCB) in age-based and ethnicity-based strata. ~ is used when cell sizes are <5
Page 40 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
14
Appendix 7: Comparison of thiazide-like diuretics with calcium channel blockers
Appendix 7 Figure: Comparison of thiazide-like diuretics (chlortalidone and indapamide) with calcium
channel blockers (CCB). Comparison of traditional thiazide diuretics (eg bendroflumethazide) with
CCB.
Data-points below the null line indicate that thiazides are better at reducing BP. Data-points above
the null line indicate that CCB are better at reducing BP.
Page 41 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
15
Appendix 8: On treatment analysis
Appendix 8 Table: Difference in systolic blood pressure since index date for CCB versus ACE-I/ARB and thiazide versus CCB in an on-treatment analysis
Difference between CCB and ACEI/ARB for change in systolic BP(mmHg) since baseline (99% CI)
n 12 weeks 26 weeks 52 weeks
Age <55* 19,228 0.22 (-0.66, 1.11) 0.35 (-0.83, 1.54) 1.01 (-0.90, 2.93)
Age>55* 54,149 -0.19 (-0.95, 0.57) 0.17 (-0.51, 0.84) -0.09 (-1.21, 1.02)
Non-black** 35,876 -0.35 (-0.94, 0.24) 0.03 (-0.61, 0.67) -0.30 (-1.28, 0.67)
Black** 648 -3.65 (-7.99, 0.70) -5.16 (-11.13, 0.81) -3.36 (-12.10, 5.37)
Difference between thiazide and CCB for change in systolic BP(mmHg) since baseline (99% CI)
Age <55* 5,056 3.04 (1.17, 4.90) 3.06 (0.72, 5.40) 0.40 (-3.03, 3.83)
Age>55* 29,965 4.14 (3.23, 5.04) 4.12 (3.05, 5.20), 3.91 (2.39, 5.44)
Non-black** 32,833 4.05 (3.21, 4.88) 3.93 (2.89, 4.97) 3.55 (2.03, 5.07)
Black** 555 2.73 (-3.51, 8.98) 1.46 (-4.95, 7.86) 3.47 (-8.51, 15.46) *In a non-black, non-diabetic population ** In a non-diabetic population All analyses are 1:1 PS matched within each stratum. Each n consists of half CCB and half ACE-i/ARB (or half thiazide and half CCB). For the CCB vs ACE-I/ARB comparison a negative result means CCB produced larger reductions in systolic BP and a positive results indicates that ACEI/ARB produced larger reductions in systolic BP. For the thiazide vs CCB comparison a negative result means thiazides produced larger reductions in systolic BP and a positive results indicates that CCB produced larger reductions in systolic BP. There was a median of 4 (IQR 2-6) BP measurements available during one year follow up for each patient. On treatment analysis: patients follow-up BP data were censored if the patient switched to another drug class, added a drug from another class or discontinued their anti-hypertensive treatment (drug classes: thiazides, CCB, ACE-I/ARB, beta-blockers, alpha-blockers, potassium sparing diuretics, aldosterone antagonists and other)
Page 42 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
16
Appendix 9: Complete case analysis
Appendix 9 Table: Complete case analysis
Difference between CCB and ACEI/ARB for change in systolic BP(mmHg) since baseline (99% CI)
n 12 weeks 26 weeks 52 weeks
Age <55* 11,660 -1.77 (-2.77, -0.77) -0.45 (-1.63, 0.73) -0.84 (-2.40, 0.72)
Age>55* 40,890 -0.46 (-1.06, 0.14) 0.52 (-0.09, 1.12) 1.01 (-0.002, 2.02)
Non-black** 49,186 -1.06 (-1.61, -0.52) -0.12 (-0.69, 0.45) -0.04 (-0.96, 0.89)
Black** 592 -2.21 (-6.58, 2.17) 0.27 (-4.21, 4.76) 0.000 (-6.78, 6.78)
Difference between thiazide and CCB for change in systolic BP(mmHg) since baseline (99% CI)
Age <55* 3,472 1.22 (-0.59, 3.03) 0.04 (-1.84, 1.92) -1.01 (-3.87, 1.85)
Age>55* 21,448 1.84 (0.98, 2.69) 0.83 (0.04, 1.63) -0.04 (-1.36, 1.28)
Non-black** 23,550 1.84 (1.01, 2.67) 0.89 (0.07, 1.70) -0.01 (-1.26, 1.25)
Black** 372 -1.11 (-7.29, 5.07) 0.88 (-5.00, 6.77) 2.75 (-6.01, 11.51)
*In a non-black, non-diabetic population
** In a non-diabetic population
All analyses are 1:1 PS matched within each stratum. Each n consists of half CCB and half ACEI/ARB (or half thiazide and half CCB).
For the CCB vs ACEI/ARB comparison a negative result means CCB produced larger reductions in systolic BP and a positive results indicates that ACE-I/ARB produced larger
reductions in systolic BP. For the thiazide vs CCB comparison a negative result means thiazides produced larger reductions in systolic BP and a positive results indicates that CCB
produced larger reductions in systolic BP. There was a median of 4 (IQR 2-6) BP measurements available during one year follow up for each patient.
Page 43 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
17
Appendix 10: RECORD-PE checklist
The RECORD statement for pharmacoepidemiology (RECORD-PE) checklist of items, extended from the STROBE and RECORD statements, which
should be reported in non-interventional pharmacoepidemiological studies using routinely collected health data
Item No STROBE items RECORD items RECORD-PE items Page No Title and abstract
1 (a) Indicate the study’s design with a
commonly used term in the title or the
abstract.
(b) Provide in the abstract an
informative and balanced summary of
what was done and what was found.
1.1: The type of data used should be
specified in the title or abstract. When
possible, the name of the databases
used should be included.
1.2: If applicable, the geographical
region and timeframe within which
the study took place should be reported in the title or abstract.
1.3: If linkage between databases was
conducted for the study, this should be
clearly stated in the title or abstract.
— 1 & 2
Introduction
Background rationale
2 Explain the scientific background and
rationale for the investigation being
reported.
— — 5
Objectives
3 State specific objectives, including
any prespecified hypotheses.
— — 5
Methods
Study design
4 Present key elements of study design
early in the paper.
— 4.a: Include details of the specific
study design (and its features) and
report the use of multiple designs if
used.
4.b: The use of a diagram(s) is recommended to illustrate key aspects
of the study design(s), including
6
Page 44 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
18
exposure, washout, lag and
observation periods, and covariate
definitions as relevant.
Setting
5 Describe the setting, locations, and
relevant dates, including periods of
recruitment, exposure, follow-up, and
data collection.
— — 6
Participants
6 (a) Cohort study—give the eligibility
criteria, and the sources and methods of selection of participants. Describe
methods of follow-up. Case-control
study—give the eligibility criteria,
and the sources and methods of case
ascertainment and control selection.
Give the rationale for the choice of
cases and controls. Cross sectional
study—give the eligibility criteria,
and the sources and methods of
selection of participants.
(b) Cohort study—for matched studies, give matching criteria and
number of exposed and unexposed.
Case-control study—for matched
studies, give matching criteria and the
number of controls per case.
6.1: The methods of study population
selection (such as codes or algorithms used to identify participants) should
be listed in detail. If this is not
possible, an explanation should be
provided.
6.2: Any validation studies of the
codes or algorithms used to select the
population should be referenced. If
validation was conducted for this
study and not published elsewhere,
detailed methods and results should be
provided. 6.3: If the study involved linkage of
databases, consider use of a flow
diagram or other graphical display to
demonstrate the data linkage process,
including the number of individuals
with linked data at each stage.
6.1.a: Describe the study entry criteria
and the order in which these criteria were applied to identify the study
population. Specify whether only
users with a specific indication were
included and whether patients were
allowed to enter the study population
once or if multiple entries were
permitted. See explanatory document
for guidance related to matched
designs.
6
Variables
7 Clearly define all outcomes,
exposures, predictors, potential
confounders, and effect modifiers.
Give diagnostic criteria, if applicable.
7.1: A complete list of codes and
algorithms used to classify exposures,
outcomes, confounders, and effect
modifiers should be provided. If these
cannot be reported, an explanation
should be provided.
7.1.a: Describe how the drug exposure
definition was developed.
7.1.b: Specify the data sources from
which drug exposure information for
individuals was obtained.
7.1.c: Describe the time window(s)
during which an individual is considered exposed to the drug(s).
The rationale for selecting a particular
6 & 7
Page 45 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
19
time window should be provided. The
extent of potential left truncation or
left censoring should be specified.
7.1.d: Justify how events are
attributed to current, prior, ever, or
cumulative drug exposure.
7.1.e: When examining drug dose and risk attribution, describe how current,
historical or time on therapy are
considered.
7.1.f: Use of any comparator groups
should be outlined and justified.
7.1.g: Outline the approach used to
handle individuals with more than one
relevant drug exposure during the
study period.
Data sources/measurement
8 For each variable of interest, give
sources of data and details of methods
of assessment (measurement).
Describe comparability of assessment methods if there is more than one
group.
— 8.a: Describe the healthcare system
and mechanisms for generating the
drug exposure records. Specify the
care setting in which the drug(s) of interest was prescribed.
6
Bias
9 Describe any efforts to address
potential sources of bias.
— — 6-8
Study size
10 Explain how the study size was
arrived at.
— — flowchart
Quantitative variables
11 Explain how quantitative variables
were handled in the analyses. If
applicable, describe which groupings
were chosen, and why.
— — 7
Statistical methods
12 (a) Describe all statistical methods,
including those used to control for
confounding.
— 12.1.a: Describe the methods used to
evaluate whether the assumptions
have been met.
7 & 8
Page 46 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
20
(b) Describe any methods used to
examine subgroups and interactions.
(c) Explain how missing data were
addressed.
(d) Cohort study—if applicable,
explain how loss to follow-up was
addressed. Case-control study—if applicable, explain how matching of
cases and controls was addressed.
Cross sectional study—if applicable,
describe analytical methods taking
account of sampling strategy.
(e) Describe any sensitivity analyses.
12.1.b: Describe and justify the use of
multiple designs, design features, or
analytical approaches.
Data access and cleaning methods
12 — 12.1: Authors should describe the
extent to which the investigators had
access to the database population used
to create the study population.
12.2: Authors should provide
information on the data cleaning
methods used in the study.
— 7
Linkage
12 — 12.3: State whether the study included
person level, institutional level, or
other data linkage across two or more
databases. The methods of linkage
and methods of linkage quality
evaluation should be provided.
— 6
Results
Participants
13 (a) Report the numbers of individuals
at each stage of the study (eg,
numbers potentially eligible,
examined for eligibility, confirmed
eligible, included in the study,
completing follow-up, and analysed).
(b) Give reasons for non-participation at each stage.
13.1: Describe in detail the selection
of the individuals included in the
study (that is, study population
selection) including filtering based on
data quality, data availability, and
linkage. The selection of included
individuals can be described in the
— See flowchart
Page 47 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
21
(c) Consider use of a flow diagram. text or by means of the study flow
diagram.
Descriptive data
14 (a) Give characteristics of study
participants (eg, demographic,
clinical, social) and information on
exposures and potential confounders.
(b) Indicate the number of participants
with missing data for each variable of interest.
(c) Cohort study—summarise follow-
up time (eg, average and total
amount).
— — See table 1
Outcome data
15 Cohort study—report numbers of
outcome events or summary measures
over time. Case-control study—report
numbers in each exposure category, or
summary measures of exposure. Cross
sectional study—report numbers of
outcome events or summary
measures.
— — See table 2
(outcome was
BP – a
continuous
outcome)
Main results
16 (a) Give unadjusted estimates and, if applicable, confounder adjusted
estimates and their precision (eg, 95%
confidence intervals). Make clear
which confounders were adjusted for
and why they were included.
(b) Report category boundaries when
continuous variables are categorised.
(c) If relevant, consider translating
estimates of relative risk into absolute
risk for a meaningful time period.
— — See Table 2
Other analyses
17 Report other analyses done—eg,
analyses of subgroups and
interactions, and sensitivity analyses.
— — 12
Page 48 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
22
Discussion
Key results
18 Summarise key results with reference
to study objectives.
— — 13
Limitations
19 Discuss limitations of the study,
taking into account sources of
potential bias or imprecision. Discuss
both direction and magnitude of any
potential bias.
19.1: Discuss the implications of
using data that were not created or
collected to answer the specific
research question(s). Include
discussion of misclassification bias,
unmeasured confounding, missing data, and changing eligibility over
time, as they pertain to the study
being reported.
19.1.a: Describe the degree to which
the chosen database(s) adequately
captures the drug exposure(s) of
interest.
13 & 14
Interpretation
20 Give a cautious overall interpretation
of results considering objectives,
limitations, multiplicity of analyses,
results from similar studies, and other
relevant evidence.
— 20.a: Discuss the potential for
confounding by indication,
contraindication or disease severity or
selection bias (healthy adherer/sick
stopper) as alternative explanations
for the study findings when relevant.
13
Generalisability
21 Discuss the generalisability (external
validity) of the study results.
— — 15
Other information
Funding
22 Give the source of funding and the
role of the funders for the present
study and, if applicable, for the original study on which the present
article is based.
— — 16
Accessibility of protocol, raw data,
and programming code
22 — 22.1: Authors should provide
information on how to access any
supplemental information such as the
study protocol, raw data, or
programming code.
— 9
Page 49 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
23
RECORD=reporting of studies conducted using observational routinely collected data; RECORD-PE=RECORD for pharmacoepidemiological research; STROBE=strengthening the
reporting of observational studies in epidemiology.
*REFERENCE: Langan SM, Schmidt S, Wing K, Ehrenstein V, Nicholls S, Filion K, Klungel O, Petersen I, Sorensen H, Guttmann A, Harron K, Hemkens L, Moher D,
Schneeweiss S, Smeeth L, Sturkenboom M, von Elm E, Wang S, Benchimol EI. The REporting of studies Conducted using Observational Routinely-collected health
Data (RECORD) Statement for Pharmacoepidemiology (RECORD-PE). BMJ 2018; 363: k3532.
Page 50 of 50
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960