Identifying and treating high blood pressure in men under 55 years with grade 1 hypertension: the TREAT CASP study and RCT

Article history

The research reported in this issue of the journal was funded by the EME programme as project number 10/90/22. The contractual start date was in June 2013. The final report began editorial review in October 2018 and was accepted for publication in May 2019. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The EME editors and production house have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the final report document. However, they do not accept liability for damages or losses arising from material published in this report.

Declared competing interests of authors

Bryan Williams has received honoraria fees from C.H. Boehringer Sohn AG & Ko. KG (Ingelheim am Rhein, Germany), Daiichi Sankyo Company, Ltd (Tokyo, Japan), Pfizer Inc. (New York, NY, USA) and Servier Laboratories (Neuilly-sur-Seine, France). All other authors declare no competing interests in relation to this study. Support was also provided by the North Thames Local Clinical Research Network and the North Central and East London Comprehensive Research Network. Bryan Williams is a National Institute for Health Research Senior Investigator Emeritus and is supported by the National Institute for Health Research Biomedical Research Centre at University College London Hospitals and University College London.

Copyright statement

© Queen’s Printer and Controller of HMSO 2019. This work was produced by Williams et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

2019 Queen’s Printer and Controller of HMSO

Hypertension and risk

High blood pressure (BP) is the leading preventable cause of premature morbidity and mortality worldwide. 1,2 In the UK, surveys suggest that at least 25% of the adult population is hypertensive (i.e. defined as having a seated clinic BP of ≥ 140/90 mmHg), rising to more than 50% in people aged over 65 years. 3 The benefits of treating hypertension has one of the largest evidence bases in medicine and is the cornerstone of national strategies to reduce cardiovascular disease (CVD) risk. Detection, treatment and monitoring of BP represents one of the commonest reasons for consultation in primary care and in the UK accounts for over £1B expenditure on drug costs alone. 4

Study setting

The TREAT CASP study was a single-centre study recruiting at multiple research sites. The co-ordinating centre and main recruiting site was at University College London (UCL) and its associated hospital, University College London Hospitals (UCLH). Study clinical procedures were carried out at a dedicated National Institute for Health Research (NIHR) clinical research facility at UCLH and/or in some primary care sites. cMRI was performed using a research-dedicated magnetic resonance imaging (MRI) scanner within UCL’s Institute of Cardiovascular Science, based at the Great Ormond Street Hospital. The study commenced on 1 November 2013 and the final study visit was on 7 February 2018. The study received ethics approval from the National Research Ethics Service London (Bloomsbury) committee in June 2013 and was conducted in accordance with the principles of good clinical practice (GCP). The UK Medicines and Healthcare products Research Authority (MHRA) judged that the TREAT CASP study was not a clinical trial of an investigational medicinal product (CTIMP). The study was run in association with the UCL Comprehensive Clinical Trials Unit (CCTU) and was overseen by an Independent Trial Steering Committee.

Study design

The TREAT CASP study used a two-stage design (Figure 1), incorporating:

• stage 1 – a cross-sectional screening study to identify patients with grade 1 hypertension who were then invited to participate in stage 2

• stage 2 – a 12-month RCT.

FIGURE 1.

Study design flow chart. ABPM, ambulatory BP monitoring; CKD, chronic kidney disease; LVMI, left ventricular mass index; ΔLVMI, change in left ventricular mass index. ABPM range represents waking (daytime) mean.

There was a stop–go checkpoint at the end of the screening study (see Stop–go: checkpoint for proceeding from stage 1 screening study to stage 2 randomised controlled trial and observational study). Participants with grade 1 hypertension completing the screening study were stratified into two groups according to their CASP status: participants with a low CASP (i.e. those with a CASP of < 125 mmHg) or participants with a high CASP (i.e. those with a CASP of ≥ 125 mmHg).

In the randomised controlled trial (RCT), participants were randomised (simple one-to-one randomisation; sealed envelope™; Sealed Envelope Ltd, London, UK; www.sealedenvelope.com; accessed 1 February 2017) to either BP lowering using a national guideline-directed treatment or a no-treatment group (i.e. usual care in the UK for participants with grade 1 hypertension). Follow-up was for 12 months with a cMRI scan at baseline and study closeout. The study utilised a prospective, randomised, open-label, blinded end-point (PROBE) design to evaluate whether or not people from the high CASP group benefit from BP-lowering drugs compared with no treatment, with regard to regression of their LV mass on cMRI. LV mass was indexed to body size, deriving the LV mass index (LVMI). The change in LVMI from baseline to 12 months, comparing the treatment versus no-treatment arms, constituted the primary outcome for the TREAT CASP study.

Alongside the RCT, the second stage of the study incorporated an observational follow-up study in those participants with grade 1 hypertension and low CASP, who remained untreated for BP over 12 months’ follow-up. The observational follow-up study was designed to evaluate whether or not there was any progression of cardiac structural change (as measured by the LVMI) in people with grade 1 hypertension and low CASP over a period of 12 months and to provide a reference value against which regression of cardiac structure in the RCT could be assessed.

Recruitment strategy: stage 1 screening study

Potential participants for the TREAT CASP study were identified from a number of sources:

• patients referred to the BP outpatient clinic at UCLH

• from general practice in association with liaison with the local North Central London Research Consortium (Noclor) or from searches of general practitioner (GP) databases within the greater London area by the NIHR-funded Clinical Practice Research Datalink (CPRD) organisation

• responses to advertisements placed in various media or prominent local buildings in London, for example railway stations, restaurants and other buildings for social functions.

Participants

The TREAT CASP study recruited male patients aged between 18 and < 55 years, who had uncomplicated, untreated grade 1 hypertension [with a seated clinic BP of 140–159/90–99 mmHg and/or a 24-hour ambulatory BP monitoring (ABPM) daytime mean of ≥ 135/85 mmHg] and, if previously treated for hypertension, for whom BP-lowering treatment had been discontinued for at least 3 months prior to screening. Study participants were also required to be low-risk grade 1 hypertension, that is, that they had no pre-existing CVD or diabetes mellitus or chronic kidney disease and no evidence of HMOD [i.e. electrocardiography (ECG) LVH, grade 2 or more hypertensive retinopathy or renal impairment or an increased urinary albumin-to-creatinine ratio (UACR)].

Inclusion criteria

Inclusion criteria were men who:

• were aged 18 to < 55 years

• had a clinical diagnosis of grade 1 hypertension based on a seated clinic BP of 140–159/90–99 mmHg, and/or ABPM

• were currently not being treated for hypertension (if previously treated with antihypertensive medication, patients will have received no treatment for > 3 months)

• were without evidence of HMOD on routine clinical screening, established CVD, diabetes mellitus or chronic kidney disease

• were willing and capable of giving informed consent.

Exclusion criteria

Exclusion criteria included:

• women of any age

• men –

  • with grade 2 hypertension on clinic BP

  • with ‘white coat hypertension’, that is, grade 1 hypertension according to clinic blood BP but a normal BP on ABPM (daytime mean BP of < 135/85 mmHg)

  • who had a prior diagnosis of a secondary cause for hypertension, for example renal artery stenosis, primary aldosteronism, phaeochromocytoma, aortic coarctation

  • who had grade 1 hypertension with evidence of HMOD on routine clinical testing (e.g. LVH, hypertensive retinopathy, renal impairment, proteinuria) or concurrent CVD (e.g. stroke, transient ischaemic attack, cardiac or peripheral vascular disease) and/or diabetes mellitus or chronic kidney disease – that is, risk factors that would mandate treatment of grade 1 hypertension according to current NICE guidance [NICE Clinical Guidance (CG) 1274]

  • in whom it was not possible to measure conventional BrBP

  • with atrial fibrillation or any other significant pulse rhythm irregularity, making BP measurement difficult and unreliable

  • who regularly consumed > 28 units of alcohol per week or used recreational drugs

  • who had chronic inflammatory diseases requiring regular concomitant steroids and/or non-steroidal anti-inflammatory drugs

  • who had severe hepatic impairment

  • who had previous hypersensitivity to the drugs to be used in the study

  • who had current malignancy

  • who were unwilling to undergo, or had a contraindication to, MRI scanning

  • who had currently or recently (i.e. in the last 6 weeks) participated in an interventional clinical trial

  • who had any clinical condition for which the investigator would consider the patient unsuitable for the trial

  • who had co-enrolled in any CTIMP or any other study of a non-investigational medicinal product (there was no prohibition on co-enrolment into observational studies).

Withdrawal criteria

Participants had the right to withdraw from the study at any time, for any reason and without giving a reason. The investigator also had the right to withdraw patients from the study in the event of intercurrent illness, adverse events (AEs), serious adverse events (SAEs), suspected unexpected serious adverse reactions (SUSARs), protocol violations or other reasons. If a patient decided to withdraw from the study, all efforts were made to report the reason for withdrawal. Participants randomised to study intervention who then declined to take any study medication were withdrawn from the study with no further study follow-up. For participants electing to withdraw, having commenced taking their study medication, efforts were made to continue to obtain follow-up data (with the permission of the patient).

All recruited participants were required to provide written informed consent after reviewing the detailed study information prior to the consent process.

Stage 1 screening study

In order to recruit men with uncomplicated grade 1 hypertension, the first stage of the TREAT CASP study comprised a screening study. Information about the study was disseminated widely via advertisements in local or national press. Participants with a history of elevated BP, who remained untreated, were sought through searches of local GP databases and through local specialist hypertension clinics. Men responding to study advertisements or who were identified via their GP records as previously having an elevated BP in the grade 1 hypertension range and no other CVD were invited to participate in the screening study.

The screenees were invited to attend for an initial BP check (i.e. a prescreen) where they underwent a measurement of seated clinic BP and CASP. Participants with elevated clinic BP were asked to undergo 24-hour ambulatory BP monitoring to confirm the diagnosis of hypertension and to exclude white coat hypertension. Prescreening was essential to filter out, at an early stage, participants who were clearly normotensive or those who had grade 2 or higher BP and would require treatment according to current national treatment guidelines. Participants identified with grade 1 hypertension and who met study inclusion criteria were then invited to undergo detailed cardiovascular screening.

Clinical assessments performed during the detailed cardiovascular screening included:

• documentation of medical and lifestyle history

• physical examination

• measurement of seated BrBP and CASP

• height and weight measurement

• waist and hip circumference measurement

• body fat composition using non-invasive bioimpedance

• blood tests – including haematological, biochemical, lipid profile, glucose and renal function [i.e. estimated glomerular filtration rate (eGFR)] assessments

• UACR analysis/urine dipstick for blood and protein

• 12-lead ECG.

The TREAT CASP screening study was designed to generate a cohort of participants with confirmed grade 1 hypertension who were then stratified according to their CASP measurement as high CASP (i.e. a CASP of ≥ 125 mmHg) or low CASP (i.e. a CASP of < 125 mmHg) prior to entry to stage 2 of the study, the high CASP RCT and the low CASP observational study.

Defining high and low central aortic systolic pressure

When the study was designed, there was no existing definition of high versus low CASP. The original plan was to screen approximately 500 patients in stage 1 of the study and then identify a high CASP cohort for the RCT from the upper half of the distribution and the low CASP cohort from the lower half of the distribution. This became unnecessary when population threshold values for the low and high CASP groups were defined following publication of a definitive analysis of all available studies worldwide, including 45,436 people who had measurements of both conventional BrBP and CASP. 28 From these data, a high CASP threshold of ≥ 125 mmHg was defined for the TREAT CASP study (see Chapter 1, Defining high and low central aortic systolic pressure values in younger people with grade 1 hypertension).

Stop–go: checkpoint for proceeding from stage 1 screening to stage 2 randomised controlled trial and observational study

At the time the study was designed, it was unknown whether or not CASP would be related to LVMI, although this seemed highly likely as CASP is mathematically derived from BrBP and CASP is the pressure closest to the left ventricle. To avoid futility in undertaking the RCT if high CASP was not associated with an increased LVMI, a stop–go checkpoint was requested by the funder and this was introduced into the study protocol. The stop–go checkpoint required confirmation (evidence) that there was a significant difference in LVMI between men in the low CASP and men in the high CASP groups prior to commencement of the RCT. This required all screenees considered suitable for the RCT and observational study (i.e. ≈ 160 participants) to undergo baseline cMRI. The cMRI scan would then be evaluated to confirm that the stop–go checkpoint had been satisfied, before any patient could be randomised into the RCT in the second stage of the study. The implications of this stop–go checkpoint soon became apparent. Patients screened and eligible for stage 2 of the study would have to wait until the entire stage 1 screening study had been completed before they could enter stage 2. Moreover, once screening had been completed, ≈ 160 cMRI scans would have to be performed and evaluated within a few weeks. This was logistically very difficult for the study team and the participants. Moreover, it meant that some of the baseline cMRIs for the RCT would have been performed a few weeks, or even months, before randomisation into the RCT, which was undesirable.

To evaluate the need for such an extensive and detailed cMRI study to confirm the stop–go criterion, we undertook an exploratory study in which we evaluated cMRI LVMI data on 11 patients from the high and low CASP groups we had identified during screening. The exploratory study showed a clear and significant association between high CASP and increased LVMI relative to the low CASP patients (Figure 2). It was agreed with the funder, based on this analysis, that it was highly likely that the stop–go criterion would be satisfied once all baseline cMRIs had been performed (which it was – see Chapter 3, Retrospective evaluation of the stop–go criterion based on baseline cardiac magnetic resonance imaging data from the stage 2 studies); to avoid further delay it was agreed that the study should proceed to stage 2.

FIGURE 2.

Relationship between LVMI and CASP (a) and LVMI and 24-hour ambulatory SBP (b) in a preliminary subset (n = 11) of stage 1 study participants undergoing cMRI.

The study proceded to the stage 2 RCT and observational study. Patients considered eligible for the high CASP RCT and low CASP observational study then had to be re-contacted to undergo baseline cMRI scans and enter the observational study or the RCT. Inevitably, following the delay in starting stage 2 of the study, some of the patients deemed eligible when they were originally screened were either not contactable or no longer interested, or were no longer eligible because their BP status had changed or they had been treated. This meant that we had to continue to recruit further patients to ensure that a sufficient number of patients were entered into the RCT.

Stipulating a stop–go requirement for this study was not without consequences. It resulted in patients eligible for the study when originally screened waiting for up to 2 years before they could enter the RCT, rather than immediately transitioning from screening to the RCT. Moreover, the need to complete all screening before any patients could be entered into the RCT, and the subsequent need to undertake further screening, resulted in an overall delay in completion of the study by approximately 2 years. This resulted in 98 patients originally identified as eligible and consented for the study subsequently being unable to enter the RCT or observational study (lost contact or no longer willing or change in BP status or treated) as a result of the delays in completing the stop–go decision. This meant that we had to undertake further recruitment alongside the initiation of the second phase of the study. The stop–go checkpoint was retrospectively evaluated in participants who went on to complete the study and confirmed the prediction obtained from the exploratory cMRI scans (see Chapter 3, Retrospective evaluation of the stop–go criterion based on baseline cardiac magnetic resonance imaging data from the stage 2 studies).

It was originally planned to screen approximately 500 participants to yield at least 50 evaluable patients for each arm of the RCT (treatment vs. no treatment) and at least a further 50 participants with low CASP to enter the observational study. The first participant was recruited into the stage 1 screening study on 15 November 2013. The first participants for the observational study and RCT were recruited on 18 August 2015 and 21 October 2015, respectively. The final participant was recruited into stage 2 of the study on 8 February 2017 and the last patient visit for the RCT was on 14 February 2018 when trial follow-up had been completed. A total of 726 men were screened for the stage 1 study and 162 men entered the stage 2 RCT and observational studies.

Stage 2 randomised controlled trial and observational study

The second stage of the TREAT CASP study incorporated a RCT of BP lowering for men with grade 1 hypertension and high CASP. The RCT was designed to evaluate whether or not treatment regresses LVMI relative to the comparator, that is, no treatment, which represents usual clinical care for these people. Participants with grade 1 hypertension and low CASP were entered into an observational study and remained untreated. Participants in both the RCT and the observational study were followed up over the course of 12 months and underwent cMRI at both baseline and at the end of the study (i.e. study closeout).

The RCT used a prospective, randomised, open, blinded, end-point (PROBE) design85 to allow the BP-lowering treatment to be titrated to achieve a CASP reduction of at least 5 mmHg in the treatment arm participants. Study team members involved in the MRI measurement and/or analysis were blinded to treatment allocation.

Baseline visit procedures for the randomised controlled trial and observational study

Participants with grade 1 hypertension and high CASP underwent a baseline cardiovascular evaluation to reconfirm their eligibility for the study and to collect baseline data. The baseline examination included:

• documentation of medical and lifestyle history

• measurement of seated BrBP and CASP

• 24-hour ABPM

• a physical examination

• height and weight measurement

• waist and hip circumference measurement

• body fat composition evaluation using non-invasive bioimpedance

• blood tests – including haematological, biochemical, lipid profile, glucose and renal function (i.e. eGFR) assessments

• UACR analysis (analysed by an accredited external laboratory, specifically by The Doctors Laboratory, London, UK)/urine dipstick analysis for blood and protein

• a 12-lead ECG

• non-invasive measurement of arterial stiffness [as assessed via the cardio-ankle vascular index (CAVI)]

• quality-of-life assessment using the Short Form questionnaire-36 items (SF-36)

• non-mydriatic retinal photography

• 1.5-T cardiovascular MRI scan.

Randomisation

Participants in the RCT were randomised to BP-lowering treatment or no intervention. Randomisation (simple, one-to-one allocation ratio using the next available treatment in the randomisation list) was performed using a web-based randomisation service (Sealed Envelope Ltd).

Blood pressure-lowering intervention

The BP-lowering treatment used for this study was as recommended by NICE4 for this age group. BP-lowering treatment was up-titrated according to NICE4 to achieve a seated CASP value of < 120 mmHg and a minimum 5-mmHg reduction in CASP from baseline. The intervention comprised the angiotensin receptor blocker (ARB) losartan [50–100 mg once daily (o.d.)] plus the calcium channel blocker (CCB) amlodipine (5–10 mg o.d.) when required (Table 1). For men of black African ethnicity, consistent with NICE guidance,4 treatment commenced with the CCB to which an ARB was added if required. Participants allocated to treatment began taking their medication 1 day after their baseline cMRI scan. There was no wash-out or run-in period as the RCT was conducted in treatment-naive participants or previously treated participants who had not taken BP-lowering medication for at least 3 months prior to randomisation. For participants unable to tolerate a higher dose of medication, treatment could be back-titrated to achieve the best CASP reduction that could be tolerated.

The comparator was no treatment (i.e. usual care); no placebo was used in this open-label study. The intervention was prescribed unblinded and was dispensed from the hospital pharmacy.

Concomitant medication was permitted with the exception of other BP-lowering medications.

Randomised controlled trial follow-up schedule and procedures

Follow-up visits for participants randomised to treatment were scheduled at 4, 8, 12, 16, 26 and 38 weeks to allow for study drug titrations. Participants randomised to no treatment attended follow-up visits at 12 and 26 weeks. Follow-up visits included measurements of seated brachial clinic BP and CASP, and documentation of AEs. Follow-up appointments were scheduled ± 2 weeks of the due date. Study closeout occurred after 12 months of study participation. This terminology (‘study closeout’) is used throughout this report to refer to the end of follow-up. Study closeout procedures, including the final cMRI, were identical to those at baseline; however, a physical examination was not performed unless indicated (Table 2).

Observational study

A 12-month observational follow-up study of participants with grade 1 hypertension and low CASP was performed in parallel with the RCT. Participants in the observational study received no intervention and the visit schedule was identical to that of the no-treatment group for the RCT. Concomitant medication (other than BP-lowering medication) was allowed.

Study outcomes

Second phase of the study: randomised controlled trial to determine if reducing central aortic systolic pressure in patients with high central aortic systolic pressure leads to a reduction of left ventricular mass index versus no treatment

Statistical analysis plan

Study procedures/experimental assessments

Retinal photography

Fundal images were acquired using a dedicated retinal camera (Canon CR-DGi; Canon Inc., Tokyo, Japan). Images were taken in a darkened room without mydriasis acquiring a 45° view in each eye, centred between the optic nerve and the macula. Images were acquired in TREAT CASP RCT and observational study follow-up participants at baseline and study closeout.

Physical examination

A standard physical examination was carried out by the study doctor.

Adverse event reporting

The pharmacological intervention used in the RCT was medication routinely prescribed according to national guidelines for the treatment of hypertension in the UK both as monotherapy and combination therapy, and, for this reason, the intervention has a good safety profile. Moreover, the UK’s MHRA determined on more than one occasion that the RCT part of this study did not constitute a CTIMP. Accordingly, an independent data monitoring and ethics committee was not required for safety monitoring. Nevertheless, data on AEs, adverse reactions (ARs) and SAEs were routinely collected at study visits and recorded in the participant’s notes. In the event of ARs inconsistent with the known safety profile of the study medication, the intention was to report these via the Yellow Card Scheme through the MHRA website106 and to report such events to the study sponsor as SUSARS. By the end of the study, no such reports were deemed necessary to be made.

The severity of AEs was assessed by the study physician and the chief investigator, with reference to the study protocol and the World Health Organization (WHO)’s toxicity grading. 107 Expectedness and causality of AEs were also assessed by the study doctor and chief investigator with regard to clinical judgement and the known safety profile of the medications as set out in national texts, for example the British National Formulary. 108

Study monitoring/Comprehensive Clinical Trials Unit involvement

The quality assurance (QA) and quality control (QC) considerations for the TREAT CASP RCT were based on standard UCL CCTU quality management policy that included a formal risk assessment, and that acknowledged the risk associated with the conduct of the trial and proposals of how to mitigate them through appropriate QA and QC processes. Risks were defined in terms of their impact on the rights and safety of participants; project concept including trial design, reliability of results and institutional risk; project management; and other considerations.

Quality assurance was defined as all the planned and systematic actions established to ensure that the trial was performed and data generated, documented and/or recorded and reported in compliance with the approved protocol, the principles of GCP and applicable regulatory requirements. QC was defined as the operational techniques and activities performed within the QA system to verify that the requirements for quality of the trial-related activities were fulfilled. A risk assessment was completed for the TREAT CASP RCT and reviewed by the UCL CCTU Quality Management Group. UCL CCTU staff reviewed case report forms and electronic data for errors and missing key data points.

Trial oversight was intended to preserve the integrity of the trial by independently verifying a variety of processes and prompting corrective action where necessary. The processes reviewed related to participant enrolment, consent, eligibility and allocation to trial groups; adherence to trial interventions and policies to protect participants, including the reporting of harms; and completeness, accuracy and timeliness of data collection. Independent trial oversight complied with the UCL CCTU trial oversight policy.

Protocol amendments and changes to recruitment strategy

The stage 1 screening study began in November 2013. By September 2014, recruitment was behind schedule and it was clear that age was a constraint to recruitment (the study’s original design had been to recruit patients up to the age of 40 years). Moreover, many recruited participants who attended the study centre for a detailed cardiovascular investigation were not suitable for further study participation because their ambulatory BP monitoring, performed at the end of the study screening visit, revealed that their BP was normal despite a high-seated clinic BP.

To address these concerns, a substantial amendment was approved in February 2014 to raise the maximum age of study participants from 40 to 55 years. This change was made in order to increase study recruitment and to maintain consistency with NICE guidelines,4 which use a cut-off point of 55 years for defining treatment strategies for younger and older patients. In addition, an amendment was made to introduce a prescreening BP check (i.e. measurement of both clinic BP and ambulatory BP) to identify participants who were normotensive or had higher grade hypertension, prior to undergoing a detailed cardiovascular evaluation. This complemented changes to the study’s recruitment strategy, which became more focused on direct advertising to the general public because searches of GP databases for patients identified as having grade 1 hypertension did not provide a reliable source of study recruits, as many were normotensive on more formal testing of their BP.

We were encouraged to engage with the CPRD, which took considerable time and effort, only to find that the return from this method of recruitment was very disappointing because, unbeknown to us at the time, CPRD relied on data solely acquired from GP records using the VISION database system. The VISION database mainly covered outer London areas and not our immediate locality (i.e. inner London), where most GPs were using EMIS (Egton Medical Information Systems) Health, which was not accessible to the CPRD at the time. Moreover, even when letters were dispatched to potential participants via the CPRD, the response was very poor. It was therefore decided to stop using the CPRD and focus our recruitment effort on local advertising, which was the most effective strategy and more straightforward.

To further improve recruitment, in 2016 the study team worked with the local NIHR network, Noclor and recruitment sites were established in local general practice surgeries, where prescreening activities (recruitment, clinic and ambulatory BP measurement) were performed by Noclor research nurses after training by our study staff. Participants undergoing successful prescreening were then invited to attend for a detailed cardiovascular evaluation at the UCL/UCLH clinical research facility. Owing to the length of time required to recruit all the study participants, compounded by delays introduced by the stop–go process (see Stop–go: checkpoint for proceeding from stage 1 screening to stage 2 randomised controlled trial and observational study), a 1-year no-cost extension was granted to the study in 2016.

Study recruitment: stage 1 screening study

Participants were recruited into the stage 1 screening study between 15 November 2013 and 19 January 2017. The final patient was recruited into the stage 1 screening study in January 2017.

From a total of 726 participants recruited into the study, 424 (58%) were excluded either at their prescreening visit or following the detailed cardiovascular examination, predominantly because their BP was outside the grade 1 BP range. Thus, 302 participants were confirmed with grade 1 hypertension based on clinic-measured BP and/or ABPM criteria. Of these 302 participants, using a CASP cut-off point of ≥ 125 mmHg, 200 had high CASP, 100 had low CASP and CASP was not measurable in two participants. Figure 3 shows the normal distribution of CASP in the patients with grade 1 hypertension identified in the screening study. Figure 4 shows the study Consolidated Standards of Reporting Trials (CONSORT) flow diagram.

FIGURE 3.

Distribution of CASP values from the screening study population with grade 1 hypertension (300 participants with grade 1 hypertension and measurable CASP values). The cut-off point for high and low CASP was ≥ 125 mmHg and is shown by the vertical line.

FIGURE 4.

The TREAT CASP RCT CONSORT flow diagram.

The TREAT CASP stage 1 screening study population

Transition to the stage 2 TREAT CASP randomised controlled trial and the observational cohort study

As indicated in Stage 1 screening study demographics, 302 patients were identified who were eligible to enter the TREAT CASP RCT and the observational study. However, the stop–go checkpoint between stage 1 and stage 2 of the study meant that the screening study had to be completed before any of these patients could transition to the stage 2 RCT and observational studies.

As a consequence, many patients had to be re-contacted and re-screened many months, and, in some cases, almost 2 years, after they were originally deemed eligible. As a result, 98 potential participants in stage 2 were not contactable, were not available or were no longer eligible for further study participation. This reduced the pool of people available for recruitment into the second stage to 204 (see Figure 4). Of these 204 participants, a further 42 were excluded on re-screening as a result of the delayed start of stage 2, predominantly because of a change in their BP status while awaiting the stop–go decision. The remaining 162 men entered stage 2 of the study. Of these, 57 with low CASP entered the observational study and 105 with high CASP were randomised into the TREAT CASP RCT. For the RCT, 51 men were randomised to receive BP-lowering treatment and 54 men were randomised to no treatment. During follow-up, 17 participants (seven randomised to treatment, five randomised to no treatment and five participants in the observational study) were excluded or lost to study follow-up. Reasons for exclusion or loss are shown in Figure 4. By the end of the study, 145 men had completed follow-up, with 44 in the treatment arm, 49 in the no-treatment arm and 52 in the observational study.

Retrospective evaluation of the stop–go criterion based on baseline cardiac magnetic resonance imaging data from the stage 2 studies

The stop–go decision was taken to proceed to the second phase of the study (i.e. the TREAT CASP RCT and the observational study) on the basis of a limited number of cMRI scans (see Figure 2) that demonstrated that patients with high CASP had a higher LVMI than patients with low CASP. This decision was then retrospectively evaluated from the much larger number of cMRI data collected at baseline for the stage 2 study in patients with high and low CASP. Comparison between participants with high CASP (i.e. TREAT CASP RCT participants) and low CASP (i.e. the observational study participants) demonstrated a significant difference in LVMI between groups [low CASP LVMI (end-systole): 64.0 ± 8.5 g/m² (n = 54); high CASP LVMI (end-systole): 67.9 ± 8.8 g/m² (n = 101); difference 4.0 g/m², 95% CI 1.1 to 6.9 g/m²; p < 0.01; low CASP LVMI (end-diastole): 63.9 ± 8.3 g/m² (n = 54); high CASP LVMI (end-diastole): 67.5 ± 8.7 g/m² (n = 101); difference 3.7 g/m², 95% CI 0.8 to 6.5 g/m²; p = 0.01]. This confirmed that the decision to proceed to the second-stage RCT was appropriate.

The relationship between CASP and LVMI by linear regression for cMRI scans performed at baseline is shown in Figures 5 and 6. A clear and positive relationship between LVMI and CASP was seen and a higher CASP was associated with a higher LVMI. The relationships with clinic-measured and ambulatory SBP and LVMI are also shown in Figures 5 and 6. Regression analysis of the relationship between LVMI and BP showed similar regression coefficients and slopes, irrespective of the modality of BP measurement (ABPM, clinic BP or CASP).

FIGURE 5.

Relationship between LVMI at end-systole and central aortic, clinic-measured and ambulatory SBP. The regression line is shown with 95% CI.

FIGURE 6.

Relationship between LVMI at end-diastole and central aortic, clinic-measured and ambulatory SBP. The regression line is shown with 95% CI.

Formal testing of regression coefficients (Fisher’s r-to-z transformation) showed no significant differences between BP modalities:

• comparison of regression coefficients for LVMI at end-systole versus BP modality (format: regression 1: regression 2; comparison) –

  • LVMI versus CASP: LVMI versus brachial SBP; z = –0.19; p = 0.9

  • LVMI versus CASP: LVMI versus ABPM 24-hour SBP; z = –0.28; p = 0.8

  • LVMI versus CASP: LVMI versus ABPM daytime SBP; z = –0.66; p = 0.5.

• comparison of regression coefficients for LVMI at end-diastole (format: regression 1; regression 2; comparison) –

  • LVMI versus CASP: LVMI versus brachial SBP; z = –0.09; p = 0.9

  • LVMI versus CASP: LVMI versus ABPM 24-hour SBP; z = –0.28; p = 0.8

  • LVMI versus CASP: LVMI versus ABPM daytime SBP; z = –0.66; p = 0.5.

Similarly, comparisons of regression slopes (seemingly unrelated regression estimates) showed no differences between BP modalities:

• comparison of regression slopes for LVMI at end-systole versus BP modality (format: regression 1; regression 2; comparison) –

  • LVMI versus CASP: LVMI versus brachial SBP; χ² = 0.20; p = 0.7

  • LVMI versus CASP: LVMI versus ABPM 24-hour SBP; χ² = 1.62; p = 0.2

  • LVMI versus CASP: LVMI versus ABPM daytime SBP; χ² = 2.96; p = 0.1.

• comparison of regression slopes for LVMI at end-diastole (format: regression 1; regression 2; comparison) –

  • LVMI versus CASP: LVMI versus brachial SBP; χ² = 0.15; p = 0.7

  • LVMI versus CASP: LVMI versus ABPM 24-hour SBP; χ² = 1.65; p = 0.2

  • LVMI versus CASP: LVMI versus ABPM daytime SBP; χ² = 2.96; p = 0.1.

Taken together, these data provide evidence to refute the hypothesis for there being a superior relationship for CASP with LVMI, compared with any other BP modality used in this study at baseline.

Baseline characteristics for the TREAT CASP randomised controlled trial and the observational study

A total of 162 participants were recruited into the stage 2 TREAT CASP RCT and the observational follow-up study. The baseline characteristics of these participants are shown in Tables 5–11 and see also Table 17. The distribution of CASP values at baseline for the TREAT CASP RCT and the observational study in relation to the cut-off point defining high and low CASP is shown in Figure 7 and was similar to that observed for participants with grade 1 hypertension in the screening cohort (see Figure 4).

FIGURE 7.

Distribution of CASP values at baseline from the pooled TREAT CASP RCT (i.e. those participants with high CASP) and the observational study (i.e. those participants with low CASP) population with grade 1 hypertension (n = 162). The cut-off point for high and low CASP (≥ 125 mmHg) is shown by the vertical line.

The baseline characteristics were well matched between the TREAT CASP RCT randomised groups. There were no significant differences in baseline demographics and characteristics between the treatment and no-treatment arms of the TREAT CASP RCT. However, men in the treatment arm of the RCT were slightly older (mean 47.7 vs. 46.6 years; p = 0.37) and slightly heavier (mean 90.4 vs. 87.2 kg; p = 0.33) than men in the no-treatment arm. There was no difference in cardiovascular risk score (i.e. QRISK 2 at age 60 years, QRISK-lifetime) between the randomised groups.

Baseline characteristics for the TREAT CASP randomised controlled trial and the observational study by high and low CASP status

The participants with low CASP in the observational study were younger than those with high CASP in the TREAT CASP RCT (age: observational study, 41.9 ± 8.7 years; TREAT CASP RCT, 47.1 ± 6.2 years; difference –5.3 years, 95% CI –7.9 to –2.7 years; p < 0.001) and had lower clinic BP (clinic SBP: observational study, 133.3 ± 8.2 mmHg; TREAT CASP RCT, 145.5 ± 6.9 mmHg; difference –12.3 mmHg, 95% CI –14.7 to –9.9 mmHg, p < 0.001; clinic DBP: observational study, 80.4 ± 5.1 mmHg; TREAT CASP RCT, 89.7 ± 6.4 mmHg; difference –9.3 mmHg, 95% CI –11.2 to –7.3 mmHg, p < 0.001) and CASP (CASP: observational study, 118.0 ± 5.4 mmHg; TREAT CASP RCT, 133.4 ± 6.3 mmHg; difference –15.4 mmHg, 95% CI –17.4 to –13.5 mmHg; p < 0.001).

The participants with low CASP in the observational study also had lower ambulatory BP (24-hour SBP: observational study, 132.1 ± 6.5 mmHg; TREAT CASP RCT, 137.2 ± 6.7 mmHg; difference –5.2 mmHg, 95% CI –7.3 to –3.0 mmHg, p < 0.001; 24-hour DBP: observational study, 83.0 ± 5.8 mmHg; TREAT CASP RCT, 87.1 ± 5.8 mmHg; difference –4.1 mmHg, 95% CI –6.0 to –2.2 mmHg, p < 0.001). Participants with low CASP in the observational study also had higher heart rates on ECG than those with high CASP in the TREAT CASP RCT (heart rate: observational study, 68.6 ± 12.9 beats per minute; TREAT CASP RCT, 61.3 ± 9.5 beats per minute; difference 7.3 beats per minute, 95% CI 3.4 to 11.2 beats per minute; p < 0.001).

Baseline blood pressure characteristics for the TREAT CASP randomised controlled trial and the observational study

Patients were entered into the study with grade 1 hypertension according to clinic BP and/or ABPM. No patient with grade 1 hypertension as determined by clinic BP had a normal ABPM (i.e. < 135/85 mmHg daytime mean). The majority of participants (81%) with low CASP had high–normal clinic BP (i.e. 130–139/80–89 mmHg) but qualified for the study because their ABPM was in the grade 1 hypertensive range. In the observation study group, 19% of participants had a clinic BP in the grade 1 hypertension range. In contrast, the majority of participants (85%) with high CASP were classified with grade 1 hypertension as determined by clinic BP. The remainder of the participants qualified because their ABPM was in the grade 1 hypertension range. There was no difference in baseline cardiovascular risk score between the observational group and TREAT CASP RCT participants (QRISK 2 at age 60 years: observational study, 8.2% ± 4.9%; TREAT CASP RCT, 7.2% ± 3.7%, difference 1.1%, 95% CI –0.4% to 2.5%; p = 0.16. QRISK-lifetime: observational study, 47.9% ± 18.0%; TREAT CASP RCT, 47.1% ± 16.0%, difference 0.8%, 95% CI –4.6% to 6.3%; p = 0.8).

Baseline pressure amplification characteristics for the TREAT CASP randomised controlled trial and the observational study

Pressure amplification at baseline differed, as expected, between men with low CASP in the observational study and men with high CASP in the TREAT CASP RCT (low CASP vs. high CASP: BrSBP minus CASP difference: 3.1 mmHg, 95% CI 1.5 to 4.8 mmHg, p < 0.001; SBP amplification difference: 0.04, 95% CI 0.02 to 0.05, p < 0.001; PP amplification difference: 0.13, 95% CI 0.08 to 0.18, p < 0.001; see Table 8).

Relationship between clinic blood pressure and central aortic systolic pressure in high and low central aortic systolic pressure patient groups

The relationship between CASP and BrSBP for 663 men recruited into the stage 1 screening study is shown in Figure 8. Figure 8 shows that there was a good relationship between CASP and BrSBP, with moderate scatter around the line of regression. This indicates that CASP and BrSBP track well with each other. It was also demonstrated that the median BrSBP for patients in the high CASP range was 148.5 mmHg and 40% of patients with high CASP had a BrSBP in the upper half of the grade 1 hypertension range (i.e. ≥ 150–159 mmHg). In this BrSBP range (i.e. ≥ 150–159 mmHg), low CASP (i.e. < 125 mmHg) was found in only seven people (≈ 1%). In fact, across the entire SBP range for grade 1 hypertension (i.e. ≥ 140–159 mmHg), low CASP was found in only 53 people (i.e. only 8% of the initial screening population). This indicates that the prevalence of high systolic pressure amplification resulting in low CASP in younger patients with grade 1 hypertension is actually quite low. This illustrates why matching the BrSBP in groups with high and low CASP was very difficult to achieve.

FIGURE 8.

The relationship between CASP and BrSBP for participants in the TREAT CASP screening study. (a) Linear regression between SBP and CASP showing the prevalence of SBP in the grade 1 hypertension range (140–159 mmHg) and low CASP (i.e. < 125 mmHg). Points in green show participants with a SBP of 140–149 mmHg and low CASP. Points in blue show participants with a SBP of 150–159 mmHg and low CASP. (b) Frequency distributions for SBP in people with low CASP (green) and people with high CASP (black shading). Blue and light-green areas show the overlap of SBP for people with low CASP and a SBP of 140–159 mmHg (light green) and 150–159 mmHg (blue).

Interestingly, the data in Figure 8a also indicate that the line of regression between SBP and CASP intersects a CASP value of 125 mmHg for a SBP of 140 mmHg (the threshold BP for grade 1 hypertension). This suggests that the level of CASP used in this study to categorise high and low CASP values was appropriate for people with grade 1 hypertension.

Seated clinic blood pressure changes during randomised controlled trial and observational study follow-up

Seated clinic BP and CASP during study follow-up are shown in Figure 9 and Table 8.

FIGURE 9.

Clinic-measured SBP (a), DBP (b) and CASP (c) (mean ± SD) throughout the TREAT CASP RCT. The SBP, DBP and CASP values are presented as means ± SDs. The AUC data (d) are displayed as means and 95% CIs. **p < 0.01. AUC, area under the curve.

The average SBP, DBP and CASP were reduced from baseline in both the treatment and no-treatment groups and these reductions were sustained throughout the course of study follow-up. By the end of the study follow-up (i.e. study closeout), the changes from baseline were:

• treatment –

  • BP change: –20.0 mmHg (95% CI –23.3 to –16.6 mmHg)/–13.0 mmHg (95% CI –15.0 to –11.1 mmHg)

  • CASP change: –21.1 mmHg (95% CI –24.4 to –17.9 mmHg).

• no treatment –

  • BP change: –9.6 mmHg (95% CI –12.9 to –6.3 mmHg)/–5.8 mmHg (95% CI –7.9 to –3.7 mmHg)

  • CASP change: –10.2 mmHg (95% CI –13.3 to –7.1 mmHg).

Using area under the curve, the changes in SBP, DBP and CASP (baseline to study closeout) were significantly greater for the treatment arm than the no-treatment arm of the TREAT CASP RCT (all p < 0.001) (see Figure 9).

By contrast, clinic BP and CASP showed no significant change during follow-up for participants in the observational study [clinic BP change –1.7 mmHg (95% CI –4.3 to 1.0 mmHg)/0.1 mmHg (95% CI –1.8 to 1.9 mmHg); CASP change –1.0 mmHg (95% CI –3.3 to 1.3 mmHg); see Table 8 and Figure 10].

FIGURE 10.

Clinic-measured SBP and DBP (a) and CASP (b) (mean ± SD) throughout the observational study.

Ambulatory blood pressure changes during TREAT CASP RCT and observational study follow-up

Ambulatory BP was measured at baseline and study closeout in the TREAT CASP RCT and observational study. For participants in the treatment arm, ambulatory BP decreased significantly between baseline and study closeout [24-hour BP change –10.3 mmHg (95% CI –12.8 to –7.9 mmHg)/–6.9 mmHg (95% CI –8.6 to –5.2 mmHg); p < 0.001 for both]. By contrast, there was no significant change for participants in the no-treatment arm [24-hour BP change –0.1 mmHg (95% CI –2.3 to 2.2 mmHg; p = 0.95)/–0.3 mmHg (95% CI –1.7 to 1.1 mmHg; p = 0.7)]. No significant change in ambulatory BP was observed in the observational study [24-hour BP change: –1.6 mmHg (95% CI –3.6 to 0.5 mmHg; p = 0.1)/–0.3 mmHg (95% CI –1.7 to 1.1 mmHg; p = 0.7)]. Although there was little change in ambulatory BP during study follow-up for participants in the no-treatment arm and observational study, there was considerable variability in the range of values for ABPM change within individual participants. These data are shown in Tables 9–11 and Figures 11 and 12.

FIGURE 11.

Ambulatory SBP during RCT and observational study follow-up. Change in 24-hour ambulatory SBP between baseline and study closeout by randomised or allocated study group (a–c – upper panel) showing absolute number (and percentage) of participants with positive and negative changes. Absolute values for 24-hour ambulatory SBP at baseline and study closeout with means ± SDs also shown (d–f – lower panel). Significant differences (p < 0.001) within participants (overtime) are indicated with ***.

FIGURE 12.

Ambulatory DBP during RCT and observational study follow-up. Change in 24-hour ambulatory DBP between baseline and study closeout by randomised or allocated study group (a–c – upper panel) showing absolute number (and percentage) of participants with positive and negative changes. Absolute values for 24-hour ambulatory SBP at baseline and study closeout with means ± SDs also shown (d–f – lower panel). Significant differences (p < 0.001) within the participants (overtime) are indicated with ***.

Medication time course

The treatment dose and number of treatments are described as a defined daily dose (DDD) in accordance with the WHO’s classification index. 109 According to the DDD method, the starting dose of any drug (i.e. 50 mg of losartan o.d. or 5 mg of amlodipine o.d. is defined as 1 DDD). One hundred mg of losartan o.d. or 10 mg of amlodipine o.d. each represent 2 DDDs, or a combination of 50 mg of losartan and 5 mg of amlodipine o.d. is 2 DDDs (see Table 4). Participants randomised to treatment, received treatment for, an average of, 373 ± 16 days. The majority of participants in the treatment arm completed the study on 2 DDDs (n = 26). Ten participants completed the study on 3 DDDs, whereas seven participants were treated with 1 DDD throughout the trial. One participant discontinued taking study medication shortly after randomisation, but was followed up to study closeout according to intention to treat. For those participants receiving more than one DDD, mean time to first up-titration of treatment was 63 ± 45 days and 136 ± 72 days to the second up-titration. For all study participants, the average time spent on 1 DDD was 118 ± 125 days, 2 DDDs was 243 ± 114 days and 3 DDDs was 244 ± 69 days. The total time spent on each DDD for participants completing the study on 1, 2 or 3 DDDs is shown in Table 12. None of the study participants allocated to treatment had their medication down-titrated by the study physician during follow-up. Those participants randomised to no treatment were followed up for, an average of, 378 ± 36 days.

Outcomes

Primary end point: change in left ventricular mass index

The primary outcome for the TREAT CASP RCT was the change in LVMI (i.e. baseline to study closeout), comparing the treatment versus the no-treatment arms. Analysis was performed according to original randomised group and LVMI was evaluated at both end-systole and end-diastole. Baseline LVMI was similar between the treatment arms [treatment vs. no treatment, baseline LVMI (end-systole): 67.9 ± 8.9 vs. 67.9 ± 8.7 g/m²; baseline LVMI (end-diastole): 67.6 ± 8.7 vs. 67.5 ± 8.8 g/m²].

Unadjusted analysis (Student’s t-test) for the primary end point showed a significantly greater reduction in LVMI with treatment than with no treatment [treatment vs. no treatment: LVMI (end-systole) change (baseline to study closeout) –3.3 g/m² (95% CI –4.5 to –2.2 g/m²) vs. –0.9 g/m² (95% CI –1.7 to –0.2 g/m²); difference –2.4 g/m² (95% CI –3.8 to –1.0 g/m²); p < 0.01] (Table 13 and Figure 13). A similar effect of treatment was observed with the evaluation of the primary end point at end-diastole [treatment vs. no treatment: LVMI (end-diastole) change (baseline to study closeout) –3.4 g/m² (95% CI –4.5 to –2.2 g/m²) vs. –1.0 g/m² (95% CI –1.7 to –0.3 g/m²); difference –2.4 g/m² (95% CI –3.7 to –1.0 g/m²; p < 0.01] (Table 13 and Figure 13). Effect sizes calculated using Cohen’s d statistic were –0.74 (LVMI end-systole) and –0.75 (LVMI end-diastole), indicating a medium-to-large effect of treatment.

TABLE 13 - Primary outcome: LVMI at baseline and study closeout for the TREAT CASP RCT and observational study, by study group

*p < 0.05, **p < 0.01, treatment vs. no-treatment arm for the change (closeout minus baseline).

LVMI diastole, LVMI at end-diastole indexed to body surface area; LVMI systole, LVMI at end-systole indexed to body surface area.

Notes

The effect size, calculated using Cohen’s d statistic, was –0.74 for LVMI evaluated at end-systole and –0.75 for LVMI evaluated at end-diastole.

Baseline data represent the intention-to-treat population.

Parameter	Time point, mean ± SD		Δ from baseline		Number of participants
	Baseline	Closeout	Mean	95% CI
LVMI systole (g/m2)
Treatment	67.9 ± 8.9	65.1 ± 8.6	–3.34**	–4.53 to –2.15	44
No treatment	67.9 ± 8.7	67.5 ± 8.7	–0.94	–1.66 to –0.22	49
Observational	64.0 ± 8.5	63.9 ± 8.1	–0.49	–1.22 to 0.23	52
LVMI diastole (g/m2)
Treatment	67.6 ± 8.7	64.8 ± 8.4	–3.37**	–4.52 to –2.23	44
No treatment	67.5 ± 8.8	67.0 ± 8.7	–1.00	–1.73 to –0.27	49
Observational	63.9 ± 8.3	63.6 ± 7.9	–0.66	–1.40 to 0.09	52

FIGURE 13.

Primary end point for the TREAT CASP RCT showing the change in LVMI (baseline to study closeout), by randomised group. Data show individual values with means ± 95% CIs at end-systole (a) and end-diastole (b). Significant differences (i.e. p-values < 0.01) between groups are indicated with **. The effect sizes calculated using Cohen’s d statistic were –0.74 for LVMI evaluated at end-systole and –0.75 for LVMI evaluated at end-diastole.

The primary end point was then analysed using an analysis of covariance model. The change in LVMI (baseline to study closeout) at end-systole was adjusted for baseline covariates, including age, LVMI, clinic BP, heart rate, smoking status, waist-to-hip ratio and trunk fat-free mass. Correlations within the model were found for baseline LVMI (p = 0.06), heart rate (p < 0.01), smoking status (p = 0.06) and trunk fat-free mass (p < 0.01). The significant effect of treatment on the change in LVMI (baseline to study closeout) was maintained after adjustment [treatment: adjusted LVMI change (end-systole) –3.2 g/m², 95% CI –4.1 to –2.3 g/m²; no treatment: –1.1 g/m², 95% CI –2.0 to –0.2 g/m²; difference –2.1 g/m², 95% CI –0.8 to –3.4 g/m²]. Similar results were obtained for the change in LVMI at end-diastole [treatment: adjusted LVMI change (end-diastole) –3.2 g/m², 95% CI –4.1 to –2.3 g/m²; no treatment: –1.1 g/m², 95% CI –2.0 to –0.3 g/m²; difference –2.1 g/m², 95% CI –3.3 to –0.8 g/m²].

Secondary end point: comparison of change in left ventricular mass indices across observational study and TREAT CASP randomised controlled trial participants

Comparing the changes in LVMI and unindexed LV mass after 12 months, for all three groups [treatment group in the TREAT CASP RCT (i.e. high CASP), no-treatment group in the TREAT CASP RCT (i.e. high CASP) and observational study (low CASP)], showed that the reduction in LVMI and unindexed LV mass was significantly greater with treatment than for either of the other two groups (i.e. no treatment and observational study; see Tables 13 and 14 and Figure 14). BP-lowering treatment in people with high CASP reduced their LVMI to values approaching those seen in the observational study group [treatment vs. observational: LVMI (end-systole) at study closeout, 65.1 ± 8.6 vs. 63.9 ± 8.1 g/m²; difference –1.2 g/m², 95% CI –4.6 to 2.2 g/m², p = 0.47; treatment vs. observational: LV mass (end-systole) at study closeout, 135.7 ± 23.2 vs. 132.4 ± 22.9 g; difference –3.3 g, 95% CI –12.7 to 6.1 g, p = 0.48]; see Table 14 and Figure 13. By contrast, at study closeout, LVMI was higher in the TREAT CASP RCT no-treatment group than in the observational group [no treatment vs. observational: LVMI (end-systole), 67.5 ± 8.7 vs. 63.9 ± 8.1 g/m²; difference –3.6 g/m², 95% CI –6.9 to –0.3 g/m²; p = 0.04].

FIGURE 14.

Secondary end point showing the change in LVMI (a and b) and unindexed LV mass (c and d) between baseline and study closeout, by study group, for the TREAT CASP RCT and the observational study. Data show individual values with means ± 95% CIs at end-systole (a and c) and end-diastole (b and d). Significant differences (i.e. p-values < 0.01) between the treatment and no-treatment groups are indicated with **.

Secondary end point: 12-lead electrocardiogram voltages at baseline and changes during follow-up

Electrocardiography voltage criteria are used clinically as a marker of LVH110 and were evaluated in this study as a prespecified secondary end point. ECG voltage was assessed by three common criteria to approximate LV mass:

1. Cornell voltage

2. voltage in lead aVL

3. Sokolow–Lyon voltage. 110

At baseline, voltage in lead aVL was significantly higher for participants with high CASP versus low CASP (high CASP vs. low CASP: aVL voltage 662.9 ± 398.6 µV vs. 501.1 ± 357.6 µV; p = 0.01). Cornell voltage was also higher at baseline for participants with high CASP, but the difference did not achieve significance (high CASP vs. low CASP: Cornell voltage 1.7 ± 0.6 mV vs. 1.6 ± 0.6 mV; p = 0.4).

In the TREAT CASP RCT, BP lowering was associated with a reduction in both Cornell and aVL voltages, although this did not quite achieve significance for voltage in lead aVL (treatment: Cornell voltage change (baseline to study closeout) –0.2 mV, 95% CI –0.2 to –0.1 mV, p < 0.01; aVL voltage change –51.5 µV, 95% CI –104.6 to 1.6 µV, p = 0.06). By contrast, Cornell and aVL voltage both increased after 12 months’ follow-up for the no-treatment group (no treatment: Cornell voltage change 0.1 mV, 95% CI 0.0 to 0.2 mV, p < 0.01; aVL voltage change 68.7 µV, 95% CI 5.0 to 132.4 µV, p = 0.035); Table 17 and Figure 15. Taken together, a highly significant difference was seen for the change in voltage for these parameters between TREAT CASP RCT groups (i.e. treatment vs. no treatment; difference in Cornell voltage change: –0.3 mV, 95% CI –0.4 to –0.2 mV, p < 0.001; difference in aVL voltage change: –120.2 µV, 95% CI –202.1 to –37.8 µV, p < 0.01). Effect sizes using Cohen’s d statistic were –0.97 for Cornell voltage change and –0.62 for aVL voltage change, thus suggesting large and medium effects of treatment, respectively. Differences were maintained after adjustment for baseline covariates (treatment: adjusted Cornell voltage change –0.1 mV, 95% CI –0.2 to –0.1 mV; no treatment: adjusted Cornell voltage change 0.1 mV, 95% CI 0.0 to 0.2 mV; difference –0.2 mV, 95% CI –0.4 to –0.1 mV; treatment-adjusted aVL voltage change –49.1 µV, 95% CI –109.2 to 11.0 µV; no treatment-adjusted aVL voltage change 66.4 µV, 95% CI 8.4 to 124.4 µV; difference –115.5 µV, 95% CI –201.0 to –30.0 µV).

TABLE 17 - Electrocardiogram parameters at baseline and study closeout for the TREAT CASP RCT and the observational study, by study group

*p < 0.05, **p < 0.01, treatment vs. no-treatment arm for the change (closeout minus baseline).

Note

Baseline data represent the intention-to-treat population.

Parameter	Time point, mean ± SD		Δ from baseline		Number of participants
	Baseline	Closeout	Mean	95% CI
ECG rate (beats per minute)
Treatment	60.8 ± 10.3	62.9 ± 10.6	2.22	–0.86 to 5.30	44
No treatment	61.7 ± 8.7	63.8 ± 10.3	2.34	–0.41 to 5.09	47
Observational	68.6 ± 12.9	65.3 ± 11.5	–2.13	–5.18 to 0.93	51
PR interval (milliseconds)
Treatment	169.0 ± 26.1	169.4 ± 24.8	–0.05*	–3.73 to 3.64	44
No treatment	164.8 ± 18.8	167.7 ± 18.6	4.04	0.57 to 7.52	47
Observational	161.2 ± 21.6	164.4 ± 21.9	1.14	–1.72 to 3.99	51
QRS duration (milliseconds)
Treatment	96.0 ± 10.7	95.4 ± 12.0	–0.45	–2.55 to 1.64	44
No treatment	93.8 ± 7.1	93.6 ± 6.9	–0.04	–1.15 to 1.06	47
Observational	99.0 ± 13.9	99.5 ± 14.5	0.59	–0.57 to 1.74	51
QTc (milliseconds)
Treatment	405.5 ± 23.6	407.0 ± 24.8	2.23	–4.26 to 8.71	44
No treatment	404.2 ± 19.0	407.8 ± 21.4	2.72	–2.29 to 7.73	47
Observational	407.0 ± 23.3	405.3 ± 17.5	–0.86	–6.34 to 4.61	51
Cornell voltage (mV)
Treatment	1.7 ± 0.6	1.5 ± 0.4	–0.15**	–0.24 to –0.06	43
No treatment	1.6 ± 0.5	1.7 ± 0.5	0.12	0.04 to 0.21	46
Observational	1.6 ± 0.6	1.6 ± 0.7	–0.02	–0.10 to 0.06	50
aVL voltage (µV)
Treatment	679.7 ± 380.5	644.9 ± 356.5	–51.51**	–104.59 to 1.57	43
No treatment	646.8 ± 418.2	652.4 ± 422.5	68.70	4.99 to 132.41	46
Observational	501.1 ± 357.6	518.8 ± 340.0	21.18	–23.37 to 65.73	51
Sokolow–Lyon voltage (mV)
Treatment	2.4 ± 0.7	2.3 ± 0.7	–0.08	–0.20 to 0.05	38
No treatment	2.5 ± 0.7	2.5 ± 0.8	–0.11	–0.23 to 0.00	42
Observational	2.4 ± 0.8	2.4 ± 0.9	–0.05	–0.17 to 0.08	49

FIGURE 15.

Change in ECG Cornell voltage (a), voltage in lead aVL (b) and Sokolow–Lyon voltage (c) between baseline and study closeout for the TREAT CASP RCT and the observational study, by study group. Significant differences (i.e. p-values of < 0.01) between the treatment and no-treatment groups are indicated with **.

In the observational study of low CASP participants, the Cornell and aVL voltages were unchanged after 12 months’ follow-up [Cornell voltage change (baseline to study closeout): 0.0 mV, 95% CI –0.1 to 0.1 mV; aVL change: 21.2 µV, 95% CI –23.4 to 65.7 µV; see Table 17 and Figure 15].

The Sokolow–Lyon voltage criteria showed no significant differences between groups (low CASP vs. high CASP) at baseline or study closeout. Furthermore, there was no change in Sokolow–Lyon voltage at study closeout after treatment for 12 months. There was, however, a small reduction in Sokolow–Lyon voltage after 12 months’ follow-up for the no-treatment arm [no treatment Sokolow–Lyon voltage change (baseline to study closeout): –0.1 mV, 95% CI –0.2 to 0.0 mV, p = 0.05; see Table 17 and Figure 15].

Urinary albumin-to-creatinine ratio

The UACR was measured at baseline and study closeout. As this study was in a relatively healthy study population without evidence of HMOD, diabetes mellitus or CKD, as expected, many urine samples were returned with undetectable albumin levels. A total of 55 participants (treatment, n = 15; no treatment, n = 19; observational, n = 21) were identified with detectable urine albumin at both baseline and study closeout. Comparison of the change in UACR in the TREAT CASP RCT showed a modest reduction with treatment relative to that seen in the no-treatment group [treatment: UACR change (baseline to study closeout) –0.5, 95% CI –1.1 to 0.1; no treatment: UACR change 0.1, 95% CI –0.2 to 0.4; difference –0.6, 95% CI –1.2 to 0.0, p < 0.05; Figure 16]. However, this change may have been influenced by one outlying data point in the treatment group, as exclusion of these data points yielded a slightly smaller change (treatment: UACR change (baseline to study closeout) –0.3, 95% CI –0.7 to 0.1; no treatment: UACR change 0.1, 95% CI –0.2 to 0.4; difference –0.4, 95% CI –0.9 to 0.1; p = 0.09). The effect size, calculated using Cohen’s d statistic, was –0.74, suggesting a medium-to-large effect of treatment. However, the difference for the change in UACR (baseline to study closeout) between treatment arms for the TREAT CASP RCT (including the outlier described above) was attenuated after adjustment for baseline variables [treatment: adjusted UACR change (baseline to study closeout) –0.5, 95% CI –1.0 to 0.0; no treatment: UACR change 0.1, 95% CI –0.4 to 0.6; difference –0.6, 95% CI –1.4 to 0.2].

FIGURE 16.

Change in UACR across the TREAT CASP RCT, by randomised group. Absolute values with means and 95% CIs for change in UACR in the TREAT CASP RCT are shown. Significant differences (i.e. p-values of < 0.05) between the treatment and no-treatment groups are indicated with *.

Eleven (73.3%) of those participants with detectable urine albumin, at both baseline and study closeout, in the high CASP group who received BP-lowering treatment showed a reduction in their UACR, whereas four (26.7%) showed an increase in UACR from baseline after 12 months. In the high CASP group, which did not receive treatment, nine (47.4%) participants showed a reduction in UACR, whereas 10 (52.6%) showed an increase in UACR from baseline after 12 months (Figure 17).

FIGURE 17.

Change in UACR across the TREAT CASP RCT and the observational study, by study group. Individual changes (a–c) and absolute values with means ± SDs (d–f) for UACR at baseline and study closeout across all study groups.

A small reduction in UACR was also observed in the observational group (change in UACR baseline to study closeout –0.3, 95% CI –0.8 to 0.1). Three-way comparison of the change in UACR showed no difference between groups.

Safety analyses

Study strengths and weaknesses

This study had a number of strengths, the most important of which is that it addressed the paucity of data on the impact of BP-lowering treatment in younger, low-risk people with hypertension. Our study clearly shows that such people with CASP levels that corresponded to a conventional clinic BP level in the grade 1 hypertensive range (i.e. 140–159/90–99 mmHg), and confirmed by an elevated ABPM, have early evidence of cardiac remodelling and increased LV mass that is most probably attributable to increased BP because it was reduced by BP-lowering treatment. An additional strength of the study was the comprehensive characterisation of clinical phenotype of participants with grade 1 hypertension. In this, BP and changes in BP were measured using both clinic and ambulatory methods. Few other studies have reported both parameters across a clinical trial. Use of cMRI as a robust primary end point is also a study strength, with a wealth of parameters regarding cardiac structure and function being collected in a single measurement. Many of these parameters remain to be analysed and will generate future reports from this study. We also performed additional phenotypic measurements at baseline and study closeout, such as CAVI, to examine large artery function, and non-mydriatic retinal photography, to examine small artery changes, which will be the subject of future reports. Our finding that a standard ECG was able to detect beneficial changes in LV mass with treatment in younger people is also of interest with regard to implementation in routine clinical practice. With regard to practice, although our hypothesis was proven (i.e. that high CASP predicted a higher LVMI and that treatment of high CASP reduced LVMI back towards levels seen in patients with low CASP), we also discovered that the use of CASP offered no additional predictive value over and above the careful characterisation of BrBP in the clinic, complemented by ABPM, suggesting that conventional BP measurement would be sufficient to stratify these patients for treatment.

Our study has some weaknesses. We discovered that the prevalence of low CASP in patients with grade 1 hypertension was very low (≈ 8%) and ≈ 1% in those with a BrSBP in the upper range of men in the high CASP cohort, making it very difficult to match BrBP levels for those with high and low CASP and, thus, impossible to dissect the independent effect, if any, of CASP on LVMI. In so doing, we confirmed that CASP has no added or practical value in stratifying treatment decisions for younger patients with grade 1 hypertension.

We recruited exclusively men because of the complexity of recruiting and exposing women of childbearing potential to treatment. However, there is no evidence that BP-lowering treatment is any less effective at reducing LV mass and preventing subsequent cardiovascular events in women. On the contrary, increased LVMI is more strongly predictive of adverse cardiovascular outcomes in women versus men126 and BP lowering has been shown to regress LV mass to a greater extent in women than in men, after adjusting for differences in baseline LV mass. 130 We were unable to use a traditional double-blind study design or a placebo in the no-treatment arm because of the need to up-titrate treatment to achieve a predefined reduction in BP. Instead we used a PROBE design in which randomised treatment was open label, but the evaluation of end points was blind to treatment allocation. In reality, the clinic and ABPM BP responses to no treatment in the TREAT CASP RCT had little difference in magnitude of change to those seen in placebo-controlled trials of BP-lowering treatment. Another consideration is whether or not studies of this kind are too small to influence clinical practice. We used cMRI to record changes in LV mass and such studies have sufficient power with the number of patients we studied. Very much larger studies would be required to evaluate major cardiovascular events and mortality in this patient group but, as discussed in Chapter 1, Evaluating the benefit of treatment in young people with grade 1 hypertension, these are unlikely to happen because of the low rates of such end points in younger people. Thus, well-established portents of future cardiovascular risk, such as LV mass, are the only feasible option to evaluate potential benefits of treatment in this age group.

Clinical implications

There is a paucity of clinical trial data on the potential benefits of BP-lowering treatment in younger people (i.e. those aged < 55 years) with grade 1 hypertension and uncertainty about whether or not BP-lowering treatment is indicated. We show that the non-invasive measurement of CASP provided no added value over carefully measured clinic BP or ABPM in stratifying the risk of an increased LVMI in these patients. However, we also show that younger men with grade 1 hypertension and high CASP (corresponding to a higher BrBP in the grade 1 hypertensive range) but otherwise defined as low risk are already developing cardiac changes (increased LVMI) due to increased BP. We also show that the increase in LVMI is reversible with BP-lowering treatment and that treatment was well tolerated. As an increase in LV mass is the most well-characterised BP-related biomarker of future risk of CVD and premature mortality, these findings suggest that younger people with grade 1 hypertension, even if otherwise characterised as low risk, are more likely to have an increased LV mass and that this increase can be regressed by BP-lowering treatment.

Research implications

Previous studies have suggested that CASP is highly variable in patients with similar BrBP measurements and may be a better predictor of HMOD, such as increased LV mass. We found that CASP was variable, but that this variability was mainly a function of the BrBP from which it was derived and only rarely resulted in grade 1 hypertensive patients who had a low CASP level. Moreover, higher CASP values corresponded to higher BrBP levels and the high CASP threshold of ≥ 125 mmHg used in this study corresponds to the current guideline-recommended threshold for grade 1 hypertension, if the clinic BP elevation was also confirmed by ABPM. We found grade 1 hypertensive patients rarely had a low CASP level. In this regard, the measurement of CASP provided no incremental value in predicting LVMI over carefully measured clinic BP. This suggests that future studies evaluating BP-lowering treatment effects in younger patients should be based on carefully measured clinic BrBP, complemented by ABPM, rather than clinic BP alone, to more accurately define the hypertensive phenotype.

A key question generated by this study is whether or not younger patients with grade 1 hypertension, as well as having cardiac structural changes (increased LVMI), also have LV functional changes (i.e. early evidence of impaired systolic function or diastolic relaxation) and, if so, is this reversible with treatment? The cMRI data we have acquired in this study will allow us to address these questions. Whether or not impaired diastolic function is already manifested in younger patients with grade 1 hypertension is unknown. This question is of particular interest because diastolic dysfunction and, ultimately, heart failure with preserved ejection fraction (HFpEF) are closely linked to the development of LVH in hypertensive patients. Future analyses from cMRI data acquired in our study will examine conventional measures of diastolic function, such as E/A ratios, as well as more contemporary measures using tissue-phase mapping, to determine if diastolic function is already impaired in younger people with grade 1 hypertension and, if so, whether or not this has improved with BP-lowering treatment. The impact of BP-lowering treatment has never been assessed before in younger hypertensive patients and is particularly important because it is well established that hypertension is a major risk factor for diastolic dysfunction leading to HFpEF in later life. Furthermore, in these older patients, HFpEF does not appear to be reversible with treatment. Thus, it will be intriguing to determine if early signs of diastolic dysfunction are present in the participants in this study and, if so, whether or not it can be reversed by early intervention with BP-lowering treatment. This study also highlights the urgent need for more research in younger people with grade 1 hypertension and deemed to be at low short-term risk to determine if there is a way to better stratify those who might benefit from early initiation of BP-lowering drug therapy.

Trial Steering Committee

The Trial Steering Committee comprised:

• Professor Tony Heagerty (chairperson), Division of Cardiovascular Sciences, University of Manchester, UK

• Professor Paul Leeson, Medical Sciences Division, University of Oxford, UK

• Professor John Potter, Foundation Chair in Ageing and Stroke Medicine, University of East Anglia, UK

• Professor Bryan Williams, Institute of Cardiovascular Science, UCL, UK

• Professor Clive Osmond, Department of Human Development and Health, University of Southampton, UK.

Contributions of authors

Bryan Williams (https://orcid.org/0000-0002-8094-1841) (Professor of Medicine, UCL, Institute of Cardiovascular Science, Director of Research, UCLH, NHS Foundation Trust, Director of the UCL/UCLH NIHR BRC) was the chief investigator. He devised the study, obtained funding, oversaw the design and conduct of the project and the interpretation and reporting of the results and drafted the manuscript with Peter S Lacy.

Ewan McFarlane (https://orcid.org/0000-0002-5183-8386) (Research Assistant, UCL, Institute of Cardiovascular Science) was involved in patient recruitment and data collection and processing and MRI data analysis. He was involved in the drafting and reviewing of the manuscript.

Dawid Jedrzejewski (https://orcid.org/0000-0001-5111-4807) (Research Assistant, UCL, Institute of Cardiovascular Science) was involved in patient recruitment and data collection and processing. He was involved in the drafting and reviewing of the manuscript.

Peter S Lacy (https://orcid.org/0000-0002-5916-6465) (Senior Research Associate, UCL, Institute of Cardiovascular Science) was involved in the design and acquisition of funding for the study. He contributed substantially to protocol design and revision, and the processing and interpretation of the study data, and drafted the manuscript.

Data-sharing statement

All data requests should be submitted to the corresponding author for consideration. Access to available anonymised data may be granted following review.

Patient data

This work uses data provided by participants and collected by the NHS as part of their care and support. Using patient data are vital to improve health and care for everyone. There is huge potential to make better use of information from people’ s patient records, to understand more about disease, develop new treatments, monitor safety, and plan NHS services. Patient data should be kept safe and secure, to protect everyone’s privacy, and it is important that there are safeguards to make sure that it is stored and used responsibly. Everyone should be able to find out about how patient data are used. #datasaveslives you can find out more about the background to this citation here: https://understandingpatientdata.org.uk/data-citation.

Disclaimers

This report presents independent research. The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, the MRC, NETSCC, the EME programme or the Department of Health and Social Care. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, NETSCC, the EME programme or the Department of Health and Social Care.

Account of the event reported to be published in the University College London Hospitals Biomedical Research Centre’s annual report and website

Photographs taken during the TREAT CASP trial’s patient and public involvement event

Programme schedule for the TREAT CASP trial’s patient and public involvement event

(PDF download)

The TREAT CASP trial’s patient and public involvement participant questionnaire and summary of responses from 25 attendees

Individual comments and question answers provided anonymously by study participants at the TREAT CASP trial’s patient and public involvement event

Sample copy of the TREAT CASP trial’s patient and public involvement event questionnaire

(PDF download)