Control, Fludrocortisone or Midodrine for the treatment of Orthostatic Hypotension (CONFORM-OH): results from an internal pilot randomised controlled trial

Abstract

Background

Orthostatic hypotension (OH) is a common debilitating condition characterised by a significant drop in blood pressure (BP) on standing upright. Adults with OH are typically offered non-pharmacologic therapies, either alone or in combination with medication. The two most used agents are fludrocortisone and midodrine. There is a lack of good quality evidence for any of these treatments, all of which are in widespread clinical use. The aim of this internal pilot trial was to evaluate recruitment, attrition, treatment crossover and quality of outcomes.

Methods

The trial was designed as a pragmatic, open-label, randomised, prospective, multicentre, superior and multi-arm internal pilot. Recruitment was for 10 months, during which we aimed to open 14 sites and recruit a target of 64 adults with OH to evaluate feasibility of recruitment. Attrition, crossover and data collection were also assessed. Participants were randomised to one of three treatments: nondrug therapies (control), fludrocortisone plus nondrug therapies or midodrine plus nondrug therapies. Outcomes measured included symptoms, quality of life, activities of daily living, postural BP, use of health and care services, falls and safety. Participants received treatment and were followed up for 12 months. Pre-planned criteria to progress from internal pilot were defined for recruitment, retention, crossover and outcome completion.

Results

Between the 3rd December 2021 and 31st August 2022, 13 participants were randomised from 4 of 9 recruiting centres. Redeployment of clinical and research staff, due to COVID-19, limited the number of available sites. Participants already receiving fludrocortisone or midodrine accounted for 120 of 233 eligible participants being excluded. Due to the low sample size, the rates of attrition and crossover are of limited value. Apart from the falls diaries, completion rates of the outcome measures were high (> 80%). Due to low recruitment rates, the pilot did not progress to the planned multi-arm multistage trial.

Conclusions

In its current design, this trial was not feasible. The main barriers to success were participants already receiving treatment and redeployment of clinical and academic staff during and after the COVID-19 pandemic.

Trial registration

ISRCTN 87213295, 23 July 2021, https://www.isrctn.com/ISRCTN87213295

Supplementary Information

The online version contains supplementary material available at 10.1186/s40814-025-01704-7.

Key messages regarding feasibility

• What uncertainties existed regarding feasibility?

  The clinical and cost effectiveness of different treatment strategies for orthostatic hypotension is unknown. People with orthostatic hypotension tend to be older, have multiple long-term conditions and take multiple medications. There is uncertainty around recruitment of this population, attrition rates, potential crossover of treatments and completion of multiple outcome measures.

• What are the key feasibility findings?

  Redeployment of clinical and research staff during COVID-19 and the post-COVID period meant that trial delivery was not feasible. Recent or current exposure to one of the trial interventions precluded a large proportion of participants being eligible. A flexible and pragmatic trial protocol was not sufficient to overcome these barriers. The number and timing of outcome measures appears to be appropriate and feasible.

• What are the implications of the feasibility findings for the design of the main study?

  The main study was not feasible as planned. Alternative study designs should be considered.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40814-025-01704-7.

Background

Orthostatic hypotension (OH) is a common and disabling condition, characterised by a significant reduction in blood pressure (BP) on standing upright [1]. It is particularly prevalent in older populations and in those with chronic disease, affecting up to one in five community-dwelling older people, one in four with diabetes and one in three with Parkinson’s disease (PD) [2–4]. In the USA, the presence of OH in people with PD is known to increase overall healthcare-related costs 2.5-fold compared to those who have PD without OH [5].

The reduction in BP during standing leads to a wide variety of symptoms including dizziness, headache, nausea, fatigue and visual disturbance; at its most severe, it can result in falls and syncope [6, 7]. People with OH have a reduced quality of life due to the difficulties in performing even simple tasks which involve standing [8]. There are also longer-term sequelae, as OH is associated with an increased risk of stroke, cognitive impairment and all-cause mortality [9, 10].

Despite the high prevalence of OH, there is very little good quality evidence to support its management [11]. The UK’s National Institute for Health and Care Excellence (NICE) provides evidence summaries for fludrocortisone and midodrine to treat OH but notes that long-term efficacy and safety are unclear [12, 13]. NICE also comments that these studies are limited by their use of disease-centred outcomes (i.e. BP) rather than patient-centred outcomes, such as symptoms and quality of life.

Following conservative measures, including physical counter-manoeuvres and fluid and salt loading, the European Society of Cardiology recommends the use of midodrine or fludrocortisone for OH but notes that the quality of evidence is based on expert opinion and/or small studies and that further research is needed [14]. Similarly, systematic reviews and meta-analyses consistently describe existing evidence as poor quality, calling for more rigorous evaluation to guide clinical practice [14–18]. Both agents have side effects, but as their relative effectiveness is unclear, it is unknown whether the benefits outweigh their harms.

The medications to treat OH are inexpensive, but the cost consequences of successful treatment or management of side effects are not known. No high-quality economic evaluations comparing these interventions, with each other or against non-pharmacologic therapies, have been identified. Therefore, the lack of evidence on the relative cost-effectiveness of fludrocortisone and midodrine means that there is insufficient evidence about which of these therapies represents a good use of scarce healthcare resources.

In 2018, the UK National Institute for Health and Care Research Health Technology Assessment (NIHR HTA) programme opened a call for commissioned research on ‘Management of orthostatic hypotension (HTA 18/32)’. CONFORM-OH (CONtrol, Fludrocortisone OR Midodrine for Orthostatic Hypotension) was funded by NIHR HTA in 2019, and the protocol was designed in line with the commissioning brief.

Methods

The full protocol is available on the International Standard Randomised Controlled Trial Number website at https://www.isrctn.com/ISRCTN87213295.

Objectives

To address uncertainties in the study design, the internal pilot evaluated the feasibility of reaching the recruitment target, the rate of attrition, the crossover rates between arms and quality of outcomes.

Progression criteria and sample size

The main aim of the internal pilot phase was to evaluate the recruitment rate to determine feasibility of recruitment. There was no formal sample size calculation for the internal pilot phase; instead, the duration was fixed at 10 months, after which the recruitment rate would be estimated. A duration of 10 months was set based on logistics and anticipated time needed to reach a steady recruitment rate. To deliver the required sample size for a full trial would require 20 sites to be opened in a staged approach, with 14 sites planned to be open by the end of the 10-month pilot, and a recruitment rate of 0.8 participants per site per month, with a target of 64 participants recruited by the end of the 10-month pilot.

A traffic light system was in place to judge feasibility at 10 months:

• Recruitment

  • ◦ Green: Sixty-four participants (or ≥ 0.8 participants per site per month)
  • ◦ Amber: A total of 40 to 63 participants (or ≥ 0.5 to < 0.8 participants per site per month) but with strategies identified to increase recruitment
  • ◦ Red:  < 40 participants (or < 0.5 participants per site per month)
• Attrition

  • ◦ Green: ≤ 15% participants withdraw before the primary endpoint.
  • ◦ Amber: A total of 16 to 35% of participants withdraw before the primary endpoint but with strategies identified to improve retention.
  • ◦ Red: > 35% participants withdraw with no feasible solutions to improve.
• Crossover

  • ◦ As there were no existing data to estimate crossover rates, these were to be monitored and reviewed during the pilot with no fixed traffic-light criteria.
• Outcome data

  • ◦ The quality and completeness of outcome data were monitored to make improvements if required.

A sample size of 64 participants at the end of the internal pilot would allow binary outcome rates such as the attrition rate to be estimated with a 95% confidence interval of width at most 24% (i.e. if the observed rate was 50%).

Progression or termination of the trial following the pilot was judged by the Trial Management Group, Data Monitoring Committee and Trial Steering Committee who made their recommendations to the funder.

The planned sample size for the full trial can be found in the statistical analysis plan which is available in supplementary material Document 1.

Trial design

CONFORM-OH was designed as a pragmatic, multi-arm, multistage, parallel group, prospective, randomised, open-label and superior trial, with a 10-month internal pilot. Participants were randomised in a 1:1:1 ratio to conservative management, conservative management plus fludrocortisone or conservative management plus midodrine. The intervention and follow-up were for 12 months.

Setting

Participants were recruited from multiple secondary care sites across the UK NHS which included falls and syncope clinics, movement disorders services, geriatrics clinics and cardiology clinics. Fully informed, written consent was obtained by a member of the research team either in person or remotely via telephone/video with consent forms posted to site.

Participants

Inclusion criteria

• Adults aged 18 years and over

• A clinical diagnosis of symptomatic OH which was either as follows:

  • ◦ Clinically significant where treatment was indicated quickly without a trial of lifestyle modification OR
  • ◦ Refractory to an adequate period of lifestyle modification (to be judged clinically)

• A drop in systolic blood pressure of ≥ 20 mmHg and/or a drop in diastolic blood pressure of ≥ 10 mmHg, within 3 min of standing upright from a supine position (or on tilt testing)

• A score of ≥ 2 on the Orthostatic Hypotension Questionnaire

• Willing and able to provide informed consent

Exclusion criteria

• OH secondary to acute or reversible causes

• Use of fludrocortisone or midodrine within the last 6 months

• Terminal illness or life expectancy < 12 months

• Supine hypertension (where the risks of treatment outweigh the benefits)

• A known allergy to study medication

• A known contraindication to fludrocortisone or midodrine which outweighs the potential clinical benefit

• Current or planned pregnancy/breast feeding during the trial

• Inability to communicate in English

• Inability to comply with the study procedures

• Currently taking part in another clinical trial that would interfere with the outcomes of CONFORM-OH

Changes to planned criteria

Originally, the inclusion criteria required participants to have a diagnosis of OH which was refractory to a minimum of 4 weeks of lifestyle modification. This criterion was specified by the funder in the commissioned call for funding. However, as a pragmatic trial and to reflect clinical practice, this was modified as above. Allowing diagnosis by tilt-table testing was an additional modification made at one site’s request.

Interventions

Control

Non-pharmacologic therapies (or conservative management) included trigger avoidance, physical counter-manoeuvres, fluid and salt intake, compression hosiery and ‘culprit’ medication review. Trial specific guidance on nondrug therapies was not provided to sites, as they were encouraged to follow their usual clinical practice. To account for sites’ differing practice, the randomisation method stratified by recruiting centre and data were collected from sites to describe which conservative measures they used.

Fludrocortisone

Sites were requested to initiate and titrate fludrocortisone tablets as they would in their usual clinical practice. Permissible dosing ranged from 50 µg per day to a maximum of 400 µg per day. There were no restrictions on the brand used.

Midodrine

Sites were requested to initiate and titrate the midodrine as they would in their usual clinical practice. Dosing ranges typically started at a dose of 2.5 mg three times per day orally, increasing to a maximum tolerated dose, not exceeding 10 mg three times per day. Lower starting doses (e.g. 2.5 mg twice per day) were permissible. There were no restrictions on the brand used.

Randomisation and allocation

Allocation was via a central, secure, web-based system (Sealed Envelope^™). Participants were assigned in a 1:1:1 ratio between control and two intervention arms using a minimisation algorithm with a random element. Age (> 80 vs ≤ 80 years), aetiology (neurogenic vs non-neurogenic) and site were used as stratification factors. Participants, site investigators, research staff and trial team members were aware of treatment allocation. As a pragmatic trial, study investigators followed their usual clinical practice once the participant was allocated a treatment. Changes to the treatment arm were allowed if clinically indicated.

Outcomes

Feasibility outcomes

Feasibility outcomes included the total number recruited, the average recruitment rate per site per month, the proportion of participants withdrawing from the trial, the proportion of participants crossing over between trial treatments and the proportion of outcome assessments completed out of those expected.

Trial outcome measures

The primary outcome of the CONFORM-OH trial was the change in OH-related symptoms from baseline to 6 months, measured using the Orthostatic Hypotension Questionnaire (OHQ) [6].

Secondary outcomes are displayed in Table 1 and are described in the published protocol and statistical and health economics analysis plans (Documents 1 and 2, respectively, in the supplementary material).

Table 1.

Secondary outcomes

	Outcome	Measurement	Time point
Clinical effectiveness	Change in symptoms	Orthostatic Hypotension Questionnaire (OHQ) Patient-reported questionnaire rating the presence and severity of OH symptoms and their impact on daily activities	3 months 12 months
	Daily function	Nottingham Extended Activities of Daily Living Scale (NEADL) questionnaire Patient-reported questionnaire assessing level of independence to perform 22 different everyday activities	3 months 6 months 12 months
	Quality of life	EQ-5D-5L questionnaire General health questionnaire evaluating five domains and including a visual analogue scale to rate health-related quality of life	3 months 6 months 12 months
	Falls and syncope	Monthly falls diary is as follows: •Time to first fall •Number of falls •Number of fallers (patients reporting ≥ 1 fall) •Fall rate per person year •Fall-related injuries •Number of syncopal events	Over 12 months
	Blood pressure	Supine and standing BP measured according to usual clinical practice as follows: •Orthostatic systolic BP drop* •Orthostatic diastolic BP drop* •Standing nadir systolic BP •Standing nadir diastolic BP	3 months 6 months 12 months
Safety	Side effects	Adverse event reporting	12 months
	Safety	•Serious adverse reactions and events •Hospitalisations •Deaths	12 months
Cost-effectiveness		•Health and social care use questionnaire •Time and travel questionnaire	12 months

Data collection

Questionnaires were completed by participants either during research visits, remotely over the telephone or returned via post. For the EQ-5D-5L, a validated proxy version was used for those participants not able to complete the questionnaire themselves.

Postural BP was performed by a clinician during either a research, usual clinic appointment or home visit. In cases where it was not possible for the participant to attend, a recent (within six weeks) postural BP recorded in the participant’s medical records was permissible. As a contingency measure, if the BP assessment was not possible, participants were able to perform their own postural BP assessment at home after being provided with a study BP monitor and instructions [19]. If participants withdrew from the study, consent was requested from participants to use routinely collected clinical data.

Participants were provided with a diary to report falls and faints during the study follow-up period. Each week, participants were asked to report the number of falls and faints they had experienced each day. Falls were defined as those events which resulted in landing on the floor or ground and did not include trips and slips without falling to the ground. Participants were asked to bring the diary to study research visits or have it available for telephone appointments.

Adverse events were identified through participant self-report or open-ended questioning during each research visit. There was no systematic assessment of any pre-specified adverse events. Adverse events were coded centrally using the MedDRA dictionary.

Feedback

Feedback was sought from site staff during remote interviews at the end of the pilot to identify barriers to site opening and recruitment; trial feedback questionnaires were sent to participants in the post following their last visit. Feedback was collated and summarised to identify barriers and solutions. These questions are listed in the supplementary material (Document 3: Feedback questionnaire).

Statistical analysis

Due to the small number of participants recruited in the internal pilot, feasibility metrics and clinical outcome data are summarised descriptively. No hypothesis testing has been performed. Continuous outcomes are reported using the mean and standard deviation (SD) and/or the median and range (minimum and maximum values). Categorical outcomes are reported as frequencies and percentages.

All clinical outcome data are reported according to randomised treatment group, following the intention-to-treat principle. To account for crossovers between intervention groups, safety data are reported according to treatment strategy received.

The OHQ and NEADL questionnaires were scored according to published scoring methods [6, 20]. Full details can be found in the statistical analysis plan which is available in supplementary material Document 1.

Health economics analyses

Pilot data are summarised descriptively; continuous outcomes are reported using the mean and standard deviation (SD) and/or the median and range (minimum and maximum values). Categorical outcomes are reported as frequencies and percentages. For the EQ-5D-5L, we report responses by the descriptive framework, the EQ-5D-VAS and health state utilities.

Full details, including the analysis methods which would have been used had the trial continued past the internal pilot stage, can be found in the health economics analysis plan which is available in supplementary material Document 2.

COVID-19

The process of setting up the CONFORM-OH study began in November 2019, with a planned recruitment start date of May 2020. In early 2020, the study set-up process was paused as COVID-19 studies were prioritised by ethical review committees and local research sponsors. In addition, several coinvestigators were redeployed to support clinical work during the pandemic. Our patient and public experts informed the study team that they would not be confident to attend clinical or research visits until they were certain that reliable systems were in place to keep them safe from COVID-19 in these settings. In mid-2020, the CONFORM-OH study protocol was modified to make it more feasible during the pandemic. These modifications included the following:

• Allowing remote consent (telephone or video link)

• Accepting routinely measured blood pressure outcomes in lieu of research visit blood pressure

• Allowing participants to measure their own postural blood pressure if research or clinic visits were not possible

• Additional questions were included in the health economics outcomes about COVID-19

Results

In total, between 3rd December 2021 and 31 st August 2022, 282 patients were screened for eligibility, 49 were found to be eligible, and 13 were randomised. Due to low recruitment rates, the trial was terminated at the end of the internal pilot phase and did not progress to the main study. The findings from each component of the feasibility assessment are summarised below.

Sites

In March 2020, 41 sites had expressed an interest to recruit. Of these, 20 did not return site feasibility questionnaires and were excluded. The process of gaining sponsorship and ethical permission took 9 months. The first site opened in December 2021, with a further eight sites open by May 2022. Twelve sites were in the process of being set up when the pilot phase ended. Specific sites and their screening and recruitment figures can be found in Table 2.

Table 2.

Screening and recruitment data by site

Site	Months open	Number screened	Number eligible	% eligible (of screened)	Number recruited	% recruited (of eligible)
Newcastle	9	52	11	21%	5	45%
Walsall	7	79	21	27%	6	29%
Norfolk and Norwich	7	26	6	23%	1	17%
Dumfries and Galloway	7	10	4	40%	1	25%
Lewisham and Greenwich	6	16	3	19%	0	0%
Gateshead	5	3	0	0%	0	0%
Devon and Exeter	4	3	1	33%	0	0%
Bath	4	87	3	3%	0	0%
Salford	3	6	0	0%	0	0%
Total	52	282	49	17.4%	13	26.5%

Recruitment

Between the 3rd December 2021 and 31 st August 2022, 13 participants were randomised from 4 of the 9 recruiting centres.

Progress against the internal pilot progression criteria was evaluated by the Trial Steering Committee on 30th August 2022. Nine sites had been open to recruitment for a total of 52 site months, and 13 participants had been randomised. The average accrual rate was 0.25 participants per site per month, which fell well below the amber progression criteria target and into the red category. On the recommendation of the Trial Steering Committee, the trial closed to recruitment on 31 st August 2022, 1 month earlier than the original planned 10-month internal pilot phase. All participants who had been recruited continued in the trial as planned.

In total, 282 patients were screened for eligibility. Forty-nine patients (17%) were found to be eligible, and of those, 13 (27%) were randomised. Reasons for ineligibility and eligible patients not taking part in the trial are summarised in Table 3. The main reason for ineligibility was due to recent use of fludrocortisone or midodrine within the last 6 months (120 of 233 ineligible patients, 52%). In most cases (116, 97%), this was due to current use of either or both study medications. A contraindication to fludrocortisone or midodrine also led to the exclusion of 25 (11%) patients; predominantly, contraindications were to midodrine.

Table 3.

Screening data

Screened	N = 282
Eligibility status
Eligible	49 (17%)
Ineligible	233 (83%)
Reasons for ineligibility	N  = 233
Use of fludrocortisone or midodrine within the last 6 months	120 (52%)
Drug used
Fludrocortisone	45 (38%)
Midodrine	64 (53%)
Both	10 (8%)
Unknown	1 (1%)
Current or previous use
Current	116 (97%)
Previous	3 (3%)
Not available	1 (1%)
A known contraindication to fludrocortisone or midodrine which outweighs the potential clinical benefit	25 (11%)
Drug contraindicated
Midodrine	19 (76%)
Both	2 (8%)
Unknown	4 (16%)
Inability to comply with the study procedures as specified in the schedule of events	21 (9%)
Supine hypertension (where the risks of treatment outweigh the benefits) at baseline	20 (9%)
Other: A score of < 2 on the OHQ	12 (5%)
Other: Too frail/pre-existing conditions	7 (3%)
Other: No BP drop as per inclusion criteria	6 (3%)
OH secondary to acute or reversible causes	5 (2%)
Other: Unable to provide informed consent	2 (1%)
Other: already started on other medication/did not want to start new	2 (1%)
Other: Unable to wait 4 weeks of lifestyle modification	2 (1%)
Terminal illness or life expectancy < 12 months	1 (< 0.5%)
Inability to communicate in English	1 (< 0.5%)
Other: Did not have OH on examination	1 (< 0.5%)
Other: Unknown reason	3 (1%)
Eligibility status unknown	5 (2%)
Reasons for eligible patients not participating	N  = 36
Participant declined	30 (83%)
I would find it difficult to participate (e.g. due to burden of assessments/questionnaires)	10 (33%)
No reason given	8 (27%)
I do not want to change my treatment plan	5 (17%)
I see no benefit in the study to me	2 (7%)
I do not want to be randomised to usual care	2 (7%)
Other: Did not want to be randomised into either IMP	2 (7%)
I am too busy	1 (3%)
Participant relocated/unable to contact after screening	3 (8%)
Patient not approached by doctor	3 (8%)

Thirty-six of the 49 (73%) eligible patients did not take part in the trial. Thirty (83%) declined participation; the burden of completing trial assessments was the most common reason given (10 of 30, 33%).

Attrition

Participant flow through the trial is provided in a CONSORT diagram in Fig. 1. Of the 13 participants randomised, 3 were allocated to the conservative management arm (control), 4 were allocated to conservative management plus fludrocortisone and 6 were allocated to conservative management plus midodrine. Of the three participants allocated to the control arm, two (67%) withdrew from the trial, one prior to the 3-month visit and one prior to the 6-month visit, and one further participant died prior to the 3-month visit. Both participants who withdrew from the trial agreed for routinely available data (e.g. BP measurements) to still be collected from medical records where available. One participant allocated to the midodrine arm was lost to follow-up prior to the 12-month follow-up visit. Overall, 11 (85%) participants remained in follow-up at the 3-month visit, 10 (77%) participants at the 6-month visit and 9 (69%) participants at the 12-month visit.

Fig. 1.

CONSORT flow diagram. LTFU, lost to follow-up

Treatment adherence and crossover

Of the four participants allocated to the fludrocortisone arm, one (25%) participant decided not to take their allocated study medication but continued to provide follow-up data. Of the six participants allocated to the midodrine arm, two (33%) participants discontinued their allocated treatment: one participant crossed over to fludrocortisone at the 3-month visit due to side effects of midodrine, and one participant stopped taking midodrine at the 6-month visit as their symptoms had improved.

Outcome measure completion

At the 6-month primary outcome time point, data were available for 10 (77%) participants: zero (0%) allocated to control, 4 (100%) fludrocortisone and 6 (100%) midodrine. Completion rates of outcome measures are summarised in Table 4. The OHQ, NEADL questionnaire and BP measurements were available for all participants who remained in follow-up at each time point. No additional BP measurements were provided from medical records for either of the two participants who withdrew from the trial but allowed for continued collection of routinely available data. Falls data were collected from all participants who remained in follow-up at each time point; however, falls diaries were only completed and returned by seven (64%) participants at the 3-month visit, six (60%) participants at the 6-month visit and five (56%) participants at the 12-month visit.

Table 4.

Completion rates of outcome measures

	Conservative management	Conservative management plus fludrocortisone	Conservative management plus midodrine	Overall
Month 31	Exp = 1	Exp = 4	Exp = 6	Exp = 11
BP2	Exp = 2	Exp = 4	Exp = 6	Exp = 12
Collected	1 (50%)	4 (100%)	6 (100%)	11 (92%)
Not collected	1 (50%)	0 (0%)	0 (0%)	1 (8%)
How was the BP obtained?
Study visit	0 (0%)	4 (100%)	6 (100%)	10 (91%)
Other medical records	1 (100%)	0 (0%)	0 (0%)	1 (9%)
OHQ	1 (100%)	4 (100%)	6 (100%)	11 (100%)
NEADL	1 (100%)	4 (100%)	6 (100%)	11 (100%)
Falls diary
Falls data collected	0 (0%)	4 (100%)	6 (100%)	10 (91%)
Falls diary completed	0 (0%)	2 (50%)	5 (83%)	7 (64%)
Month 61	Exp = 0	Exp = 4	Exp = 6	Exp = 10
BP2	Exp = 2	Exp = 4	Exp = 6	Exp = 12
Collected	0 (0%)	4 (100%)	6 (100%)	10 (83%)
Not collected	2 (100%)	0 (0%)	0 (0%)	2 (17%)
How was the BP obtained?
Study visit	NA	4 (100%)	5 (83%)	9 (90%)
Other medical records	NA	0 (0%)	1 (17%)	1 (10%)
OHQ	NA	4 (100%)	6 (100%)	10 (100%)
NEADL	NA	4 (100%)	6 (100%)	10 (100%)
Falls diary
Falls data collected	NA	4 (100%)	6 (100%)	10 (100%)
Falls diary completed	NA	2 (50%)	4 (67%)	6 (60%)
Month 121	Exp = 0	Exp = 4	Exp = 5	Exp = 9
BP2	Exp = 2	Exp = 4	Exp = 5	Exp = 11
Collected	0 (0%)	4 (100%)	5 (100%)	9 (82%)
Not collected	2 (100%)	0 (0%)	0 (0%)	2 (18%)
How was the BP obtained?
Study visit	NA	3 (75%)	5 (100%)	8 (89%)
Self-assessed at home	NA	1 (25%)	0 (0%)	1 (11%)
OHQ	NA	4 (100%)	5 (100%)	9 (100%)
NEADL	NA	4 (100%)	5 (100%)	9 (100%)
Falls diary
Falls data collected	NA	4 (100%)	5 (100%)	9 (100%)
Falls diary completed	NA	1 (25%)	4 (80%)	5 (56%)

¹The number of expected CRFs at this visit (unless otherwise stated). ²Also expected for those that withdrew but allowed use of routinely collected medical data

The barriers to recruitment, assessment of the feasibility components and lessons learnt from the internal pilot are summarised in Table 5.

Table 5.

Summary of learning from feasibility components

Feasibility issues	Key findings & learning
Permissions	Ethical permissions were delayed due to prioritisation of COVID-19 studies
Site opening	Clinical and research staff redeployed during and after COVID-19
Eligible patient population	The biggest barrier to recruitment was the exclusion criteria, predominantly by excluding those already taking one of the trial interventions. Potential solutions include having a washout period before entering the trial, although this may not be acceptable to patients and clinicians and prior experience of treatment may introduce bias
Acceptance rate	One-third of eligible participants declined due to the burden of assessments. Given the level of frailty and multiple long-term conditions in this population, this may be expected. Feedback from trial participants was that the number of assessments was not burdensome
Intervention crossover	Given the low numbers of participants, the estimate is unlikely to be of a value. Participants and clinicians reported reluctance to be randomised to the conservative arm; had the trial continued, this may have led to increased crossover from the control arm to one of the intervention arms
Retention	Retention was 100% at the primary time point of 6 months for the intervention arms but was 0 of 3 for the control arm. Retention remained good (90%) for the intervention arms at 12 months. As described above, the control arm appears to have been a barrier to recruitment and retention, although numbers in all arms were small to make definitive conclusions
Availability of outcome measures	With the exception of the falls diaries, completion rates of the outcomes were excellent. Of those remaining in the trial, the primary outcome was available for all participants. Of secondary outcomes, with the exception of falls diaries, these were all available for all participants at all time points Return rates for falls diaries are known to be poor in falls studies. Strategies such as telephone and text reminders may help. One participant stated they would have preferred an online falls diary. In one participant who did not return their falls diary, a review of their medical records identified an injurious fall. Reviewing medical records will likely improve identification of injurious falls rather than relying on self-report

Demographic data

Participant characteristics, medication at baseline and conservative measures advised at baseline are summarised descriptively in Table 6 and supplementary data Table S1. The average age of all participants was 74 years (SD 9.6); 7 (54%) were men, and all (100%) were of white background. Five (38%) participants had orthostatic hypotension caused by neurologic disorders, four of which were Parkinson’s disease. At baseline, the mean OHQ score was 6.4 (SD 1.5), the mean systolic postural BP drop was 40 mmHg (SD 16) and the mean diastolic BP drop was 10 mmHg (SD 16).

Table 6.

Participant demographics and characteristics at baseline

	Conservative management	Conservative management plus fludrocortisone	Conservative management plus midodrine	Total
	N = 3	N = 4	N = 6	N = 13
Age
Mean (SD)	75.3 (14.6)	70.3 (12.0)	75.0 (6.0)	73.6 (9.6)
Median (range)	80 (59, 87)	74 (53, 80)	73 (70, 84)	73 (53, 87)
< 80 years	1 (33%)	3 (75%)	4 (67%)	8 (62%)
≥ 80	2 (67%)	1 (25%)	2 (33%)	5 (38%)
Sex at birth
Female	2 (67%)	1 (25%)	4 (67%)	7 (54%)
Male	1 (33%)	3 (75%)	2 (33%)	6 (46%)
Ethnicity
Any white background	3 (100%)	4 (100%)	6 (100%)	13 (100%)
BMI
No. with data	3 (100%)	3 (75%)	6 (100%)	12 (92%)
Mean (SD)	22.5 (6.1)	26.1 (2.5)	23.8 (1.5)	24.1 (3.3)
Median (range)	22.4 (16.4, 28.7)	25.8 (23.7, 28.7)	24.2 (21.7, 26.0)	24.2 (16.4, 28.7)
Disease aetiology
Neurogenic OH	1 (33%)	1 (25%)	3 (50%)	5 (38%)
Non-neurogenic OH	2 (67%)	3 (75%)	3 (50%)	8 (62%)
Medical history
Diabetes	0 (0%)	0 (0%)	0 (0%)	0 (0%)
Pure autonomic failure	0 (0%)	0 (0%)	0 (0%)	0 (0%)
Parkinson’s disease	1 (33%)	0 (0%)	3 (50%)	4 (31%)
OHQ score
Mean (SD)	6.8 (1.7)	6.8 (1.6)	6.0 (1.6)	6.4 (1.5)
Median (range)	7.5 (4.9, 8.0)	6.7 (5.1, 8.8)	5.9 (3.8, 8.7)	6.2 (3.8, 8.8)
NEADL score
Mean (SD)	13.7 (6.0)	16.0 (9.4)	14.2 (5.3)	14.6 (6.4)
Median (range)	13 (8, 20)	20 (2, 22)	15 (6, 21)	17 (2, 22)
Blood pressure
How was the BP obtained?
Study visit	2 (67%)	2 (50%)	2 (33%)	6 (46%)
Other medical records	1 (33%)	1 (25%)	4 (67%)	6 (46%)
Self-assessed at home	0 (0%)	1 (25%)	0 (0%)	1 (8%)
Supine blood pressure
Systolic
Mean (SD)	134.7 (20.7)	128.5 (26.2)	141.3 (22.9)	135.8 (22.3)
Median (range)	131 (116, 157)	129 (96, 160)	151 (102, 160)	132 (96, 160)
Diastolic
Mean (SD)	74.7 (4.0)	70.5 (24.6)	74.5 (10.6)	73.3 (14.3)
Median (range)	77 (70, 77)	69 (42, 102)	79 (53, 80)	77 (42, 102)
Lowest standing BP
Systolic
Mean (SD)	88.3 (23.6)	88.8 (10.6)	105.0 (18.4)	96.2 (18.3)
Median (range)	80 (70, 115)	93 (73, 96)	105 (80, 130)	94 (70, 130)
Diastolic
Mean (SD)	60.0 (20.9)	61.5 (13.0)	65.3 (8.4)	62.9 (12.2)
Median (range)	50 (46, 84)	62 (46, 77)	66 (55, 76)	60 (46, 84)
Postural blood pressure drop
Systolic
Mean (SD)	46.3 (4.5)	39.8 (19.6)	36.3 (18.2)	39.7 (15.9)
Median (range)	46 (42, 51)	34 (23, 68)	34.5 (11, 64)	39 (11, 68)
Diastolic
Mean (SD)	14.7 (19.6)	9.0 (23.6)	9.2 (10.4)	10.4 (15.9)
Median (range)	20 (−7, 31)	−1 (−6, 44)	7.5 (−6, 22)	7 (−7, 44)
Conservative measures advised at baseline
Trigger avoidance	2 (67%)	3 (75%)	5 (83%)	10 (77%)
Hydration	3 (100%)	4 (100%)	6 (100%)	13 (100%)
Bolus water drinking	1 (33%)	2 (50%)	3 (50%)	6 (46%)
Salt intake	3 (100%)	2 (50%)	2 (33%)	7 (54%)
Caffeine intake	1 (33%)	0 (0%)	1 (17%)	2 (15%)
Physical counter-manoeuvres	3 (100%)	2 (50%)	3 (50%)	8 (62%)
Provision of compression stockings (below knee)	0 (0%)	1 (25%)	0 (0%)	1 (8%)
Provision of compression stockings (full length)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
Abdominal compression	0 (0%)	0 (0%)	0 (0%)	0 (0%)
Medication review	1 (33%)	2 (50%)	3 (50%)	6 (46%)
Other—advised to consider stockings from GP	1 (33%)	1 (25%)	2 (33%)	4 (31%)

Trial outcome measures

Data collected on primary and secondary clinical outcome measures are summarised descriptively in supplementary materials. No data were available for any participant allocated to the control arm at 6- or 12-month follow-up.

OHQ, activity of daily living, nadir standing blood pressure and orthostatic blood pressure drop.

For each outcome measure, the change from baseline to each follow-up time point is summarised descriptively by randomised treatment group in supplementary data Table S2. Individual participant data are also plotted in supplementary data Fig. S1.

The mean change in the OHQ score from baseline to 6 months was −4.5 points (N = 4, SD 2.0) in those allocated to conservative management plus fludrocortisone and −1.7 points (N = 6, SD 2.0) in those allocated to conservative management plus midodrine.

Summary data for each outcome measure at each time point are also provided in supplementary data Table S3. The correlation between baseline and follow-up OHQ scores is presented in supplementary data Table S4.

Falls

Six participants reported at least one fall during the 12-month follow-up period: one (25%) participant allocated to conservative management plus fludrocortisone and five (83%) allocated to conservative management plus midodrine (supplementary Table S5). In total, two falls were reported in the fludrocortisone arm and 46 in the midodrine arm (34 falls were reported by one participant). No syncopal events were reported by participants allocated to fludrocortisone, and nine syncopal events were reported by one participant allocated to the conservative management plus midodrine arm. No data on falls or syncopal events were available for participants allocated to the control arm, as the one participant who remained in follow-up at the 3-month visit did not return their falls diary. Three participants reported fall-related injuries during the 12-month follow-up, summarised in supplementary Table S6.

Safety

Safety data are summarised in supplementary Table S7. In total, 26 adverse events were reported: 4 in participants allocated to conservative management, 16 in the conservative management plus fludrocortisone arm and 6 in the conservative management plus midodrine arm. Of those, five adverse reactions (adverse events considered to be possibly, probably or definitely related to trial treatment) were reported from three (50%) participants exposed to midodrine. All were mild in severity; however, for one participant, this led to a permanent discontinuation of trial treatment and a temporary interruption of treatment for a further participant. Of the 26 adverse events reported, 9 were reported as serious adverse events from 4 participants; all were deemed to be unrelated to trial treatment. One adverse event resulted in death. Further detail is provided in the supplementary Tables S8, S9 and S10.

Health economics

Supplementary Tables S11-S30 report the summary data for the use of health and care services. These are reported by the randomised arm and for all time points. Due to the limited data, no comparisons are made between trial arms, and no totals are presented for each area of resource use. For contacts with general practitioners, there is a right skew in the data, with a small number of participants having a larger number of contacts. The same pattern is not clearly observed for most other areas of resource use because the use of services is very low. Similarly, the use of private healthcare and personal social services is likewise infrequent in every group. However, the small number of participants means firm conclusions cannot be drawn.

Supplementary Table S31 reports response rates for the EQ-5D-5L which describes the pattern of responses to the healthcare services data. The EQ-5D-5L could be either self-completed or completed by proxy. Very few participants had proxy completion for the EQ-5D-5L at any time point (Table S32). Tables S33-S37 provide summary data for each of the five dimensions of the EQ-5D-5L separately. At baseline, no participant was in the worst state of health for mobility. Overall, the mean and median scores were 1.75 and 1.5, respectively, indicating participants had slight problems with mobility (Table S33). For self-care, the mean score was 1.67, indicating that overall participants had slight problems with washing or dressing themselves; similar inferences can be made from the median score, which was 1 (Table S34). Responses for usual activity are shown in Table S35. At baseline, the mean score for usual activities was just over 2, and the median score was 1. The mean score indicates that participants had slight problems doing their usual activities, and the median score for usual activities indicated that participants had no problems doing them. Table S36 shows responses for the pain and discomfort dimension. This table shows that the mean and median scores were 2.5 and 3, respectively, suggesting that participants had moderate pain and discomfort. Table S37 reports responses by the anxiety and depression dimensions. The mean and median scores were 2.33 and 2, respectively. These scores suggest that participants were slightly anxious or depressed. Table S38 shows the responses for the EQ-5D-VAS. At baseline, the mean and median scores were 63 and 60, respectively (100 is best imaginable health, and 0 is the worst possible health). The mean and median baseline EQ-5D-5L utility values, which can be found in Table S39, were 0.58 and 0.62, respectively (1 is the value of full health, and 0 is equivalent to dead). For all dimensions, it was unclear if the severity of responses varied over the follow-up. This is due to the small number of participants completing the EQ-5D-5L.

One participant provided EQ-5D-5L by proxy at baseline, and at 12 months, two participants provided proxy data at 6 months. Tables S40, S41, S42, S43, S44 and S45 show the responses for each dimension for proxy response. As so few data were available, summary text is not provided. Table S46 shows the responses for the EQ-5D-VAS proxy responses at baseline, which was 45. As so few data were available, summary text is also not provided.

Discussion

This randomised controlled internal pilot did not meet its progression criteria, suggesting the trial was not feasible as planned. The timing of the trial, originally planned to open in 2020, was unfortunate as the effects of the COVID-19 pandemic had a severe impact on research sites. Several sites were unable to recruit due to staff redeployment, either to other clinical duties, or to focus on COVID-19 clinical trials. COVID-19 also had an impact on clinical pathways. Sites noted that an increase in virtual consultations resulted in fewer postural BP measurements and fewer diagnoses of OH. From a recruitment perspective, the biggest barrier was the eligibility criteria which excluded participants who were taking or had recently tried the medications under investigation; this amounted to just over half of screened participants.

Of the 13 participants recruited, the primary outcome was available in 10. With such small numbers, it is difficult to estimate whether these figures would be helpful to estimate attrition in a larger study. One potential solution to reduce attrition at the primary endpoint would be to move this from 6 to 3 months. Indeed, looking at the individual results in Supplement Fig. S1, it appears that most of the change seen from baseline occurs at the 3-month time point, with little additional gain at 6 months. However, feedback from participants and from site investigators was that 6 months allows adequate time to titrate medication doses. At the primary outcome time point (6 months), all participants from the control arm had withdrawn or died, in contrast to no withdrawals in the medication arms. Being randomised to the control arm was raised as a concern by some sites in their posttrial feedback, as some participants did not want to be allocated to control. This is an important factor to consider for future trials. However, it should be stressed that it remains unknown whether medication is superior to nondrug therapy, and this is an important question which needs answering. The participant and clinician equipoise could be addressed through novel trial design. Had the trial progressed from the pilot, the main trial was designed as an adaptive multi-arm multistage (MAMS) trial. An interim analysis was planned after the 200th participant had been randomised which would have allowed the study to drop an intervention arm if the interim analysis suggested it was no more effective than control. However, this was precluded by low recruitment. However, other novel designs could be considered. For example, a Personalised Randomised Controlled Trial (PRACTICAL [21]) design allows randomisation to acceptable arms only, such that if a contraindication/exclusion to one intervention existed, the participant would not be randomised to that arm. This may address the issue of clinicians and participants not wanting to be randomised to control or participants having a contraindication to one of the medications.

The completion rates of the secondary outcomes were excellent at all time points for symptoms, activities of daily living and BP. There were two exceptions. One was the completion rates of falls diaries. However, this is not unusual in the field of falls-related research. One participant fed back that they would have preferred an online falls diary, and this should be explored as an option for future trials, with the possibility of completing all the outcomes remotely. The number of outcome measures in the trial was not considered too many by most participants. In fact, feedback from participants was very positive, with no negative experiences reported by respondents in the posttrial feedback. The second exception was the reporting of BP for those that withdrew but consented to the use of routinely collected medical data; this applies to two participants. For these participants, BP measurements were recorded while enrolled; however, after withdrawing, no measurements were reported. It is not clear if no measurements were available or if these were not obtained. For future trials, consideration may be needed on how to incorporate routinely collected data.

An additional uncertainty when developing the trial protocol was the expected level of crossover between intervention arms. Only one participant changed their allocated intervention arm. Unfortunately, the numbers of participants in the pilot were too low to provide robust data for future estimates, but as far as we are aware, it is the only current data available to inform this.

It is uncertain whether the trial may now be more feasible in the post-COVID-19 period. Modified clinical pathways are now established, and there is a greater recognition that non-COVID-19 clinical services and research should continue even during future outbreaks (with safety measures in place). The trial was designed to be as pragmatic and flexible as possible, to match usual clinical pathways and to make recruitment and trial processes simpler for participants and clinician investigators. However, one of the limitations which became apparent was that there was considerable heterogeneity in usual practice between sites, so the study pathways were not always easily deliverable for sites. In addition, while the pragmatic nature attempted to reduce the workload for sites, there was a lack of available research support at sites to help deliver the trial in clinical settings. There may be an argument for increased funding to provide more research support, rather than depending on clinicians to deliver pragmatic clinical trials.

It is more difficult to address the exclusion criteria of currently taking treatment which led to the exclusion of large numbers of people with OH. To include these participants would require a ‘washout’ period off treatment, to measure baseline outcomes, which may be considered unacceptable by some patients and clinicians. Furthermore, their inclusion would introduce bias, as previous exposure to treatment will influence an individual’s assessment of treatment response; this is particularly relevant with the primary outcome being subjective. One solution would be to conceal treatment allocation in a placebo-controlled trial. However, participants and clinicians may be unwilling to stop treatment with the chance of being randomised to placebo. Alternatively, identifying treatment-naïve people with OH could be done in primary care. However, the coding of orthostatic hypotension in primary care is very poor (24, 973 cases among 2,911,260 records over 10 years) [22]. BP screening for OH in primary care is also unlikely to be an efficient recruitment method. Not all people with OH would require medication and screening for OH does not reflect current clinical practice and may produce skewed results. This has been seen in falls intervention research where interventions have been shown to be successful in those recruited from falls services, but not in those screened in primary care [23, 24].

Conclusions

The CONFORM-OH internal pilot showed that a definitive study would not be feasible as planned. The main barriers to success were participants already receiving treatment and redeployment of clinical and academic staff during and after the COVID-19 pandemic.

Abbreviations

OH

Orthostatic hypotension

BP

Blood pressure

PD

Parkinson’s disease

NICE

National Institute for Health and Care Excellence

NIHR HTA

National Institute for Health and Care Research Health Technology Assessment

CONFORM-OH

Control, fludrocortisone or midodrine for orthostatic hypotension

OHQ

Orthostatic hypotension questionnaire

NEADL

Nottingham extended activities of daily living

MEDDRA

Medical Dictionary for Regulatory Activities

QALY

Quality adjusted life year

Authors’ contributions

All authors have contributed to writing, reviewing and revising the manuscript. JF, SP, MW, JW, HM, LV, AY and RAK secured financial support for the trial and developed the trial protocol. SAA and HM performed the statistical analysis. TS and LV performed the health economic evaluation. EC, GW, NM and HH were responsible for the management of all research activity and execution including provision of trial resources. JP curated the trial database.

Funding

This study was funded by the UK NIHR Health Technology Assessment (NIHR127385) programme. The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.

Data availability

A de-identified dataset will be prepared and stored by the Newcastle University. Requests for data sharing will be subject to request which should provide a clear purpose, analysis plan, how the results will be disseminated and who the authors will be. Data transfer will be subject to completion of a data sharing agreement between the Newcastle University and the end users.

Declarations

Ethics approval and consent to participate

All participants provided fully informed, written consent. The study was approved by the UK NHS Health Research Authority’s Newcastle & North Tyneside 1 Research Ethics Committee (reference 21/NE/0083).

Consent for publication

Not applicable.

Competing interests

None.