Original Article
Effect of vitamin D supplementation on orthostatic hypotension: data from theVitamin D in Isolated Systolic Hypertension randomized controlled trial Miles D. Witham a, Rosemary J.G. Price a, Allan D. Struthers b, Peter T. Donnan c, Martina Messow d, Alex McConnachie d, Ian Ford d, and Marion E.T. McMurdo a
Objective: Orthostatic hypotension commonly accompanies supine hypertension, and is associated with low 25-hydroxyvitamin D levels. We tested whether highdose intermittent oral vitamin D therapy could ameliorate orthostatic hypotension in older patients with isolated systolic hypertension. Methods: We conducted a subgroup analysis of data from a parallel-group, double-blind, randomized, placebo-controlled trial. Patients aged over 70 years with supine office SBP above 140 mmHg and DBP below 90 mmHg received 100 000 units oral vitamin D3 or matching placebo every 3 months for 1 year. Office supine and standing blood pressure were measured at baseline, and 3, 6, 9 and 12 months, along with arterial stiffness and flow-mediated dilatation of the brachial artery. Results: Of 159 patients randomized to the main trial, 75 patients with orthostatic hypotension at baseline were included in this analysis. The mean age was 78 (SD 5) years, baseline blood pressure was 162/76 mmHg and the mean baseline orthostatic fall in blood pressure on standing was 32/5 mmHg. After adjustment for baseline age, 25-hydroxyvitamin D, SBP and orthostatic fall, the fall in SBP was less in the vitamin D group at 3 months [treatment effect 6 mmHg, 95% confidence interval (CI) 0 to 12], but repeated-measures analysis showed no significant treatment effect (3 mmHg for systolic fall, 95% CI 1 to 8; 1 mmHg for diastolic fall, 95% CI 1 to 3). Conclusion: Twelve months of intermittent, high-dose oral vitamin D3 did not significantly improve orthostatic hypotension in older patients with isolated systolic hypertension. Keywords: older, orthostatic hypotension, randomized controlled trial, vitamin D Abbreviations: 25OHD, 25-hydroxyvitamin D; FMD, flowmediated dilatation of the brachial artery; PTH, parathyroid hormone; VitDISH, Vitamin D in Isolated Systolic Hypertension trial
INTRODUCTION
O
rthostatic hypotension is defined as a fall in SBP of more than 20 mmHg or a fall of DBP of more than 10 mmHg on standing [1]. It is commonly accompanied by a sensation of light-headedness, which may result in syncope. Orthostatic hypotension is common in older people, particularly those with vascular disease including hypertension. Community-based studies suggest a prevalence of 30% in those aged 65 and over living in the community [2–8], rising to 70% of those in nursing homes. Although antihypertensive medications are frequently blamed, the evidence for this is surprisingly weak [9,10], and it is likely that a combination of vascular stiffness, impaired autonomic function, reduced cardiac output and impaired cerebral autoregulation contribute to the phenomenon and symptoms of orthostatic hypotension [11]. Existing therapeutic strategies for orthostatic hypotension are disappointing. Non-pharmacological manoeuvres, including stockings, abdominal compression and head-up sleeping posture are either ineffective, impractical or lacking in evidence [12,13]. Pharmacological interventions, including fludrocortisone, midodrine and other pressor agents, usually rely on increasing both supine and standing blood pressure. This strategy risks increasing baseline blood pressure with consequent increased risk of vascular events, and may not reduce the fall in blood pressure with erect posture. A recent systematic review concluded that the evidence for these interventions was of very poor quality and it was insufficient to recommend any of them for routine use [14]. Low 25-hydroxyvitamin D (25OHD) levels are common in older people, and observational studies have shown Journal of Hypertension 2014, 32:1693–1699 a Ageing and Health, bDepartment of Clinical Pharmacology, cEpidemiology and Biostatistics Unit and dRobertson Centre for Biostatistics, University of Dundee, Dundee, UK
Correspondence to Dr Miles D. Witham, Clinical Reader in Ageing and Health, Ninewells Hospital, Dundee DD1 9SY, UK. Tel: +44 1382 383086; fax: +44 1382 644972; e-mail: [email protected] Received 13 January 2014 Revised 26 March 2014 Accepted 26 March 2014 J Hypertens 32:1693–1699 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. DOI:10.1097/HJH.0000000000000223
Journal of Hypertension
www.jhypertension.com
1693
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Witham et al.
associations with an increased risk of hypertension and of cardiovascular events [15]. Vitamin D is known to have a number of potentially beneficial effects on vascular biology, including reducing renin secretion, reducing PTH, which is known to be vasculotoxic, reducing proinflammatory cytokines, and improving myocardial cell contractility [16]. Vitamin D supplementation improves endothelial function in some studies, and may also reduce blood pressure in selected groups [17–19]. Unlike some vascular interventions, vitamin D reduces the risk of falls and fractures in selected patient groups [20]. It is therefore possible that by improving arterial and myocardial function via a range of possible mechanisms, vitamin D could help to ameliorate orthostatic hypotension, and observational work supports a relationship between low 25OHD levels and orthostatic hypotension [21]. To test this hypothesis, we analysed data from patients with orthostatic hypotension enrolled into a randomized controlled trial of vitamin D supplementation for older patients with isolated systolic hypertension.
METHODS We performed a secondary analysis of data collected as part of the Vitamin D in Isolated Systolic Hypertension (VitDISH) trial of vitamin D supplementation to reduce blood pressure in older patients with isolated systolic hypertension. Detailed methods have been published previously [22]. The trial was a double-blind, placebocontrolled, parallel-group randomized trial, performed on 159 patients aged 70 and over, with isolated systolic hypertension and serum 25OHD less than 75 nmol/l. Exclusion criteria were: DBP above 90 mmHg, SBP above 180 mmHg, hypertension known to be due to a correctable underlying medical or surgical cause; estimated glomerular filtration rate below 40 ml/min (by four-variable Modified Diet in Renal Disease equation [23]); any liver function test (alanine aminotransferase, bilirubin, alkaline phosphatase) greater than 3x upper limit of local normal range; and albumin-adjusted serum calcium above 2.60 mmol/l or below 2.15 mmol/l. We also excluded patients with known metastatic malignancy or sarcoidosis, a history of renal calculi, diagnosis of heart failure with left-ventricular systolic dysfunction, atrial fibrillation, and those already taking prescription vitamin D supplements. Consumption of fish oils and low-dose (200 IU per day or less) over-the-counter vitamin D supplements were not a contraindication to enrolment. Three routes of recruitment were used. The majority of participants were recruited from the community via primary care practices. Patients were identified via General Practice lists of collaborating practices within the East Node of the Scottish Primary Care Research network and letters were sent from the patient’s primary care physician. Patients expressing interest in the study were put in touch with the research team and were sent study information. Participants were also recruited after responding to an article in the local paper about the research study; a small number of patients were recruited after being approached at secondary care clinics (Cardiovascular and Medicine for the Elderly). Written informed consent was obtained from 1694
www.jhypertension.com
all participants. Research Ethics approval was gained from Fife and Forth Valley NHS Research Ethics Committee (ref: 08/S0501/90). Clinical Trials authorization was obtained from the UK Medicines and Healthcare Regulatory Authority (EudraCT number 2008-004534-24). The trial sponsor was the University of Dundee, and the trial was registered at www.controlled-trials.com (ISRCTN92186858). The full protocol is available from the authors.
Intervention Participants entering the study were allocated to intervention or placebo in a 1 : 1 ratio using a minimization algorithm, administered by the Robertson Centre for Biostatistics (Glasgow Clinical Trials Unit, University of Glasgow, UK), using a telephone-based system to conceal study allocation from investigators and participants. Minimization variables used were baseline 25OHD level above or below 50 nmol/l, baseline SBP above or below 160 mmHg, baseline age above or below 80 years and the presence or absence of diabetes mellitus. Identical blinded medication bottles were used; patients were observed ingesting 100 000 units of oral vitamin D3 (cholecalciferol) (Vigantol oil; Merck KgAA, Darmstadt, Germany) or matching placebo (Mygliol oil; Merck KgAA), after completion of baseline, and 3, 6 and 9-month assessments.
Outcomes For this analysis, we included participants who displayed a fall in office blood pressure on standing of at least 20 mmHg systolic or 10 mmHg diastolic at baseline assessment, regardless of symptoms. Blood pressure measurements were taken at 0, 3, 6, 9 and 12 months using an OMRON HEM-705CP oscillometric machine. After a minimum 5 min supine rest in a quiet room, three readings were taken in the supine position, with the mean of the second and third readings taken as the measure of supine blood pressure. Standing blood pressure was then recorded at 0, 1 and 3 min after standing. A 22–32-cm cuff was used for the majority of patients, with a large (32–42 cm) cuff used for larger participants. The same cuff size was used for all visits by an individual participant. Parathyroid hormone (PTH) was collected, spun and analysed within 4 h using a Roche multichannel analyser system as part of routine NHS biochemistry laboratory workflow. 25OHD samples were collected, spun and stored at 208C, and assayed using the Immunodiagnostic Systems (IDS) ELISA (IDS, Boldon, UK). Intra-assay coefficient of variability was 6.3% and the laboratory participated in the vitamin D External Quality Assessment Scheme (DEQAS) quality assurance scheme for vitamin D assay. Screening samples were analysed in small batches prior to entry into the trial to ascertain trial eligibility; follow-up samples were stored and batch analysed at the end of the trial.
Endothelial function Endothelial function was assessed at 0, 3 and 12 months by measuring flow-mediated dilatation (FMD) of the brachial artery in response to hyperaemia (endothelium-dependent Volume 32 Number 8 August 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Vitamin D and orthostatic hypotension
vasodilation) following 5 min of forearm cuff occlusion [24]. An ultrasound system (Sequoia 512; Siemens, Camberley, UK), using an 8-MHz linear transducer, was used for imaging. Images gated to the ECG R-wave were acquired for 60 s before cuff inflation and for 2 min after the end of cuff occlusion. Endothelium-independent vasodilatation using sublingual glyceryltrinitrate was not measured for most patients due to an unacceptably high rate of syncopal side effects in this older population during the study (>5%). Artery diameter was measured using Vascular Research Tools software (Medical Imaging Applications LLC, Coralville, Iowa, USA). Mean diameter was measured during baseline acquisition, and was compared to the maximum diameter achieved after cuff deflation. FMD was then expressed as the percentage change from baseline diameter.
Pulse wave velocity Carotid-radial pulse wave velocity and augmentation index were measured using the Sphygmocor applanation tonometry system (AtCor Medical, Gloucester, UK) at 0, 3 and 12 months to give indices of arterial stiffness. Measurements were taken in the supine position after a minimum of 5 min of rest. For augmentation index, at least 15 good-quality waveforms derived from the radial artery were averaged and transformed to central aortic pressure using the supplied generalized transfer function. Augmentation index is presented normalized to a heart rate of 75 per min. Carotid and radial tonometry measures were obtained in conjunction with 3-lead ECG monitoring to compare the onset of systole (R-wave peak) with the onset of the pulse wave, detected by Sphygmocor software using the intersecting tangents method [25], which has previously been shown to have good reproducibility. Comparison of delay between R-wave peak and carotid pulse wave onset vs. R-wave peak and radial pulse wave onset was used to derive carotidradial pulse wave velocity.
Statistical analysis Analyses were performed using SPSS version 21 (SPSS, Chicago, Illinois, USA). A two-sided P value of 0.05 was taken as significant for all analyses. Correlations were performed using Pearson’s chi-square test. Between-group comparisons of postural drop were performed using analysis of variance (ANOVA) at each time-point, with adjustment for a range of baseline variables: baseline age, 25OHD level, baseline blood pressure, and baseline blood pressure fall. Repeated-measures ANOVA was also undertaken to estimate the overall treatment effect across all time-points; these analyses were then adjusted for a range of baseline variables as above. A sensitivity analysis was performed, excluding participants who changed their antihypertensive medication over the study period.
RESULTS A total of 159 participants were randomized into the trial. Figure 1 gives the flow of participant recruitment and follow-up through the trial, focussing on those participants fulfilling the case definition of orthostatic hypotension analysed in this study. The case definition of orthostatic Journal of Hypertension
hypotension was fulfilled by 75 of 159 (47%) of the participants. Details of this group are given in Table 1. This group had similar baseline 25OHD levels to those without orthostatic hypotension (45 vs. 44 nmol/l; P ¼ 0.61), similar baseline supine SBP (162 vs. 163 mmHg; P ¼ 0.86), slightly lower supine DBP (76 vs. 79 mmHg; P ¼ 0.04) and were slightly older (78 vs. 76 years; P ¼ 0.04). Analyses of the relationships between baseline 25OHD, blood pressure, orthostatic hypotension and other markers of vascular function are given in Table 2. For the 75 patients fulfilling the case definition of orthostatic hypotension at baseline, 25OHD levels were significantly higher in the vitamin D treatment group at 3 months (64 vs. 52 nmol/l; P ¼ 0.03); the overall treatment effect on 25OHD levels calculated using repeated-measures ANOVA was 17 nmol/l [95% confidence interval (CI) 8–25, P < 0.001]. Details of the drop in SBP and DBP on standing at each time-point are given in Table 3. A reduction in systolic drop of 7 mmHg was seen in the vitamin D group at 3 and 6 months compared to placebo, but the overall treatment effect by repeated measures over the 12-month study period was non-significant at 4 mmHg. There was no significant correlation between the increase in 25OHD level between baseline and 3 months, and the degree of improvement in systolic drop between baseline and 3 months (r ¼ 0.04, P ¼ 0.75). The number of patients reporting symptoms on standing did not change between baseline and 3 months in either group (four in vitamin D group vs. five in placebo group; P ¼ 0.73), but fell in both groups by 12 months (one in vitamin D group vs. two in placebo group; P ¼ 0.61). Sensitivity analyses omitting the 13 patients with a change in antihypertensive medication over the course of the trial showed little change to the effect size for systolic drop (4.3 mmHg, 95% CI 0.7 to 9.2 by repeated measures) or diastolic drop (0.8 mmHg, 95% CI 3.1 to 1.4 by repeated measures). Further sensitivity analysis using additional potential confounders was performed adjusting for baseline pulse wave velocity, baseline 6-min walk test and baseline FMD in addition to baseline age, SBP, SBP fall and 25OHD levels. These additional adjustments did not change the treatment effect estimate using repeated-measures analysis (systolic drop 4.2 mmHg, 95% CI 0.2 to 8.7; diastolic drop 0.8 mmHg, 95% CI 2.9 to 1.3). No interaction was observed between baseline 25OHD level and the effect of treatment (systolic drop P for interaction ¼ 0.34 at 3 months; P for interaction ¼ 0.99 at 12 months).
DISCUSSION The results of this analysis suggest that vitamin D may produce a transient improvement in orthostatic blood pressure fall, but that the overall effect with longer-term follow-up is neither statistically nor clinically significant. We did not demonstrate a significant relationship between the degree of orthostatic postural drop and either 25OHD levels, endothelial function or arterial stiffness. A previous study in older people reported lower 25OHD levels in patients with orthostatic hypotension [21], but no relationship was seen between 25OHD levels and orthostatic haemodynamics. Orthostatic tachycardia was associated www.jhypertension.com
1695
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Witham et al.
Invited to participate from 13 GP practices: 2255
Given information via clinics: 8
Expressed interest in study: 730
Attended screening: 341
Expressed interest in study via news article: 22
Failed screening: BP too low (83) BP too high (30) 250HD >75 (39) eGFR
Effect of vitamin D supplementation on orthostatic hypotension: data from theVitamin D in Isolated Systolic Hypertension randomized controlled trial Miles D. Witham a, Rosemary J.G. Price a, Allan D. Struthers b, Peter T. Donnan c, Martina Messow d, Alex McConnachie d, Ian Ford d, and Marion E.T. McMurdo a
Objective: Orthostatic hypotension commonly accompanies supine hypertension, and is associated with low 25-hydroxyvitamin D levels. We tested whether highdose intermittent oral vitamin D therapy could ameliorate orthostatic hypotension in older patients with isolated systolic hypertension. Methods: We conducted a subgroup analysis of data from a parallel-group, double-blind, randomized, placebo-controlled trial. Patients aged over 70 years with supine office SBP above 140 mmHg and DBP below 90 mmHg received 100 000 units oral vitamin D3 or matching placebo every 3 months for 1 year. Office supine and standing blood pressure were measured at baseline, and 3, 6, 9 and 12 months, along with arterial stiffness and flow-mediated dilatation of the brachial artery. Results: Of 159 patients randomized to the main trial, 75 patients with orthostatic hypotension at baseline were included in this analysis. The mean age was 78 (SD 5) years, baseline blood pressure was 162/76 mmHg and the mean baseline orthostatic fall in blood pressure on standing was 32/5 mmHg. After adjustment for baseline age, 25-hydroxyvitamin D, SBP and orthostatic fall, the fall in SBP was less in the vitamin D group at 3 months [treatment effect 6 mmHg, 95% confidence interval (CI) 0 to 12], but repeated-measures analysis showed no significant treatment effect (3 mmHg for systolic fall, 95% CI 1 to 8; 1 mmHg for diastolic fall, 95% CI 1 to 3). Conclusion: Twelve months of intermittent, high-dose oral vitamin D3 did not significantly improve orthostatic hypotension in older patients with isolated systolic hypertension. Keywords: older, orthostatic hypotension, randomized controlled trial, vitamin D Abbreviations: 25OHD, 25-hydroxyvitamin D; FMD, flowmediated dilatation of the brachial artery; PTH, parathyroid hormone; VitDISH, Vitamin D in Isolated Systolic Hypertension trial
INTRODUCTION
O
rthostatic hypotension is defined as a fall in SBP of more than 20 mmHg or a fall of DBP of more than 10 mmHg on standing [1]. It is commonly accompanied by a sensation of light-headedness, which may result in syncope. Orthostatic hypotension is common in older people, particularly those with vascular disease including hypertension. Community-based studies suggest a prevalence of 30% in those aged 65 and over living in the community [2–8], rising to 70% of those in nursing homes. Although antihypertensive medications are frequently blamed, the evidence for this is surprisingly weak [9,10], and it is likely that a combination of vascular stiffness, impaired autonomic function, reduced cardiac output and impaired cerebral autoregulation contribute to the phenomenon and symptoms of orthostatic hypotension [11]. Existing therapeutic strategies for orthostatic hypotension are disappointing. Non-pharmacological manoeuvres, including stockings, abdominal compression and head-up sleeping posture are either ineffective, impractical or lacking in evidence [12,13]. Pharmacological interventions, including fludrocortisone, midodrine and other pressor agents, usually rely on increasing both supine and standing blood pressure. This strategy risks increasing baseline blood pressure with consequent increased risk of vascular events, and may not reduce the fall in blood pressure with erect posture. A recent systematic review concluded that the evidence for these interventions was of very poor quality and it was insufficient to recommend any of them for routine use [14]. Low 25-hydroxyvitamin D (25OHD) levels are common in older people, and observational studies have shown Journal of Hypertension 2014, 32:1693–1699 a Ageing and Health, bDepartment of Clinical Pharmacology, cEpidemiology and Biostatistics Unit and dRobertson Centre for Biostatistics, University of Dundee, Dundee, UK
Correspondence to Dr Miles D. Witham, Clinical Reader in Ageing and Health, Ninewells Hospital, Dundee DD1 9SY, UK. Tel: +44 1382 383086; fax: +44 1382 644972; e-mail: [email protected] Received 13 January 2014 Revised 26 March 2014 Accepted 26 March 2014 J Hypertens 32:1693–1699 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. DOI:10.1097/HJH.0000000000000223
Journal of Hypertension
www.jhypertension.com
1693
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Witham et al.
associations with an increased risk of hypertension and of cardiovascular events [15]. Vitamin D is known to have a number of potentially beneficial effects on vascular biology, including reducing renin secretion, reducing PTH, which is known to be vasculotoxic, reducing proinflammatory cytokines, and improving myocardial cell contractility [16]. Vitamin D supplementation improves endothelial function in some studies, and may also reduce blood pressure in selected groups [17–19]. Unlike some vascular interventions, vitamin D reduces the risk of falls and fractures in selected patient groups [20]. It is therefore possible that by improving arterial and myocardial function via a range of possible mechanisms, vitamin D could help to ameliorate orthostatic hypotension, and observational work supports a relationship between low 25OHD levels and orthostatic hypotension [21]. To test this hypothesis, we analysed data from patients with orthostatic hypotension enrolled into a randomized controlled trial of vitamin D supplementation for older patients with isolated systolic hypertension.
METHODS We performed a secondary analysis of data collected as part of the Vitamin D in Isolated Systolic Hypertension (VitDISH) trial of vitamin D supplementation to reduce blood pressure in older patients with isolated systolic hypertension. Detailed methods have been published previously [22]. The trial was a double-blind, placebocontrolled, parallel-group randomized trial, performed on 159 patients aged 70 and over, with isolated systolic hypertension and serum 25OHD less than 75 nmol/l. Exclusion criteria were: DBP above 90 mmHg, SBP above 180 mmHg, hypertension known to be due to a correctable underlying medical or surgical cause; estimated glomerular filtration rate below 40 ml/min (by four-variable Modified Diet in Renal Disease equation [23]); any liver function test (alanine aminotransferase, bilirubin, alkaline phosphatase) greater than 3x upper limit of local normal range; and albumin-adjusted serum calcium above 2.60 mmol/l or below 2.15 mmol/l. We also excluded patients with known metastatic malignancy or sarcoidosis, a history of renal calculi, diagnosis of heart failure with left-ventricular systolic dysfunction, atrial fibrillation, and those already taking prescription vitamin D supplements. Consumption of fish oils and low-dose (200 IU per day or less) over-the-counter vitamin D supplements were not a contraindication to enrolment. Three routes of recruitment were used. The majority of participants were recruited from the community via primary care practices. Patients were identified via General Practice lists of collaborating practices within the East Node of the Scottish Primary Care Research network and letters were sent from the patient’s primary care physician. Patients expressing interest in the study were put in touch with the research team and were sent study information. Participants were also recruited after responding to an article in the local paper about the research study; a small number of patients were recruited after being approached at secondary care clinics (Cardiovascular and Medicine for the Elderly). Written informed consent was obtained from 1694
www.jhypertension.com
all participants. Research Ethics approval was gained from Fife and Forth Valley NHS Research Ethics Committee (ref: 08/S0501/90). Clinical Trials authorization was obtained from the UK Medicines and Healthcare Regulatory Authority (EudraCT number 2008-004534-24). The trial sponsor was the University of Dundee, and the trial was registered at www.controlled-trials.com (ISRCTN92186858). The full protocol is available from the authors.
Intervention Participants entering the study were allocated to intervention or placebo in a 1 : 1 ratio using a minimization algorithm, administered by the Robertson Centre for Biostatistics (Glasgow Clinical Trials Unit, University of Glasgow, UK), using a telephone-based system to conceal study allocation from investigators and participants. Minimization variables used were baseline 25OHD level above or below 50 nmol/l, baseline SBP above or below 160 mmHg, baseline age above or below 80 years and the presence or absence of diabetes mellitus. Identical blinded medication bottles were used; patients were observed ingesting 100 000 units of oral vitamin D3 (cholecalciferol) (Vigantol oil; Merck KgAA, Darmstadt, Germany) or matching placebo (Mygliol oil; Merck KgAA), after completion of baseline, and 3, 6 and 9-month assessments.
Outcomes For this analysis, we included participants who displayed a fall in office blood pressure on standing of at least 20 mmHg systolic or 10 mmHg diastolic at baseline assessment, regardless of symptoms. Blood pressure measurements were taken at 0, 3, 6, 9 and 12 months using an OMRON HEM-705CP oscillometric machine. After a minimum 5 min supine rest in a quiet room, three readings were taken in the supine position, with the mean of the second and third readings taken as the measure of supine blood pressure. Standing blood pressure was then recorded at 0, 1 and 3 min after standing. A 22–32-cm cuff was used for the majority of patients, with a large (32–42 cm) cuff used for larger participants. The same cuff size was used for all visits by an individual participant. Parathyroid hormone (PTH) was collected, spun and analysed within 4 h using a Roche multichannel analyser system as part of routine NHS biochemistry laboratory workflow. 25OHD samples were collected, spun and stored at 208C, and assayed using the Immunodiagnostic Systems (IDS) ELISA (IDS, Boldon, UK). Intra-assay coefficient of variability was 6.3% and the laboratory participated in the vitamin D External Quality Assessment Scheme (DEQAS) quality assurance scheme for vitamin D assay. Screening samples were analysed in small batches prior to entry into the trial to ascertain trial eligibility; follow-up samples were stored and batch analysed at the end of the trial.
Endothelial function Endothelial function was assessed at 0, 3 and 12 months by measuring flow-mediated dilatation (FMD) of the brachial artery in response to hyperaemia (endothelium-dependent Volume 32 Number 8 August 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Vitamin D and orthostatic hypotension
vasodilation) following 5 min of forearm cuff occlusion [24]. An ultrasound system (Sequoia 512; Siemens, Camberley, UK), using an 8-MHz linear transducer, was used for imaging. Images gated to the ECG R-wave were acquired for 60 s before cuff inflation and for 2 min after the end of cuff occlusion. Endothelium-independent vasodilatation using sublingual glyceryltrinitrate was not measured for most patients due to an unacceptably high rate of syncopal side effects in this older population during the study (>5%). Artery diameter was measured using Vascular Research Tools software (Medical Imaging Applications LLC, Coralville, Iowa, USA). Mean diameter was measured during baseline acquisition, and was compared to the maximum diameter achieved after cuff deflation. FMD was then expressed as the percentage change from baseline diameter.
Pulse wave velocity Carotid-radial pulse wave velocity and augmentation index were measured using the Sphygmocor applanation tonometry system (AtCor Medical, Gloucester, UK) at 0, 3 and 12 months to give indices of arterial stiffness. Measurements were taken in the supine position after a minimum of 5 min of rest. For augmentation index, at least 15 good-quality waveforms derived from the radial artery were averaged and transformed to central aortic pressure using the supplied generalized transfer function. Augmentation index is presented normalized to a heart rate of 75 per min. Carotid and radial tonometry measures were obtained in conjunction with 3-lead ECG monitoring to compare the onset of systole (R-wave peak) with the onset of the pulse wave, detected by Sphygmocor software using the intersecting tangents method [25], which has previously been shown to have good reproducibility. Comparison of delay between R-wave peak and carotid pulse wave onset vs. R-wave peak and radial pulse wave onset was used to derive carotidradial pulse wave velocity.
Statistical analysis Analyses were performed using SPSS version 21 (SPSS, Chicago, Illinois, USA). A two-sided P value of 0.05 was taken as significant for all analyses. Correlations were performed using Pearson’s chi-square test. Between-group comparisons of postural drop were performed using analysis of variance (ANOVA) at each time-point, with adjustment for a range of baseline variables: baseline age, 25OHD level, baseline blood pressure, and baseline blood pressure fall. Repeated-measures ANOVA was also undertaken to estimate the overall treatment effect across all time-points; these analyses were then adjusted for a range of baseline variables as above. A sensitivity analysis was performed, excluding participants who changed their antihypertensive medication over the study period.
RESULTS A total of 159 participants were randomized into the trial. Figure 1 gives the flow of participant recruitment and follow-up through the trial, focussing on those participants fulfilling the case definition of orthostatic hypotension analysed in this study. The case definition of orthostatic Journal of Hypertension
hypotension was fulfilled by 75 of 159 (47%) of the participants. Details of this group are given in Table 1. This group had similar baseline 25OHD levels to those without orthostatic hypotension (45 vs. 44 nmol/l; P ¼ 0.61), similar baseline supine SBP (162 vs. 163 mmHg; P ¼ 0.86), slightly lower supine DBP (76 vs. 79 mmHg; P ¼ 0.04) and were slightly older (78 vs. 76 years; P ¼ 0.04). Analyses of the relationships between baseline 25OHD, blood pressure, orthostatic hypotension and other markers of vascular function are given in Table 2. For the 75 patients fulfilling the case definition of orthostatic hypotension at baseline, 25OHD levels were significantly higher in the vitamin D treatment group at 3 months (64 vs. 52 nmol/l; P ¼ 0.03); the overall treatment effect on 25OHD levels calculated using repeated-measures ANOVA was 17 nmol/l [95% confidence interval (CI) 8–25, P < 0.001]. Details of the drop in SBP and DBP on standing at each time-point are given in Table 3. A reduction in systolic drop of 7 mmHg was seen in the vitamin D group at 3 and 6 months compared to placebo, but the overall treatment effect by repeated measures over the 12-month study period was non-significant at 4 mmHg. There was no significant correlation between the increase in 25OHD level between baseline and 3 months, and the degree of improvement in systolic drop between baseline and 3 months (r ¼ 0.04, P ¼ 0.75). The number of patients reporting symptoms on standing did not change between baseline and 3 months in either group (four in vitamin D group vs. five in placebo group; P ¼ 0.73), but fell in both groups by 12 months (one in vitamin D group vs. two in placebo group; P ¼ 0.61). Sensitivity analyses omitting the 13 patients with a change in antihypertensive medication over the course of the trial showed little change to the effect size for systolic drop (4.3 mmHg, 95% CI 0.7 to 9.2 by repeated measures) or diastolic drop (0.8 mmHg, 95% CI 3.1 to 1.4 by repeated measures). Further sensitivity analysis using additional potential confounders was performed adjusting for baseline pulse wave velocity, baseline 6-min walk test and baseline FMD in addition to baseline age, SBP, SBP fall and 25OHD levels. These additional adjustments did not change the treatment effect estimate using repeated-measures analysis (systolic drop 4.2 mmHg, 95% CI 0.2 to 8.7; diastolic drop 0.8 mmHg, 95% CI 2.9 to 1.3). No interaction was observed between baseline 25OHD level and the effect of treatment (systolic drop P for interaction ¼ 0.34 at 3 months; P for interaction ¼ 0.99 at 12 months).
DISCUSSION The results of this analysis suggest that vitamin D may produce a transient improvement in orthostatic blood pressure fall, but that the overall effect with longer-term follow-up is neither statistically nor clinically significant. We did not demonstrate a significant relationship between the degree of orthostatic postural drop and either 25OHD levels, endothelial function or arterial stiffness. A previous study in older people reported lower 25OHD levels in patients with orthostatic hypotension [21], but no relationship was seen between 25OHD levels and orthostatic haemodynamics. Orthostatic tachycardia was associated www.jhypertension.com
1695
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Witham et al.
Invited to participate from 13 GP practices: 2255
Given information via clinics: 8
Expressed interest in study: 730
Attended screening: 341
Expressed interest in study via news article: 22
Failed screening: BP too low (83) BP too high (30) 250HD >75 (39) eGFR
