Evidence review: efficacy

One 2013 systematic review of 2 randomised controlled trials (RCTs) and 1 additional RCT not included in the systematic review were identified that met inclusion criteria.

Systematic review

One systematic review was identified (Ong et al. 2013) that included double-blind RCTs evaluating the efficacy of drug treatments for orthostatic (postural) hypotension in adults. It included trials assessing efficacy of short-term (less than 24 hours) and long-term use (more than 24 hours and with blood pressure measurements taken over at least 48 hours).

The review included 3 RCTs that assessed the long-term effects (24 hours or more) of fludrocortisone. One of these RCTs compared midodrine plus fludrocortisone with placebo plus fludrocortisone (Kaufmann et al. 1988). Because both trial arms in this study received fludrocortisone, its efficacy cannot be determined, so the study has not been further described. The other 2 RCTs (Campbell et al. 1975 and Schoffer et al. 2007) included in the systematic review are summarised below.

Randomised controlled trial by Campbell et al. (1975)

Campbell et al. (1975) performed a double-blind, placebo-controlled, crossover RCT that included 6 adult males (mean age 52 years) with symptomatic postural hypotension (with a fall of systolic blood pressure of 30 mmHg or more) as a result of diabetic autonomic neuropathy. None of the included participants had ischaemic heart disease or cardiac failure. Two had intermittent proteinuria and a reduced serum albumin. Four participants are reported to have been receiving insulin and 2 as receiving an oral antidiabetic drug (not further specified).

Participants completed a 3‑week observation period and were then randomised to 3 weeks of oral 100 micrograms fludrocortisone tablets twice daily or placebo. This was followed by a 3‑week washout period, then crossover to 3 weeks of the alternative treatment. A total of 5 participants were evaluated, with 1 excluded because of 'default' while taking fludrocortisone. No further information is provided about why this participant dropped out.

Measurements were carried out at the end of each 3‑week period during a physical examination. Systolic and diastolic blood pressure (SBP and DBP) were measured during standardised tilt-table testing. Mean values for SBP, DBP and heart rate in the supine position were compared with mean values in the tilted position.

After 3 weeks of treatment, there was a statistically significant increase in mean supine and tilted SBP with fludrocortisone compared with placebo:

  • Mean supine SBP was 180 mmHg (±26) with fludrocortisone compared with 149 mmHg (±21) with placebo, p<0.05.

  • Mean tilted SBP was 154 mmHg (±29) with fludrocortisone compared with 110 mmHg (±16) with placebo, p<0.005.

There was a statistically significant difference between mean supine and tilted position SBP with placebo (p<0.001). However, there was no statistically significant difference between mean supine and tilted position SBP with fludrocortisone (p<0.10).

After 3 weeks of treatment, there was a statistically significant increase in tilted position DBP with fludrocortisone compared with placebo. There was no statistically significant difference between fludrocortisone and placebo for mean supine DBP:

  • Mean supine DBP was 95 mmHg (±17) with fludrocortisone compared with 87 mmHg (±10) with placebo; reported as non-significant, p value not reported.

  • Mean tilted DBP was 88 mmHg (±11) with fludrocortisone compared with 76 mmHg (±4) with placebo, p<0.05.

There was a statistically significant difference between mean supine and tilted position DBP with placebo (p<0.02). However, there was no significant difference between mean supine and tilted position DBP with fludrocortisone (p value not reported).

There was no statistically significant difference between fludrocortisone and placebo for supine heart rate. However, there was a statistically significant reduction in heart rate in the tilted position with fludrocortisone compared with placebo (86 beats/min [±20] with fludrocortisone compared with 97 beats/min [±18] with placebo, p<0.001). Participant-reported symptoms of postural hypotension were assessed regularly during the trial, and 4 out of 5 participants evaluated reportedly noticed 'a marked improvement' in symptoms while on fludrocortisone. However, it is not reported how this was assessed and no symptom scores or between-group comparisons were reported.

Randomised controlled trial by Schoffer et al. (2007)

Schoffer et al. (2007) was a double-blind, crossover RCT that included 17 adults (mean age 69 years; 76% male) with symptomatic postural hypotension (postural drop of at least 20 mmHg in SBP and/or 10 mmHg in DBP) and idiopathic Parkinson's disease. Participants with SBP more than 200 mmHg or DBP more than 100 mmHg were excluded as were participants with acute coronary syndrome or other causes of autonomic failure. At baseline, 1 of the 17 participants did not show a postural blood pressure drop that would fit the description of postural hypotension described above; however, they were still included in the study because they had a previous history of confirmed postural hypotension. Participants had an average time since Parkinson's diagnosis of 6 years. The study was carried out in 2 phases: phase 1 evaluated non-pharmacological treatment for postural hypotension and phase 2 evaluated drug treatment (fludrocortisone and domperidone).

The 3 primary outcome measures were: the orthostatic domain of the Composite Autonomic Symptom Scale (COMPASS-OD; maximum score of 16, with higher scores indicating more severe symptoms); clinical global impression of change (CGI) score focusing on orthostatic symptoms (+3=very much improved, +2=much improved; +1=minimally improved, 0=no change; −1=minimally worse, −2=much worse, −3=very much worse) and postural blood pressure. Comparisons were made between baseline and non-pharmacological treatment and between non-pharmacological treatment alone and non-pharmacological treatment plus the designated drug treatment. No direct comparisons between fludrocortisone and domperidone were made.

At baseline, COMPASS-OD score and supine and standing blood pressure were obtained. Participants were then asked to comply with 12 non-pharmacological treatments (including increasing dietary salt intake, elevating the head of the bed and wearing thigh-high pressure stockings) for 3 weeks. At the end of the 3‑week period, COMPASS-OD score and blood pressure testing were repeated and the CGI score was obtained.

Participants then entered into phase 2 of the study; postural blood pressure was assessed by tilt-table testing. Participants lay supine for at least 15 minutes, then had blood pressure and heart rate changes recorded during 5 minutes lying supine; during 5 minutes at an 80° head-up tilt; and during a further 5 minutes of lying supine. Maximal drop in SBP and DBP over 5 minutes and drop in SBP and DBP at 3 minutes were calculated.

Participants were then randomised to receive either 3 weeks of 100 micrograms once-daily fludrocortisone tablets plus 2 placebo tablets (given at lunch and dinner) or 10 mg 3 times daily domperidone tablets. This was followed by a 1‑week washout period, then crossover to 3 weeks of the alternative treatment. Participants were instructed to continue with the 12 non-pharmacological treatments throughout the study. At the end of each 3‑week treatment period, COMPASS-OD, CGI and tilt-table testing were repeated.

There was no significant change in postural blood pressure or COMPASS-OD score after 3 weeks of non-pharmacological treatment.

Withdrawal from the study was higher while receiving domperidone, with 3 people withdrawing in the first week of treatment compared with 1 withdrawal in the first week with fludrocortisone. Therefore, in phase 2, a total of 13 of 17 participants were evaluated in a per protocol analysis.

The mean COMPASS-OD score after non-pharmacological treatment was 9±3 (median 9; range of scores 5 to 15). The mean CGI score after non-pharmacological treatment was 0.4±1 (median 0; range of scores −2 to 2).

The mean COMPASS-OD score with fludrocortisone plus non-pharmacological treatment was 6±3 (median 6; range 1 to 10). This was a statistically significant improvement compared with non-pharmacological treatment alone (p=0.02). The mean CGI score was 0.6±1.2 (median 1; range −1 to 2) after fludrocortisone treatment; no p value reported.

The mean COMPASS-OD score with domperidone plus non-pharmacological treatment was 7±2 (median 6; range 3 to 11). Again, this was a statistically significant improvement compared with non-pharmacological treatment alone (p=0.04). The mean CGI score was 0.9±1.2 (median 1; range −2 to 2) after domperidone treatment; no p value reported.

There were inconsistencies in the reporting of SBP and DBP in the figures and in the narrative text of the study. Findings reported here are taken from the figures, because this included results for all 3 treatment periods, whereas the narrative text only reported results for baseline. The authors concluded that there was a trend towards reduced blood pressure drop on tilt-table testing with fludrocortisone and domperidone. No statistical analysis is reported for the tilt-table test results. The results below are expressed as mean±standard deviation (SD) (median; range):

  • Drop in SBP mmHg at 3 minutes: before treatment 21±20 (17; −15 to 48), after fludrocortisone 18±24 (8, −8 to 64), after domperidone 18±23 (5; −4 to 57)

  • Maximal drop in SBP mmHg over 5 minutes: before treatment 35±23 (32; 6 to 68), after fludrocortisone 30±23 (24; −2 to 64), after domperidone 28±21 (19; 5 to 61)

  • Drop in DBP mmHg at 3 minutes: before treatment 7±7 (8; −4 to 18), after fludrocortisone 8±13 (8, −11 to 33), after domperidone 7±15 (0, −10 to 36)

  • Maximal drop in DBP mmHg over 5 minutes: before treatment 17±10 (20; 3 to 37), after fludrocortisone 18±12 (20, 0 to 35), after domperidone 14±15 (6; −1 to 40).

Mean supine SBP before drug treatment was 138 (±23) mmHg compared with 134 (±24) mmHg after fludrocortisone, and 138 (±27) mmHg after domperidone.

Randomised controlled trial by Rowe et al. (2001)

Rowe et al. (2001) performed a double-blind, placebo-controlled RCT that included 100 adults aged 18–50 years with chronic fatigue syndrome (CFS) and neurally mediated hypotension (NMH) diagnosed during a 2‑stage tilt-table test. NHM was defined as a drop of 25 mmHg in SBP from baseline supine values, sustained for at least 1 minute, accompanied by symptoms of presyncope (defined as the presence of premonitory symptoms and signs of imminent syncope, such as severe weakness, light-headedness, nausea or diaphoresis [sweating]), with no increase in heart rate.

The study aimed to assess whether treatment with fludrocortisone would improve general wellbeing and orthostatic intolerance among people with CFS and NMH. It has been included in this evidence summary as part of the review of evidence on the use of fludrocortisone to treat postural hypotension.

To be included, participants had to have at least moderate severity of illness as determined by a score of 65 or less (out of 100) on a global wellness scale, with higher scores indicating greater wellness.

Participants were randomised in equal numbers to fludrocortisone titrated to 100 micrograms daily or placebo for 9 weeks with follow-up for a further 2 weeks after stopping treatment. If adverse effects occurred, participants were advised to reduce the dosage to the most recently tolerated dose. Both groups received potassium chloride tablets from the onset of treatment.

Tilt-table testing was performed in 2 stages. Participants lay supine for 15 minutes followed by head-up tilt to 70° for up to 45 minutes. Blood pressure and heart rate were recorded every 5 minutes while supine, 1 minute after head-up tilt and then every 5 minutes. If NMH was not provoked at stage 1, participants were returned to the supine position, received an infusion of 2 microgram/min isoproterenol hydrochloride for 10 minutes, followed by head-up tilt to 70° for a maximum of 15 minutes. Baseline tilt-table testing was performed 2 weeks before the start of treatment and again in the 9th week of treatment, while participants were still taking the study medication.

Total dropout from the study was higher in the fludrocortisone group (26%, n=13) than in the placebo group (16%; n=8), but no statistical comparison was reported. Dropout was reported up until the end of week 8, although treatment lasted for 9 weeks.

There was no statistically significant difference between fludrocortisone and placebo for the main primary outcome of the study, the proportion of participants with at least a 15‑point improvement on global wellness scores over the course of the study (7/50 compared with 5/50; p=0.76). This analysis was reported to be by intention to treat and appeared to include all 100 participants; however, only 83 of 100 participants were reported to have adequate outcome data for analysis, so it is unclear if this analysis was in fact performed by intention to treat.

Tilt-table testing found no statistically significant differences for mean supine SBP, DBP or heart rate between the fludrocortisone and placebo groups:

  • Mean supine SBP was 117.5 (SD 9.6) while on fludrocortisone compared with 113.7 (SD 10.0) on placebo, p=0.11.

  • Mean supine DBP was 73.3 (SD 6.6) while on fludrocortisone compared with 73.4 (SD 7.4) on placebo, p=0.93.

At baseline, 67% of participants had NMH provoked during stage 1 of the tilt test. The remaining 33% had NMH provoked during stage 2. At the second tilt test, which was carried out in the 9th week of treatment, there was no statistically significant difference between fludrocortisone and placebo for the number of participants who had NMH provoked during stage 1 (20/33 compared with 17/41; p=0.16). Nine participants in the placebo group had a normal tilt test in both stages of the second tilt test compared with 4 in the fludrocortisone group (no p value reported). However, this is not based on the whole population because complete results from the second tilt test are not available for 20 participants in the fludrocortisone group and 10 participants in the placebo group.