Background
Orthostatic hypotension (OH) is a relatively common problem in the elderly, a significant cause of disability, and an independent risk factor for falls and mortality [
1‐
5]. Neurogenic OH (nOH) is a less common but more severe form of the condition, and is the result of an impaired sympathetic nervous system response during transition to an upright posture associated with reduced norepinephrine release [
6].
Despite the clinical importance of nOH, there has been a paucity of treatment options for this orphan condition [
7]. For nearly 2 decades, no new treatments were introduced; in 2014, droxidopa was approved by the US Food and Drug Administration (FDA) for the treatment of symptomatic nOH caused by primary autonomic failure (Parkinson disease [PD], multiple system atrophy [MSA], pure autonomic failure), dopamine β-hydroxylase deficiency (DBHD), and nondiabetic autonomic neuropathy (NDAN) [
8]. Droxidopa is a prodrug that is converted into norepinephrine by dopa decarboxylase, the same enzyme that converts levodopa into dopamine [
9].
Accelerated approval by the FDA under Subpart H was based on 3 randomized controlled double-blind clinical studies that showed the efficacy of droxidopa for relief of nOH symptoms [
10‐
13]. In Studies NOH301 and NOH306, patients were randomized to placebo or droxidopa to determine if their symptoms would improve during active treatment [
11,
12]. In Study NOH302, patients treated with open-label droxidopa were randomized to continue receiving active treatment or to switch to placebo to determine if withdrawal from active treatment was associated with worsening of symptoms [
10].
Cumulatively, these trials represent the largest experience to date in the treatment of nOH. Here, we analyzed the integrated data from these 3 trials. The increased number of observations provided us with a unique opportunity to perform a pooled efficacy analysis and a post hoc targeted subgroup analysis to improve our understanding of nOH and to provide physicians with more extensive information about the use of droxidopa for the treatment of patients with nOH.
Discussion
Here, we report the analysis of the combined dataset of the 3 pivotal studies that support the efficacy and safety of droxidopa for the treatment of nOH in patients with chronic autonomic failure. Combined, these studies represent the largest randomized clinical trial experience in the treatment of nOH (a total of 460 patients) and provide a unique opportunity to learn about the condition and its treatment with droxidopa.
Consistent with findings of the individual studies, the combined analyses showed that droxidopa significantly reduced the cardinal symptom of nOH, orthostatic dizziness/lightheadedness, feeling faint or feeling like you might black out (the primary outcome recommended by the FDA), by −3.0 ± 2.9 units versus 1.8 ± 3.1 units for placebo (P < 0.001). Furthermore, the combined analysis provided increased power to detect the effect of droxidopa on the less frequently reported symptoms of nOH and on symptom impact on activities of daily living. Droxidopa significantly improved 3 of the other 5 individual measures of orthostatic symptoms (visual disturbances [P < 0.001], weakness [P < 0.001], and fatigue [P = 0.010]). Conversely, OHSA Items 5 (trouble concentrating) and 6 (head/neck discomfort) were not particularly useful in discriminating between the treatment groups. Droxidopa also significantly improved 3 of the 4 individual measures of nOH symptoms: interference with the ability to carry out activities of daily living (standing a long time [P < 0.001], walking a short time [P = 0.001], and walking a long time [P = 0.003]); the improvement in OHDAS Item 1, standing a short time, did not reach statistical significance. Overall, the pooled data show that patients treated with droxidopa derived a broad range of symptomatic benefits. Importantly, it appears that these symptom improvements translated into an increased ability of patients to perform activities of daily living.
Clinical improvements in symptoms of nOH and in symptom impacts were seen in all 3 studies, and in 2 of the 3 studies, the improvements were statistically significant. Nonetheless, proving treatment efficacy was more challenging than we had expected; recognition of these challenges will inform the design of future randomized clinical trials in patients with nOH. Several factors may explain this difficulty in addition to the relatively small size of the studies, a consequence of the challenges in recruiting adequate numbers of patients with this orphan condition. First, there was a large placebo effect. Because many of the outcome variables were subjective symptom scores, this finding is perhaps not surprising. However, even upright BP improved in patients randomized to placebo, raising the possibility that participation in the studies improved patients’ adherence to general non-pharmacologic treatment recommendations for nOH (increased salt and water intake, compression garments, and sleeping with the head of the bed elevated).
Second, there was significant variability in the quantification of symptoms of OH. This is likely due a combination of factors, including the subjective nature of an unanchored scale such as the OHQ, bias in the patient’s recall of symptoms occurring in the past week toward significant individual events (such as a bad or good day), and shifts in the perception/recollection of baseline condition as a response to improvement (i.e., the “treadmill effect” [
16]), which can result in underestimation of the overall magnitude of the reported benefit.
Third, there was significant variability in the magnitude of individual patients’ response to droxidopa. Because patients with nOH have varying degrees of blunted baroreflex buffering, depending on the underlying diagnosis and severity of disease, these findings are not unexpected.
In these clinical trials, treatment with droxidopa was initiated at 100 mg TID and was escalated in 100-mg TID increments every 24 to 48 h, while monitoring BP for safety. The integrated data suggest that simply asking patients to grade their symptom burden, using OHSA Item 1 (dizziness, lightheadedness, feeling faint or feeling like you might black out), can be used to assess efficacy. As with other treatments, droxidopa is not universally effective; 8.7% of enrolled patients were unable to complete the dose optimization phase because of AEs or hypertension. Almost 40% of patients reached the maximal approved dose (600 mg TID). It is possible that some patient would have benefited from larger doses; however, the safety and efficacy of higher doses have not been tested.
With regard to post hoc subset analyses, treatment with droxidopa appears to be effective in all primary diagnosis subsets (PD, MSA, and pure autonomic failure). It should be noted that the results for patients with MSA were not significantly different from placebo; however, this could be due to a larger than expected placebo effect and the small sample size for the MSA group (n = 55). These reasons may also explain the differences noted by sex (i.e., droxidopa appearing to be less effective in women, although a strong trend was still noted [P = 0.054]).
The efficacy of droxidopa in the subset of patients taking a DDCI is of particular interest because dopa decarboxylase (l-aromatic amino acid decarboxylase), the enzyme that converts droxidopa into active norepinephrine, also converts levodopa to dopamine. Inhibitors of this enzyme are routinely used together with levodopa to reduce the side effects associated with peripheral conversion of levodopa to dopamine in the treatment of PD. Thus, the pharmacologic effects of DDCIs should render droxidopa ineffective. Indeed, a previous proof-of-concept study found that a single 200-mg dose of the DDCI carbidopa prevented the pressor effects of a dose of droxidopa administered 90 min later [
17]; however, this dose is 8- to 10-fold higher than those used in combination with levodopa for the treatment of PD. Another study found that DDCIs, at doses used clinically, did not prevent the therapeutic effect of droxidopa [
18]. Similarly, it is reassuring that droxidopa produced significant symptomatic improvement in patients receiving a DDCI in our studies. The effect, however, seems to be lower than in patients not receiving a DDCI. Information on the dosage of DDCIs was not collected in the studies but it is likely that the effect is dose dependent. It is possible that patients receiving high doses of DDCIs may have a reduction in the pressor effect of droxidopa and may require titration to higher doses [
9]. Nonetheless, in these studies, the average dose of droxidopa in patients receiving DDCIs (439 mg TID) was not markedly higher than in those not receiving DDCIs (421 mg TID). Nonetheless, when modifying dosages of levodopa/carbidopa in patients with PD, it may be necessary to re-titrate droxidopa.
There is less information about other drug-drug interactions that can potentiate the effects of droxidopa, but it is reassuring that drugs commonly used in the treatment of patients with PD were allowed in these trials. Nevertheless, any medication that reduces the metabolism of norepinephrine could theoretically potentiate the pressor effects of droxidopa [
9]. There is limited experience with NRIs in these studies; only 28 patients received concomitant treatment with an NRI. Although there was no increase in the number of cardiovascular AEs reported in patients receiving concomitant NRIs, there remains the possibility that these drugs will potentiate the actions of droxidopa, which is converted to norepinephrine. However, this may be less critical when titrating droxidopa in patients on stable doses of NRIs or other drugs that reduce the metabolism of norepinephrine. Even so, care should be taken when adding NRIs to the treatment regimen of patients already receiving droxidopa.
Our results also underscore the need to develop objective endpoints to assess symptomatic burden in nOH. In the past, an increase in upright BP has been used as a surrogate for reduced symptom burden. A recent study in patients with PD noted that the presence of symptomatic OH was associated with an upright mean BP less than 75 mmHg [
19]. In our studies, treatment with droxidopa resulted in improved upright BP; however, our patient population was less homogeneous. There was a significant relationship between symptoms and BP at a population level; however, the individual variation in both BP and OHQ scores made it impossible to predict symptomatic benefit for any given patient based on BP (data not shown).
Droxidopa appears to be well-tolerated. In the pooled analysis of studies with short-term exposure to droxidopa (Studies NOH301 and NOH302; 1–2 weeks), there were no obvious differences in AEs between groups, with the exception of headaches in the droxidopa group (6% vs 3% for placebo). In the study with a longer duration of drug exposure (Study NOH306; 8–10 weeks), headaches, dizziness, nausea, and hypertension were more frequently reported by patients receiving droxidopa, and the incidence of supine hypertension was greater with droxidopa. Although this finding is not unexpected, caution is required for patients with severe supine hypertension (who were excluded from these studies). Patients should be advised against lying down after taking droxidopa, and, instead of using a rigid TID regimen, the last dose should be given at least 3 h before bedtime. During the individual studies, no evidence of significant cardiac AEs and no trend for clinically meaningful electrocardiogram changes were noted; heart rate increases were minimal (approximately 1 beat per minute). Interestingly, the typical pattern of AEs seen with a selective alpha-1 agonist (i.e., piloerection, dysuria, and pruritus) [
20] were not noted with treatment with droxidopa, suggesting that the mechanism of droxidopa involves more than just alpha agonistic effects.
It should be noted that these studies only enrolled patients with nOH associated with alpha synucleinopathies (pure autonomic failure, PD, and MSA), NDAN, and DBHD. The efficacy of droxidopa has not been studied in the treatment of OH due to other causes. Also, evidence of efficacy was limited to 1–2 and 8–10 weeks in these studies, although longer-term safety and efficacy data have been reported [
21,
22]. Droxidopa received accelerated FDA approval, with commitment from the sponsor to undertake a postmarketing trial to assess the durability of efficacy; such a study is underway [
23]. Future research can also determine if droxidopa is efficacious on other conditions characterized by central or peripheral deficits of norepinephrine.
Acknowledgments
The authors received editorial assistance from CHC Group (North Wales, PA), which was supported by Lundbeck LLC.