Midodrine Is an Effective Therapy for Resistant Intradialytic Hypotension: CON

Intradialytic hypotension (IDH), often defined as systolic BP <90 mm Hg (or <100 mm Hg) or a BP drop during dialysis of at least 20 mm Hg, is a common complication for 20%–50% of patients undergoing in-center hemodialysis (ICHD).1 The drop in BP can result in a variety of symptoms, including nausea, vomiting, dizziness, chest pain, cramps, and syncope. In addition, due to repeated (often subclinical) instances of hypoperfusion to various organs, IDH can lead to ischemic brain disease, myocardial stunting, and gut endotoxin translocation, which likely contributes to the elevated mortality risk associated with IDH.2 Because IDH is associated with higher mortality and morbidity in ICHD patients, various strategies have been employed to limit its occurrence, including use of cool dialysate, the withholding of antihypertensive medications on dialysis days, food avoidance during dialysis, and administration of vasoconstrictive medications. The most commonly used vasoconstrictive medication to prevent and/or alleviate the symptoms of IDH is midodrine, an α-1 adrenergic receptor agonist, which causes constriction of both veins and arteries, preventing venous pooling while increasing arterial vascular tone. It is typically given to the patient 15–30 minutes before dialysis start, reaches peak active metabolite levels around 1 hour after administration, and is cleared from the body during the dialysis process (half-life of 3 hours).3 Midodrine was initially approved by the US Food and Drug Administration in 1996 for the treatment of orthostatic hypotension. Since that time, dialysis prescribers have used midodrine off-label to treat IDH. Yet, despite decades of anecdotal observation, we contend that there is insufficient data to declare that midodrine is an effective therapy for IDH. To date, there have been few published studies on the use of midodrine in treating IDH. Although these studies did find that midodrine caused an increase in BP during dialysis and a decrease in some symptoms associated with IDH, the efficacy of midodrine use for IDH must be carefully examined with respect to study limitations. A literature search was able to identify a total of nine reports of prospective studies (described in Table 1) and one retrospective study. Whereas the existence of nine prospective studies may sound encouraging, optimism must be tempered by the realization that cumulative enrollment for these studies was 105 patients, an average of approximately 12 per study. (Note: these figures do not account for overlap between two of these studies, one of which was a longer-term extension of the other.) Given the small number of patients and various patient characteristics and comorbidities included in each study, caution should be exercised when extrapolating the results to a larger clinical population. Table 1 - Study characteristics Publication Number of Participants Method Blinded Outcome/Findings Duration of Midodrine Exposure (5) 10 Randomized crossover Yes Difference in mean BP fall during dialysis: SBP and DBP fell less during dialysis with midodrine versus placebo 5 tx (8) 10 Pre-/postintervention No Difference in mean BPs: Nadir and postdialysis SBP, DBP, and MAP higher with midodrine versus control 10 tx (7) 13 Pre-/postintervention No Difference in mean BPs: Nadir and postdialysis SBP, DBP, and MAP higher with midodrine versus control 8 months (14) 11 Pre-/postintervention a No Difference in mean BPs: Nadir and postdialysis SBP, DBP, and MAP higher with midodrine versus control 3 weeks (9) 10 Pre-/postintervention No Difference in mean BPs: Predialysis and postdialysis SBP, DBP, and MAP higher with midodrine 4 weeks (10) 21 Pre-/postintervention No Difference in mean BPs: Nadir and postdialysis SBP, DBP, and MAP higher with midodrine versus control 5 tx (13) 6 Pre-/postintervention b No Difference in mean CO, CBV, MAP Mean CO, CBV, and MAP decreased less with midodrine versus control 24 tx (12) 12 Pre-/postintervention No Difference in mean nadir SBP and DBP: Mean nadir SBP and DBP higher with midodrine versus control 18 tx (11) 12 Pre-/postintervention c No Difference in mean MAP and PVR: Mean MAP and PVR higher with midodrine versus control 4 weeks CO, cardiac output; CBV, central blood volume; DBP, diastolic BP; HD, hemodialysis; MAP, mean arterial pressure; PVR, postvoid residual; SBP, systolic BP; tx, treatments.aFirst two stages of a larger protocol (phases consisted of control, midodrine, cool dialysate, and midodrine+cool dialysate).bPart of a larger protocol including two phases (phase 1A control, phase 1B cool dialysate; phase 2A control, phase 2B midodrine).cExternal controls: age- and sex-matched normotensive patients served as controls. Aside from immediate concerns regarding generalizability of findings, studies of this size very much invite the possibility of publication bias (i.e., editors might very well publish a 12-person study that shows a therapy “works” but would be much less inclined to publish a 12-person study that shows the therapy does not).4 Moreover, although these studies were prospective, only one (total enrollment 10 patients) was actually randomized.5 The remainder crossed patients unidirectionally from usual care/control to midodrine. This design is problematic, considering the principle of regression to the mean, which states that in any series of complex phenomena that are dependent on many variables and chance is involved, extreme outcomes tend to be followed by more moderate ones.6 In other words, if you enroll patients in a study on the basis that they had been experiencing low BPs during dialysis, their BPs on dialysis are likely to become less low as a function of time (i.e., as they happen to cross over to midodrine). Additionally, of the studies that only included one therapy (midodrine) for treatment of IDH, various doses of midodrine and administration times before dialysis start were used.5,7–12 Many of the studies administered a dose of 10 mg roughly 30 minutes before dialysis start; however, doses ranged from 5 to 10 mg on dialysis days, and one study even included midodrine administration on nondialysis days for uremic patients.9 Two of the studies investigated midodrine and/or cool dialysate for the treatment of IDH. Both midodrine and cool dialysate showed similar increases in intradialytic BP over control, whereas the combination therapy had no added benefit over each individual therapy to IDH patients.13,14 Finally, only one of these reports explicitly described having used blinding, which invokes the possibility of the placebo effect.5 And although the placebo effect is often thought of more in instances where the outcome is subjective, there is nonetheless ample evidence that it can impact studies of objective outcomes such as BP.15 It is also important to note that none of these prospective studies examined IDH directly. Rather, the studies looked at mean differences in predialysis, nadir intradialytic, and postdialysis BPs. Although these phenomena are related, it is not necessarily the case that impacting the former will translate into an impact on the latter. Figure 1 demonstrates two hypothetical scenarios by which a variable might influence a continuous outcome and yet fail to influence a derivative event-based outcome.Figure 1: Example BPs before, during, and after dialysis with/without midodrine treatment for intradialytic hypotension (IDH). Each circle represents a dialysis treatment; red circles indicate those treatments during which IDH occurred (using a definition of nadir systolic BP <90),2 and white circles those in which IDH did not occur. Beneath each treatment are the corresponding predialysis, nadir, and postdialysis systolic BPs. (A) In all instances, corresponding BPs are 6 mm Hg greater in the midodrine arm versus placebo; yet, in both arms, the same number of treatments are affected by IDH. (B) Again, mean predialysis, nadir, and postdialysis systolic BP are 6 mm Hg different between groups, but the percent of treatments effected by IDH is the same.The above is not meant as a blanket condemnation of the studies in question. In fact, we commend the investigators of each for their ability to perform meaningful studies that address an important clinical topic largely in the absence of dedicated research funding. Neither do we hold up our prior study as a definitive counterpoint. Although our study was larger (1043 midodrine patients and 2033 controls) and did consider IDH as a dedicated outcome, it was retrospective and observational.16 We noted at the time and reiterate here that despite maximal efforts, our findings were likely residually confounded, and such confounding could account for the absence of association detected between midodrine use and reductions in IDH. So, rather than holding up one set of evidence versus the other and debating the relative merits, we instead focus on the one important commonality among all of the studies: each indicates in some manner that it is not definitive and that certainty regarding whether midodrine ameliorates IDH must await a large, rigorous clinical trial.3,7,8,10,13 Moreover, for midodrine to be an effective therapy for IDH requires more than just proof of efficacy. To receive a strong clinical endorsement, a therapy must be proven, above all, safe. In this regard, it is critical to note that none of the studies, prospective or retrospective, were designed or powered to make a rigorous safety assessment. And although some modicum of reassurance regarding safety can be extrapolated from studies in other populations or be on the basis of anecdotal experience, definitive proof of safety must wait. Finally, any endorsement of midodrine as an effective therapy for IDH must take into account durability of effect. Among susceptible patients, IDH affects treatment on a long-term, recurrent basis. In other contexts, direct manipulation of the adrenergic system has resulted in tolerance over time.17 Thus, it is plausible that midodrine’s effect (if any) on IDH might diminish over time. In this regard, it is noteworthy that only two published studies have reported on midodrine exposure for more than 2 months, and one of these showed no effect.7,16 In summary, IDH is a common and bedeviling condition for many patients receiving chronic maintenance hemodialysis. We would all very much like there to be a reliable safe and effective silver bullet to assist such patients. Unfortunately, despite a lot of good work by scientists and clinicians, we lack the necessary evidence to ascribe to midodrine the mantle of an effective therapy for IDH.