Midodrine initiation criteria, dose titration, and adverse effects when administered to treat shock: A systematic review and semi-quantitative analysis

A structured search of MEDLINE (PubMed) identified all English-language publications with “midodrine” in the title or abstract from inception through May 10, 2023. Publications that met predefined patient, intervention, comparator, and outcome (PICO) criteria were screened for full-text review: Patients (adults ≥18 years of age with shock); Intervention (midodrine); Comparator (not required; studies with and without control groups were included); and Outcomes (midodrine dosing and adverse drug effects).

Two investigators (TDS and DJG) independently screened titles and abstracts for evaluation with a third investigator (RRR) available for disagreements. Publications were excluded if they treated a diagnosis other than shock (e.g., orthostatic hypotension) or were a case report, trial protocol, letter to the editor, conference abstract, systematic review or meta-analysis. Publication references were evaluated during the full-text review.

Data were extracted by two investigators (MMP and KJA) using a standardized template (Figure 1). No protocol was published for this systematic review, but consensus guidance was followed including the Preferred Reporting Items for Systematic review and Meta-Analysis Protocols (PRISMA-P).

Fig. 1.

Preferred reporting items for systematic reviews and meta-analyses (PRISMA) flow diagram

Demographic and clinical characteristics included study design, country, patient population including rurality, severity of illness (e.g., APACHE II), shock etiology, renal function at the time of midodrine initiation and during therapy, phase of care (i.e., emergency department or ICU), ICU and hospital length of stay, and mortality.

Midodrine administration data included use of a midodrine dosing protocol, initial and maximum dose and frequency, dosing strategy (titrated or fixed), renal dose adjustments, timing of initiation (before, with, or after IV vasopressors), duration of therapy, route of administration (oral or feeding tube) and continuation at ICU and hospital discharge.

Vasopressor data included dosage and frequency of administration, weaning protocols, number of patients on IVPs at the time of midodrine initiation, central venous catheter duration and complications (e.g., central line-associated bloodstream infections), complications related to IVPs (e.g., extravasation), and time to IVP discontinuation. Vasopressor doses were converted to norepinephrine equivalents as previously described [11].

Potential adverse drug effects were determined a priori including bradycardia, bowel or limb ischemia, and stroke. Definitions were according to the study under review and are referred to in this manuscript as present or absent, accordingly.

Cost analysis data included direct medical costs per day of patients administered midodrine versus those receiving standard care.

If a data point was not evaluated in a publication, it was classified as “not reported,” and if it was evaluated for but not observed, it was classified as “not observed.” Corresponding authors for publications with missing data were contacted by e-mail, when appropriate.

Study quality was assessed by two blinded reviewers (MMP and TDS) using the JBI Critical Appraisal Checklists for RCTs, case control studies, case series, and cohort studies [12,13,14]. Studies were evaluated for their methodologic rigor and for potential bias in their design, conduct, and analysis. Initial, pre-discussion interrater agreement on quality appraisal was 0.84, 95% CI: 0.758–0.918 using Cohen's kappa statistic. A consensus process was then used to come to a final decision on initial disagreements.

Continuous data are reported as median (interquar-tile range 25^th – 75^th percentile), and categorical or dichotomous data as number and percentage. This study reports semi-quantitative data; quantitative analyses were not performed given the study objectives and the heterogeneity of aggregated data.

A total of 698 publications were identified and 15 (2%) were included (Figure 1) [2,3,4, 15,16,17,18,19,20,21,22,23,24,25,26]. Midodrine was administered to 1,714 patients with a median of 31 (20–79) patients per study. The first study included patients treated as early as February 2012, with the most recent study including patients treated through April 2021 [15,26]. Seven (47%) studies were retrospective, two (13%) were prospective observational, and six (40%) were RCTs; four of the six (67%) RCTs were open-label. Most studies (12/15; 80%) were single center and conducted in the United States (9/15; 60%) (Table 1). The primary outcome was time to IVP discontinuation in nine (60%) studies.

The most common admitting unit was a medical or mixed ICU (11/15; 73%) followed by a trauma/surgical ICU (7/15; 47%); many included both ICU types (Table 2). The most common shock type was “mixed” which included cardiogenic, spinal, septic, and postoperative shock/hypotension cases into one category (7/15; 47%) followed by septic only (5/15; 30%). One (7%) study did not report shock type. Severity of illness was defined using APACHE II, III or IV in ten (67%) studies, Euroscore in one (7%), and SOFA score in one (7%); severity of illness was not reported in three (20%) studies. Patients with renal insufficiency, ranging from chronic kidney disease to acute kidney injury, were excluded from seven (47%) studies.

A starting dose of 10 mg every 8 hours (4/15; 27%) or three times daily (3/15; 20%) was most common, with the exception of two (13%) studies that reported a starting dose of 20 mg every 8 hours, and six (40%) that did not report an initial dose (Table 3). A protocol for midodrine dosing was present in six (40%) studies. Seven (47%) studies used fixed dosing of 10 mg every 8 hours or three times daily and two (13%) used a fixed dose of 20 mg every 8 hours. Doses ranged from 2.5 mg every 12 hours (5 mg total daily dose) to 20 mg every 6 hours to 40 mg every 8 hours (120 mg total daily dose).

No study adjusted the midodrine dose for renal function though one did recommend a lower starting dose for patients with kidney dysfunction [27]. Most studies reported administering midodrine orally (7/15; 47%), but many did not specify if this was given by mouth or through a feeding tube, and only two explicitly stated they crushed or administered it via gastrostomy tube [18,19]. Midodrine was initiated in the ICU in a majority of publications (14/15; 93%). No studies included patients in the emergency department.

Two (13%) studies specified weaning protocols for midodrine including decreasing the dose from 10 mg to 7.5 mg every 8 hours for 24 hours, then 5 mg every 8 hours for 24 hours, then discontinuation or decreasing the dose every 1–2 days from 20 mg to 10 mg every 8 hours, then 5 mg every 8 hours, then discontinuation [16,23]. Six studies (40%) reported midodrine continuation past ICU discharge (range: 13–67% of patients) and three (20%) studies reported it was continued at hospital discharge (range: 13–52% of patients).

Midodrine was initiated to wean off IVPs during shock resolution in most studies (11/15; 73%) with a minority describing its use before or with IVPs during the early phase of shock (5/15; 30%). Nine (60%) studies reported that all patients were on IVPs when midodrine was initiated, and in the other six studies, 48% to 59% of patients were being treated with IVPs (Table 4). The number of IVPs administered at midodrine initiation was reported in ten (67%) studies and most (8/10; 80%) reported only one IVP (norepinephrine, phenylephrine, or metaraminol). The median dose of IVPs, expressed as norepinephrine equivalents, was 0.08 (0.05–0.14) mcg/kg/min. No study exclusively looked at patients not on IVPs and only two (13%) specified a weaning procedure for IVPs [15,18].

There were no reports of IVP-related or central venous catheter-related complications. Only one (6%) study reported that the midodrine group required a shorter duration of central venous catheterization, but the finding was not statistically significant [21]. The time to IVP discontinuation, the most common primary outcome studied, was 26 (20.1–59.4) hours for the midodrine patients and 78.5 (23.3–105.6) for controls.

Thirteen (87%) studies reported the incidence of bradycardia with six (46%) reporting it was present (Table 5). The definition for bradycardia varied and was generally defined as a heart rate <40–60 beats per minute. Of the 204 individual patients with bradycardia, only one (0.5%) required midodrine discontinuation and none required a medical intervention (e.g., atropine) [4].

Three (20%) studies reported the incidence of hypertension using various definitions, most commonly a systolic blood pressure >160 mmHg. The incidence of hypertension ranged from 5.6%–10.6% in the studies that reported it. None of the studies reported hypertension as a reason for midodrine discontinuation.

Four (27%) studies assessed for ischemia, either mesenteric or peripheral, with limited description on how it was assessed. Five (5/1128; 0.4%) patients in the four studies developed mesenteric ischemic requiring midodrine discontinuation. Three of the five had alternative explanations (e.g., multiple high-dose IVPs) but two did not. No peripheral (e.g., digits and limb) ischemia was observed.

One study conducted a cost analysis and reported direct medical cost per day in midodrine patients was $2,776.50 compared to $2,454.00 in control patients. Indirect medical costs were not considered.¹⁸

Evaluation using the JBI Critical Appraisal Checklist criteria for randomized controlled trials, case control studies, case series, and cohort studies, revealed varied adherence to bias-reducing strategies within individual study designs (Table 6). Few studies (5/15, 33.3%) met all bias-reduction criteria for their study type, with the majority of studies (10/15, 66.7%) being at risk for the introduction of bias in at least one facet of the study [2,22,23,24,26]. Importantly, five of six randomized controlled studies were at significant risk of bias with only one study employing all assessed methods of bias reduction [15,16,18,19,21,23].Full details of quality assessments are provided in Table 6.