Studies were eligible for inclusion in our review if they were English language articles that included patients diagnosed with POTS being treated with an oral medication for a period of seven days or longer. Studies that specifically evaluated treatments in POTS patients in the setting of post cardiac ablation or based on failure of multiple first line therapies were excluded. All age ranges of patients were considered. Studies with or without a comparator group were included to include observational (i.e., cohort, case series), randomized controlled trials (RCT), and previous systematic reviews and metanalyzes. Published articles as well as articles available on pre-print servers were eligible for inclusion. Conference abstracts and methods papers were not eligible for inclusion. Studies were excluded if they evaluated medications in animal models, if they included only individual case reports of management of POTS, if the medication was delivered in a route other than oral administration, or if the duration of administration was < 7 days.

A literature search was conducted in LitCOVID, Web of Science, Ovid ALL EBM Reviews, Embase, and PubMed on 26 APR 2024. A total of 1,675 results were retrieved and 649 duplicates removed, leaving 1,026 articles to review. Literature published from the inception date of each database to the date of search and limited English language were considered for inclusion in the review. A search query was developed in consultation with a reference librarian (RA) to include a combination of keywords and subject headings that fully represented each concept. The full query of the search strategy is included in Supplementary Figure S1. The tool Covidence was used for the management of the review process. Covidence is a web-based collaboration software platform that streamlines the production of systematic and other literature reviews.

After removal of duplicate articles, two authors (BP and KA) independently performed a screening of all titles and abstracts of identified studies and determined if based on the information provided, they met the criteria for study selection. Any disagreements between authors at this stage were decided upon by a separate author, a cardiologist experienced in the management of POTS (MF). Studies that were selected for full text review underwent screening for inclusion by two independent authors (BP and KA). Any disagreements were discussed with a senior author experienced in performing systematic reviews (TK) to reach a consensus on final inclusion of the study into the review.

All studies included in the review had relevant study variables extracted independently by two authors (BP and KA). Some administrative study details (i.e., author, year of publication) were extracted through an autonomous process by Covidence, however all data related to study design, population, intervention, or outcomes was manually extracted by the reviewers separately in a standardized fashion. Any discrepancies found in the data extraction between the two reviewers was discussed with a senior author (TK) to reach consensus. The variables extracted included the year of investigation, location of investigation, if PASC associated POTS was evaluated, if there were other specific groups under investigation (i.e., pediatric patients, or only those with hyperadrenergic POTS), the inclusion and exclusion criteria of the study, the enrollment of the study and number/reason for dropouts, the study design, the medication under investigation (including dose, and duration of treatment), the proportion of the study group that was female, the mean age of the study group. Specific outcome variables extracted as available included proportion of the treatment group meeting study criteria for treatment success (and the definition of that success), changes in reports of symptom score tools after treatment, and changes in heart rate variability on positional change testing after treatment.

All studies included in the review underwent independent critical appraisal and assessment of bias independently by two authors (BP and KA) utilizing critical appraisal tools from the Joanna Briggs Institute (15–18). Each study was evaluated on a variety of domains relevant to their individual study design on a scale of low, high, or unclear risk of bias. Any disagreements between the reviewers on the risk of bias in any study was discussed with a senior author (TK) to reach consensus. Visualizations of the assessed bias in each individual study and amongst all studies of each type were prepared using the Robvis tool (19).

A narrative synthesis was performed of the included studies. Information was segregated by medication under investigation and study design. A table was constructed of administrative data for each study to include information such as study location, funding sources, and authorship. Further tables reporting the outcomes of specific interventions for each study design were constructed. Outcome data was evaluated in terms of single arm analysis for the intervention under study, as well as relative to comparator groups as available. For case-series, cohort, and RCT, interventions with ≥2 studies evaluating their outcomes were included in quantitative outcome analysis with tables reporting results of symptomatic and heart rate response among participants. Interventions which were only trialed in one study, and specific differences in outcomes among subgroups were narratively synthesized. Heterogeneity and sensitivity analyses were performed evaluating differences in outcomes between the subgroups of PASC associated POTS as compared to the overall outcomes for each intervention as the data allowed. Publication bias was assessed using funnel plots of the outcome of symptomatic treatment response. A GRADE approach was used to assess the confidence in the studies following the guidance in the Cochrane Handbook (20). Statistical analyses and funnel plot creation were performed using SAS version 9.4 (SAS Institute Inc.). This review was registered, and the full protocol presented on PROSPERO (CRD42024505967). This review followed the PRIMA report guidelines for systematic review, no funding was received to conduct this research.

A total of 1,675 articles were initially identified in the literature search, with 649 identified as duplicates leading to 1,026 articles advancing to title and abstract screening. At this stage 980 articles were found to be irrelevant to the research question and did not meet inclusion criteria for the review. Of the remaining 46 studies undergoing full text review, an additional 14 were excluded, leaving 32 studies remaining in our review. The flowchart of study selection and rationale for study exclusion are presented in Figure 1. The included studies are presented in Table 1, included are 8 case series, 14 cohort studies, 5 RCT, and 5 systematic reviews. Overall, the studies evaluated were generally published recently, with 30 of the 32 identified articles published after 2010. The primary counties in which observational studies and RCT studies were performed were the United States (11) and China (10). The study populations comprised a mix of age ranges, with approximately half (16) of the observational or RCT studies including only children or adolescents in their study population, a smaller number including only adults (2), and the remainder including all age ranges.

The most common medications evaluated in original research were cardioselective beta-blockers (9 articles) (21–29), midodrine (8 articles) (21, 29–35), ivabradine (6 articles) (36–41), non-cardioselective beta-blockers (2 articles) (28, 42), and fludrocortisone (2 articles) (23, 31), with several other interventions evaluated in one article. Previous review articles have evaluated these agents, as well as pyridostigmine, selective serotonin reuptake inhibitors, and methylphenidate (43–47). Two articles specifically evaluated the treatment of PASC associated POTS, one with ivabradine, and the other with a cardioselective beta-blocker (22, 36).

The assessment of bias risk for each study is presented in Figure 2. Utilizing the critical appraisal bias assessment for each study and incorporating general strengths and weaknesses of study approach, the overall GRADE assessment for each study is included in Table 1. Overall, the majority of studies were found to be generally of high or moderate quality. Some areas of concern highlighted in the quality analysis review included several RCTs with unclear methodology regarding randomization and blinding (28, 29), cohort studies with unclear management of confounding variables and concerns over incomplete follow up (21, 36, 42), case series with incomplete reporting of demographics (31, 38–40, 48), and systematic reviews with incompletely described methodology for critical appraisal and data extraction (45). A synthesis of the overall proportion of studies with specific bias concerns are presented in the Supplementary Figure S2. Funnel plots evaluating for risk of publication bias are presented in Figure 3, with no significant concerns identified overall, or for any of the interventions with ≥5 studies included (midodrine, beta-blockers, and ivabradine).

The primary endpoints reviewed as available were the proportion of participants meeting the study criteria for treatment success, and the mean change in heart rate variability upon positional change. As available mean changes in symptom score were sparingly reported as well. The uncontrolled response of each medication trialed in at least two studies is reported in Table 2. Given the significant heterogeneity in study design, treatment duration, and definition of treatment success between studies, interpretation of combinations of these measures must be undertaken cautiously. When reviewing treatment success in terms of patients' symptomatic response (either qualitatively assessed as symptomatic improvement, or quantitatively as having a decrement in symptoms score above some threshold) midodrine and ivabradine have response rates of 77.76% and 74.51% respectively, while beta-blockers have a 64.45% response rate. When performing subgroup analysis by study design, midodrine had a higher response rate in the lone RCT evaluating it (89.47%, binary qualitative symptomatic response outcome) than in cohort studies (77.01%, mix of outcome definitions). Beta-blockers had higher response rates in cohort studies (65.75%, mix of outcome definitions), then in the one case series (59.38%, binary qualitative symptomatic response outcome), and RCT (57.89%, binary qualitative symptomatic response outcome) in which they were evaluated. Ivabradine had similar treatment responses in the case series (74.65%, binary qualitative symptomatic response outcome) and cohort studies (74.39%, binary qualitative symptomatic response outcome) with no RCT evidence evaluating this endpoint. Another variable that appeared to potentially skew symptomatic response was the duration of the study. When comparing studies of maximum duration of at least 6 months to those < 6 months, for all medications longer studies had lower response rates (midodrine 82.87% vs. 71.72%, beta-blockers 65.40% vs. 57.57%, and ivabradine 78.18% vs. 72.45%). Fewer studies reported changes in heart rate variability, however there were striking decreases in the pooled changes seen with both cardioselective [15.7 beats/min (bpm)] and non-cardioselective (24.3 bpm) beta-blockers. Midodrine and ivabradine respectively had pooled changes in heart rate variability of 10.3 bpm and 6.1 bpm. Study subtype analysis did find that RCTs reported greater improvements in heart rate variability than other study types, potentially owing to more rigorous methodologies around consistency in measurement. Additionally, longer study duration tended toward lower response in heart rate variability for ivabradine and beta-blockers, but for midodrine studies at least 6 months in duration reported improved heart rate responses. Other medications evaluated in only one observational study with study endpoint of patient symptomatic improvement were naltrexone, which was evaluated individually in a case series, with 50% of participants reporting symptomatic improvement, and bupropion was evaluated in a case series with 58.3% of participants reporting improvement in orthostatic intolerance (48, 49). Yozgat et al. did not evaluate the proportion of participants reporting successful treatment response, but instead reported mean changes in orthostatic intolerance symptom score between a group receiving conventional therapy and a group receiving a combination of conventional therapy, propranolol, and oral rehydration solution for 3 months of treatment (42). In this study the active group had a mean 1.84 point improvement in symptom score as compared to the conventional therapy group which had a mean improvement of 0.42 points in symptom score (42).

When considering other treatment response outcomes, Boris et al published two retrospective analyses of prescription data for POTS patients, one focusing on fatigue and other cognitive symptoms, and one evaluating physical symptoms including orthostatic intolerance (31, 50). Treatment success was defined as repeated prescription of the medication at least 5 times. McDonald et al. used a similar approach in evaluating the treatment efficacy of ivabradine, evaluating whether medication was continued at the end of an observational period (40). In general these methodologies reported lower proportions of treatment success when compared to studies using patient reported outcomes. McDonald reported a 55% treatment success for ivabradine, while Boris found a 33.91% success for midodrine, and 42.78% success for fludrocortisone. Other medications evaluated by this methodology include atomoxetine (16.5%), desmopressin (38.9%), methylphenidate (51.2%), mixed amphetamine salts (44.9%), and modafinil (43.6%) (31, 50).

Table 3 presents comparisons of study endpoints in placebo-controlled studies. Midodrine had slightly greater performance over placebo compared to metoprolol in treatment success and reduction in heart rate variability, while use ivabradine had striking improvements in SF-36 physical functioning scores compared to placebo (29, 35, 41). Table 4 presents comparisons of study endpoints in a study using active comparators. Endpoints were largely similar between bisoprolol and propranolol but adding pyridostigmine to these agents did not significantly improve symptom scores (28).

Table 5 presents the findings of previous systematic reviews on a study level, while Table 6 presents the findings of previous systematic reviews with participant level results. The findings of previous systematic reviews generally appear to be in line with the findings from the original research identified in this review. Negative findings for pyridostigmine and fludrocortisone as compared to other interventions under investigation are striking, with our review failing to find a positive study of fludrocortisone and another finding symptomatic improvement in only 51% of patients using pyridostigmine (44, 46).

Studies where PASC associated POTS was evaluated are presented in Table 7, with comparison of outcomes of studies reporting participant symptomatic improvement in which patients were identified as having PASC associated POTS to the pooled effect of studies that did not evaluate the treatment in the setting of PASC. Ivabradine slightly outperformed its historical use (78.2% of participants meeting study criteria for successful symptomatic improvement vs. 72.5%, while bisoprolol underperformed the historical performance of beta blockers in general (59.4% vs. 65.1%) (22, 36).