National Prescribing Practices for Pediatric Dystonia Among Providers in the United States

Authors: Sarah Paige W. Davis, Natalie Kane, Haley E. Botteron, Rose Gelineau‐Morel

Abstract

While multiple oral medications are used to treat dystonia, limited information exists on current prescribing practices. This study analyzes real‐world prescribing practices for pediatric dystonia in the United States, evaluating prescription frequency, dosing, and the impact of comorbidities. Oracle electronic health record real‐world data were queried from 2014 to 2019 for encounters of patients under age 18 with a dystonia diagnosis and available medication records. Information was extracted on prescriptions for dystonia medications (baclofen, clonidine, carbidopa‐levodopa, gabapentin, tetrabenazine, trihexyphenidyl, and select benzodiazepines), dosing, and comorbid diagnoses of cerebral palsy (CP), epilepsy, or spasticity. A total of 4010 pediatric patients with dystonia were included. Benzodiazepines were most commonly prescribed (midazolam in 53.5% of patients, diazepam 46.7%, lorazepam 41.9%, clonazepam 28.3%). This was followed by baclofen (33.4%), clonidine (26.3%), and gabapentin (19.7%). Dystonia patients with epilepsy were more commonly prescribed benzodiazepines than patients without epilepsy (diazepam 79.1% vs. 29%; clonazepam 50.9% vs. 16%) and baclofen was more often prescribed in patients with CP (59.4%) or spasticity (63.8%) than those without (17%). All medications showed decreased milligram per kilogram dosage as patient weight increased. Benzodiazepines, baclofen, and clonidine were the most common medications prescribed to pediatric patients with dystonia in the United States, although medical comorbidities impact prescribing practices. There was significant variability in weight‐based dosing of all medications. There remains a need to determine which dystonia medications are most effective for each patient and the necessary drug exposure to maximize therapeutic efficacy and minimize adverse effects.

Summary What is the current knowledge on the topic? ○Multiple oral medications are used in dystonia management; however, there is currently little evidence or consensus for effective medication selection and dosing. Physician surveys of prescribing practices exist in the literature, but real‐world data regarding dystonia medication prescriptions is lacking. What question did this study address? ○This study addresses the questions of which medications are being prescribed most frequently at a national level, what dosages are being prescribed, and how other medical diagnoses impact which dystonia medications are prescribed. What does this study add to our knowledge? ○This study highlights the impact of medical comorbidities on dystonia prescriptions and suggests that the use of benzodiazepines and clonidine may be underestimated. The significant variation in weight‐based dosing for pediatric patients with dystonia is illustrated. How might this change clinical pharmacology or translational science? ○These results emphasize the need for further dystonia treatment trials with a focus on exposure‐controlled comparative effectiveness in order to optimize patient care.

Introduction

Dystonia is defined as “a movement disorder in which involuntary sustained or intermittent muscle contractions cause twisting and repetitive movements, abnormal postures, or both” [1]. It is a heterogeneous disease with diverse etiology. Pediatric dystonia is found in various conditions including cerebral palsy (CP), acquired brain injury, and metabolic and genetic disorders. By nature, it is a dynamic disorder with fluctuating symptom severity that can vary greatly between patients and even within the same patient from day to day. Treatment for dystonia can include the use of an intrathecal baclofen pump, botulinum toxin injections, or deep brain stimulation (DBS), but oral medications are usually the first line of treatment.

Effective dystonia management greatly impacts quality of life, as its symptoms can result in painful dystonic movements and high energy expenditure. This can lead to severe limitations in daily function, chronic pain and discomfort, and malnutrition. However, there is currently little evidence or consensus for dystonia medication selection and dosing.

There are no FDA‐approved oral medications for pediatric dystonia; therefore, medications are used off‐label. In the absence of randomized controlled trials for pediatric dystonia, prescribers may use adult data, expert opinion, and guidelines from influential societies to inform their practice. The American Academy of Cerebral Palsy and Developmental Medicine (AACPDM) created a care pathway for dystonia management through a systematic review of pharmacological and neurosurgical interventions in the management of dystonia in cerebral palsy [2]. This pathway is largely based on expert opinion, with minimal evidence available to guide medical management of dystonia in children. There are no FDA‐approved oral medications for pediatric dystonia in cerebral palsy, and the available literature is largely focused on dystonia in CP rather than all pediatric dystonia. There is a need to evaluate current prescribing practices for all dystonia diagnoses to guide clinical management and determine medications that should be targeted for future clinical trials.

A recent survey of physicians in the AACPDM and the Child Neurology Society in the US and Canada reported that baclofen, trihexyphenidyl, and gabapentin were the most commonly prescribed medications for dystonia in patients with CP [3] in line with previously published studies [4, 5, 6, 7, 8]. Reported dosing regimens varied between providers in the studies, but generally follow published guidelines for dystonia or other CP symptoms such as spasticity and seizures. However, these studies are limited by recall bias and lack generalizability to all prescribers.

In this study, we used real‐world clinical data from the electronic medical record to examine the prescribing practices for dystonia medications in pediatric patients in the United States. We evaluated common medications and dosing used in pediatric dystonia and the impact of medical comorbidities, with the aim of informing clinical guidelines and identifying medications to target for future clinical trials.

Methods

This study used the September 2022 refresh of Oracle electronic health record real‐world data (OERWD), a large de‐identified data warehouse of electronic health records (EHR) from health organizations throughout the United States [9]. This warehouse includes data from a wide range of health organization types, including inpatient and outpatient records from regional integrated delivery networks (IDN), rural and community hospitals, and academic medical centers. Oracle EHR Real‐World Data are extracted from the EMR of hospitals in which Oracle has a data use agreement. Encounters may include pharmacy, clinical, and microbiology laboratory, admission, and billing information from affiliated patient care locations. All admissions, medication orders and dispensing, laboratory orders, and specimens are date and time stamped, providing a temporal relationship between treatment patterns and clinical information. Oracle has established Health Insurance Portability and Accountability Act‐compliant operating policies to establish de‐identification for Oracle EHR Real‐World Data.

There were 688,552,891 encounters for 52,478,804 unique patients at 134 diverse health organizations during the 2014–2019 study period. OERWD was queried through the Health Data Lab (HDL) cloud‐based platform, which is provisioned with R (v4.0.2) and Spark (v 2.4.4). Data processing and visualization of the query output was performed locally using R (v. 4.3.2) and the Tidyverse packages (v2.0.0).

Diagnoses

OERWD was queried to identify all patient encounters with a dystonia ICD‐9 or ICD‐10 diagnosis code during the study period, 2014–2019. Please see the supplemental materials for a complete list of all diagnosis codes included in the study. This included any patient encounter with an active dystonia diagnosis at the time of the visit. The sample was limited to pediatric patients under age 18 at the time of the encounter. The base dystonia patient sample was then used to query RWD for cerebral palsy, epilepsy, and spasticity‐related diagnoses. These comorbid diagnoses were considered due to their high prevalence in patients with dystonia and their potential impact on the medications prescribed for these patients. Flags were created for each comorbid diagnosis to create condition cohorts for comparison.

Medications

Medication records for the dystonia patient sample during the 2014–2019 study period were queried from OERWD. All medication records for patients with a record of a dystonia diagnosis were included, given the potential for missing diagnosis data associated with medication encounters. Medication records were subset to only include those for patients in the sample who were under the age of 18 at the time of the encounter. OERWD medication records represent both orders for medications to be dispensed by retail pharmacies and medications administered at the point of care. They do not include pharmacy dispensing records.

The medication sample was reviewed to select records for medications by both brand and generic drug name for the dystonia medications included in the baclofen, clonidine, carbidopa‐levodopa, gabapentin, tetrabenazine, trihexyphenidyl, and benzodiazepines, including diazepam, clonazepam, midazolam, and lorazepam. These medications are all included in the AACPDM care pathway, except tetrabenazine [2]. Tetrabenazine was added to the medication list to provide information regarding the vesicular monoamine transporter (VMAT) inhibitor mechanism of action, which would not otherwise be represented. Medication usage was summarized for each patient across the study period in order to reflect overall prescribing practices in this population. A patient was recorded as receiving the medication if the medication was prescribed to them in any encounter. If a patient had multiple encounters in which they were prescribed the same drug, the drug was only counted once for that patient.

Medication records for each drug included different dose units, quantities, and drug routes, some of which corresponded with “tablets” or other measurements with the actual drug concentration recorded in the drug name. In these cases, records were processed to extract dosage and standardize to a milligram concentration where possible. Observations were excluded if there was insufficient information to convert the drug dosage to milligrams. The final sample of medication records included distinct drug and dose data recorded for an encounter. Information on total daily dose was not extracted due to inconsistent records of dose frequency.

Weight measurements for the patient encounters associated with included medication records were also queried from OERWD. The weight records were processed to exclude observations with missing values, with measurement types indicative of a weight value that was not recorded at the time of the encounter (e.g., “ideal body weight”), or observations with values ≤ 0. The weight unit was used to standardize weight measurements to kilograms where possible. Observations were excluded if no weight unit was recorded. Additional observations were dropped if the measurement was extreme and indicative of a possible unit miscoding, including a weight measurement ≥ 400 US or British pounds or ≥ 200 kg.

Demographics

OERWD provides consolidated demographics data, incorporating all patient records to identify the likely race and ethnicity categories for the patient. Race and ethnicity values were mapped to standard census categories, where possible, and summarized for the patient encounter sample with medication records. Age at encounter was also summarized for the main sample of encounters with medication records.

Results

There were a total of 6754 patients with dystonia identified, of which 4010 had medication information and were included in the study. Demographic information for these patients is summarized in Table 1. There were 1852 patients with dystonia without a diagnosis of cerebral palsy, spasticity, or epilepsy and 2158 patients with dystonia and at least one of these comorbid diagnoses (see Figure 1 for further details on dystonia patient groups).

Medication Prescriptions

The distribution of medications prescribed among all patients with dystonia can be seen in Figure 2. Benzodiazepines were the most prescribed class of medications in patients with dystonia. Midazolam was prescribed to the highest number of patients overall, with 2145 patients (53.5%) receiving at least one midazolam prescription during the study period. Diazepam was prescribed to 1874 patients (46.7%), lorazepam to 1680 (41.9%), and clonazepam to 1133 (28.3%). Excluding benzodiazepines, baclofen was the most prescribed medication, prescribed to 1338 dystonia patients (33.4%) within the study period. This was followed by clonidine at 1053 (26.3%). The remaining medications in descending order are as gabapentin prescribed to 788 patients (19.7%), trihexyphenidyl to 581 patients (14.5%), carbidopa‐levodopa to 482 patients (12.0%), and tetrabenazine to 64 patients (1.6%).

The impact of various comorbidities, including epilepsy, cerebral palsy (CP), and spasticity, on prescribing practices was observed. Total of 1852 patients included in the study did not have any of these 3 comorbid diagnoses. In these patients, the top 3 most prescribed medications were midazolam (prescribed to 45.2% of patients), lorazepam (33.4%) and clonidine (24.8%). These were followed by clonazepam (13.4%), baclofen (12.9%), gabapentin (11.8%), and carbidopa‐levodopa (10.2%).

Dystonia patients with epilepsy (1409 patients) received a higher number of prescriptions for certain benzodiazepines than patients without epilepsy, most notably diazepam (prescribed to 79.1% of dystonia patients with epilepsy vs. 29% of dystonia patients without a comorbid diagnosis of epilepsy) and clonazepam (50.9% vs. 16%).

While baclofen was the most common non‐benzodiazepine prescription overall for patients with dystonia, patients with a comorbid diagnosis of either CP or spasticity were prescribed baclofen in even higher percentages. Among dystonia patients without CP or spasticity, baclofen was prescribed to approximately 17%. This percentage increased to 59.4% in patients with dystonia and CP and 63.8% in patients with dystonia and spasticity.

Medication Dosing

The mean medication dosages for all weight groups across all medication records are shown in Table 2. This included multiple medication records for patients with multiple encounters; however, only encounters with up‐to‐date patient weights were used. The mean dosage for baclofen was 13.03 mg per dose with a mean weight‐based dosage of 0.58 mg/kg/dose. Clonidine's mean dosage was 0.10 mg, or 0.004 mg/kg/dose. Gabapentin's mean dosage among all weight groups was 240.49 mg, or 9.66 mg/kg/dose.

All medications showed variability in dosing between weight groups, with decreased mean dosing in mg/kg/dose as patient weight increased. Mean dosages by weight group for the three most prescribed medications, excluding benzodiazepines, are demonstrated in Figure 3. Mean weight‐based dosing for baclofen varied by weight group from 0.162 mg/kg/dose (110–120 kg group, mean dose 18.3 mg) to 0.662 mg/kg/dose (0–20 kg group, mean dose 9.3 mg). Mean weight‐based dosing for clonidine varied from 0.0009 mg/kg/dose (110–120 kg group, mean dose 0.117 mg) to 0.005 mg/kg/dose (0–20 kg group, mean dose 0.063 mg). Mean weight‐based dosing for gabapentin varied from 2.13 mg/kg/dose (140–150 kg group, mean dose 300 mg) to 11.40 mg/kg/dose (0–20 kg group, mean dose 147.7 mg). Weight‐based dosing by weight groups for other medications is included in the supporting Information S1.

Discussion

This retrospective study found that benzodiazepines were the most prescribed medication class in patients with dystonia. Excluding benzodiazepines, the most prescribed medications in patients with dystonia were baclofen and clonidine, followed in descending order by gabapentin, trihexyphenidyl, carbidopa‐levodopa, and lastly tetrabenazine. The high percentage of benzodiazepine prescriptions may be affected by the high frequency of short‐term benzodiazepine prescriptions used in dystonia as medication bridges during illness and dystonia exacerbations, though the duration of each prescription was not included in our study. While our study did not evaluate potential sleep or behavioral disorders as comorbidities, benzodiazepines may also be used for this purpose (ex. pre‐procedural, anxiolysis). The higher numbers of clonazepam and diazepam prescriptions in patients with dystonia and epilepsy may be at least partially attributed to the use of both agents as common seizure abortive therapy, such as clonazepam oral dissolving tablets and rectal diazepam.

The high frequency of benzodiazepine prescriptions is in contrast with previous studies. In a recent survey, benzodiazepines were not reported in the top tier of medications selected for dystonia treatment in cerebral palsy, listing clonazepam and diazepam as the fourth and sixth most commonly prescribed medications, respectively [3]. Similarly, a prospective cohort study of 278 patients with childhood dystonia found that the most common medications prescribed were baclofen (42.4%), trihexyphenidyl (35.2%), and levodopa (20.5%), followed by diazepam (19%) [4].

The discrepancy in these results highlights that, while providers may not perceive benzodiazepines as the most important component of the treatment regimen for dystonia, their use may be underestimated in this patient population. Although benzodiazepines may not be prescribed for the treatment of dystonia, they can impact tone and may play a larger role in dystonia management than is currently appreciated. At the very least, these results suggest that benzodiazepines are a component of the care plan that should be considered.

Excluding benzodiazepines, the most prescribed medications were baclofen, clonidine, and gabapentin. In patients with comorbid CP or spasticity, they were even more likely to be prescribed baclofen than other patients with dystonia. Baclofen is likely chosen more frequently for these patient populations with a goal of dual‐purpose therapy for both spasticity and dystonia. This is in agreement with other studies that report baclofen as the most commonly prescribed medication in pediatric dystonia. A recent survey of providers in the United States and Canada found that baclofen was the most common first‐line treatment for pediatric dystonia in cerebral palsy [3]. In a 2018 survey of physicians at a tertiary care center, oral baclofen was described as the first‐choice medication option for the treatment of pediatric dystonic cerebral palsy in 10 of the 11 total physician participants [5]. In the same study, prospective surveys from 57 pediatric patients showed that gabapentin and baclofen were the most frequently prescribed (21/57 and 19/57, respectively) for this patient population. In a 2021 prospective surveillance study of 16 Australian physicians who prescribed new medications for dystonia in patients with CP over the 6‐month study period, baclofen and gabapentin were prescribed most, followed by levodopa, trihexyphenidyl, and diazepam. The most common medication combinations were baclofen and diazepam or baclofen and gabapentin [6].

Interestingly, while other studies have shown baclofen and gabapentin are commonly prescribed in dystonia, the high frequency of clonidine prescriptions has not been previously reported. As clonidine is often prescribed for the management of sleep disturbances or attention deficits, this could explain the higher frequency of use in children with dystonia. However, as one of the medications recommended for dystonia management on the AACPDM dystonia care pathway, clonidine is another medication that warrants further investigation as a component of the dystonia treatment plan.

The average milligram dosing for medications was consistent with recommended dosing in prescribing guidelines and previous studies. While our study found substantial variability in dosing between patients, a similar degree of dosing variability has been observed in prior studies. In a 2021 prospective surveillance study, diazepam ranged in starting dose from 0.5 to 5 mg, with the maximum dose ranging from 1 to 15 mg per day [6]. In the 2018 survey, dosage regimens varied among the physicians and between patients, with the most consistent dosage regimen being for baclofen (14 patients with starting dose of 5 mg daily, additional 4 patients with starting dose of 2.5 mg daily), while the least consistent dosage regimens were for diazepam and gabapentin [5].

A novel finding in our study was that weight‐based dosing was consistently higher in smaller pediatric weight groups compared with higher weight groups. Every medication included in the study demonstrated a decrease in weight‐based dosing as the patient weight increased. For example, when comparing the 10–20 kg weight group with the 50–60 kg weight group, all medication doses in milligrams per kilogram decreased by approximately 50% or more. In the pediatric patient population, typical practice is to maintain an appropriate weight‐based dosage in milligrams per kilogram. This allows patients to achieve a relatively stable medication exposure as they grow. With such a prominent downward trend in milligrams per kilogram dosing in all dystonia medications, it is likely that patients with dystonia are exposed to different drug levels depending on their weight, potentially contributing to variability in medication efficacy and side effects. It will be important to further examine the relationship between medication dosage, systemic exposure, and patient response to maximize the clinical benefit of dystonia medications. Additionally, future clinical trials in dystonia will need to carefully control for drug exposure across patients before they can accurately evaluate drug response.

Our study has some limitations. The diagnoses for patients in this study are reliant on accurate, up‐to‐date ICD‐9 or ICD‐10 code diagnoses at the time of patient encounter; patients with dystonia may lack a formal diagnosis, may be classified under broader classification (ex. dyskinetic CP), or may be misdiagnosed (ex. diagnosis of spasticity rather than dystonia or vice versa). Because of the reliance on accurate diagnosis codes, the full population of pediatric patients with dystonia may be underrepresented in the study. While the results shed light on which medications commonly used for dystonia are most prescribed in pediatric patients with dystonia, it is possible that these medications were primarily prescribed for a different indication, such as sleep disturbance or another tone or movement disorder in some instances. All encounter types were included in this study; however, future studies may benefit from comparing dystonia prescribing practices between settings, such as an intensive care stay versus a clinic visit. Information on the total daily dose, duration of each prescription, and treatment indication for each medication was not available in our data set. A longitudinal assessment would be helpful to further understand medication duration, dosing, and treatment course. Future studies could include dosing information over time to further compare dosing regimens between patients and to differentiate which medications are used as daily maintenance therapy versus short courses used for dystonia or seizure exacerbations. Additionally, the dosage data do not differentiate between starting dose, titration, and goal dose, likely contributing to the variability in milligram dosing across patients. However, this applies to all weight groups equally and should not significantly impact the difference in milligram per kilogram dosing that was observed between weight groups.

Conclusion

Prescribing practices for dystonia across the United States are quite variable, especially when accounting for medical comorbidities. Further guidelines are needed regarding dystonia medication selection and dosing to optimize patient care and to inform the design of future clinical trials. Benzodiazepines were the most prescribed medications in patients with dystonia. Excluding benzodiazepines, the most frequently prescribed medications were baclofen, clonidine, and then gabapentin. These results largely align with prior survey‐based studies but suggest that benzodiazepines and clonidine may play a larger role in dystonia care than was previously appreciated. Importantly, weight‐based dosing substantially varied between groups, likely resulting in inconsistent medication exposure. It will be important to perform exposure‐controlled comparative effectiveness trials of the most commonly prescribed dystonia medications to determine which are most effective and to establish the appropriate drug exposure that will maximize efficacy with minimal adverse effects.

Author Contributions

S.P.W.D., N.K., H.E.B., and R.G.‐M. wrote the manuscript. R.G.‐M. and N.K. designed the research. N.K. performed the research. S.P.W.D., N.K., H.E.B., and R.G.‐M. analyzed the data. All authors approved the final manuscript as submitted and agreed to be accountable for all aspects of the work.

Ethics Statement

Ethical approval was not required for this study.

Consent

Patient consent was not required for this study.

Conflicts of Interest

R.G.‐M. is one of the co‐inventors on PCT international patent application publication number WO2024/059631 titled “R‐Trihexyphenidyl for Treatment of Movement Disorders”. The other authors declare no conflicts of interest.

Supporting information