A retrospective observational cohort study of the anesthetic management and outcomes of pediatric patients with Alexander disease undergoing lumbar puncture or magnetic resonance imaging

Abstract

Background:

Alexander disease is a rare, progressive leukodystrophy, which predisposes patients to complications under general anesthesia due to clinical manifestations including developmental delay, seizures, dysphagia, vomiting, and sleep apnea. However, study of anesthetic outcomes is limited.

Aims:

Our aim was to describe patient characteristics, anesthetic techniques, and anesthesia-related complications for Alexander disease patients undergoing magnetic resonance imaging and/or lumbar puncture at a quaternary care children’s hospital.

Methods:

We performed a retrospective review of anesthetic outcomes in patients with Alexander disease enrolled in a prospective observational study. Included patients had diagnosed Alexander disease and underwent magnetic resonance imaging and/or lumbar puncture at our institution. We excluded anesthetics for other procedures or at outside institutions. Collected data included patient characteristics, anesthetic techniques, medications, and complications under anesthesia and in the subsequent 24 hours. We performed descriptive statistics as appropriate.

Results:

Forty patients undergoing 64 procedures met inclusion criteria. Fifty-six procedures (87.5%) required general anesthesia or monitored anesthesia care (MAC) and eight (12.5%) did not. The general anesthesia/MAC group tended to be younger than non-anesthetized patients (median age 6 years [IQR 3.8; 9] vs 14.5 years [IQR 12.8; 17.5]). In both groups, dysphagia (78.6% vs 87.5%, respectively), seizures (62.5% vs 25%), and recurrent vomiting (17.9% vs 25%) were frequently reported pre-procedure symptoms. Inhalational induction was common (N = 48; 85.7%), and two (3.6%) underwent rapid sequence induction.

Serious complications were rare, with no aspiration or seizures. Hypotension resolving with ephedrine occurred in eight cases (14.3%). One patient each (1.8%) experienced post-procedure emergence agitation or vomiting. Fifty-three (94.6%) were ambulatory procedures. No inpatients required escalation in acuity of care.

Conclusions:

In this single-center study, patients with Alexander disease did not experience frequent or irreversible complications while undergoing general anesthesia/MAC. Co-morbid symptoms were not increased post-anesthesia. Some patients may not require anesthesia to complete short procedures.

Introduction:

Alexander disease is a progressive, life-limiting leukodystrophy whose primary pathology involves astrocytes in the central nervous system^1–4. Pathogenic variants in GFAP, the gene encoding glial fibrillary acidic protein, contribute to a toxic excess of this protein in astrocytes^4–6. The disorder is rare, with an estimated 5-year incidence of 1 in 2.7 million⁷. Patients with Alexander disease present across the age spectrum, from infancy through adulthood^1–3,7. The phenotype of Alexander disease is highly variable, and symptoms can have a significant impact on global function, including impaired cognitive abilities, frequent seizures, dysphagia, and abnormal autonomic nervous system control over vascular tone. Symptoms may vary with patient age, as seizures are more common in younger children and apneic disorders more common in adolescents and young adults^4,8.

Patients with Alexander disease often require general anesthesia or monitored anesthesia care (MAC) for routine diagnostic procedures, including magnetic resonance imaging (MRI), to monitor progression of white matter disease and ventricular size, as well as for symptom management, such as placement of a gastrostomy tube or ventricular shunt, or correction of scoliosis. Currently, one antisense oligonucleotide therapy, administered intrathecally every three months, is under investigation to reduce GFAP production^9–12 and improve signs and symptoms of disease, which necessitates frequent exposure to anesthetic agents for certain trial patients. However, due to the rarity of leukodystrophies, anesthetic safety data are currently limited and, to our knowledge, do not include patients with Alexander disease^13–17. Though case series data report low rates of complications under general anesthesia, it is not known if these outcomes can be extrapolated to all patients with Alexander disease, particularly those with primary brainstem pathology and associated symptoms.

To better understand how patients with Alexander disease tolerate general anesthesia or monitored anesthesia care, we analyzed patient characteristics, anesthetic management techniques, and short-term outcomes of Alexander disease patients undergoing MRI and/or LP at our institution over a seven-year study period. We aimed to describe the disease manifestations that most impacted anesthetic care, techniques for induction and maintenance of anesthesia, and incidence of serious complications. We hypothesized that (1) adverse events under general anesthesia or MAC would be rare, (2) patients would not require deviations from our institution’s typical anesthetic practice for MRI and/or LP, and (3) patients would return to their pre-anesthetic baseline prior to same day discharge.

Methods:

We performed a retrospective observational study of anesthetic techniques and complications of patients with Alexander disease who were enrolled in a prospective, observational study of the outcomes of Alexander disease at the Children’s Hospital of Philadelphia (CHOP), a quaternary-care children’s hospital, from May 2017 to August 2023 [ClinicalTrials.gov NCT02714764 registered February 19, 2016; principal investigator Amy T. Waldman, MD, MSCE]. The prospective study was approved by the Institutional Review Board (IRB 16–012649). Written informed consent in addition to verbal assent (where appropriate from minor patients) were obtained for all procedures, including administration of anesthesia or monitored anesthesia care, and retrospective medical record review. Study design and reporting were guided by the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement¹⁸.

Inclusion criteria for the prospective natural history study included a documented diagnosis of Alexander disease. Patients with other leukodystrophies were excluded. While enrolled, patients (or their parents, if the patient was less than 18 years of age) could consent to undergo two optional tests for the purpose of correlating imaging findings and cerebrospinal fluid (CSF) composition with functional status over time: (1) up to two lumbar punctures (LP) at least 12 months apart and (2) one MRI of the brain. Monitored anesthesia care (MAC) and/or general anesthesia were offered as needed to facilitate tolerability of all procedures. Anesthetic techniques, including the option to forego anesthesia, were not dictated by the study protocol, but rather, left to the discretion of the anesthesiologist based on clinical history, assessment of risk on the day of the procedure, and patient and family preference. Diagnosis and treatment of intra-procedural complications were also subject to clinician expertise and judgment.

For this analysis, we included anesthetics for all MRIs and LPs completed within our institution during the study period and excluded anesthetics for study patients done at outside institutions and for other procedures during the study period. While enrolled, some patients were hospitalized and underwent additional clinical MRIs and/or LPs to evaluate and treat symptomatic disease progression. These clinical studies were also included as the study protocol permitted the inclusion of medical records for non-research procedures. We considered each anesthetic as an independent data point, as some patients had more than one anesthetic during the study period. We included data from patients who completed their procedures with an anesthesia team on standby.

The electronic health record and anesthesia information management system databases (CompuRecord, Philips, Andover, MA; Epic Systems Corporation, Verona, WI; ChartMaxx, Quest, Secaucus, NJ) were queried to obtain the patient characteristics and peri-anesthetic data for patients meeting the inclusion criteria. International Classification of Diseases Code, Ninth Revision-Clinical Modifications (ICD-9-CM) and Tenth Revision-Clinical Modifications (ICD-10-CM) were used to identify patients with Alexander disease. Current Procedural Terminology (CPT) codes were used to identify all relevant research procedures, including MRI of the brain with or without contrast and lumbar puncture. MRI with CSF flow study was performed on some patients for clinical evaluation. We manually reviewed charts and validated data for study patients to ensure that all patients were correctly identified.

The collected data included patient characteristics, procedure type, anesthetic techniques, and intra- and post-anesthetic complications. Patient medical history and comorbidities were populated into the anesthesia record per institutional practice by both an anesthesia nurse practitioner and an anesthesiologist using a combination of manual chart review and patient interviews on the day of the procedure. Peri-anesthetic data collected included induction strategy, airway devices used, maintenance of anesthesia, and medications. Table 1 describes the most common manifestations of Alexander disease and the potential sequelae that patients may experience under general anesthesia^1–4. Anesthetic records were reviewed manually for the adverse events described in Table 1, in addition to other well-known risks of general anesthesia or MAC such as laryngospasm, bronchospasm, unspecified hypoxia, use of vasopressors to treat clinician-diagnosed hypotension, brady- or tachyarrhythmias, unexpected difficult mask ventilation or intubation, oral injuries, emergence agitation, and delayed emergence. Any other unexpected adverse events were also documented. We also reviewed patient medical records for evidence of complications within 24 hours after the end of an anesthetic, such as phone calls from parents expressing concerns, emergency room visits or readmission after discharge, or escalation in acuity of care for inpatients.

Table 1:

Manifestations of Alexander disease and implications under anesthesia^1–4

System	Symptom	Considerations under anesthesia
Neurologic (supratentorial or general)	Seizures	Breakthrough seizures on induction or emergence
	Megalencephaly	Difficult ventilation and/or intubation
	Hydrocephalus	Elevated intracranial pressure, impaired cerebral perfusion
	Hypotonia	Residual weakness during recovery
Neurologic (bulbar/pseudobulbar)	Dysphagia, dysphonia, dysarthria	High burden of oral secretions, aspiration
Gastrointestinal	Gastroesophageal reflux, vomiting	Aspiration, delayed gastric emptying
	Failure to thrive	Pressure injuries, challenges with weight-based dosing of medications, glycemic management, hypothermia
Musculoskeletal	Neuromuscular scoliosis	Pressure injuries, difficulty positioning, possible thoracic insufficiency
Pulmonary/airway	Sleep apnea – central and/or obstructive	Difficulty ventilation, may require prolonged invasive ventilation or non-invasive ventilatory support during recovery
Cognitive/developmental	Developmental delay or regression	Delayed emergence, challenges with cooperation during IV placement or inhalational induction
Dysautonomia	Temperature instability	Hypothermia or hyperthermia
	Postural hypotension, tachycardia, and/or bradycardia	Tachycardia, bradycardia, and/or hypotension

Statistical analysis:

As this study was intended to be a descriptive analysis, there was no statistical power analysis conducted before it. Basic summary statistics were generated, as appropriate for the data, including counts and percentages, mean and standard deviation, or median and interquartile range (IQR) for all patient characteristics and outcomes.

Results:

Forty patients undergoing 64 procedures met inclusion criteria from 2017–2023 (Figure 1). Fifty-six procedures for 35 patients were completed with general anesthesia or monitored anesthesia care, while eight procedures were done for eight patients without anesthesia. Of note, three patients underwent one procedure with general anesthesia and one procedure without anesthesia, which accounts for the total of 40 patients. Demographics and clinical information for patients subdivided by anesthesia/MAC or no anesthesia are shown in Table 2. Subjects in the study population were evenly distributed by gender (53.1% male).

Figure 1:

Patient enrollment

^† 1 patient only had a non-MRI/LP procedure and was excluded; 4 anesthetics for non-MRI/LP procedures were also excluded but patients met inclusion criteria on other occasions

^‡ 5 anesthetics were for clinical procedures

^§ 3 patients each underwent procedures under anesthesia and with anesthesia standby

TABLE 2:

Patient demographics

	GA/MAC (n = 56 cases)	Anesthesia Standby (n = 8 cases)
Age (years), median [IQR]	6 [3.8; 9]	14.5 [12.8; 17.5]
Body mass index, median [IQR]	16.7 [15.4; 17.7]	21.5 [17; 23.8]
Associated comorbidities, n (%)
Seizure disorder	35 (62.5)	2 (25)
Obstructive or central apnea	5 (8.9)	4 (50)
Arrhythmia	5 (8.9)	1 (12.5)
Scoliosis	11 (19.6)	2 (25)
Dysphagia	44 (78.6)	7 (87.5)
Vomiting	10 (17.9)	2 (25)
Adrenal insufficiency	1 (1.8)	1 (12.5)

Abbreviations: GA: general anesthesia; MAC: monitored anesthesia care; IQR: inter-quartile range

Age was a notable difference between the general anesthesia/MAC group and the anesthesia standby group. Anesthetized patients had a median age of 6 years [IQR 3.8; 9]. Standby patients tended to be older, with a median age of 14.5 years [IQR 12.8; 17.5]. Many of the common manifestations of Alexander disease were reported by patients and families at similar rates in both groups, including dysphagia (78.6% anesthesia/MAC vs 87.5% standby), recurrent vomiting (17.9% vs 25%), scoliosis (19.6% vs 25%), and dysautonomia, including cardiac arrhythmia (8.9% vs 12.5%). Seizure disorders tended to be more frequently reported in the anesthesia group (62.5% vs 25%), whereas apneic disorders were more common in the anesthesia standby group (8.9% vs 50%). Importantly, concern for post-procedure apnea was cited by an anesthesiologist as a reason for not offering general anesthesia to one patient in the standby group.

Procedure length likely factored into the decision to proceed without anesthesia. The anesthesia standby procedures tended to be shorter in duration, with a median procedure length of 26 minutes [IQR 23; 28.5] versus 110.5 minutes [IQR 98.8; 127] for the anesthesia or MAC group. Whereas 49 of the 56 general anesthesia or MAC procedures (87.5%) included MRI or MRI combined with LP, most of the standby procedures (62.5%) were LP only, without MRI, which can account for the significant disparity in duration.

Since anesthetic technique was not dictated by the study protocol, induction and maintenance techniques varied slightly (Table 3). To facilitate a calm induction, oral midazolam was administered for pre-procedure anxiolysis in 21.4% of cases and parental presence in 100% of cases. In most of the 56 cases, an inhalational agent (sevoflurane) was used for induction (85.7%) and maintenance (87.5%) of anesthesia. Propofol infusions were used in five cases as part of a general anesthetic with a natural airway. Two patients received intermittent propofol and midazolam boluses for monitored anesthetic care. Propofol bolus doses were also used to facilitate airway placement and transfers between anesthetizing locations. Dexmedetomidine was used in three cases to reduce the incidence of emergence delirium in patients considered high risk by history or significant concern from the clinician or family. In the anesthesia standby group, none of the patients received an oral pre-medication. A topical eutectic mixture of local anesthetic was used for comfort during LPs.

TABLE 3:

Anesthetic features/events (all general anesthetics or sedation, n = 56)

Induction strategy, n (%)
Midazolam pre-medication	12 (21.4)
Parental presence	56 (100)
Induction method, n (%)
Inhalational	48 (85.7)
Intravenous without rapid sequence	6 (10.7)
Rapid sequence induction and intubation	2 (3.6)
Maintenance of anesthesia, n (%)
Volatile anesthetic (sevoflurane)	49 (87.5)
Propofol infusion	5 (8.9)
Propofol and midazolam bolus (monitored anesthesia care only)	2 (3.6)
Airway type, n (%)
Supraglottic airway	42 (75)
Invasive airways	7 (12.5)
Endotracheal tube	6 (10.7)
Tracheostomy	1 (1.8)
Natural airway	7 (12.5)
Documented intra-procedural complications, n (%)
Low blood pressure requiring ephedrine	8 (14.3)
Hypoventilation	1 (1.8)
Difficult ventilation and/or intubation	0 (0)
Vomiting or aspiration	0 (0)
Seizure	0 (0)
Documented post-procedural complications, n (%)
Persistent emergence agitation	1 (1.8)
Vomiting	1 (1.8)
Post-anesthetic disposition, n (%)
Home	53 (94.6)
Return to pre-procedure inpatient bed	3 (5.4)

Though patients in 10 of the 56 anesthesia/MAC cases (17.9%) self-reported a history of vomiting, only two (3.6%) were considered significant enough by their anesthesiologist to undergo a rapid sequence induction and intubation (RSI). Succinylcholine and high-dose rocuronium were used once each for RSI. Despite the prevalence of dysphagia in the study population, supraglottic airway use was common (75%) and natural airways were as frequently used as invasive airways (12.5% each).

Serious complications of general anesthesia were rare (Table 3). One patient had a brief period of hypoventilation after induction requiring bag mask ventilation. There were no instances of vomiting, aspiration, or seizures at any time during the anesthetic, and no documented reports of difficult airway management. There were no cases of post-anesthetic apnea in the patients with a history of sleep apnea. Ephedrine was given during eight anesthetics (14.3%) for treatment of hypotension. Median low systolic and diastolic blood pressures treated with ephedrine were 68 mm Hg (IQR 62; 69) and 25 mm Hg (24; 28), respectively. Five patients accounted for all eight anesthetics in which ephedrine was administered, and two patients required ephedrine during more than one anesthetic.

Recovery for most patients was also uncomplicated, as 94.6% of patients were discharged home on the same day (Table 3). Three patients, who had clinical imaging for symptom progression, were inpatient at the time of their anesthetic exposure and returned to the same level of care after anesthesia. One inpatient, induced by RSI for history of vomiting, had post-anesthesia nausea and vomiting (without clinical concern for aspiration) after return to the inpatient floor. One outpatient experienced emergence delirium, with self-limited behavioral changes after discharge to home.

Discussion:

Alexander disease is a rare, progressive leukodystrophy of infants, children, and adults with significant supra- and infratentorial pathology, which contributes to significant disability and predisposes patients to complications under general anesthesia.

This study provides evidence that patients with Alexander disease can be anesthetized for low-risk procedures without significant complications or worsening of underlying disease. Despite our cohort having many identifiable risk factors for complications under general anesthesia, serious events were rare and reversible. With time and routine interventions, all patients returned to their pre-anesthetic baseline state of health. Some patients seemed to be predisposed to certain complications, as hypotension occurred in two patients during more than one anesthetic. Additionally, in the case of emergence delirium, the patient’s mother attributed the event to oral midazolam for pre-procedural anxiolysis; emergence delirium did not occur during a later anesthetic when midazolam was not given. As such, modifying anesthetic technique based on history can help to decrease the risk for subsequent events in Alexander disease patients undergoing repeated anesthetics.

Our findings are consistent with the small body of literature on anesthetic and sedation safety in other leukodystrophies. In a prospective study comparing sedative dosing and complication rates for 18 patients with metachromatic leukodystrophy (MLD) compared to 20 healthy controls all undergoing MRI, Mattioli et al. showed no significant differences in mean induction or maintenance doses of thiopental or propofol between groups¹⁹. None of the MLD patients or healthy controls had complications of sedation. Recovery times were also comparable between groups. In a retrospective analysis of 96 patients with Krabbe disease or MLD undergoing general anesthesia for surgical procedures, Bascou et al. demonstrated a complication rate in 11 of 287 cases (3.8%), which, though not a high absolute number, was higher than the overall complication rate for general anesthetics at their institution (0.246%, p < 0.0001). Hypoxemia (n = 2) and respiratory distress (n = 2) were most common¹³. Bascou et al.’s study did not limit patients to mostly optional procedures, as our study did, which may suggest a more acutely ill population at baseline and explain their findings.

As in our population, anesthetic techniques varied in the literature, with frequent use of both inhalational inductions and supraglottic airways¹³. Volatile and intravenous techniques have been used widely for maintenance in patients with leukodystrophy, with no technique showing superiority^{13–16,19,20}.

While rapid sequence induction and intubation (RSI) can mitigate the risks associated with Alexander disease, such as vomiting, this was not deemed necessary by the clinical judgment of the anesthesiologists caring for most patients. In this study, there were no episodes of vomiting or aspiration even when airway-protective strategies were not used. Of note, there are little data on the safety of succinylcholine for RSI in leukodystrophies. Much of the evidence, though limited, against succinylcholine has been extrapolated from adult case reports of massive hyperkalemia in different upper motor neuron disorders such as multiple sclerosis leading to a general avoidance of succinylcholine in leukodystrophies^16,20–22. In this study, succinylcholine was used in one patient for RSI without adverse events, though absence of adverse events in this small sample does not necessarily validate its use in patients with Alexander disease.

Importantly, all anesthetics in this study were conducted outside of an operating room. This is relevant given the rising proportion of pediatric anesthetics taking place outside of an operating room and the higher American Society of Anesthesiology Physical Status Scores found in patients treated in non-operating room anesthetizing locations²³. Though this study did not compare Alexander disease patients to the general population, Bascou et al. showed a significantly higher rate of complications in the MRI scanner among Krabbe disease and MLD patients versus the general population¹³. Thus, it is essential to emphasize that care for Alexander disease patients outside of the operating room, where resources may be scarce, is not trivial, and requires staff with expertise in the complications that patients with Alexander disease may experience under general anesthesia.

Older patients with Alexander disease were sometimes able to tolerate procedures without general anesthesia or monitored anesthesia care, particularly for procedures of short duration. Mild cognitive disabilities may not prohibit completing short procedures without general anesthesia, especially with the help of child life specialists. Though patients in this study tolerated general anesthesia well, minimizing exposure to its inherent risks and discomforts, including prolonged fasting and recovery time, is undoubtedly beneficial.

There are some limitations to this study. Only those patients and families who agreed to optional procedures were included in our anesthetic analysis. This can introduce selection bias, as patients who opted into the elective procedures may be different demographically or clinically than those who opted out. Despite this potential confounder, we sought to analyze and present these data due to the rarity of Alexander disease and the limited anesthetic safety data available for this rare disease. These data are particularly relevant in the setting of one ongoing clinical trial for an intrathecal disease-modifying therapy that requires general anesthesia for administration to most patients with Alexander disease. To our knowledge, this is one of the largest populations of patients with Alexander disease under general anesthesia studied to date and includes a diverse group of patients referred from outside our regional catchment and abroad.

Medical record-keeping is subject to bias and human error, which may include missing or inaccurate data, particularly as related to high-acuity events during an anesthetic²⁴ and/or patient or parental perceptions of disease severity. Relevant comorbidities, particularly those that might dictate anesthetic decision-making such as frequent vomiting, may have been underreported during the pre-anesthetic interview. We also relied on medical records for evidence of post-anesthetic complications. For the three inpatients, the electronic medical record provided robust data for the subsequent 24 hours. For outpatients, however, there was little evidence for post-anesthetic complications, but an absence of data may not account for all events. In-depth analysis of vital sign data was not included in this study. Notably, MRI-compatible monitors at our institution vary from those used in other locations, leading to discrepancies in vital sign recordings between anesthetizing locations. Though some patients received ephedrine for low blood pressure, variability in measurements and differences in clinician threshold for its treatment are a source of potential bias and limit our ability to comment on specific vital sign abnormalities. As a result, some subjective and objective complications may have been under-reported.

Since anesthetic techniques were not dictated by study protocol, this analysis was not designed to evaluate the safety or efficacy of medication dosages, nor to directly compare volatile anesthetic to intravenous anesthetic for maintenance. Patients in this study were anesthetized per routine practice at our institution, which is subject to provider preference and clinical judgement. Twenty anesthesiologists managed all general anesthesia and MAC cases, which can introduce variability in care, though most cases were still done with sevoflurane induction and maintenance and a supraglottic airway, as is common at our institution for patients cared for in the MRI suite. Complication rates were low overall, though no technique appeared to be associated with greater risk.

At present, care for patients with Alexander disease remains concentrated at highly-specialized institutions with experienced anesthesiologists who can mitigate disease-specific risk factors. Therefore, the outcomes of this study may not be generalizable to other institutions where Alexander disease is less prevalent. This study was not designed to analyze outcomes for operating room procedures or surgical emergencies, which require different techniques in patients who may be more acutely ill, thus limiting the applicability of this study’s results to the operating room.

Our results indicate that patients with Alexander disease can undergo short duration general anesthesia for low-risk, routine procedures such as MRI and LP without serious, irreversible complications. Further study is needed to understand the long-term safety outcomes of repeat anesthetic exposures in the Alexander disease population, particularly for administration of future treatments for Alexander disease. Identifying which anesthetic techniques are the best tolerated and most risk-reducing would be especially relevant for future investigation.

Clinical Implications:

The study of anesthetic safety in pediatric patients with rare disorders such as Alexander disease is critical to minimizing risk during necessary exposures to general anesthesia for diagnostic and therapeutic procedures. Owing to the rarity of these conditions, however, anesthetic safety is often limited to case reports. In this single center study of forty patients with Alexander disease, we have shown that these patients are able to undergo brief procedures under general anesthesia or monitored anesthesia care without significant complications or deviations from standard anesthetic practice at our institution. Select patients may be considered for procedures without anesthesia.