Pompe disease (PD), also known as glycogen storage disease type II and acid maltase deficiency, is an autosomal recessive progressive muscle disorder (1). The condition is caused by deficiency of acid-α-glucosidase (GAA) associated with pathogenic variants in the GAA gene (1). Deficiency of GAA results in the accumulation of glycogen in muscle tissues leading to irreversible muscle damage and a range of clinical symptoms (1, 2). PD symptoms occur on a spectrum, but are typically categorized into two broad forms, infantile and late-onset. Within the same form of PD patients can have a variable progression of the disease, involvement of body tissues, and presentation of symptoms (1–3).
Late-onset Pompe disease (LOPD) is generally defined as presenting after the first year of life (2, 3). Symptoms of LOPD typically include weakness in trunk and pelvic muscles, high creatine kinase levels, and respiratory insufficiency. The average time from symptom onset to diagnosis of LOPD has been found to be as long as ten years (2–5). When a patient has been misdiagnosed with another neuromuscular disease the delay in correct diagnosis of LOPD can be prolonged by 2.5 to 10.5 years (3–6). Delayed diagnosis of LOPD is due in part to overlapping symptoms that exist with other neuromuscular diseases (4–6). For example, LOPD presents similarly to many of the limb girdle muscular dystrophies (LGMD) where pelvic and proximal muscles are weaker than distal arm and leg muscles (4–6).
In 2006, treating PD became a possibility with the first FDA licensing of an enzyme replacement therapy (ERT) for PD followed by specific licensing approval for use in LOPD in 2010 (7, 8). Additional products to treat PD have since entered the market. Although ERT cannot cure LOPD, it has the potential to help stabilize a patient’s condition and slow the progression of the disease, allowing for a better quality of life (7, 8). A review of ERT by Hagemans et al showed that for each year a patient with LOPD goes untreated after diagnosis, the likelihood of wheelchair requirement increases by thirteen percent and the need for either invasive or non-invasive ventilation increases by eight percent (8). Therefore, it is imperative to identify patients efficiently and provide them the greatest possible quality of life.
Most research suggests that the combined incidence of both infantile and late-onset PD is about 1:40,000 (9). It is believed that LOPD is more common than the infantile form due to the increased likelihood of pathogenic variants with milder phenotypes to be passed through generations. In 2011, a research study on newborn screening in Taiwan found the incidence of LOPD to be 1:26,466 (10). In Austria, a newborn screening study found an incidence of LOPD to be as high as 1:8,684 (11). In 2013, Missouri was the first state in the United States to implement newborn screening for PD. Over the six years, 36 infants were diagnosed with LOPD for an approximate incidence of 1: 14,567 (12). In Georgia, a state pilot screening project on 59,332 newborns found 2 cases of confirmed LOPD for an incidence of 1 in 29,666 (13).
Prior to our study, there were fifteen known adult cases of LOPD in Georgia based on information from the Emory Lysosomal Storage Disease Center that theoretically receives all referrals for LOPD patients in the state and none treated at Lafayette General Hospital System (Now Ochsner Lafayette General Medical Center). However, using the conservative incidence estimate of 1:29,666, approximately 125 people in Georgia should be living with LOPD, suggesting significant under diagnosis of this treatable condition.
Efforts have been made to establish guidelines to shorten the time to LOPD diagnosis. In 2009, the American Association of Neuromuscular and Electrodiagnostic Medicine established an algorithm where GAA enzyme testing should be considered when a patient presents with limb girdle weakness, axial weakness of the paraspinal muscles, mild scapular winging, and symptoms of orthopnea (2, 3). Vissing et al. encouraged the utilization of GAA enzyme assay as a first tier test and urged physicians to conduct retrospective chart reviews of patients presenting with elevated creatine kinase levels and inconclusive limb girdle muscle weakness to rule out LOPD (2).
In 2013, Preisler et al. conducted the first retrospective chart review looking for patients with undetected LOPD (6). The authors identified 3 of 103 patients with LOPD, all of whom had elevated creatine kinase and unclassified LGMD diagnosis (6). This study focused on two neuromuscular centers and one respiratory center to recruit possible participants. Preisler et al. selected participants that had elevated creatine kinase, unexplained myopathy on muscle biopsy, unexplained respiratory insufficiency, or unspecified myopathy. In 2013, Keever et al completed a retrospective chart review looking for patients with undetected LOPD using the ICD-9 code code 359.1 in a neuromuscular center at Emory Healthcare (14). They identified 637 patients that led to the genetic counseling and testing of 23 subjects. In the end, 2 patients and several additional family members (1 sibling and 2 children) were diagnosed with LOPD (14). In 2023, a group in Spain took a slightly different approach by reviewing all adult internal medicine patients seen in 13 hospitals who presented with: polymyositis or any type of myopathy of unknown etiology; diagnosis of obstructive sleep apnea syndrome by polysomnography together with a body mass index (BMI) ≤ 30 kg/m2; asymptomatic or pauci-symptomatic elevated creatine kinase. After dried blood spot testing for GAA enzyme and confirmatory molecular testing of the GAA gene, they found 3/322 individuals were affected by LOPD (15).
The overall objective of this study was to build on this literature by expanding the scope of high-risk patient review to all patients seen in two very different health care systems: a large academic medical center in Georgia and a small regional hospital in the underserved medical region of Louisiana. The study goal was to find patients at-risk for LOPD via Electronic Medical Records (EMRs) data review starting with an automated data pull for specific ICD-10 codes and then utilizing a symptom-based scoring system using observational clinical information, laboratory results, and billing information.