Multiple acyl-Coa dehydrogenase deficiency: an underdiagnosed disorder in adults

Multiple acyl-CoA dehydrogenase deficiency (MADD) is an autosomal recessive disorder of fatty acid oxidation that exhibits three phenotypes – types 1 and 2 occur in childhood and type 3, or late-onset, in adulthood.^{1, 2} Pathogenic variants in the electron transfer flavoprotein dehydrogenase (ETFDH) genes primarily occur in late-onset MADD, whereas mutations in ETFA and ETFB genes, which encode for the alpha and beta subunits of the electron transfer flavoprotein (ETF) and ETF-ubiquinone oxidoreductase, respectively, are more common in types 1 and 2.³ There are no established clinical criteria for the diagnosis of MADD.³ Elevated plasma acylcarnitines and altered urinary organic acids are observed during acute decompensations but offer limited utility in non-acute settings.¹ Instead, the combination of non-specific clinical (muscle weakness, seizures and behavioural disturbances) and laboratory findings (metabolic acidosis, hypoglycaemia, rhabdomyolysis and transaminitis) can support a diagnosis of MADD, prompting molecular genetic testing for known pathogenic variants.^2-4 However, in some cases, these variants are not identified, suggesting the presence of unknown aetiologies, leading to equivocal diagnosis of a MADD-like disorder that responds to MADD treatment (Table 1).

Here, we describe a series of cases of a MADD-like disorder to highlight the clinical heterogeneity and diagnostic challenges of this disorder. These cases reinforce the importance of early recognition, which leads to life-altering therapy.

Case 1 is a 27-year-old male who presented with generalised tonic–clonic seizure after 3 days of vomiting and fever. His medical history included focal epilepsy with previous left temporal cortical dysplasia resection. On the index presentation, he was found to be profoundly unrousable with hypoglycaemia (2.1 mmol/L), with no improvement following glucagon administration. Marked ketonuria (3+) and lactic acidosis were observed (nadir pH 7.21 and lactate 11.1 at its highest). He had an acute myopathy with a peak creatinine kinase of 979 μmol/L (30–190 U/L). He was intubated for possible status epilepticus but remained non-responsive to midazolam, clonazepam and propofol infusions. A subsequent EEG demonstrated a generalised encephalopathic process without any seizure-like activity. Cranial imaging demonstrated previous left mesial temporal lobectomy for known history of cortical dysplasia. A septic screen was negative. Ammonia was subsequently measured and found to be markedly elevated at 927 μmol/L, with rapid improvement to 68 μmol/L (<60 μmol/L) within 8 h of initiating intensive haemofiltration, sodium benzoate and sodium phenylbutyrate therapy. He was also commenced on carnitine 2 g three times a day (TDS), riboflavin 100 mg TDS and thiamine 100 mg BD as empirical therapy at that time. Investigations for a metabolic cause of hyperammonaemic encephalopathy were performed. Plasma acylcarnitine profile demonstrated elevations of multiple acylcarnitine species, from very long to short chains, but in a non-specific profile. Riboflavin level was low (130 nmol/L (ref 174–471)), which may be associated with a similar phenotype to MADD. Urinalysis revealed lactate, 3-hydroxypyruvate, acetoacetate, glutarate, dicarboxylic acids, lactate, fumarate, malate and 2-oxoglutarate and 4-hydroxyphenyllactate. He also developed mixed hepatocellular and cholestatic liver disease during his admission, attributed to acute fatty liver disease, and responded to riboflavin and l-carnitine therapy. Recovery occurred following 9 days in the intensive care unit and 8 days in rehabilitation. He was discharged on a high-complex carbohydrate, low-protein and low-fat diet with riboflavin and l-carnitine, in addition to his regular antiepileptic medications (phenytoin, lamotrigine and levetiracetam). Subsequent outpatient reviews saw the patient return to baseline function, and he was walking unaided by 12 months after discharge. He had declined genetic testing for MADD initially but has subsequently consented to this.

Case 2 is a 29-year-old female who presented with a history of progressive muscle weakness over several weeks, ultimately becoming wheelchair bound. This occurred following a viral infection. Her past medical history included benign intracranial hypertension, previous sleeve gastrectomy and gastric bypass. She had severe sensory neuropathy and ataxia, proximal muscle weakness and areflexia of the lower limbs. Thorough work-ups for infective, malignant and autoimmune conditions were negative and included blood and cerebrospinal fluid cultures, viral serology, autoimmune screen, cardiac magnetic resonance imaging (MRI), positron emission tomography scan, echocardiography, electromyography and cranial and spinal MRI. She had normal ammonia (24 μmol), riboflavin levels (289 nmol) and creatine kinase. Pelvic and thigh MRI demonstrated extensive patchy myositis of the gluteal and thigh regions. A subsequent muscle biopsy was consistent with a lipid storage disorder. Subsequent plasma acylcarnitine profile revealed increased medium- and long-chain acylcarnitine species and significantly reduced free carnitine levels. A urine metabolic screen demonstrated increased hexanoyl and suberylglycine consistent with a diagnosis of MADD. Genetic testing for ETFDH, ETFA and ETFB did not identify any pathogenic variants. She was commenced on l-carnitine 1000 mg twice daily, riboflavin 50 mg twice daily and co-enzyme Q10 150 mg daily with resolution of biochemical abnormalities. She improved clinically for 6 months in a rehabilitation facility, was able to mobilise with walking aids on discharge and eventually returned to full independence.

Case 3 is a 38-year-old female who presented with 24 h of vomiting, fever and global dysphasia. Her previous history included a 2-year period of acute neuropsychiatric disturbances occurring concurrently with infection or menstruation. Autoimmune, vascular and anti-neural screens were normal; cranial imaging and lumbar puncture were also unremarkable. EEG later demonstrated focal epileptiform activity leading to a diagnosis of focal epilepsy. Her subsequent presentations were managed as seizures with any unusual behaviour attributed to post-ictal psychosis. Prior to these episodes, she was a high-functioning, minimally comorbid (apart from congenital deafness) adult. On the index presentation, she was clinically mute with otherwise normal neurology. Initial management with carbamazepine and clonazepam in addition to her regular olanzapine did not lead to a clinical improvement. A subsequent metabolic screen revealed elevated short- and medium-chain acylcarnitine species with normal carnitine levels. Riboflavin level was 234 nmol/L (ref 174–471). Urinary metabolic screen showed trace glutarylcarnitine. Ammonia level was normal. Genetic testing did not identify any pathogenic variants. She was commenced on riboflavin 100 mg BD and carnitine 1 g BD. After an 18-day admission, her behaviour improved and she remained seizure-free and was discharged. On outpatient follow-up, the patient had regained her functional independence with improved neurocognition.

Discussion

MADD, a genetically heterogeneous metabolic disease, poses diagnostic challenges. Non-specific clinical findings and variable biochemical abnormalities, which can be mild, atypical or only evident during acute decompensations, further complicate diagnosis.⁵ Consequently, we encounter a spectrum of MADD-like phenotypes with unknown genetic aetiologies, highlighting the need for further research to uncover these underlying factors.

An established diagnosis of MADD is by identification of pathogenic variants and/or elevated plasma acylcarnitines and increased excretion of urinary organic acids.³ Of those cases who consented to genetic testing, none had an established genetic diagnosis of MADD despite clinical and biochemical findings suggestive of MADD. Moreover, cases assigned MADD of unknown genetic aetiology have been described in the literature highlighting difficulty and lack of consensus with the diagnostic criteria.³

The primary differential diagnosis is a disorder of riboflavin metabolism, which can mimic MADD both biochemically and clinically.^{3, 5} These defects include riboflavin transporters genes (SLC52A1, SLC52A2 and SLC52A3), mitochondrial flavin adenine dinucleotide (FAD) transporter gene (SLC25A32) and FLAD1 gene encoding the FAD enzyme.^{5, 6} Large clinical overlap is seen particularly in the myopathic symptoms and differs from MADD by associated swallowing, speech and respiratory difficulties.⁵ Thus, patients presenting with MADD-like phenotype should be investigated for defects in riboflavin metabolism. Notably, mutations in FLAD1 are regarded by some as part of MADD, whereas others classify it as a distinct MADD-like illness.^{3, 7, 8}

Secondary riboflavin deficiency from malabsorptive states because of bariatric surgery is rarely observed and is accompanied by multiple nutrient deficiencies.⁹ Dermatological manifestations such as stomatitis, seborrhoeic dermatitis and anaemia should be sought.⁹ A micronutrient panel comprising water- and fat-soluble vitamins and a mineral panel should be performed to exclude this differential. In Case 2, this panel was normal. However, in Case 2, the acylcarnitine profile coupled with findings from a muscle biopsy suggested a lipid storage disorder, thereby supporting a diagnosis of a MADD-like disorder.

Rhabdomyolysis can elevate plasma acylcarnitine levels resembling a MADD profile.¹⁰ Furthermore, hyperammonaemia (up to 500 μmol) secondary to rhabdomyolysis has been described in cases of severe agitation and seizures.^{11, 12} In Case 1, the modest creatine kinase elevation was insufficient to explain the severe hyperammonaemia, altered plasma acylcarnitine profile and urinary organic acids. Although ketonuria is typically considered to exclude fatty acid oxidation defects, metabolic acidosis and ketonuria, although rare, have been reported in cases of MADD.^{13, 14} Additionally, the patients' seizures were refractory to empirical treatment however responded well to riboflavin therapy, supporting a diagnosis of a MADD-like disorder. Case 1 also had a mildly low riboflavin which may have worsened the presentation.

Furthermore, some pharmaceutical agents, such as valproic acid and antibiotics (e.g. pivalic acid), can mimic parts of the MADD biochemical profile by inducing secondary carnitine deficiency.⁵ Elevated levels of short-chain acylcarnitines and reduced medium- and long-chain acylcarnitines have been observed in some patients after 8 weeks of treatment with selective serotonin re-uptake inhibitors and anti-psychotics,^{15, 16} although other studies have reported increased medium-chain and long-chain acylcarnitines in schizophrenia patients with comorbid metabolic syndrome.¹⁷ Thus, the role of acylcarnitines in the pathophysiology of neuropsychiatric disorders has not been fully clarified. Furthermore, acylcarnitine supplementation therapies in psychiatric disorders are conflicting.^18-20 In Case 3, a singular psychiatric disorder was unlikely to explain the patient's presentation when compared to a MADD-like disorder. The lack of response to empirical therapy, contrasted with the positive response to riboflavin treatment, along with an altered biochemical profile, supports a diagnosis resembling MADD. This underscores the importance of considering the broader clinical context rather than relying solely on a single metabolic test.

The majority of late-onset MADD cases respond to riboflavin treatment, although there are reported cases of riboflavin unresponsive MADD.² The cases discussed here did not respond to standard treatment and exhibited dramatic response when a MADD-like disorder was suspected and patients were treated with riboflavin. Acute administration of intravenous dextrose therapy, to provide energy, reduce lipolysis and restore anabolism, alongside riboflavin therapy, is crucial for the management of MADD (or suspected MADD) and, as demonstrated, can result in rapid recovery.

MADD is a rare inherited metabolic disorder characterised by significant genetic heterogeneity, some of which remains undefined. We emphasise the importance of correlating clinical features and treatment response to biochemical findings even in the absence of genetic confirmation to support a diagnosis of a MADD-like disorder. Furthermore, we recommend clinicians consider metabolic diseases in the differential diagnosis of any adult patient presenting with atypical features of ‘common’ conditions, thereby presenting the possibility of effective and potentially life-saving therapies.