Should Kids With Autism Be Screened for Mitochondrial Disorders?

Another question prompted by the case of Hannah Poling is whether children with autistic features should be screened for mitochondrial disorders. When Hannah exhibited autistic regression after a series of vaccinations at the age of 19 months, she was ultimately found to demonstrate mitochondrial dysfunction (Poling JS et al. J Child Neurol. 2006;21:170-172). The authors (including Hannah's neurologist father) suggested that "oxidative stresses from immune activation" may lead to autistic regression in children with compromised mitochondrial function.

Poling and company then conducted a follow-up, retrospective study of patients with autism, who had not been diagnosed with metabolic disorders. They found that AST levels were elevated in 38% (vs 15% in age-matched controls; P < .0001), and that serum creatinine was elevated in 47%. The authors concluded that "further metabolic evaluation is indicated in autistic patients and that defects of oxidative phosphorylation might be prevalent."

The recommendation to evaluate autistic children metabolically was made by Pons et al (J Pediatr. 2004;144:81-85), who described several autistic children with the mitochondrial DNA mutation that typically causes MELAS.* These authors concluded that "autistic features may be the expression of mitochondrial dysfunction in a developing brain" and that "analysis of mitochondrial metabolism should be considered in autistic patients with associated neurologic findings or evidence of maternal transmission."

However, in a review article published that same year (Lerman-Sagie T et al. J Child Neurol. 2004;19:397-381), investigators at Tel Aviv University concluded that "mitochondrial diseases are probably a rare and insignificant case of pure autism" and further advised that care should be taken when ascribing autism to identified mitochondrial abnormalities. Detection of mitochondrial dysfunction may be the result of laboratory shortcomings, or mitochondrial dysfunction may be unrelated to clinically manifest autism.

The following year, a population-based Portugese study of autistic school children (Oliveira G et al. Dev Med Child Neurol. 2005;47;185-189) found elevated plasma lactate levels in approximately 20% of those tested and mitochondrial dysfunction in a surprisingly high 7%. Unfortunately there were no distinguishing features, such as clinical phenotype or family history, to aid the differentiation between autistic children with mitochondrial dysfunction and those without.  

The bulk of the data to date (and there is little "bulk") indicate that the likelihood of identifying mitochondrial dysfunction in an autistic child is low, but probably not insignificant. Also Oliveria and colleagues remind us that the tissue distribution of mutant mitochondrial DNA is variable and could explain possible false-negative evaluations for mitochondrial dysfunction in cases of autism.

*MELAS = mitochondrial myopathy, encephalopathy, lactic acidosis, and stroke.