Treatable inborn errors of metabolism presenting as cerebral palsy mimics: systematic literature review

A systematic review was performed to identify all reports of inborn errors of metabolism (IEMs) presenting with CP symptoms by searching the PubMed database, checking reference lists of relevant articles and consulting with experts in the field. We considered only articles that were published in English, described findings in humans, and those where full text publication was available electronically through our institution’s subscription. No publication date restrictions were imposed and articles included in analyses were published between 1957–2014. The last search was performed October 15, 2014.

We included only studies describing diseases that are IEMs, which we have previously defined as “genetic disease involving a disorder of metabolism with confirmation based on the internationally accepted diagnostic test(s) for that IEM (gene mutations, enzyme deficiency, or specific biochemical marker); this term excludes endocrine disorders” [19]. To identify diseases most likely to be misdiagnosed as CP (i.e., to exclude adult-onset), we limited our inclusion to reports where CP symptoms (Table 1) present before 5 years of age in at least one reported case.

The goal of this systematic review was to identify all treatable IEMs that present with CP symptoms, rather than to identify every report. We selected the most reliable or comprehensive reference for this publication; this implies that additional case reports of CP symptoms for a given IEM beyond those cited here may exist.

The search strategy is outlined in Table 1. Primarily, combinations of search terms that describe cerebral palsy symptoms and inborn errors of metabolism were used to identify relevant articles in the PubMed database (see Table 1). Different combinations yielded common articles, indicating our approach is conservative and our search terms are effective. Results from all key word combination searches were pooled and sorted to identify 472 unique articles (i.e., without duplicate results), which were then manually screened based on inclusion/exclusion criteria by the first author (ELL), with duplicate publications and those that did not meet eligibility criteria being removed. As well, reference lists of these publications were screened for additional relevant articles and these were judged for inclusion using the same criteria. Finally, to ensure that we have identified all salient IEMs, we performed another search using known treatable IEMs listed in a previous publication of treatable IEMs that present with intellectual disability (ID) [19] in combination with our search terms describing CP. Initial searches were performed independently by ELL following review protocol and all possible studies were compiled into a table. The generated list of IEMs was reviewed periodically by experts SS and CvK to confirm data, refine the search strategy and inclusion criteria.

The following information was extracted from each article: disease name, cerebral palsy symptoms as reported by the original authors, age of symptom onset, any treatment used (with the dose and frequency) and the treatment outcome as described by the authors. For completeness, we later assigned each disease to a biochemical category and retrieved OMIM identifiers, underlying genes and pattern of inheritance, appropriate diagnostic test, and standard of care treatment; all were agreed upon by expert clinicians. Diagnostic tests were compared with a recent diagnostic algorithm of treatable IEMs [20] and with accepted clinical practices (e.g., GeneReviews®). Given the positive experience with the 2-tiered protocol for the work-up the child with IDD for treatable IEMs, we have used this as a basis for the evaluation of the child with CP for similar conditions.

As most reviewed studies were case reports, the type of outcome measures varied. We noted whether or not treatment was implemented and the original author’s observations on treatment effect. However, due to the time span of the reviewed publications, some treatments are now available that were not at the time of the original publication. Therefore, we took a more objective approach of employing clinical expertise to determine the standard of care treatment for the systematically identified CP mimics and categorized them as treatable versus non-treatable. Causal treatment for this proportion of conditions has been thoroughly evaluated by us already [19], and we apply the same treatment recommendations here. For conditions where treatment was not previously reviewed by us, we adhered to guidelines for the specific IEMs and where needed provided expert input, with consideration for treatments applied in the original case reports of CP mimics.

Treatment of IEMs can be either ‘primary treatment’, which targets the pathophysiology at a cellular level and improves at least the CP features (muscle tone, rigidity, etc.) plus/minus the cognitive, behavioural, and MRI features; or ‘stabilizing/preventative treatment’, which halts decline and/or prevents further damage, especially during metabolic crises. For example, creatine supplementation in creatine deficiencies targets the underlying cause of the IEM and can reverse the features. Several disorders caused by deficiencies in vitamins or co-factors can show improvement in primary features with appropriate supplementation. Examples of stabilizing/preventative treatment include emergency regimen for fatty acid oxidation disorders, HSCT for lysosomal disorders and dietary restriction of amino acids in hyperhomocysteinemias and amnio acid disorders.

Levels of evidence for treatments were evaluated based on existing level classification by the Centre of Evidence-Based Medicine (www.​cebm.​net): Level 1a = Systematic Review of RCTs, 1b = Individual RCT, 1c = ’All or None’ [=(prolongation of) survival with therapy]; Level 2a = Systematic Review of Cohort Studies, 2b = Individual Cohort Study, 2c = ’Outcomes Research’ [focussed on end results of therapy for chronic conditions, including functioning and quality of life]; Level 3 = Systematic Review of Case–Control Studies; Level 4 = Individual Case–Control Study or Case-series/report; Level 5 = Expert opinion without critical appraisal; based on physiology, bench research or first principles.

The authors who are expert clinicians in pediatric IEMs (CvK & SS) identified 13 additional diseases which they have designated as CP mimics in their clinical experience. For the majority of these, the pathophysiology is identical to IEMs in the same category for which literature evidence does exist; for example PSPH and PSAT deficiency which are both characterized by lack of serine in the brain similar to PGHDH deficiency (primarily treatable by serine supplements), or urea cycle defects in which hyperammonemic crises cause irreversible brain damage (which are preventable via dietary manipulation and medication), just like OTC deficiency for which CP-like features have been described. Likely such diseases have not (yet) been described to present as CP in the literature, likely due to the rare frequency and the increasing challenge of publishing case reports alone. The expert opinion approach was used, so that also for these IEMs, affected individuals might also receive the benefit of early detection and intervention.

To provide guidance for the identification of treatable CP-mimics, we used the 2-tiered algorithm published by van Karnebeek et al. [19] to identify treatable IEMs in children presenting with IDD: First tier tests are generally accessible and offered by most biochemical genetics laboratories around the world with reasonable turn-around times and affordable prices (total costs $567.97 CAD), including tests in blood (lactate; ammonia; copper; ceruloplasmin; plasma total homocysteine; plasma amino-acids, and bloodspot quantitative acylcarnitine profile) and in urine (creatine metabolites; purines and pyrimidines; organic acids; oligosaccharides; and glycosaminoglycans). Each of these screening tests has the potential to specifically identify treatable IEMs, which is then often confirmed via molecular and/or enzymatic analysis. The 2^nd tier requires a more directed, ‘single test per disease’ approach based on signs and symptoms. In general these tests are more invasive and more expensive.

Based on the defined inclusion/exclusion criteria, we identified 54 treatable IEMs reported to mimic CP (Table 1). These are categorized alphabetically into 13 biochemical categories (Table 2): amino acids (n = 5), cerebral glucose transport (n = 1), creatine (n = 1), fatty acid-related processes (n = 3), hyperhomocysteinemia (n = 2), lipids (n = 1), lysosomal (n = 4), metals (n = 2), mitochondrial (n = 3), neurotransmission (n = 9), organic acids (n = 12), urea cycle (n = 4) and vitamins/co-factors (n = 7). The CP symptoms for each IEM and corresponding reference are described in Additional file 1: Table S1. A further 13 treatable IEMs were identified by expert clinicians on our team (Table 3) for a total of 67 treatable CP mimics. Treatment modalities included: dietary restriction/supplement, co-factor/-enzyme, vitamin, substrate inhibition, substrate reduction, bone marrow and hematopoietic stem cell transplant, gene therapy. The majority of these treatments are accessible and affordable. The total number of different treatments is 50, and evidence varies for the various treatments from Level 1b, c (n = 2); Level 2a, b, c (n = 16); Level 4 (n = 35); to Level 4–5 (n = 6); Level 5 (n = 8). For 26 (39%) of these IEMs, a treatment is available that targets the primary underlying pathophysiology with the potential to improve CP symptoms; while for the remaining 41 (61%) IEMs, treatment is available that stabilizes disease or prevents further damage (e.g., treatment of Succinic semi-aldehyde dehydrogenase deficiency (SSADH) with Vigabatrin can stabilize symptoms [22],[23]).

Thirty-eight of the 67 disorders (57%) of the treatable IEMs described in this review can be identified by ‘1^st tier’ metabolic screening tests in blood or urine (Table 4). The other 29 (43%) require more specific and sometimes invasive methods (‘2^nd tier tests’; Table 5). Of the 1^st tier tests in the Treatable IDD protocol, urine MPS and urine oligosaccharides are not required for the diagnostic evaluation of CP for treatable IEMs. Most 1^st tier tests which detect treatable IEMs described in the literature as CP mimics can also identify ‘treatable IEMs identified by expert opinion’ (e.g., urine creatine metabolites identify GAMT deficiency, but also Creatine transporter deficiency and AGAT deficiency).

There are several IEMs that can present as CP mimics that are not (yet) treatable. These include:

Disorders of amino acids (e.g., Hyperprolinemia type I); cholesterol (e.g., Mevalonic aciduria); lipids (e.g., FAHN, Pelizaeus-Merzbacher disease); lysosomal disorders, such as sphingolipidoses (e.g., GM1/2 gangliosidoses); mitochondrial diseases (e.g., Leigh’s disease, sulfite oxidase deficiency, respiratory chain deficiencies); metals (e.g., NBIA1, PLAN); organic acids (e.g., Fumarate Hydratase Deficiency); peroxisomes (e.g., NALD); and purine and pyrimidine disorders (e.g., Adenylosuccinase deficiency, Purine nucleoside phosphorylase deficiency). Some of these conditions have emerging treatments, but not an established standard of care treatment; for example, Canavan disease, [25] and Gaucher disease, type 3 [26].