Early developmental delay in Leigh Syndrome spectrum disorders is associated with poor clinical prognosis

Abstract

Background.

Leigh spectrum syndrome (LSS) is a primary mitochondrial disorder characterized by neurodevelopmental regression and metabolic stroke typically in early life. Developmental delay (DD) is known to follow episodes of neurologic regression in LSS, although primary developmental delay (pDD) has been rarely reported. We hypothesized that pDD precedes regression in a broader subset of LSS individuals and may associate with worse long-term educational outcomes.

Methods.

From a retrospective cohort, subjects with pathogenic variant(s) in a nuclear or mitochondrial gene associated with LSS and consistent clinical manifestations and neuroradiological findings. Detailed developmental histories and neurologic outcomes were extracted.

Results.

Of 69 LSS subjects, 47 (68.1%) had a history of pDD and 53 (76.8%) had neurodevelopmental regression. We identified 3 distinct developmental phenotypes: [1] pDD followed by regression (N=31/69, 44.9%), [2] pDD without subsequent regression (16/69, 23.2%), [3] regression without pDD (N=22/69, 31.9%). A history of pDD was associated with earlier disease onset (p=0.0003) and worse educational outcomes (OR 22.14).

Conclusion.

LSS is associated with multiple developmental phenotypes and pDD is associated with negative educational outcomes. pDD occurring prior to neurologic regression suggests that mitochondrial energetics impact developmental trajectories prior to acute metabolic failure and regression, providing an opportunity for earlier diagnosis and/or therapeutic intervention.

Introduction

Leigh syndrome (LSS) (OMIM # 256000) is the most common clinical pediatric presentation of primary mitochondrial disease. The diagnosis of LSS is made clinically based on a constellation of symptoms, neuroimaging or neuropathology findings demonstrating brain stem and/or basal ganglia involvement, and elevated lactate concentration in blood and/or cerebrospinal fluid ^1,2. Revised diagnostic criteria proposed in 2014 allowed for the inclusion of patients with normal lactate levels when a molecular etiology supporting impairment of mitochondrial energy production was identified ^3,4. Also known as subacute necrotizing encephalomyopathy, classic LSS is a severe and progressive disorder with regression typically occurring by age two years ⁵. Historic reports estimate mortality of 35%−50% within the first three years of life ^1,6,7. While current standard of care management provides supportive therapy in both chronic and acute settings, no FDA-approved therapies or cures currently exist ^4,8–12. Indeed, developmental disability was prioritized both as one of the top disabling symptoms for pediatric mitochondrial disease patients and as one of the top study outcomes to motivate their participation in clinical trials ¹³.

Widespread use of massively parallel sequencing, together with revised diagnostic criteria for LSS, has broadened the phenotypic spectrum and associated neurologic outcomes to encompass a broader spectrum of disease than previously appreciated ^{3,14 6,15–18 19,20}. Indeed, it is now recognized that LSS is genetically heterogenous and may result from pathogenic variants in over 110 genes located both in mitochondrial DNA (mtDNA) and nuclear DNA (nDNA) genomes ²¹. An international ClinGen-approved mitochondrial disease expert panel has been established that curated published literature to support the strength of pathogenicity for each gene associated with LSS ²². Most LSS-associated genes encode proteins directly involved in mitochondrial energy production ¹⁴.

Classically, LSS disorders present with sudden onset of neurologic regression in the first year of life followed by resultant developmental delay. It is unknown if primary developmental delay (pDD) may precede the first episode of regression ³. Regression often results from episodic metabolic decompensation that induce metabolic strokes. Resultant symptoms may include loss of developmental skills, central respiratory failure, oculomotor abnormalities (e.g. nystagmus and ophthalmoparesis), ataxia, dystonia, and dysarthria ⁶. Additional multi-systemic involvement may occur and result in significant morbidity, although great variability may be seen between different genetic etiologies of LSS disorders. Individual case reports describe milder adult cases, further expanding the phenotypic spectrum of LSS ^23,24. Here, we retrospectively investigated the pre-regression developmental history of LSS to attempt to identify prodromal signs that could result in earlier diagnosis in a single-site cohort of 69 subjects with molecularly-confirmed LSS disease.

Methods

Patient Ascertainment and Enrollment

This retrospective study was conducted under Children’s Hospital of Philadelphia (CHOP) IRB#08-6177 (PI: MJF) of mitochondrial disease subjects who lived across 3 countries, had been clinically evaluated in the CHOP Mitochondrial Medicine Frontier Program, and consented to research data analysis. Electronic medical records for 78 subjects were reviewed. Inclusion criteria for study analysis were: (1) neurologic symptoms with neuroimaging findings consistent with the diagnosis of LSS disorders including subacute necrotic lesions in the thalamus, brainstem, and/or posterior spinal column; (2) metabolic evidence of mitochondrial dysfunction on blood, urine, and/or tissue testing; (3) molecular confirmation of pathogenic variants in genes associated with LSS; (4) and sufficient information available on early development. LSS subjects with POLG disease or overlapping features of either Mitochondrial Encephalomyopathy, Lactic acidosis, and Stroke-like episodes (MELAS) (OMIM # 540000) or Neuropathy, ataxia, and retinitis pigmentosa (NARP) (OMIM # 551500) were included if subjects met the neuroimaging inclusion criteria. LSS subjects who harbored single large-scale mtDNA deletions (SLSMD) were excluded because of the clinically distinct, multi-systemic phenotypes seen in Pearson syndrome or Kearns-Sayre syndrome relative to other LSS disorders (N=5) ²⁵. Ultimately, 69 individuals met full criteria for study inclusion. Four of the 69 LSS subjects studied were deceased at the time of last medical records collection, with a median age at death of 4.2 years. The median age at last contact was 10.9 years (IQR 4.9–14.7).

Data extraction and database platforms

Data were extracted directly from the electronic medical record (EPIC, Madison, WI, USA) using a custom CHOP ‘MMFP-Tableau’ pipeline based on curating and modelling data exported in Clarity into an Alteryx server (Irvine, CA) and visualizing data via Tableau version 9.1 (Seattle, WA, USA). Data were confirmed and supplemented by direct electronic medical record review. The diagnosis of developmental delay and the age at developmental milestone acquisition was collected via retrospective chart review of clinically-obtained developmental histories. Epilepsy classifications were derived from the clinical documentation. Data was collected in Research Electronic Data Capture (REDCap) ²⁶. SPSS (Windows version 23.0; SPSS Inc., Chicago, IL, USA), and Microsoft Excel 2019 (Microsoft Cooperation Inc, Seattle, WA, USA). GraphPad Prism Software Version 9 (San Diego, California, USA) was used for statistical analyses and graphical representations.

Ages at attainment and/or loss of common developmental milestones were collected for the following: head control, rolling over, sitting without support, independent ambulation, bringing hands together in midline, reaching for objects, pincer grip, social smile, non-specific speech (mature babbling), single word speech, two-word phrases, and six-word speech. Based on the availability of specific milestone data in the retrospective medical records, not all milestones were available across the full population; the number of individuals for whom information was available is noted for each parameter. When age of acquisition was noted to be “normal” in the medical records, the 50^th percentile values (p50) from the Denver Developmental Screening Test II (DDST-II) normative data were used, and when the specific age at which a skill was acquired was unknown but noted as “delayed” a value of p90 plus 2 standard deviations was used ²⁷. The ages at first presentation to health care with signs or symptoms related to LSS (e.g., developmental concerns, failure to thrive, hypotonia), age at MRI radiological diagnosis, and the age at genetic diagnosis were collected. The age of onset of developmental delay was defined as noted in the medical records or at the age at which the 90^th percentile plus two standard deviations was exceeded. Transient neurologic symptoms were noted when a child experienced temporary, non-progressive loss of neurologic function not requiring hospitalization. Neurologic regression was defined as a persistent loss of neurologic function or a loss of neurologic function that required hospitalization. For the list of variable definitions used for data extraction, see Supplemental Table 1. Highest educational status attained was recorded as the least restrictive educational environment attended from kindergarten to 12^th grade based on the medical records: children with no individualized educational plan (IEP) support were documented as “mainstream” (even if physical accommodations were needed for mobility, etc.). Children who spent most of the day in a typical classroom with IEP supports (e.g., increased test-taking time, one-on-one) were classified as “mainstream with support”, while children who spent the majority of the day in a special education classroom were documented as “non-mainstream”. Modified Rankin scores for neurologic severity were retrospectively assigned to the last neurologic encounter, where a score of 1 represents minimal neurologic impairment and 6 represents death.

Statistical Analysis

Descriptive analyses were conducted using SPSS for Windows, version 24 (IBM, Armonk, NY, USA) and GraphPad Prism Software Version 9 (San Diego, California, USA). Mean with standard deviation (SD) or median with interquartile range (IQR) was provided for continuous variables, and percentage and frequency for categorical variables. Wilcoxon rank sum test was used to compare continuous variables between groups since data were not normally distributed. Pearson’s chi-square test or Fisher’s exact test was used to compare categorical variables, as appropriate. Kaplan-Meier curves were used to represent the gain and loss of developmental milestones. Log-rank test was used to compare the age at acquisition of developmental milestones between subjects who were enrolled in mainstream education and those who were not. Logistic regression was applied to compare the likelihood of having an outcome of mainstream education placement between groups. Two-sided tests were used with a p-value < 0.05 as the criterion for statistical significance.

Results

The LSS cohort of individuals who met inclusion criteria for this study (n=69; 66 kindreds) represented a diverse genetic etiology and developmental spectrum (Table 1–2). Subjects came from a geographically diverse region. Four of the 69 individuals (5.7%) were deceased at the time of last medical records collection (median age at death of 4.0 years with IQR 2.5–10.5). Among all individuals, the median age at last contact was 10.9 years (IQR 4.9–14.7). This LSS cohort was genetically heterogeneous, with pathogenic variants in 34 different causal genes identified. SURF1 (N=9) was the most common causal nDNA gene. MT-ATP6 (N=11) was the most common causal mtDNA gene (Table 1). Of the 69 subjects, the causative gene was nuclear in 55% (N=38) and in the mtDNA genome in 45% (N=31) individuals.

Table 1.

Descriptive analysis of the cohort

Cohort Characteristics
Sex (N,%)	Female (35, 51%) Male (34, 49%)
Clinical Classifications (N)	Leigh (58, 84%) Leigh/MELAS (7, 10%) Leigh/NARP (4, 6%)
Genetic classification	nDNA (n)				mtDNA (n)
	SURF1	9	NUBPL	1	MT-ATP6	11
	MRPS34	3	NDUFS8	1	MT-ND3	7
	FBXL4	3	NDUFS4	1	MT-ND5	4
	PDHA1	2	MTFMT	1	MT-ND4	3
	USMG5	1	MRPL3	1	MT-ND6	2
	TTC19	1	HIBCH	1	MT-ATP8	1
	SUCLG1	1	GTPBP3	1	MT-ND1	1
	SLC19A5	1	ECSH1	1	MT-ND2	1
	SDHB	1	EARS2	1	MT-TL1	1
	POLG	1	DLD	1
	PDHB	1	C12ORF65	1
	PDHA1	1	ATP50	1
	PARS2	1

Table 2. Neurologic outcomes in Leigh Syndrome.

Only genotypes with 3 or more subjects are represented.

	Neurologic Trajectory			Age at last medical encounter			Age at presentation (years)			Age at onset of primary developmental delay (years)			Age at first neurologic regression (years)			Age at genetic diagnosis (years)			Age at radiographic diagnosis (years)			Best educational status (n)			Modified Rankin at last encounter (n)
	pDD (n)	pDD + R (n)	R only (n)	Ave	Min	Max	Ave	Min	Max	Ave	Min	Max	Ave	Min	Max	Ave	Min	Max	Ave	Min	Max	MS	MS + Support	non-MS	Mild-Mod (0–3)	Mod-Sev (4–6)
nDNA	11	14	13	11.1	0.5	48.7	2.5	0.0	24.0	0.7	0.0	1.7	5.8	0.2	24.0	7.1	0.2	42.0	3.7	0.0	25.3	6	8	11	13	25
SURF1		4	5	17.0	2.8	48.7	6.0	0.2	24.0	0.9	0.5	1.5	7.9	1.4	24.0	12.0	1.5	42.0	8.2	1.1	25.3	3	1	2	4	5
FBXL4	1	2		7.8	3.2	16.0	0.0	0.0	0.0	0.3	0.0	0.5	3.3	3.3	3.3	3.1	0.2	8.9	1.8	0.6	2.6			1		3
MRPS34	1	2		16.2	5.8	31.9	0.7	0.4	1.1	0.7	0.4	1.1	9.8	1.7	18.0	9.3	1.8	20.0	2.4	1.6	4.0		1	2		3
mtDNA	5	17	9	12.4	2.0	37.9	1.4	0.0	7.5	0.9	0.3	3.0	4.6	0.4	17.7	5.7	0.6	15.0	4.3	0.5	15.6	6	5	14	14	17
MT-ATP6	2	8	1	12.9	2.0	37.9	0.4	0.0	0.8	0.6	0.4	0.8	4.5	0.5	17.7	5.6	0.6	14.8	3.2	0.5	7.3		1	6	5	6
MT-ND3	2	2	3	13.6	6.5	22.8	2.6	0.4	7.5	1.0	0.4	1.6	7.3	5.0	12.4	7.4	3.3	8.7	7.5	3.4	15.6	3		3	5	2
MT-ND5		3	1	9.8	4.4	13.2	1.1	0.4	2.0	1.4	0.7	2.0	2.8	0.4	4.1	2.7	1.3	4.3	2.1	1.2	4.2		1	2	1	3
MT-ND4	1	2		16.1	14.3	18.5	0.7	0.2	1.5	0.8	0.3	1.5	5.6	2.3	9.0	6.1	2.0	10.8	5.3	1.0	9.5	1		2	1	2

pDD: primary developmental delay; R: Regression; MS: mainstream education

Early clinic features and age at first neurologic regression were identified. The most common presenting features were primary developmental delay (pDD, N=28) and feeding concerns (N=13). Six individuals first presented with ocular concerns, including nystagmus and dysconjugate gaze. Four subjects had clinical presentations of neurologic tone abnormalities, neonatal acidosis (N=2), or epilepsy (N=1). Nine individuals presented with isolated neurologic regression. Clinical presentation noted among all individuals occurred at a median age of 0.5 years (IQR 0.2–1.4). In our cohort 6% (N=4) presented with their first neurological symptom of LSS after the age of 10. Of the 4 individuals with later-onset LSS, 2 were adults at time of the first symptoms (defined as onset above the age of 18). These subjects were brothers, both of whom presented with ocular complaints. Their molecular cause was a homozygous SURF1 pathogenic variant (C.845_846delCT; P.Ser282Cysfs). The molecular cause of LSS in the two other late-onset cases was PARS2 (N=1) and PDHB (N=1). Median age of onset of developmental delay was 0.50 years (IQR 0.4–1.3). Most individuals (N=53/69, 76.8%) were noted to have experienced an episode of neurodevelopmental regression, with this first neurologic decline having occurred at a median age of 3.5 years (IQR 1.3–6.2). Twenty-two of 63 individuals (31.9%) for whom the information was available also experienced periods of transient neurologic symptoms, of whom 14 (63.6%) went on to later experience a neurologic regression. Seizures were a common clinical complication noted in 27 of 69 individuals (39%). Among those affected by seizures, the median age at onset was 4.0 years of age (range 0.25 – 35; IQR 0.83–9.00). The most common initial epilepsy subclassification was complex partial seizures (N=9), although many experienced evolving semiologies, with 16 individuals developing multiple seizure types. Among the children affected by infantile spasms (N=8), 5 were from the MT-ATP6 subcohort, of which 4/5 were noted to have harbor the pathogenic variant m.8993T>G.

A history of primary developmental delay was associated with an earlier age at onset. Among the individuals who exhibited pDD prior to regression (N=31, 44.9%) or who had pDD with no history of regression (N=16, 23.2%), the median age at clinical presentation was 0.5 years (IQR 0.2–0.8) versus 1.3 years (IQR 0.5–6.7) in those without a history of pDD (N=22, 31.9%) (Figure 1A; Mann Whitney test two-tail p = 0.0003). Among the subcohort of individuals who experienced a neurodevelopmental regression within the period of data collection (N=53), pDD preceded the first clinical regression in 58.3% (N=31). While age at regression was not associated with the presence of pDD (Mann Whitney test two-tail p = 0.21; Figure 1B), younger age at first neurologic symptom was significantly associated with the presence of pDD (Mann Whitney test two-tail p = 0.0003) (Figure 1C). No difference was seen between the age at last data collection of subjects who did or did not experience neurodevelopmental regression (Mann Whitney test two-tail p = 0.13; Figure 1D).

Figure 1.

Primary developmental delay (pDD) and neurologic regression in Leigh Syndrome.

(A) Forty-seven individuals had a history of developmental delay with a known age at presentation (pDD; noted as +), and 22 individuals had no history of DD prior to regression (−). The age at presentation was compared between the cohorts (Mann Whitney test, 2 tail P value 0.0003). Grey line represents the median with bars as 25^th–75^th quartile range.

(B) Of the individuals with a history of regression, 31 had a history of prior DD (pDD) versus 22 individuals with no history abnormal development. The age at first regression was compared between the cohorts (Mann Whitney test, 2 tail p = 0.30). Grey line represents the median with bars as 25^th–75^th quartile range.

(C) The age at first neurologic feature (developmental delay, neurologic fluctuations, or regression) was compared between the cohort of individuals who were noted to have prior DD versus those who presented first with regression (+pDD n=47; −pDD n=22; Mann Whitney test, 2 tail p = 0.0003). Grey line represents the median with bars as 25^th–75^th quartile range.

(D) The age at last available medical record was compared between those who did or did not experience regression. There was no difference in the ages (Mann Whitney test, 2 tail P = 0.13). Grey line represents the median with bars as 25^th–75^th quartile range.

(E) Detailed chronologic information was available from 69 individuals. The age at last data collection is shown by grey bar with age at pDD onset (open square), first neurologic regression (X), genetic diagnosis (open circle), and radiographic diagnosis (closed triangle) overlayed on top. (D) Detailed chronologic information was available for 67 subjects with LS disorder. Duration of data collection is shown by the grey bar, overlayed with age at pDD onset (circle), first neurologic regression (open square), genetic diagnosis (open triangle), and radiographic diagnosis (X).

Subject-level timelines were created to reflect the onset of neurologic symptoms (pDD and regression), clinical diagnosis by MRI and genetic etiology, and duration of medical records collection available (Figure 1E). The first diagnostic MRI was obtained at a median age of 2.2 years (IQR 1.2–4.6). A definitive molecular genetic diagnosis of LSS was reached at a median age of 4.3 years (IQR 1.6–8.8), which represented a diagnostic delaye by a median of 2.3 years (IQR 0.74–7.3) from initial presentation. The median duration of time between radiographic and genetic diagnoses of LSS was 0.52 years (IQR 0.1–4.5).

To characterize the role of genotype in the neurologic trajectory, the LSS cohort was further divided by genome location of causal gene pathogenic variants: nDNA versus mtDNA. There was no significant difference between these groups in age at overall presentation, age at onset of developmental delay, or age at regression by variant location (Figure 2).

Figure 2.

Comparison of neurologic events by location of variants in Leigh Syndrome.

(A) The proportion of children with pDD by variant location as presented by bar graph (dark grey = history of pDD; light grey = no history of pDD).

(B–D) Comparison of age at pDD, overall presentation, and first regression by variant location (Mann Whitney test, 2 tailed p > 0.05). Grey line represents the median with bars as 25^th–75^th quartile range.

Ages at acquisition of each of the developmental milestones were collected from the medical records. The LSS cohort was divided based on peak educational attainment. Compared to LSS subjects who participated in mainstream education, subjects with non-mainstream education experienced significant early delays in key milestones [acquisition of head control (Mantel-Cox log rank p = 0.018), social smile (p= 0.0068), sitting (p = 0.0009), independent ambulation (p = 0.0001), single word communication (p = 0.0096), and six-word vocabulary (p = 0.0020)] (Figure 3). No significant differences between these subcohorts were seen in the ages of acquisition of rolling (p = 0.22) or reach (p = 0.19; Figure 3). Subjects who experienced neurodevelopmental regression without pDD were more likely to be receive mainstream education compared to those who experienced pDD, irrespective of later regression (Odds ratio = 22.14; 95% CI 3.84 to 108.6) (Figure 3c). Modified Rankin scores were assigned respectively to the last clinical encounter (Figure 3D). There were no statistical differences based on Rankin outcomes and history of developmental delay or regression.

Figure 3.

Neurologic outcomes in Leigh Syndrome.

A–B. The age at acquisition of early developmental milestones was compared by maximum education status using the log-rank (Mantel-Cox) test. Children with non-mainstream educational status were delayed in acquisition of all milestones except rolling and reaching (2 tailed-p value <0.05 for all other milestones).

C. For children 5 years or older at the last clinical encounter, the best educational status was collected and compared across neurologic history (pDD alone, pDD followed by a regression, and regression only). Educational outcomes were categorized as mainstream (black), mainstream with supports (light grey), and learning support classes (white).

D. A Modified Rankin score was assigned to the last medical encounter and compared across neurologic history categories.

Discussion

We performed detailed analyses of the spectrum of early neurodevelopmental trajectories in a large cohort of 69 individuals with molecularly- and radiographically-confirmed LSS. This work provides novel insights into the neurodevelopmental phenotypic spectrum of LSS by delineating three developmental subgroups: pDD followed by regression (45%), isolated pDD (23%), and isolated regression without preceding pDD (32%). Early onset of clinical disease symptoms in LSS disorders was previously recognized to be associated with poor outcome ^6,18. Our data specifically demonstrate that the occurrence of pDD, regardless of subsequent neurodevelopmental regression, predicts worse educational attainment (non-mainstream as opposed to mainstream). pDD was also associated with earlier onset of disease, defined at age of first presenting symptom (including lactic acidosis, neurologic disease, cardiac disease, ophthalmic disease, etc.) In many of these cases, developmental delay itself was the presenting symptom, which suggests that mitochondrial energetics impact developmental trajectories prior to acute metabolic failure and regression, providing an opportunity for earlier diagnosis and/or therapeutic intervention.

The most common neurologic phenotype in this cohort was a history of pDD preceding regression. The finding of early pDD in the majority of our cohort (68%) critically expands understanding of the developmental history of LSS. This also has important pathophysiological implications, suggesting that disrupted mitochondrial bioenergetics in LSS disorders play a key role in early pediatric neurodevelopmental trajectories that is independent of subsequent regression. This finding is also important clinically as diagnostic delay was seen in this LSS cohort between clinical symptom onset and molecular genetic confirmation. This prolonged delay may often be stressful to families, result in a costly diagnostic odyssey, and may be a barrier to medical management ^28–30. Identification of pDD as one of the earliest features of LSS allows the opportunity for earlier genetic diagnosis and medical interventions, including mitochondrial medicines, acute stress prevention, metabolic stroke recognition and earlier therapeutic intervention, and clinical trial participation^11,12.

As diagnostic MRIs were often obtained years after symptom onset, it is unknown whether these same radiographic changes would have been present at disease onset. Within this cohort, some individuals had initially non-diagnostic MRIs. Future investigations will be required to better understand the relationship and evolution of early neuroradiographic features to neurodevelopmental trajectories and clinical outcomes in LSS disorders ². Recognizing the occurrence of early pDD in LSS subjects may also potentially lead to earlier genetic diagnostic testing and accurate diagnosis of LSS disorders.

In addition to its diagnostic importance, early pDD had significant prognostic implications. The presence of early developmental delay was associated with a need for greater educational support. Delayed age at attainment of individual early milestones was associated with later non-mainstream educational enrollment. This was true of LSS patients with pDD regardless of whether they experienced a subsequent neuroregression episode. Interestingly, no clear association was seen between neurologic phenotype and Rankin scale outcome. Whereas the Rankin scale is a tool for measuring neurologic outcomes after stroke that focuses on gross motor function, educational status was defined here based on cognitive function. Future prospective studies are needed to better characterize neurologic outcomes and identify variables associated with specific cognitive and motor outcomes.

This cohort is distinct from previously published LSS cases in several core ways. Our cohort includes a significant subset who had pDD without any subsequent regression (n=16; 23%), which is a newly recognized developmental subpopulation of individuals affected by LSS. While this could be potentially attributed to ascertainment bias, there was no difference ween between phenotypes throughout the duration of medical records available. As more individuals are diagnosed with LSS with isolated pDD and followed prospectively, we will be able to better understand the rate of neurodevelopmental regression in the pDD cohort as well as precipitating or preventative factors. Our cohort is also more inclusive of a broad range of ages compared to previously published reports. While cases of adult-onset LSS have been reported previously in the literature ^23,24, our cohort includes an additional 4 individuals (6%) who presented with their first symptom after the age of 10.

By using all available medical records and a retrospective study design, early events, such as developmental milestones, were able to be extracted, irrespective of the timing of diagnosis. To reduce reporting bias, the earliest medical records with specific milestone information were used. To reduce the potential for exclusion-related biases, such as when a specific age was not available, but a milestone was noted to be ‘on time’ or ‘delayed’, imputed values based on normative data of acquisition of milestones were used. Interestingly, no differences in developmental trajectory were seen based on genetic etiology occurring in the nuclear genome (N= 38) vs mtDNA genome (N= 31).This study was not sufficiently powered to perform detailed genotype-specific analyses given the breadth of causal genotypes now recognized for LSS disorders. Observed trends, such as the increased rate of infantile spasms associated with LSS secondary to the pathogenic m.8993T>G variant in MT-ATP6, will need to be further studied. While this study is based on care provided at a single center, the subjects enrolled in this study represents a broad national and international geographic distribution, and these findings represent an expansion of the LSS phenotype resulting from improved utilization of broadscale next-generation sequencing diagnostic technologies, such as whole mtDNA genome sequencing and whole exome sequencing. By contrast, we hypothesize that non-classical, milder, and/or later-onset LSS cases may be underrepresented in historical published LSS cohorts due to bias in single gene or targeted panel-based diagnostic testing patterns ³¹. As symptomatic care and increasingly broader molecular diagnostic testing technologies become clinical standards, our understanding of the full neurodevelopmental phenotypic spectrum of LSS is likely to continue to expand.

Collectively, these data demonstrate that LSS disorders should be considered in the differential diagnosis of early developmental delay, regardless of an individual having sustained a neurodevelopmental regression episode. Indeed, the common occurrence of pDD that we identified to occur in LSS cases creates an opportunity to identify, potentially intervene, and/or prevent subsequent early life neurodevelopmental regression and its associated high morbidity and mortality ^7,11,12. Overall, this study creates the foundation for future prospective natural history studies to further discern specific neurodevelopmental outcomes and inform the design of future clinical trials to evaluate therapies that improve neurodevelopmental outcomes in LSS disorders.

Key Points.

Leigh syndrome spectrum (LSS) disorders have diverse genetic etiologies and variable occurrence of primary developmental delay (pDD) preceding neurodevelopmental regression.

Most individuals presented with signs and symptoms of mitochondrial dysfunction prior to regression.

A history of pDD was associated with worse educational outcomes in LSS subjects and an earlier onset of neurological symptoms.

Data availability statement:

deidentified data is available upon request.