Introduction

Deficiency of beta-ketothiolase (alternate name: mitochondrial acetoacetyl-CoA thiolase [T2] deficiency; EC 2.3.1.9) (Online Mendelian Inheritance in Man [OMIM] 203750, 607809) is an inherited disease of isoleucine and ketone body metabolism caused by mutations in the ACAT1 gene (Fukao et al 2014; Hori et al 2015). Since the first report of T2 deficiency by Daum et al (1971), more than 100 patients with this disorder have been identified worldwide, with no ethnic predisposition (Abdelkreem et al 2016). T2 deficiency is clinically characterized by intermittent ketoacidotic episodes of variable severity; no symptoms exist under stable patient conditions, except for possible consequences of a previous episode. Characteristic laboratory findings include significant ketonuria and increased urinary excretion of isoleucine catabolic intermediates 2-methyl-3-hydroxybutyrate (2M3HB), 2-methylacetoacetate (2MAA), and tiglylglycine (TIG) (Fukao et al 2014). Blood acylcarnitine analysis commonly reveals elevated levels of C5:1 and C5-OH carnitines. However, atypical clinical and laboratory manifestations have also been reported (Abdelkreem et al 2016).

Current knowledge about T2 deficiency rests mainly on published individual case reports; only a few studies have described the collective features of patient groups. The two largest published studies depend on either a review of the literature (Søvik 1993) or questionnaires returned from supervising physicians (Fukao et al 2001). From 2005 to 2016, we diagnosed a total of 41 Vietnamese patients with T2 deficiency. To our knowledge, this is the largest number of patients from a single country or ethnic group investigated thus far.

In this work, we report the clinical phenotypes, biochemical and molecular characteristics, and outcomes of 41 Vietnamese patients with T2 deficiency. The direct management and long-term follow-up of a large number of patients enabled us to study the natural history of this rare disease. Investigation of haplotypes linked to the most commonly reported ACAT1 mutation among patients provided evidence for a founder effect as a plausible explanation of the genetic epidemiology of T2 deficiency in Vietnam.

Participants, materials, and methods

Participants

We included all (41) patients with T2 deficiency identified at Vietnam National Children’s Hospital (Hanoi) during the period between 2005 and 2016. Eight patients had been previously reported (Fukao et al 2010). Thirty unrelated healthy participants of Kinh ancestry served as controls for haplotype investigation and modeling of linkage disequilibrium (LD) decay over generations. This study was approved by the Ethical Committee of Vietnam National Children’s Hospital. Informed consent was obtained from all adult participants and the parents of minors included in the study.

Patient data were retrospectively collected from medical records at Vietnam National Children’s Hospital. These included clinical and demographic data reported by physicians on initial presentation and at each follow-up visit (every 6–12 months). An assessment and evaluation of patients’ growth and development were conducted using the World Health Organization child growth standards (2006 release) (http://www.cdc.gov/growthcharts/who_charts.htm) and the Denver II developmental screening test (http://denverii.com), respectively. Routine laboratory tests, including blood counts, gases, electrolytes, glucose, ammonia, and urinary ketones levels were performed according to standard methods. Brain imaging was obtained when clinically indicated. Urinary organic acid and blood acylcarnitine analyses for 37 patients were carried out at Shimane University (Izumo, Japan) using dried urine and blood samples absorbed on filter paper, as previously described (Fu et al 2000). Fresh urine samples from seven patients were analyzed for organic acids at Vietnam National Children’s Hospital, according to standard methods.

Genotyping

Thirty-two patients from 27 different families were screened for ACAT1 mutations at Gifu University (Gifu, Japan) using DNA samples purified from patients’ blood, as previously described (Fukao et al 2010). Identified mutations are described according to the latest recommendations, HGVS version 15.11 (http://varnomen.hgvs.org/), (den Dunnen et al 2016), using ACAT1 NCBI Reference Sequences: NM_000019.3 and NG_009888.1. Five dinucleotide (D11S1325, D11S1781, D11S1343, D11S4206, and D11S1793) and one tetranucleotide (D11S1391) short tandem repeats (STRs) were genotyped by PCR amplification and capillary electrophoresis at the Center for the Study of Rare Hereditary Diseases, Niguarda Ca’ Granda Metropolitan Hospital (Milan, Italy). DNA samples for haplotype investigation and LD analysis were obtained from 16 patients, their parents, and 30 controls. Details of LD analysis and methods for estimation of the age of mutation are provided in Supplementary file 1.

Results

Table 1 summarizes the demographic, clinical, and laboratory features of patients included in our study. Further detailed data on patients are reported (Supplementary files 2 and 3).

Table 1 Summary of demographic, clinical, and laboratory findings of 41 Vietnamese patients with beta-ketothiolase deficiency
Full size table

Demographic data

All patients (23 males and 18 females) were born between 2002 and 2016. Most patients (N = 35) resided in northern Vietnam; accordingly, the estimated incidence of T2 deficiency is one in 190,000 newborns in this area. All patients were born to nonconsanguineous parents of Kinh ancestry, which is the main Vietnamese ethnic group (86% of population). Patients belonged to 34 different Vietnamese families; seven participants are sib pairs. Two unrelated families had a history of unexplained sibling death.

Clinical data of first ketoacidotic episode

Three of 41 patients were diagnosed through screening the siblings of patients; to date, two siblings have been asymptomatic. In 34 of the 39 symptomatic patients, the age at onset of the first ketoacidotic episode was between 6 and 18 months (Fig. 1). Ketoacidotic episodes were associated with an intercurrent illness (often gastroenteritis and/or respiratory infections) in 38 of 39 patients; in one case, excessive protein intake on the day preceding the episode was reported. The clinical status on admission was generally severe. Nearly all patients were dehydrated and suffered from major central nervous system (CNS) manifestations (altered mental status, convulsions); hypotension was evident in 10% of patients (Table 1).

Fig. 1
figure 1

Ketoacidotic episodes by age among 41 Vietnamese patients with beta-ketothiolase deficiency. Vertical axis indicates patients (females are underlined) and horizontal axis indicates age in years. Sib pairs are denoted by b (brother) and s (sister) following patient codes. Black and white circles depict first and recurrent ketoacidotic episodes, respectively. Solid lines show the duration of follow-up, from diagnosis through age at last follow-up. Patient outcomes were favorable unless specified otherwise, as follows: C, complication (neurological); D, died; M, missing. GK number is an internal identifier for patients with beta-ketothiolase deficiency whose ACAT1 mutations were identified at Gifu University (Gifu, Japan)

Full size image

Laboratory data of first ketoacidotic episode

Severe high anion gap metabolic acidosis (pH: 6.8–7.26; HCO3 −: 1–10.7 mmol/L; base deficit: 17–30 mmol/L) and ketonuria were detected during acute ketoacidotic episodes. Blood glucose levels were normal in 23 cases whereas hyperglycemia (>7.8 mmol/L) and hypoglycemia (<2.5 mmol/L) were observed in 11 and three cases, respectively. Mild hyperammonemia (>170 μmol/L) was present in seven cases. Patients had no significant electrolyte disturbances except for two patients who experienced hypokalemia. Brain imaging abnormalities were identified in four of 28 cases who underwent brain imaging. Magnetic resonance imaging in two patients showed abnormalities affecting the basal ganglia (hyperintensity on T2W and hypointensity on T1W) and cerebral cortex (cerebral atrophy, hyperintensity on T2W, and hypointensity on T1W). Computed tomography in the other two patients showed bilateral hypodensities affecting posterior limb of the internal capsule. Of these four patients, one died during his second acute episode, and the other three patients developed clinically apparent neurodevelopmental impairments (Table 1).

Except for one case, all patients who underwent urinary organic acid and/or acylcarnitine analyses showed typical abnormalities; an increase in 2MAA levels could only be detected using fresh urine samples. Interestingly, patient no. 38 (GK118s) showed normal urinary organic acid and acylcarnitine profiles analyzed under stable conditions. Eight ACAT1 mutations were identified in 32 patients from 27 different families. The c.622C>T (p.Arg208*) mutation was the most common, representing 66% of all mutant alleles, followed by c.1006-1G>C (p.Val336fs), which accounted for 19% of identified mutations (Fig. 2). Details of patients’ mutations are reported in Supplementary file 2.

Fig. 2
figure 2

Relative frequency of ACAT1 mutations identified in 32 Vietnamese patients with beta-ketothiolase deficiency. Mutations are described according to the latest HGVS recommendations (version 15.11) using ACAT1 NCBI Reference Sequences: NM_000019.3 and NG_009888.1

Full size image

Management

Acute management for ketoacidotic episodes and long-term preventive measures were implemented, as previously described (Hori et al 2015). Patients received intravenous glucose during acute ketoacidosis for correcting hypoglycemia and suppressing ketogenesis; target blood glucose level was the upper limit of normal. Clinical and laboratory parameters of patients dictated individualized infusion of appropriate fluids and electrolytes. Ketoacidosis gradually improved to sufficient glucose infusion; other measures (e.g., sodium bicarbonate, dialysis) were used for exceptional cases with severe acidosis. Most patients required intensive care management during acute episodes; of the 39 symptomatic patients, mechanical ventilation and hemodialysis were required for 16 and four of them, respectively. Avoiding long fasting was the mainstay for preventing severe ketoacidotic episode. Patients were advised to have frequent carbohydrate-rich meals, particularly during ketogenic stresses such as infections; glucose infusion was given when an intercurrent illness precluded oral intake. Home monitoring of urinary ketones using test strips was useful. Patients were advised to avoid excess fat and protein intake.

Follow-up and outcome

Figure 1 outlines patient ketoacidotic episodic events and outcomes during the course of the study. Sixteen of 37 patients had recurrent ketoacidotic episodes (recurrent episodes: N = 25; frequency: 1–5 times per patient). A total of five patients died: two during their first episode of ketoacidosis, two during their second episode, and one patient died owing to neurological sequelae following the first episode. Patients’ growth parameters were appropriate to age; however, significant neurodevelopmental impairment was observed in two cases.

Age of the founder mutation

The occurrence of the same ACAT1 mutation (c.622C>T) in 26 T2-deficient patients belonging to 22 independent families of Kinh ancestry is striking. This raised the question of whether unrelated identical mutations have been transmitted through these families or whether they share a common ancestor who introduced the mutation into the population by a founder effect. We addressed this intriguing question by performing haplotype analysis using 6 STR polymorphisms and the c.622C>T mutation. The presence of an ancestral seven-marker haplotype (3-4-2-T-5-2-6) and a number of related haplotypes derived from recombination events in the mutation-bearing chromosomes are compatible with the hypothesis that the pathogenic variant c.622C>T can be traced to an ancient common founder belonging to the Kinh ethnic population.

The mean (± SD) overall age estimate (in generations [g]) for the most recent common ancestor (MRCA) of the ACAT1 c.622C>T mutation is 88.7 ± 17.2 g (95% CI: 76.8–100.6 g) using the algorithm of Risch et al (1995), and 83.7 ± 16.8 g (95% CI: 75.7–91.7 g) by the method of Reich and Goldstein (1999) (Supplementary file 4). From a Bayesian perspective, applying a mean population growth rate of 0.07, a Markov chain Monte Carlo algorithm (Rannala and Reeve 2001) yielded an age estimate of 96.4 g (95% CI: 68.5–183.1 g).

Discussion

The incidence of T2 deficiency has been estimated to be less than one per million newborns worldwide. Ongoing identification of an increasing number of cases, particularly as a consequence of extended newborn screening programs, has revealed a higher prevalence in some geographic areas, such as one in 232,000 newborns in Minnesota, USA (Sarafoglou et al 2011). Based on our study, the estimated incidence of T2 deficiency could be one in 190,000 newborns in northern Vietnam. This may be underestimated because undiagnosed asymptomatic, mildly symptomatic, or deceased cases may have been missed (Abdelkreem et al 2016). The vast majority of T2-deficient families were outbred, which further supports the hypothesis of higher incidence. These data support newborn screening in Vietnam for T2 deficiency.

As has been previously reported (Fukao et al 2001; Søvik 1993), the age at onset of ketoacidotic episodes in Vietnamese T2-deficient patients usually ranges between 6 and 18 months of age. Because mitochondrial medium-chain 3-ketoacyl-CoA thiolase (T1) can partially compensate for T2 deficiency in ketone body utilization, ketogenic triggers such as prolonged fasting, infection, and protein-rich foods play an important role in disrupting this compensation and precipitating acute ketoacidosis (Fukao et al 2014). Such triggers are less likely to be present in the first 3–6 months after birth because of frequent feeding, protective maternally acquired immunoglobulins, and the relatively lower protein content of breast and ordinary milk formulas. On the other hand, older children develop ketosis slower than younger children (Bonnefont et al 1990) thanks to decreased energy demands relative to body weight coupled with increased muscle mass, which provides a protein reservoir for gluconeogenesis (Fukao et al 2014). Accordingly, ketoacidotic episodes rarely manifest in patients with T2 deficiency during early infancy or after early childhood. In contrast, the frequent neonatal presentation and persistent ketosis in succinyl-CoA:3-oxoacid CoA transferase (SCOT) deficiency, another ketolytic disorder, can be attributed to a lack of any compensatory enzymes for SCOT deficiency in ketone body utilization (Fukao et al 2014).

Blood glucose is generally normal during ketoacidotic episodes in patients with T2 deficiency; a few cases might show hypoglycemia or mild hyperglycemia (0.6–14.1 mmol/L) (Abdelkreem et al 2016). Notwithstanding, 28% of the patients investigated in this study had hyperglycemia during acute episodes; one case had the highest recorded blood glucose level reported thus far in a T2-deficient child (23.3 mmol/L). Significant hyperglycemia associated with ketoacidosis has been documented in certain organic acidurias such as propionic acidemia, methylmalonic acidemia, and isovaleric acidemia, and also in SCOT deficiency (Erdol et al 2016). Hyperglycemia may evolve as a reaction to metabolic stress. Given the different approaches to treatment required, under these circumstances, it is crucial to avoid an initial misdiagnosis of diabetic ketoacidosis (Erdol et al 2016).

Unlike the genetic heterogeneity usually found in patients with T2 deficiency (Hori et al 2015), only two mutations (c.622C>T and c.1006-1G>C) accounted for the vast majority (85%) of ACAT1 mutations identified in our Vietnamese patients. Although all patients were born to nonconsanguineous parents, half were homozygous for c.622C>T. This is consistent with a high carrier rate for c.622C>T among the Vietnamese population. However, previous screening of 400 healthy Vietnamese participants could not identify any heterozygotes for c.622C>T (Fukao et al 2010). Further investigation on a larger scale is warranted. In addition, it is noteworthy that c.622C>T is not a specific Vietnamese ACAT1 mutation; it has been reported in Dutch, American, and Chinese patients as well (Fukao et al 2014; Sarafoglou et al 2011; Wen et al 2016).

In Vietnamese patients of the Kinh ethnic group, we identified an ancestral haplotype spanning 3.4 Mb and encompassing the c.622C>T ACAT1 mutation on chromosome 11q22.3. The decay of LD over generations, modeled using different algorithms, would date the age of the MRCA of the mutation-bearing chromosomes to 1900–2500 years ago, if we assume 25 years per generation. The estimated age is consistent with the historically plausible scenario in which a founder effect originated in northern Vietnam during the last centuries BCE or at the beginning of the CE. The effect may have arisen among a successful group of conquerors or migrants who arrived in the region and underwent demographic expansion in the colonized lands before spreading over the area and intermixing with the pre-existing population or other groups of migrants. This mainly occurred when Âu Việt peoples migrated from southern China to the Red River Delta and mixed with the indigenous Văn Lang people (3rd century BCE) or around 111 BCE when Han troops invaded and annexed the kingdom of Nam Việt. After that, a large number of Chinese people including ordinary people, mandarins, and scholars moved to the conquered lands and settled there (Nguyễn 2012; Taylor 1983).

The unique genetic characterization among T2-deficient Vietnamese patients has a remarkable impact on patients’ biochemical profiles. Almost all patients demonstrated the characteristic increased urinary excretion of isoleucine catabolic intermediates, resulting from the presence of homozygous or compound heterozygous null mutations (c.622C>T and c.1006-1G>C) and the consequent loss of T2 enzyme activity. Because 2MAA is readily decarboxylated non-enzymatically to 2-butanone (Aramaki et al 1991), it could be detected in our patients only by analyzing fresh urine samples. On the other hand, the only patient who showed no increased urinary excretion of 2MAA, 2M3HB, or TIG was homozygous for a mild mutation (c.1A>G) with residual (11%) T2 enzyme activity (Fukao et al 2003). T2-deficient patients with mild mutations may have atypical or subtle biochemical profiles not only under metabolically stable conditions but also during acute episodes of decompensation (Fukao et al 2012). Therefore, a prospective newborn screening program may have a higher yield in Vietnam than in other countries where mild mutations may be more common. Indeed, typical abnormalities were identified in nearly all (38/39) T2 deficient Vietnamese patients who underwent blood acylcarnitine analysis. Nevertheless, it should be pointed out that a normal newborn screening result does not absolutely exclude a diagnosis of T2 deficiency (Fukao et al 2012).

Symptomatic patients in our study manifested relatively severe episodes of ketoacidosis associated with dehydration and CNS dysfunction. The brain lesions identified in four Vietnamese patients after ketoacidotic episodes may be attributed to a potential nonspecific consequence of severe acidosis; this may also complicate mitochondrial respiratory chain disorders and certain organic acidurias (Akella et al 2014; O’Neill et al 2014). Nevertheless, some reported cases of T2 deficiency had brain imaging abnormalities with clinical neurological impairment, even in the absence of a previous episode of ketoacidosis (Ozand et al 1994; Buhaş et al 2013). Accumulated isoleucine catabolic intermediates may have a direct neurotoxic effect even in the absence of frank metabolic decompensation (Buhaş et al 2013; Leipnitz et al 2010). T2 deficiency is a potentially serious disease. In addition to other intensive care measures, mechanical ventilation was required for management of acute ketoacidotic episodes in 41% of patients. Seven of 41 patients had unfavorable outcomes. Most adverse consequences were related to the first ketoacidotic episode, when the diagnosis had not yet been confirmed. This is also quite supportive for a newborn screening program in Vietnam for T2 deficiency. Neither the age at onset nor frequency of ketoacidotic episodes affected the eventual outcome. Vietnamese T2-deficient patients had ketoacidotic episodes of variable onset, severity, and outcome, despite sharing not only a common genotype but also the same ancestry and living conditions. This underscores a lack of genotype–phenotype correlation in T2 deficiency. Appropriate acute and preventive management appear to be critical for a favorable outcome (Hori et al 2015; Fukao et al 2001).

In conclusion, T2 deficiency likely has a high incidence in Vietnam where c.622C>T mutation seems to have an ancient common founder of Kinh ancestry. This disease has variable clinical phenotypes even when patients share a common genotype and ancestry, as well as similar personal histories and daily life environments. Most patients could have a favorable outcome with early diagnosis and proper management. Looking ahead, a newborn screening program for beta-ketothiolase deficiency may have a high yield in Vietnam.