Main

Glutaryl-CoA dehydrogenase (GCDH; EC 1.3.99.7) deficiency (glutaric aciduria type I) is a “cerebral” organic acid disorder first described in 1975 (1). The estimated overall prevalence is 1 in 100,000 newborns (2), but is considerably higher in some genetic isolates (3,4). More than 150 disease-causing mutations have been identified (5,6). GCDH is a mitochondrial key enzyme in the final degradative pathways of L-lysine, L-hydroxylysine, and L-tryptophan (7). The prognostically relevant event is the onset of acute encephalopathic crises which are usually precipitated by episodes that are likely to induce catabolic state (e.g., febrile illness) during a vulnerable period in infancy resulting in irreversible striatal injury and, subsequently, movement disorders (4,8,9). Life expectancy is strongly reduced in symptomatic patients (10,11). In two variant forms of GCDH deficiency (insidious onset, late onset), neurologic disease develops without overt crises, suggesting chronic neurodegenerative changes. If treatment starts before the onset of irreversible neurologic symptoms, encephalopathic crises can be prevented in the majority of patients (4,8,11–13).

The initial presentation of affected children is nonspecific, and macrocephaly is most frequently found (9). Therefore, metabolic screening techniques have been developed. The key metabolites can be detected by gas chromatography/mass spectrometry (GC/MS) of glutaric acid (GA) and 3-hydroxyglutaric acid (3-OH-GA; ref. 14) or tandem mass spectrometry (MS/MS) of glutarylcarnitine (C5DC; refs. 2,15). MS/MS-based newborn screening (NBS) programs have been established in some countries (15–17). In Germany, MS/MS-based NBS has been initiated in 1999 in some federal countries (Baden-Württemberg, Bavaria, Lower Saxony) and, recently, it has been implemented into the nationwide NBS. The major aims of the present study were to investigate whether implementation of NBS for GCDH deficiency and intensive management influences the disease course and neurologic outcome of affected individuals.

PATIENTS AND METHODS

Study population.

We studied 100 patients (44 female, 56 male) with GCDH deficiency. Diagnosis was confirmed by analysis of GCDH gene mutations (n = 82) and/or enzymatic analysis of GCDH activity (n = 70; Prof. Christensen, Copenhagen, Denmark). Patients were anonymized by initials, birthdates, and gender, and double entries were excluded by these parameters. Data were only included if written informed consent was obtained from patients and/or parents. The study was approved by the Institutional Ethics Committee of the University of Heidelberg (# 314/2002).

Patient groups and study protocol.

From 1999–2005, NBS in Germany identified 38 patients with GCDH deficiency (“NBS group”). Prospective follow-up was performed using a standardized protocol (URL: www.metabnet.de/index.php?lang=en&ID=4&sID=14&psID=4; subproject “Glutaric aciduria type I”, Link: “Instruments”) for therapy and therapy monitoring (18). Briefly, for maintenance treatment a combination of dietary treatment (lysine restriction plus lysine-free, tryptophan-reduced amino acids supplements) and L-carnitine supplementation was used. Treatment was intensified during episodes that are likely to precipitate encephalopathic crises, e.g., febrile illness. Stop of natural protein intake, administration of high-caloric fluids, increasing the carnitine dosage, and administration of antipyretics and/or antibiotics were transiently used during these episodes and maintenance treatment was re-introduced stepwise. Similar treatment protocols (including lysine-restricted diet) have been used at German metabolic centers since the early 1990s. In general, dietary treatment and pharmacotherapy was implemented by an interdisciplinary team including metabolic pediatricians, dieticians, nurses and occupational therapists, and parents were trained by this team.

In addition, 62 patients were identified before the start of NBS (“historical patients”). In analogy to a previous study (11), data from this cohort of patients were collected by a standardized questionnaire.

Statistical analysis.

The cut-off date for statistical analysis was October 31, 2006. We used nonparametric tests to compare outcome variables, and Kaplan-Meier analysis to calculate survival and encephalopathic crises. SD scores of anthropometrical data were calculated (19) using LMS values for body mass index (20), length and weight (kindly provided by Dr. K. Kromeyer-Hauschild), and head circumference (21).

We calculated the prevalence of GCDH deficiency in three different birth cohorts, i.e., cohort I (year of birth 1975–1990), cohort II (1991–1998), and cohort III (1999–2005). This division was made to mark important ‘milestones' in the medical history of this disease, including the first description in 1975 (1), publication of the first major case series in 1991 (22–24), and the start of NBS in Germany in 1999. Numbers for annual birth rates in Germany were obtained from the Statistisches Bundesamt Deutschland (URL: www.destatis.de) and numbers of screened neonates were kindly provided by the Deutsche Gesellschaft für Neugeborenenscreening (URL: www.neoscreening.de). Statistical analysis was performed using R (25).

RESULTS

Study population.

One hundred patients with confirmed diagnosis of GCDH deficiency were included into this study. Patients were identified presymptomatically by NBS (n = 38), or due to high-risk screening of macrocephalic infants (n = 4) or siblings of previously identified patients (n = 5), whereas 53 patients were diagnosed by selective screening after the onset of neurologic symptoms. The majority of patients was of German origin (59/100), whereas 41 patients were from migrating families, mostly of Turkish ancestry. The cumulative follow-up time of all study patients was 685 y. The median age at latest report in the NBS group was 49 mo (range: 10–196 mo) and 133 mo (range: 5–826 mo) in historical patients. Diagnosis was made at a median age of 0.25 mo in the NBS group and at 15 mo of age in historical patients.

Estimated prevalence in neonatally screened and historical patients.

We calculated the prevalence of GCDH deficiency in three different birth cohorts, i.e., cohort I (year of birth 1975–1990), cohort II (1991–1998), and cohort III (1999–2005). This division was made to mark important ‘milestones' in the knowledge and awareness of this disease, including the first description in 1975 (1), publication of the first major case series in 1991 (22–24), and the start of NBS in Germany in 1999. The lowest prevalence (mean: 1:611,400 newborns; 95% CI: 1:925,800-403,800) was found in cohort I, whereas it markedly increased in cohorts II (mean: 1:187,200 newborns; 95% CI: 1:261,600-134,000) and III (mean: 1:100,200 newborns; 95% CI: 1:138,100-72,700). These results suggest that selective screening (cohorts I and II) is highly dependent on the awareness of physicians for GCDH deficiency and thus includes a high risk to miss patients. It is not unexpected that the highest mean prevalence was found in cohort III (NBS group), since only NBS allows reliable identification of these patients. However, the number of historical patients (cohorts I and II) who have died undiagnosed is unknown due to the lack of a national tracking system.

Disease course and survival: NBS versus historical patients.

It has been demonstrated that the onset of encephalopathic crises is the prognostically relevant event in GCDH deficiency (4,8–11). These studies have highlighted that the majority of first crises occur by age 24 mo, which allowed estimating the disease course after a relatively short observation period. This was confirmed by the present study showing that 95% of first crises occurred by age 24 mo (median: 10 mo; range: 1–37 mo). Importantly, all study patients except three children from the NBS group were older than 24 mo.

Forty-six patients had encephalopathic crises; the majority (67%) had only a single crisis. Kaplan-Maier analysis (Fig. 1) showed that the frequency of encephalopathic crises was lower in the NBS group (4/38 patients) and in presymptomatically diagnosed patients from the historical cohort (1/9 patients) than in symptomatic patients from the historical cohort (41/53 patients). The absolute risk reduction for the manifestation of encephalopathic crises between the NBS group (4/38 patients) and historical patients diagnosed after the onset of neurologic symptoms (41/53 patients) was 0.67, the number needed to treat to prevent one encephalopathic crisis was 1.50 confirming the efficacy of this management. In migrating families, however, neonatally screened children had more often crises (3/16 patients) than in families of German origin (1/22 patients).

Figure 1
figure 1

Encephalopathic crises in neonatally screened and historical patients. Kaplan-Meier analysis shows a lower frequency of encephalopathic crises (χ2 (2) = 41.6, p < 0.001; log-rank test) in children identified presymptomatically by NBS (“neonatal”; n = 38) or in the historical cohort (“Hist-pre”; n = 9) than in historical patients identified after the onset of neurologic symptoms (“Hist-sympt”; n = 53). Most crises occurred by age 24 mo. No crisis has been reported after age 37 mo.

Full size image

In the four patients of the NBS group, who had had encephalopathic crises, start of emergency treatment was delayed for more than 24 h or even missed despite the onset of alarming symptoms, such as vomiting. In contrast, if emergency treatment was immediately started during threatening episodes (n = 68), no encephalopathic crisis occurred. Patients who had suffered encephalopathic crises developed dystonia (100%), and feeding problems (76%). Movement disorders were considered as severe (71%) or moderate (29%). As a consequence of severe neurologic symptoms, 11 children from the historical group died, whereas all presymptomatically diagnosed children have survived (Fig. 2). The most frequent known cause of death was pneumonia (36%). In contrast, all presymptomatically diagnosed children survived.

Figure 2
figure 2

Survival rate in neonatally screened and historical patients. Kaplan-Maier analysis shows reduced life expectancy of symptomatic patients from the historical cohort (“Hist-symp”; n = 53) compared with children identified presymptomatically by NBS (“Neonatal”; n = 38) or in the historical cohort (“Hist-pre”; n = 9) who all have survived (χ2 (2) = 6.5, p < 0.05; log-rank test). Note that the follow-up period in the NBS group is shorter than in the historical group.

Full size image

A subgroup of historical patients (n = 15) developed neurologic disease without known encephalopathic crises. Ten of these patients were classified as insidious-onset and 5 patients as late-onset type. In patients with insidious-onset type mild movement disorders (dystonia, ataxia) and learning difficulties were common, whereas patients classified as late-onset type presented with cephalgies, gait disturbance, slowing of fine motor function and tremor. The frequencies of disease courses in different patients groups are summarized in Fig. 3.

Figure 3
figure 3

Clinical presentation in neonatally screened and historical patients. Neurologic symptoms were common in the majority of historical patients (“Hist-sympt”; n = 53), whereas only a few patients, who have been diagnosed presymptomatically (“Hist-pre”; n = 9), remained asymptomatic. Asymptomatic disease course was common in the NBS cohort. Encephalopathy, encephalopathic crises; insidious, insidious-onset type; late, late-onset type; asymptomatic, asymptomatic patients.

Full size image

Gross motor development in the neonatal screening cohort.

Since it has been reported that some infants with GCDH deficiency develop transient neurologic ‘soft' signs (8), we investigated whether gross motor development was affected in neonatally screened infants. Transient trunkal hypotonia was found in 53% of infants. Between age 0 to 24 mo, most patients without an encephalopathic crisis (n = 34) reached six gross motor developmental milestones within the normal range according to the WHO motor development study (26), whereas some showed a mild to moderate delay (Fig. 4). In contrast, children with an encephalopathic crisis (n = 4) did not reach any of these motor milestones by age 24 mo.

Figure 4
figure 4

Gross motor development in the NBS group. Box plot diagram shows the achievement of six gross motor milestones in the NBS group from age 0 to 24 mo. Circles represent outliers and grey boxes indicate the windows of motor milestone achievement according to ref 25.

Full size image

Anthropometrics in the neonatal screening cohort.

Presymptomatically diagnosed patients were treated by oral restriction of lysine intake plus lysine-free, tryptophan-reduced amino acids supplements, which also contained minerals and micronutrients, according to protocol for maintenance treatment (URL: www.metabnet.de/index.php?lang=en&ID=4&sID=14&psID=4; subproject “Glutaric aciduria type I”, Link: “Instruments”; 18). Since dietary restriction of natural protein includes a theoretical risk for malnutrition and failure to thrive, we investigated anthropometrical data in the NBS group by the age of 24 mo. Height, weight, and body mass index of these patients (Fig. 5A–C) were within the range expected in unaffected children (20). No patient was dystrophic, including four children who had had encephalopathic crises. Furthermore, an increased rate of obesity was not found in these children. As expected, the mean head circumference was increased owing to the disease-related high frequency of macrocephaly (Fig. 5D).

Figure 5
figure 5

SD scores for height (A), weight (B), body mass index (C), and head circumference (D) in the NBS group. Grey boxes indicate the mean (grey lines) ±1.96 SD of measures within a 6-mo interval, whereas circles indicate single measures of asymptomatic patients. Black lines represent individuals with encephalophatic crises.

Full size image

Genotypes and biochemical phenotypes. NBS versus historical patients.

In some diseases, NBS has been shown to select patients with a mild biochemical phenotype. Therefore, we compared the genotypes and biochemical phenotypes in the NBS group and historical patients. First, GCDH gene analysis showed that the same mutations (E365K, P248L, R402W, A421V) were most frequently found in both groups (NBS group: 48%; historical patients: 55%). E365K and P248L were exclusively found in patients of Turkish origin and A421V in patients of German origin, whereas R402W, the most frequent mutation in Caucasians, was present in all ethnic groups. In line with mutation analysis, distribution of biochemical phenotypes was similar in both groups. According to the definition of Baric et al. (14), we found 82% high and 18% low excretors in the NBS group and 79% high and 21% low excretors in historical patients. Notably, NBS has reliably identified seven low excretors (Table 1), and, to our knowledge, no patient has been missed by NBS. GCDH analysis in fibroblasts demonstrated a median GCDH residual activity of 1% (range: 0–25%) in the NBS group and 0% (range: 0–30%) in historical patients. In conclusion, the distribution of genotypes and biochemical phenotypes in both subgroups were similar.

Table 1 Characterization of low excretors identified by neonatal screening
Full size table

Comparison of siblings.

There is no evidence that the genotype correlates with the clinical phenotype in GCDH deficiency owing to a high variation of the disease course among individuals (11). To investigate whether the disease course of affected siblings was influenced by the identification of index patients, we compared 24 siblings from 10 families (Table 2). Siblings shared the same GCDH gene mutations (not shown) and the same biochemical phenotype. The majority of index patients was identified after the onset of neurologic symptoms (8/10) at a median age of 19 mo, whereas only two index patients were identified presymptomatically due to macrocephaly (age at diagnosis: 1 mo, 24 mo). Three older siblings already had neurologic symptoms when being diagnosed. In contrast, the majority of younger siblings (n = 11), who were diagnosed presymptomatically, remained asymptomatic (9/11), whereas one suffered an encephalopathic crisis and one presented with insidious-onset type. These results underline that identification of presymptomatic patients is neuroprotective.

Table 2 Comparison of siblings
Full size table

DISCUSSION

The major results of the present study are that 1) NBS in combination with early intensive management decreases the frequency of acute encephalopathic crises, 2) lysine-restricted dietary treatment supports normal growth and thrive, 3) the genotypes and biochemical phenotypes of patients from the NBS group and historical patient group are similar, and 4) NBS for GCDH deficiency is reliable for the identification of low excretors.

NBS and early intensive management are neuroprotective.

Since the first report on GCDH deficiency 32 y ago (1) more than 400 patients have been identified worldwide (3,4,8–11,13). As a result of variations in the natural history, and differences in the diagnostic criteria and therapeutic protocols in different countries, the outcome has been varied. This has hampered our understanding of this disease and its treatability. Recently, we have demonstrated in 279 patients that presymptomatically diagnosed patients, who have received a combined maintenance therapy with lysine-restricted diet using natural protein with a low lysine content and amino acid supplementation (lysine-free, tryptophan-reduced) and oral carnitine supplementation, had the best outcome (11). However, the cross-sectional design of this study includes a moderate possibility for a confounder bias. For this purpose, we have conducted the present prospective follow-up study on children identified by neonatal screening in Germany (n = 38) and, subsequently, we compared outcome parameters with a historical patients group (n = 62).

The present study shows that the onset of prognostically relevant encephalopathic crises was markedly reduced (11% of patients) in the majority of children diagnosed by neonatal screening who have received intensive management. The same result was achieved in a small subgroup (n = 9) of historical patients diagnosed presymptomatically (11%). Most of these children achieved gross motor milestones without or with a mild delay, and dietary management did not affect growth and maturation. In contrast, historical patients, who already had symptoms when diagnosed, showed a high frequency of complications and reduced life expectancy owing to acute encephalopathic crises (76%) and variant disease forms, i.e., insidious-onset (16%) and late-onset types (8%). Similar findings have been reported in a previous study (4), demonstrating that screening of asymptomatic newborns and, subsequently, careful follow-up management reduced the frequency of acute encephalopathic crises from 90% (n = 57) to 35% of patients (n = 20). Compared with this study, the frequency of acute crises was somewhat lower in our study, i.e., 11% of neonatally diagnosed patients. Whether this discrepancy in the outcomes reflects differences in the natural history of included study patients, dietary treatment (no lysine-free, tryptophan-reduced amino acid supplements were used by Strauss et al. (4)), emergency treatment (in our study, intensified emergency treatment was usually performed at hospital), or environmental or socio-economic parameters is not yet known.

Does NBS for GCDH deficiency select patients with a mild disease variant? From 1999 to 2005, neonatal screening for GCDH deficiency has identified 38 asymptomatic neonates (including 7 low excretors) in Germany and, as far as it is known, no patient has yet been missed. This is in contrast with previous findings from other countries suggesting that some patients with a mild biochemical phenotype may be missed (27). However, this discrepant experience most likely reflects the optimization process while establishing a novel screening method. We have recently demonstrated using data mining strategies that the sensitivity and specificity of MS/MS-based neonatal screening for GCDH deficiency can be increased by implementation of ratios (i.e., C5DC to palmitoylcarnitine or free carnitine) as second tier (28). Thus it is likely to assume that a second tier strategy would improve the identification of low excreting neonates with GCDH deficiency.

It is subject of current debates whether neonatal screening increases the selection of individuals with a mild disease variant who may remain asymptomatic – even if undiagnosed and untreated. Mild disease forms are known for many inborn errors of metabolism, such as medium-chain acyl-CoA dehydrogenase deficiency (29), isovaleric acidemia (30), or hyperphenylalaninemia (31). However, the natural history of patients with GCDH deficiency does not correlate with the genotype or the biochemical phenotype (32) resulting in a similar a priori risk for all untreated patients to suffer acute encephalopathic crises (4,10,11,13). Furthermore, no variables are known that allow predicting the disease course of presymptomatically diagnosed patients and the necessity for treatment. Despite this, we cannot exclude that NBS for GCDH deficiency may identify patients with a mild disease variant. The major source of uncertainty is the discrepancy of the mean prevalences within the historical patients (cohorts I and II) and the NBS group (cohort III). This may be explained by missed patients in the historical cohorts, by selection of patients with a mild disease variant in the NBS cohort or a combination of both. Unfortunately, the number of missed patients in the historical cohort is unknown. Studies on sudden infant death syndrome cannot substitute for this (33) and are likely to underestimate the number of missed cases, since most patients have died after infancy (this study, median age at death: 84 mo).

Since the number of missed patients remains obscure, the ascertainment bias of this study can be estimated by the suggested maximal proportion of patients who may do well without treatment. Assuming that the 1.87-fold higher mean prevalence in cohort III (NBS group) than cohort II (latest historical patients cohort) results from a selection of patients with a mild disease variant by NBS, it can be calculated that 18 patients from the NBS group may have done well without treatment. Even if this was true, absolute risk reduction (0.57 instead of 0.67) and the number needed to treat to prevent one encephalopathic crisis (1.74 instead of 1.50) would still show a beneficial effect of NBS and intensive management. However, it seems doubtful that NBS for GCDH deficiency increases the number of patients who may not require treatment, since 1) presymptomatically diagnosed siblings of symptomatic index patients from the historical patients group show a similar decline of encephalopathic crises as patients from the NBS group, 2) the genotypes and biochemical phenotypes of patients from the NBS group and historical patients group are similar and 3) late-onset type GCDH deficiency has been demonstrated in untreated adolescents or adults (11,12).

In conclusion, the present study demonstrates that NBS in combination with intensive management reduces the risk for the onset of prognostically relevant acute encephalopathic crises in the majority of children with GCDH deficiency and thus confirms recent recommendations for the diagnosis and management of this disease (18). This guideline also describes in detail how maintenance and emergency treatment can be applied. Since no variables are known that reliably predict the clinical course of untreated patients and since treatment efficacy is high in presymptomatically diagnosed patients, there is no known evidence base for a risk stratification and thus it is recommended to treat all patients. Further studies are required to identify the major single aspects of clinical management that influence the outcome of affected patients. However, despite these promising results the long-term outcome of presymptomatically diagnosed children remains to be elucidated.