Background

Genetic disorders (GDs) are a significant concern in the neonatal intensive care unit (NICU), contributing to approximately 20% of infant deaths [1, 2]. The incidence of GDs diagnoses has risen in recent years, largely due to advancements in genomic sequencing [1, 3]. GDs are diverse and have serious clinical implications, affecting patient diagnoses and their quality of life. Early diagnosis is crucial, benefiting from clinical management and treatment. However, clinical signs are often subtle in the early stages, particularly in newborns in the NICU, where symptoms may not be fully apparent. Sometimes, the phenotype of GDs can be obscured by other clinical symptoms, complicating diagnosis despite numerous routine and specialized tests, including invasive procedures and repeated blood sampling. These processes can cause significant distress to patients, financial strain on families, and yet may not elucidate the underlying pathogenesis [4, 5]. The rapid progression of GDs can lead to death or disability if diagnosis and treatment are delayed or missed.

The evolution of molecular diagnostic techniques has significantly enhanced the role of genetic testing in GDs diagnosis [2, 6,7,8]. Recent studies highlight the benefits of genetic testing in clinical settings. For example, a study from China on the largest cohort of neonates with congenital heart defects demonstrated that next-generation sequencing facilitated precise genetic diagnoses, enabling earlier intervention by specialists [9]. A study at Bei**g Children’s Hospital [10] revealed that exome sequencing as an initial test for pediatric respiratory diseases had a diagnostic yield of 34.6%, proving its efficacy in rapidly diagnosing and guiding treatment. Additionally, a prospective study [11] showed that using whole-exome sequencing (WES) as a primary test in infants suspected of monogenic disorders could streamline the diagnostic process, offering a higher diagnostic yield than standard approaches. Recent research has also discussed the application of WES in NICU settings and among neonate populations in China from various perspectives [2, 9, 12, 13], including the study of molecular defects in neonates conceived through assisted reproductive technology. Despite these advancements, the application of WES in diagnosing neonatal genetic diseases warrants further exploration. This paper presents a retrospective observational study on a Chinese cohort of neonates, aiming to discuss the significance of WES for patients in the NICU.

Methods

Recruitment and data collection

From January 2018 to April 2021, we analyzed 132 neonates hospitalized in the NICU at Tian** Children’s Hospital in China, who presented with acute illness and were suspected of having GDs identified through WES. After excluding 11 cases due to clear etiology, incomplete information, or duplicate collections, 121 cases were included in our study, as illustrated in Fig. 1. The Tian** Children’s Hospital Ethics Committee approved this study, and informed consent was obtained from the guardians or parents. We gathered demographic, prenatal, and intranatal information, along with clinical manifestations, physical examination results, accessory examination outcomes, and family histories of the enrolled patients.

Fig. 1
figure 1

Flow chart of enrolled cases

Full size image

WES and bioinformatic analysis

Genomic DNA was extracted from the peripheral venous blood of the patients and their parents. WES was conducted by Kingmed Company (Guangzhou, China) and MyGenostics Inc. (Bei**g, China), achieving read lengths of 150 bp and an average coverage depth of 100-200X for over 95% of targeted regions. These regions included the coding areas of more than 20,000 genes and the exon–intron boundaries. Exome capture was performed using the xGen Exome Research Panel v2 (Integrated DNA Technologies, US), and sequencing was carried out on the NovaSeq 6000 system (Illumina, US). The raw data were aligned to the human reference genome hg19 using the Burrows-Wheeler Aligner software. Variant annotations were conducted using the ANNOVAR software, updated bi-monthly, with integration from databases [14] such as RefSeq Gene, dbSNP150, ClinVar, HGMD, and allele frequencies from 1000G, ESP6500, and the ExAC database. Copy Number Variants (CNVs) calling was executed using ExomeDepth software [15] with default settings and batch consistency in sequencing and bioinformatics procedures. The common cause of neonatal hypotonia, homozygous deletion of SMN1, was analyzed using established protocols [16, 17]. Verification of exome results, when necessary, was done through Sanger sequencing and/or Multiplex Ligation-dependent Probe Amplification (MLPA) using the BigDye™ Terminator v3.1 (Applied Biosystems™, US.) on the ABI3530Dx platform (Applied Biosystems™, US), and standard reagents from MRC-Holland (Netherlands) for MLPA.

Pathogenicity assessment

The analysis of genetic reports aimed to explore the connection between phenotype and genetic variation. The pathogenicity of variants was classified according to the criteria set by the American College of Medical Genetics and Genomics (ACMG) [18], which includes five levels: pathogenic, likely pathogenic, variant of uncertain significance, likely benign, and benign.

Results

Clinical information of neonates in the NICU

This study enrolled 121 neonates, comprising 66 male and 55 female infants, with ages at enrollment ranging from 1 h to 28 days. The majority of the infants were full-term, accounting for 86.8% (105/121), while preterm and extremely preterm infants represented 12.4% (15/121) and 0.8% (1/121), respectively, as detailed in Table 1. The study focused on multiple organ or system involvements, primarily in the metabolic, neurologic, and endocrine systems, with incidence rates of 24% (29/121), 15.7% (19/121), and 12.4% (15/121), respectively, also shown in Table 1. In the metabolic system, the clinical manifestation included poor feeding, vomiting, disturbance of consciousness, metabolic acidosis, hyperammonemia, hyperkalemia, hyperhomocysteinemia, hypotonia, abnormal electroencephalogram, anemia, cardiomyopathy, metabolic alkalosis, hyperphenylalaninemia, lethargy, seizures, coma, poor growth, and intrahepatic cholestasis. The neurologic system's manifestations included conditions such as convulsion, weakness of limbs, hypotonia, loss of tendon reflexes, muscle weakness, inability to suck, axial hypotonia, hearing loss, peculiar hair, weak cry, recurrent bronchopneumonia, swallowing difficulties, respiratory distress, while the endocrine system showed signs such as electrolyte disorders, dark areola, vomiting, poor feeding, and malnutrition were observed in the endocrine system. Additional details on organ involvement are provided in Supplemental Table 1.

Table 1 Clinical characteristics of neonates
Full size table

Variants information of neonates in the NICU

The overall detection rate of pathogenic or likely pathogenic genomic variants via WES was 52.9% (64/121). We identified 84 genomic variants in 64 neonates, comprising 13 copy number variations (CNVs) and 71 single-nucleotide variants (SNVs), detailed in Table 2. The CNVs mostly involved deletions and duplications, including cases of spinal muscular atrophy and global developmental delay or multiple malformations due to various CNVs. SNVs were primarily missense, nonsense, frameshift, and splicing variations. De novo variations numbered 16 in this cohort, as shown in Table 2. Parental samples were collected simultaneously with the offspring samples for verification by Sanger sequencing upon detection of variant genes. The inheritance patterns are detailed in Fig. 2, with autosomal recessive being the most common (57.8%, 37/64), followed by autosomal dominant (29.7%, 19/64). There was one case of X-linked dominant inheritance, two of X-linked recessive, and five of unknown inheritance patterns. In 4 cases of CNVs, testing at the corresponding loci was recommended for both parents, though it was not performed due to a lack of parental permission. In one case, patient 34 exhibited only one heterozygous pathogenic variant in the FGA gene, suggesting the possibility of undetected mutations by WES.

Table 2 Details of genetic findings in 64 neonates of the current cohort
Full size table
Fig. 2
figure 2

Proportion of inheritance patterns. Abbreviations: AD, autosomal recessive; AR, autosomal dominant; XLD, X-linked dominant; XLR, X-linked recessive

Full size image

Genetic disorders diagnosed by WES

WES identified 40 diseases within the study cohort. The most prevalent disorders were methylmalonic acidemia (MMA) (12.5%, 8/64), epilepsy (10.9%, 7/64), spinal muscular atrophy (6.25%, 4/64), and congenital adrenal hyperplasia (6.25%, 4/64), as shown in Fig. 3. Notably, the audiologic system showed a 100% positive rate (3/3), followed by the neurologic system at 68.4% (13/19) in this cohort.

Fig. 3
figure 3

Distribution of diagnosed diseases. Abbreviations: CAH, congenital adrenal hyperplasia; CS, Cohen syndrome; Ep, epilepsy; MMA, methylmalonic acidemia; PA, propionic acidemia; PKU, phenylketonuria; SMA, spinal muscularatrophy; Note: the details of “Other” category shown in Table 2 with superscripted “#”

Full size image

Discussion

WES has become a primary clinical diagnostic tool for children suffering from developmental delays, intellectual disabilities, respiratory disease [10, 19], and more. Recent discussions have highlighted its use in NICU settings and among neonatal populations in China, viewing it from various perspectives [2, 9, 12, 13]. Despite these discussions, the potential of WES in diagnosing neonatal GDs remains underexplored. In our study, WES was conducted on 121 infants in the NICU at Tian** Children's Hospital of China, yielding a diagnostic rate of 52.9% (64/121). This rate surpasses the 37.9% diagnostic yield of a similar study in the USA [20] and significantly exceeds the 12.3% yield reported in Lin Yang's studies in China [13]. For the discrepancy between this study and two other studies, we analyzed potential reasons based on differences in patient selection and cohort size. Our cohort, selected with stricter inclusion criteria, demonstrated a higher diagnostic yield. We focused on critically ill patients in the NICU, with strong indications for GDs assessed by experienced experts, such as evidence of metabolic disorders identified by mass spectrometry. Additionally, this was a single-center study conducted at Tian** Children’s Hospital, which also accepted critically ill patients from surrounding districts. The cohort size was relatively small, with fewer patients than in the study by Lin Yang et al. [13], which included a cohort of 2,303 neonates in China. In our cohort, a significant proportion of genetically diagnosed patients had metabolic disorders, such as methylmalonic acidemia (MMA), hyperphenylalaninemia, and congenital adrenal hyperplasia. Despite most patients undergoing newborn screening techniques, some reports were negative or showed suspected positive results. In these cases, WES was performed to confirm diagnoses. Among the positive cases, epilepsy and MMA were the most frequently identified diseases, accounting for 10.9% (7/64) and 12.5% (8/64) of cases, respectively.

MMA arising from either a deficiency in methylmalonyl-CoA or abnormal cobalamin metabolism, is a rare, inherited metabolic disorder, primarily passed down through autosomal recessive inheritance. It stands as the most common form of organic acidemia [21]. The disease's genetic underpinnings include mutations in several genes, such as MMACHC, MMADHC, and MMUT, with the prevalence of these mutations varying across different countries and populations [22,23,24]. In China, for instance, the most common mutations in children with MMA are c.609G > A, c.658_660delAAG, and c.80A > G in the MMACHC gene, occurring at frequencies of 34.09%, 13.64%, and 13.64%, respectively [22]. Our study of 8 MMA cases revealed 12 mutations across two genes (MMACHC and MMUT), which included a novel mutation in MMUT c.2131G > T/p.(Glu711*) and eleven inherited mutations. We identified nine gene variations in the MMACHC gene and three mutations in the MMUT gene. The frequency of c.658_660delAAG and c.80A > G mutations was 16.7% (2/12) and 25.0% (3/12), respectively, aligning with findings from previous studies in China [22, 23]. Among eight cases, one was homozygous, while the others were compound heterozygous. All eight MMA patients exhibited an autosomal recessive inheritance pattern. The manifestations of MMA can be nonspecific and vary among patients, especially in newborns and young infants. Previous research has shown that the clinical course of MMA can progress rapidly in neonates, sometimes resulting in death [25, 26] if not treated promptly. Fortunately, MMA is a treatable genetic disorder for most patients [27]. Our team [28] reported a case of a neonate with MMA metabolic decompensation (severe metabolic acidosis and hyperammonemia (> 1,000 µg/dl) who was successfully treated with automatic peripheral arteriovenous exchange transfusion and L-carnitine. The diagnosis was confirmed by WES, leading to a decrease in serum ammonia levels and an improvement in the child’s clinical status. Therefore, WES is crucial for diagnosing this disease, enabling timely treatment and improving the prognosis for NICU patients.

Neonatal seizures are a common manifestation of neurological dysfunction, with an incidence of about 1–5 per 1,000 births [29, 30]. Despite a decrease in mortality from 40 to 20%, the prognosis for neurodevelopmental outcomes, such as cerebral palsy, intellectual disability, and secondary epilepsy, has not significantly improved [30]. Therefore, identifying the cause of neonatal seizures and initiating timely medical treatment is crucial for managing these conditions. The causes of neonatal seizures are diverse, including acute symptomatic seizures, electrolyte imbalances, and cerebral deformity, and so on [30, 31]. Recent advancements in molecular diagnostic technologies, such as WES, have increased the detection rate of genetic disorders causing neonatal seizures [32]. In this study, we identified seven neonatal patients with seizures, uncovering seven mutations in three pathogenic genes. These included five missense mutations, one deletion mutation, and one splicing mutation, comprising one inherited and six de novo variations. For instance, in patient 9, the SCN2A c.781G > A /p.(Val261Met) mutation was identified, leading to a diagnosis of benign familial neonatal convulsions, a form of epilepsy with a favorable prognosis. Thus, WES is valuable for pinpointing genetic causes and guiding precise treatments in NICUs.

In another case, patient 28, a female term infant, was admitted to the NICU of Tian** Children’s Hospital at 25 days old due to jaundice and elevated liver enzymes. The patient, born to a 32-year-old mother, had a normal birth history. Apart from jaundice, the physical examination was unremarkable. Lab tests revealed elevated creatine kinase and glutamic-pyruvic transaminase levels, and an ultrasound cardiogram showed ventricular hypertrophy, suggesting neonatal jaundice, liver dysfunction, and potential hypertrophic cardiomyopathy. Given the unclear etiology, she opted for symptomatic treatment and underwent WES detection. Genetic testing revealed two heterozygous mutations in the GAA gene on chromosome 17q25: c.859-2A > T (p.?) and c.1861 T > G (p.Trp621Gly), confirming the diagnosis of glycogen storage disease type II, also known as Pompe disease. This diagnosis was significantly different from the initial assumption. Pompe disease, a rare autosomal recessive disorder caused by mutations in the GAA gene, leads to a chronic and progressive pathology, predominantly featuring limb-girdle muscle weakness and respiratory failure [33]. Early diagnosis is crucial to mitigate or prevent the irreversible organ damage that progresses with Pompe disease [34]. However, our patient presented without the typical clinical phenotype at admission, posing a diagnostic challenge for clinicians. Thus, WES served as a critical diagnostic tool for patients with unexplained symptoms, ranging from isolated hyper-CKemia to varying degrees of muscular impairment. This case underscores the importance of WES in diagnosing patients with suspected genetic disorders in the NICU, particularly when clinical phenotypes vary widely.

WES analysis presented a negative diagnostic yield of 47.1% (57/121) within this cohort, potentially constrained by the limitations inherent to WES. While WES can detect a wide array of variants, it has a restricted capability in identifying non-coding region variants, abnormal genomic structures, and genomic methylation [35, 36]. However, WES offers significant advantages. Firstly, it covers a broad range of detection and has become more affordable, making it accessible to most parents. Secondly, WES is invaluable for precise diagnosis and treatment strategies, potentially increasing the number of diagnosed infants in NICUs with GDs, thereby reducing infant mortality and morbidity through early neonatal diagnosis. Lastly, WES plays a crucial role in genetic counseling for parents of infants with GDs, enabling informed reproductive decisions. As a vital supplement to standard diagnostics, WES is indispensable for diagnosing GDs in NICU patients.

Conclusions

In conclusion, our findings underscore the critical role of WES in uncovering the etiology, offering targeted therapy, and enhancing the prognosis for patients with suspected GDs in NICUs, especially when diagnoses are complicated by diverse clinical phenotypes. Reflecting on this study and the evidence gathered from this cohort, we advocate for WES as the primary testing approach for suspected GDs cases in NICUs. This recommendation aligns with the evolving trend towards precision medicine, highlighting WES's clinical utility and its importance to patients and their families in cases lacking a clear diagnosis.