Severity Scoring Cutoff for MLPA and Its Diagnostic Yield in 332 North Indian Children with Developmental Delay

Abstract

Chromosomal aberrations/rearrangements are the most common cause of intellectual disability (ID), developmental delay (DD), and congenital malformations. Traditionally, karyotyping has been the investigation of choice in such cases, with the advantage of being cheap and easily accessible, but with the caveat of the inability to detect copy number variations of sizes less than 5 Mb. Chromosomal microarray can solve this problem, but again the problems of expense and poor availability are major challenges in developing countries. The purpose of this study is to find the utility of multiplex ligation-dependent probe amplification (MLPA) as a middle ground, in a resource-limited setting. We also attempted to establish an optimum cutoff for the de Vries score, to enable physicians to decide between these tests on a case-to-case basis, using only clinical data. A total of 332 children with DD/ID with or without facial dysmorphism and congenital malformations were studied by MLPA probe sets P245. Assessment of clinical variables concerning birth history, facial dysmorphism, congenital malformations, and family history was done. We also scored the de Vries scoring for all the patients to find a suitable cutoff for MLPA screening. In our study, the overall detection rate of MLPA was 13.5% (45/332). The majority of patients were DiGeorge's syndrome with probe deletion in 22q11.21 in 3.3% (11/332) followed by 15q11.2 del in 3.6% (12/332, split between Angelman's and Prader–Willi's syndromes). Also, 3.0% (10/332) of patients were positive for Williams–Beuren's syndrome 7q11.23, 1.8% (6/332) for Wolf–-Hirschhorn's syndrome 4p16.3, 1.2% (4/332) for 1p36 deletion, and 1% for each trichorhinophalangeal syndrome type I 8q23.3 duplication syndrome and cri du chat syndrome. The optimum cutoff of de Vries score for MLPA testing in children with ID and/or dysmorphism came out to be 2.5 (rounded off to 3) with a sensitivity of 82.2% and specificity of 66.7%. This is the largest study from India for the detection of chromosomal aberrations using MLPA common microdeletion kit P245. Our study suggests that de Vries score with a cutoff of 3 or more can be used to offer MLPA as the first tier test for patients with unexplained ID, with or without facial dysmorphism and congenital malformations.

Introduction

Clinicians are often faced with the scenario of cases with global developmental delay (DD)/intellectual disability (ID), with or without facial dysmorphism and congenital malformations, for whom a diagnosis has not been achieved despite extensive workup, and a genetic etiology is finally looked for. ^{1 2} In most such cases, chromosomal disorders have been found to be the culprit, as early as the 1970s. ^{3 4 5 6 7} The structural chromosomal aberrations including deletion or duplication of chromosomal segments result in a wide range of manifestations in the affected individuals. ⁶ There may be a spectrum of findings from readily classified specific phenotypes to nonspecific ones. Some indicators such as facial dysmorphism and multiple organ malformations may be of aid. Affected individuals have delayed growth and most children are born small for gestational age. In others, dysmorphism may be subtle, with DD or ID as the only predominant finding. ⁸

Traditionally, karyotyping has been the investigation of choice in such cases ^{4 5 7} as recommended by the American Academy of Neurology and the Practice Committee of the Child Neurology Society in 2003, to include routine cytogenetic studies (along with fragile X testing in males), in cases with unexplained DD, with or without facial dysmorphism. ⁹ The diagnostic yield of karyotyping has been reported to be between 3 and 7%, excluding Down's syndrome. ^{7 10 11} However, karyotyping had the glaring inadequacy of being able to detect changes of more than 5 to 10 Mb only, which are large enough to be seen via a microscope after banding. ¹² Literature review suggests the role of microdeletions in idiopathic ID to be from 0.98 to 7.9%, which are easily missed with karyotyping. ¹³

Fluorescent in situ hybridization (FISH) was promising since it could detect such smaller deletions and duplications, but it is used with targeted probes only. This meant that unless physicians already had a clinical diagnosis, FISH would not be of much help. ⁷ The arrival of array comparative genome hybridization (aCGH)/chromosomal microarray (CMA) filled this gap by providing a nontargeted solution with the ability to detect copy number variations (CNVs) as small as 10 Kb. CMA has replaced karyotype as the investigation of choice in children with unexplained DD or ID, in Western countries and well as in modern medical literature. The yield of CMA has been proven in many large studies to be much higher (up to 20%) than that of traditional cytogenetic investigations. ^{11 14} However, the higher cost and poor availability of CMA have left developing countries, such as India, still in search of a solution. This is where multiplex ligand-dependent probe amplification (MLPA) assay has tried to make a difference. Although not as efficient as CMA, relative ease of access and lower cost have some promises to make. Like FISH, MLPA is a targeted investigation, but the ability to incorporate multiple probes into a single panel to look for different CNVs simultaneously makes up for this shortcoming. ¹⁵

The specific MLPA probe mix used in this study comes from MRC-Holland, called the SALSA MLPA Probemix P245-B1 for Microdeletion Syndromes-1A. ¹⁶ This probe kit includes probes for the detection of more than 30 common microdeletion syndromes. The reason of the choice for these specific panel of syndromes has not been well documented by the company, other than the fact that these are “clinically well-described syndromes, for which the involvement of multiple disease genes has been established or is strongly suspected.” ¹⁶ Albeit this, the clinical utility of this MLPA probe sets has been well established in multiple studies, including ones from India. ^{13 17}

Though MLPA has been praised as the first-line test for many specific clinical scenarios such as methylation disorders and some hematological malignancies, ^{18 19 20} the case is not the same with cases of DD, congenital malformations, and facial dysmorphism. This is due to the obvious inability to detect aneuploidies and large structural changes, which are not covered in the MLPA kit. Karyotype, however, can aid this as discussed earlier.

John et al ¹³ studied 122 Indian children with unexplained mental retardation and a normal karyogram and found MLPA to be positive in 9% of the cases. They used four different MLPA probe sets: P245-A2 in all cases; P070 and P036 in 75 cases each of suspected telomeric CNVs; and P106-B1 in 60 cases of likely X-linked ID. John et al concluded their study by recommending the use of karyotyping in combination with MLPA in such cases until CMA becomes economically feasible in India. ¹³

In contrast to John et al's study, we intended to study the diagnostic yield of a single probe set (P245) but in a larger sample size ( N  = 332). We have also attempted to retrospectively analyze the utility of a clinical scoring system (de Vries score), to screen patients before performing the MLPA, to increase the output of MLPA in such karyotype normal cases.

The de Vries score was first put forward by de Vries et al in 2001, ²¹ as a five-point checklist to improve the diagnostic pickup rate of subtelomeric deletion studies in mentally retarded subjects. This was also a retrospective analysis with 29 cases of known microdeletions. The study failed to mention a final cutoff for the score citing its small sample size, but a review of the study's data suggests that a cutoff of 4 would yield sensitivity and specificity of 89 and 43%, respectively.

The validity of this scoring system, however, has not been evaluated in any large studies. Jehee et al performed a similar study to ours in 209 cases using three different MLPA kits, to establish the yield of de Vries score as a screening tool. ²² He found that cases with chromosomal aberrations detected by karyotyping and MLPA have a mean de Vries score of 5.3 and 4.4, respectively, and recommended a cutoff of 4 for choosing MLPA. This cutoff is similar to the one from de Vries' original article but was not based on statistical evidence. Feenstra et al ²³ used a modified version of the de Vries score to correlate chromosomal imbalances in their study cohort with their clinical phenotype. Since our study also deals with subtelomeric deletions and the same phenotype as in these studies, we chose the de Vries scoring system as the tool to retrospectively evaluate a similar yield of the P245 probe set. Table 1 shows the checklist in its entirety.

Table 1. Checklist with de Vries scoring.

Items	Score
Family history of intellectual disability Compatible with Mendelian inheritance	1
Incompatible with Mendelian inheritance (including discordant phenotypes)	2
Prenatal onset growth retardation	2
Postnatal growth anomalies For each of the following 1 point (maximum 2) Microcephaly (1), short stature (1), macrocephaly (1), tall stature (1)	2
≥2 facial dysmorphic features Notably hypertelorism, nasal anomalies, ear anomalies	2
Nonfacial dysmorphism and congenital anomalies (For each of the following 1 point (maximum 2) Notably hand anomaly (1), heart anomaly (1), hypospadias +/− , undescended testis (1)	2

In our study, we analyzed 332 children presenting to the genetic clinic with DD/ID with or without facial dysmorphism or congenital malformations, and a normal karyogram, using MLPA technique. Chromosomal etiology was identified in 13.5% of such cases. de Vries scoring was applied to all the patients to find out the optimum de Vries cutoff score, which could lead to a higher diagnostic yield and be used as a clinical screening tool before choosing cases for MLPA testing.

Methodology

Study setting : This retrospective analytical cohort study was conducted in the Department of Pediatrics, Postgraduate Institute of Medical Education and Research (PGIMER), Chandigarh, India, in between 2018 and 2020.

Inclusion criteria : Patients with DD/ID with or without dysmorphism or congenital anomalies, whose G-banding karyotyping from peripheral blood had a normal result.

Exclusion criteria : Patients whose parents refused to give informed consent for the study and patients whose phenotype was not well described in the requisition form.

All eligible cases were clinically evaluated by trained clinical geneticists at the Genetic and Metabolic Unit, PGIMER, Chandigarh, India. Informed consent was obtained in each case from the patients, or the parents/guardians when participants were younger than 18 years or unable to provide informed consent.

MLPA : Genomic DNA extraction was performed using 2 to 3 mL of peripheral blood with ethylenediaminetetraacetic acid vials and QIAamp DNA Blood Midi Kit (250) (QIAGEN, Valencia, California, United States). The quantity and quality of the extracted DNA were checked using spectrophotometry. MLPA was performed using SALSA MLPA Probemix P245-B1 Microdeletion Syndromes-1A. ¹⁶

Analysis : Coffalyser stand-alone α version ²⁴ is the standard software recommended by the manufacturer for the analysis of MLPA data and was used for the analysis of peak values obtained from GENESCAN software (ABI 3500 instrument, Applied Biosystems, United States). Coffalyser software has a built-in setup for peak height normalization and reaction quality control calculations. All the statistical analyses were performed according to Coffalyser software manual instructions supplied by the manufacturer. The mean cutoff for normalized peak height ratio of patient to the control samples was less than 0.65 in case of deletions and more than 1.40 in case of duplications.

de Vries scoring : All the patients were assessed by the clinical geneticists for the presence of prenatal growth retardation, postnatal growth abnormalities, facial and nonfacial dysmorphism, congenital abnormalities, and familial occurrence of ID and were scored based on de Vries' checklist. ²¹ For cases in which the details were not available, a clinical examination and history taking were done to fill in the gaps in data. The de Vries score was used as an abnormality evaluation, and not as criteria for inclusion or exclusion in molecular screening.

Statistical analysis : The demographics of the enrolled population, the comparison of characteristics, and analysis of the de Vries sensitivity and specificity were done using the Statistical Package for the Social Suite (SPSS) version 25 by IBM. ²⁵

Results

Demography of enrolled cases : A total of 332 patients were enrolled in this retrospective study, including 265 males and 67 females. The age of the patients varied from 1 month to 14 years, with a mean age of 4.3 years.

Clinical features in the enrolled cases : Most patients presented with dysmorphism (91%) and ID/DD (58%) followed by congenital heart defects in 38% cases, failure to thrive in 64% cases, with seizures and behavioral abnormalities seen in around 24 and 22% cases, respectively.

MLPA result : Pathogenic CNVs were identified in 45 out of 332 patients (13.5%). Among affected individuals, 30 were males and 15 were females, with a mean age of 5.2 years. The distribution of all enrolled cases with their clinical parameters are summarized in Fig. 1 .

Fig. 1.

Distribution of enrolled cases with their clinical parameters. CNS, central nervous system; DD, developmental delay; ID, intellectual disability; MLPA, multiplex ligation-dependent probe amplification.

Among MLPA positive cases, 11 were positive for DiGeorge's syndrome, 12 for Prader–Willi's syndrome/Angelman's syndrome, 10 for Williams–Beuren's syndrome, 6 for Wolf–Hirschhorn's syndrome, 4 for 1p36 deletion syndrome, and 1 each for trichorhinophalangeal syndrome type I (TRPS1) duplication and cri du chat syndromes ( Fig. 2 ).

Fig. 2.

Microdeletions detected in our cases.

DiGeorge's syndrome (22q11.2 deletion) : Eleven cases were seen with 22q11.2 deletion. Six out of the 11 cases presented with cardiac disease, the most common being the atrial septal defect, while other defects seen were ventricular septal defect, tetralogy of Fallot, and tricuspid regurgitation. Failure to thrive and facial dysmorphism ( Fig. 3D ) were present in almost all cases, whereas hypocalcemia was present in three cases. Peculiar findings worth noting seen in two patients were café au lait macule and capillary hemangiomas.

Fig. 3.

( A ) 1p36 deletion showing facial dysmorphism in the form of microcephaly, low-set ears, bilateral esotropia, depressed nasal bridge, and long smooth philtrum. ( B ) 4p16.3 deletion with facial dysmorphism in the form of wide forehead, prominent glabella, antimongoloid slant, and low-set ears. ( C ) 7q11.23 deletion showing per orbital puffiness and large mouth. ( D ) 22q11.2 deletion with cleft palate and square face. ( E ) 8q23.3 duplication with sparse hair, blue sclera, proptosis, coarse square facies, pear-shaped nose, prominent long philtrum, and thin upper lip.

Prader–Willi's syndrome/Angelman's syndrome (15q11.2 deletion) : Twelve cases were picked up with 15q11.2 deletion, of which five and three cases each had clinical features of Angelman's syndrome and Prader–Willi's syndrome, respectively. The clinical profile of patients with Angelman's syndrome includes DD, microcephaly, and fair-colored skin. Three patients presented with seizures, while three had behavior abnormalities in the form of autism. Children with Prader–Willi's syndrome presented with hypotonia and feeding difficulties initially, followed by obesity and facial dysmorphism in the form of bitemporal narrowing and almond-shaped eyes.

Williams' syndrome (7q11.23 deletion) : We detected 10 patients with Williams' syndrome, all having the deletion on chromosome 7q11.23 region. This region includes three probes: ELN gene (exons 1 and 20/2 probes) and the LIMK1 gene (1 probe). The presenting complaints were DD and cardiac disease (aortic stenosis and pulmonary stenosis) in two cases each. Facial features such as periorbital puffiness and thick lips were observed in all eight cases ( Fig. 3C ).

Wolf–Hirschhorn's syndrome (4p16.3 deletion) : In our study, we identified six cases of 4p16.3 deletion indicating a Wolf–Hirschhorn's syndrome. The common presentations in all cases were DD and facial dysmorphism, while four cases had seizures, and three cases had a cardiac defect (atrial septal defect). Three cases had abnormal findings in magnetic resonance imaging, while the facial gestalt in all cases was akin the typical “Greek warrior helmet appearance” with broad forehead, prominent glabella continuing with nasal root and low-set ears ( Fig. 3B ).

1p36 deletion : We detected four cases of 1p36 deletion. The presenting complaints in all were DD, failure to thrive, and facial dysmorphism such as microcephaly, straight eyebrows, deep-set eyes, mid-face retrusion, and pointed chin ( Fig. 3A ). Two out of the cases had seizures, while two cases had some forms of ophthalmological findings such as decreased visual acuity and squint. One patient had hearing difficulty and behavior problems.

TRPS duplication (8q23.3) : One case was seen with duplication of 8q23.3 region. A 4-week-old female infant had a child presented with failure to thrive, congenital heart disease (atrial septal defect, ventricular septal defect, and patent ductus arteriosus), facial dysmorphism (coarse facies, proptosis, and pear-shaped nose), and ectodermal features (sparse hairs) ( Fig. 3E ).

Cri du chat syndrome : One patient was identified as 5p deletion syndrome. The 8-year-old girl presented with delayed milestones and abnormal behavior. On examination, she had facial dysmorphism in the form of triangular face, upward eye slant, hypertelorism, epicanthal folds, long thin philtrum, overhanging columella, thin upper lip, anteverted nares, low-set prominent ears, prominent supraorbital ridges, bilateral clinodactyly, and bilateral simian crease. She was also noticed to have hyperactive behavior.

Applying de Vries score in the cohort : Among the MLPA positive cases with ID and/or dysmorphism, the mode value of the de Vries score was 3, while in the MLPA negative cases, it was found to be 2. Each of the de Vries scoring criteria, along with the number of cases scoring positively to it has been summarized later (both MLPA-positive and MLPA-negative cases) ( Fig. 4 ).

Fig. 4.

Distribution of de Vries checklist points across MLPA-positive cases. ID, intellectual disability; MLPA, multiplex ligation-dependent probe amplification.

To objectively calculate the cutoff value of de Vries score with maximum clinical utility, receiver operating characteristic (ROC) curve was created using the data. The cutoff for MLPA testing in children with ID/dysmorphism/congenital anomalies, as calculated from our cohort came out to be 2.5 with sensitivity of 82.2% and specificity of 66.7% ( Fig. 5 ). The area under ROC curve was 0.83, depicting the good ability of de Vries score to differentiate between positive and negative cases.

Fig. 5.

ROC curve analysis for de Vries score. ROC, receiver operating characteristic.

Discussion

Ours is the largest study from India for the detection of chromosomal aberrations using MLPA common microdeletion kit P245, with a sample size of 332. Sex distribution in our cohort was 4:1 (males to females), which contrasts with the 3:2 ratio seen in most similar studies in children with DD and ID, facial dysmorphisms, and congenital anomalies. ^{5 21} The reason of difference between male to female ratio could be attributed to the well-documented gender differences in care-seeking behavior in South Asian population. ²⁶ The distribution of various clinical features in our cohort was however unique compared with other major studies, which in turn showed high heterogenicity among themselves in the same. ^{5 10 21}

In our study, chromosomal etiology has been identified in 13.5% of cases with ID/DD, and/or dysmorphism and malformations, with normal karyotypes; 55% of cases had ID/DD, while 89% had dysmorphism, but the clinical diagnoses of specific syndromes were not made at the time of sample collection. We believe this is because the clinical features are subtle in many cases and clinical diagnosis may become obvious as the child grows.

Our study showed an overall higher detection rate (13.5%) compared with other studies from India ( Table 2 ). One study for idiopathic ID has shown a diagnostic yield of 4.6% (P070 and P036B), ²⁷ 6.9% (P245), and 7.4% (P070 and P036B) from North India. ²⁸ John et al in 2013 have used P064 and P036 kits to detect subtelomeric rearrangements in the genome which are implicated in the pathogenesis of ID and found a diagnostic yield of 9% from South India. ¹³ Another study from South India by Mohan et al showed 8.4% diagnostic yield. ²⁹ An Iranian study showed 12% yield using P245 MLPA kit. ⁶

Table 2. Other studies from India showing diagnostic yields of MLPA kits.

Studies	Clinical indications	Population	Probe set	Sample size	Positive cases	Detection rate (%)
Present study	ID/DD with or without malformation	North India	P245	332	45	13.5
Mandal et al (2009) 27	Idiopathic mental retardation	North India	P070 and P036B	65	3	4.6
Boggula et al (2014) 28	Unexplained mental retardation	North India	P245	203	14/203	6.9
			P070/P036		15/203	7.4
John et al (2013) 13	Developmental delay/MR	South Indian	P036, P070, P245 A2, and MR X 106B1	122	11	9
Mohan et al (2016) 29	ID/DD	South Indian	P064, P036	107	9	8.4

Abbreviations: DD, developmental delay; ID, intellectual disability; MLPA, multiplex ligation-dependent probe amplification; MR, magnetic resonance.

The microdeletions picked up in our cohort using the P245 kit was 3.6% (12/325) cases of 15q11.2 deletion, followed by 3.4% (11/325) of 22q11.2, then 3.1% (10/325) cases of Williams' syndrome (7q11.23), 1.8% (6/325) Wolf–Hirschhorn's syndrome cases (4p16.3), and 1.2% (4/325) cases of 1p36 deletion syndrome, while only 1 case each was positive for 8q23.3 duplication syndrome (TRPS1) and cri du chat syndrome ( Fig. 3 ). Literature review reported sparse results on the relative distribution of various microdeletion syndromes in such a cohort, for comparison.

Subtelomeric CNVs can be present in 3 to 6% of the cases with isolated ID, and in cases with ID along with any congenital anomaly, the percentage of CNVs is around 10%. ^{30 31} In our cohort, congenital heart defect was more prevalent, including 22q11.2 microdeletion in 11 patients and Williams–Beuren's syndrome in 10 patients. A congenital heart defect is known to be a common feature of both syndromes. Conversely, Monteiro et al have reported CNVs in 25.5% of patients with congenital heart defects. ²⁰

Microduplications are difficult to suspect clinically because of their subtle phenotypes. ³² One patient was detected to have duplication of the 8q23.3 region including TRPS1 gene. The patient presented with failure to thrive, a cyanotic congenital heart disease, ectodermal features in the form of sparse hair and facial dysmorphism in the form of proptosis, coarse square facies, pear-shaped nose, prominent long philtrum, thin upper lip, micrognathia, as well as soft and hard cleft palate with clinodactyly. TRPS1 is a rare disorder that causes distinctive ectodermal, facial, and skeletal features affecting the hair (tricho-), nose (rhino-), and fingers and toes (phalangeal) and is inherited in an autosomal dominant pattern. Recently, Zepeda-Mendoza et al (2019) reported intragenic tandem duplication of TRPS1 in a patient with a classic TRPS1 phenotype. ³³

Utility of the de Vries Score

The original article by de Vries et al ²¹ failed to mention a final cutoff for the selection of cases for subtelomeric testing in children with ID, owing to the poor sample size. However, their data suggested that a cutoff of 4 would yield a high sensitivity (89%), with satisfactory specificity (43%). According to that study of 29 positive cases of subtelomeric deletion, they concluded that a five-item checklist ( Table 1 ) including prenatal onset of growth failure, postnatal growth anomalies, positive family history of mental retardation, and other parameters such as facial dysmorphism and other congenital anomalies could improve the diagnostic pick up rate of subtelomeric deletions in mentally retarded patients. ²¹

On applying the same checklist to our cases, the cutoff for de Vries score for MLPA testing in children with ID and/or dysmorphism came out to be 2.5 with sensitivity of 82.2% and specificity of 66.7%, which has comparable sensitivity to de Vries et al' original study ²¹ and much better specificity ( Fig. 5 ). Since the scoring system is point based (not continuous), this would translate to a score of 3 or more, which can be used as a reasonable clinical cutoff, in clinical settings.

In a study conducted by Jehee et al in 2011, de Vries score was done for 209 patients, where they found that chromosomal abnormalities were present in 87 (33.3%) patients, but only 57 (21.8%) were considered causative. ²² They found that cases with chromosomal aberrations detected by karyotyping and MLPA have a mean de Vries score of 5.3 and 4.4 each and recommended a cutoff of 4 for choosing MLPA. This cutoff is similar to the one from de Vries et al' original article but was based on intuition from the above finding and not on statistical evidence. Our study, using a larger sample size and standard statistical tools to reach the optimal cutoff, helps in revising the clinically relevant score to 3.

The literature review suggests a diagnostic yield of up to 7% using karyotype and up to 20% using aCGH in children with DD/ID with or without dysmorphism or congenital anomalies. ^{7 10 11 14} As already described, our study yielded 13.5% positive results in karyotype-negative cases. Given that karyotype detects aneuploidies and structural aberrations which are not detectable by MLPA, it is evident that the latter can only be used in combination with karyotype and not as a stand-alone tool. aCGH remains the gold standard in such cases in a resource-sufficient setting, since the karyotype plus MLPA combination cannot detect CNVs that are not included in the MLPA kit used, or more than 5 Mb. Thus, based on our study, we recommend that MLPA can be chosen as the investigation of choice over CMA in any similar case with a de Vries score of more than or equal to 3 and a normal karyotype.

In summary, considering the relatively low cost of the technique, short turnaround time and the detection of alterations in 13.5% of patients, the investigation of microdeletions/duplications in our patients with DD/ID with or without dysmorphism and congenital malformations by MLPA using kit P245, yielded valuable results. We conclude that in developing countries where CMA is not easily accessible to such cases, clinical evaluation followed by karyotyping and MLPA is a helpful and affordable solution.

Limitations and Future Directions

• Our cohort was heterogenous with cases having DD/ID, congenital malformations, and dysmorphisms in varying proportions. Future studies can be focused on finding the yield of MLPA in each of these subgroups.

• The cutoff suggested in this study needs to be validated using a case–control study.

• The cutoff of 3 is a scoring system with a maximum of seven points, which shows the inadequacy to the screening tool. This finding portrays the need for a more granular screening tool.

Conclusion

Need for identifying the exact etiology of the syndrome helps both the parents and the physician understand the nature and course of the disease better. It provides an opportunity for genetic counselor/physician to screen other members of the family and deliver an estimated risk of recurrence. For some disorders, therapeutic intervention or a surveillance plan may be started, while for some, a scaffold for future research could be established. However, financial constraints often hold the family from obtaining such a molecular diagnosis. Our study suggests that de Vries score with a cutoff of 3 can be used in resource-limited settings, to offer MLPA over CMA as an affordable diagnostic test, in patients with DD/ID with or without facial dysmorphism and congenital anomalies.