Capkova Zuzana, Capkova Pavlina, Srovnal Josef, Staffova Katerina, Becvarova Vera, Trkova Marie, Adamova Katerina, Santava Alena, Curtisova Vaclava, Hajduch Marian, Prochazka Martin
Department of Medical Genetics, University Hospital Olomouc, Olomouc, Czech Republic.
Department of Medical Genetics/Faculty of Medicine and Dentistry, Palacky University Olomouc, Olomouc, Czech Republic.
PeerJ. 2019 Nov 15;7:e7979. doi: 10.7717/peerj.7979. eCollection 2019.
Autism spectrum disorders (ASD) and intellectual disabilities (ID) are heterogeneous and complex developmental diseases with significant genetic backgrounds and overlaps of genetic susceptibility loci. Copy number variants (CNVs) are known to be frequent causes of these impairments. However, the clinical heterogeneity of both disorders causes the diagnostic efficacy of CNV analysis to be modest. This could be resolved by stratifying patients according to their clinical features.
First, we sought to assess the significance of particular clinical features for the detection of pathogenic CNVs in separate groups of ID and ASD patients and determine whether and how these groups differ from each other in the significance of these variables. Second, we aimed to create a statistical model showing how particular clinical features affect the probability of pathogenic CNV findings.
We tested a cohort of 204 patients with ID ( = 90) and ASD ( = 114) for the presence of pathogenic CNVs. We stratified both groups according to their clinical features. Fisher's exact test was used to determine the significance of these variables for pathogenic CNV findings. Logistic regression was used to create a statistical model of pathogenic CNV findings.
The frequency of pathogenic CNV was significantly higher in the ID group than in the ASD group: 18 (19.78%) versus 8 (7%) ( < 0.004). Microcephaly showed a significant association with pathogenic findings in ID patients ( < 0.01) according to Fisher's exact test, whereas epilepsy showed a significant association with pathogenic findings in ASD patients ( < 0.01). The probability of pathogenic CNV findings when epilepsy occurred in ASD patients was more than two times higher than if epilepsy co-occurred with ID (29.6%/14.0%). Facial dysmorphism was a significant variable for detecting pathogenic CNVs in both groups (ID = 0.05, ASD = 0.01). However, dysmorphism increased the probability of pathogenic CNV detection in the ID group nearly twofold compared to the ASD group (44.4%/23.7%). The presence of macrocephaly in the ASD group showed a 25% probability of pathogenic CNV findings by logistic regression, but this was insignificant according to Fisher's exact test. The probability of detecting pathogenic CNVs decreases up to 1% in the absence of dysmorphism, macrocephaly, and epilepsy in the ASD group.
Dysmorphism, microcephaly, and epilepsy increase the probability of pathogenic CNV findings in ID and ASD patients. The significance of each feature as a predictor for pathogenic CNV detection differs depending on whether the patient has only ASD or ID. The probability of pathogenic CNV findings without dysmorphism, macrocephaly, or epilepsy in ASD patients is low. Therefore the efficacy of CNV analysis is limited in these patients.
自闭症谱系障碍（ASD）和智力障碍（ID）是具有显著遗传背景且遗传易感性位点存在重叠的异质性复杂发育疾病。已知拷贝数变异（CNV）是这些障碍的常见病因。然而，这两种疾病的临床异质性导致CNV分析的诊断效能一般。可通过根据患者临床特征进行分层来解决这一问题。
首先，我们试图评估特定临床特征在检测ID和ASD患者不同组中致病性CNV方面的意义，并确定这些组在这些变量的意义上是否以及如何彼此不同。其次，我们旨在创建一个统计模型，展示特定临床特征如何影响致病性CNV发现的概率。
我们对204例ID患者（n = 90）和ASD患者（n = 114）组成的队列进行检测，以确定是否存在致病性CNV。我们根据他们的临床特征对两组进行分层。采用Fisher精确检验来确定这些变量对致病性CNV发现的意义。使用逻辑回归创建致病性CNV发现的统计模型。
ID组中致病性CNV的频率显著高于ASD组：18例（19.78%）对8例（7%）（P < 0.004）。根据Fisher精确检验，小头畸形在ID患者中与致病性发现显著相关（P < 0.01），而癫痫在ASD患者中与致病性发现显著相关（P < 0.01）。当癫痫发生在ASD患者中时，致病性CNV发现的概率比癫痫与ID共发时高出两倍多（29.6%/14.0%）。面部畸形是两组中检测致病性CNV的显著变量（ID组P = 0.05，ASD组P = 0.01）。然而，与ASD组相比，畸形使ID组中致病性CNV检测的概率增加了近两倍（44.4%/23.7%）。ASD组中巨头畸形的存在通过逻辑回归显示致病性CNV发现的概率为25%，但根据Fisher精确检验这并不显著。在ASD组中，若不存在畸形、巨头畸形和癫痫，检测致病性CNV的概率降低至1%。
畸形、小头畸形和癫痫增加了ID和ASD患者中致病性CNV发现的概率。每个特征作为致病性CNV检测预测指标的意义因患者仅患有ASD还是ID而有所不同。在ASD患者中，不存在畸形、巨头畸形或癫痫时致病性CNV发现的概率较低。因此，CNV分析在这些患者中的效能有限。
