MT-FISH is a precise tool for detecting STRs, but its high cost means that in many developing countries it is first necessary to exclude other more common conditions before choosing this technique. For this reason, we have developed a clinical protocol for investigating the causes of MR [12, 13] that can be applied to children suspected of having MR before opting for expensive modern techniques.
Our patients were selected from a large group in which developmental delay was the principal diagnosis and was associated with dysmorphic features and/or malformations. This relationship was initially established in a large screening study in which STRs were found to contribute significantly to unexplained developmental delay and MR with or without a family history of developmental delay or MR [14, 15]. In the last decade, several studies using unselected and selected patients have shown that STRs are a prominent cause of MR, especially in idiopathic forms or unexplained developmental delay. Yu et al.  reported that the frequency of truly cryptic subtelomeric abnormalities in selected patients was 2.6%. In this study, we examined whether MT-FISH could be used as a routine diagnostic test before the onset of MR. However, this question could not be answered because no STRs were detected in our patients.
Scoring based on clinical criteria has been used in many studies [17–22] to increase the sensitivity of STR detection when MT-FISH is used. De Vries et al.  re-examined 29 patients with STRs with regard to their shared characteristics and compared the findings with an MR group without STRs (110 patients). Our group of subjects contained five patients with a low birth weight and nine with a family history of MR. A genetic cause is presumed to underlie half of the cases of undiagnosed patients with idiopathic forms of MR and in many cases there is a family history of MR . Riegel et al.  stated that family history was an important selection criterion, especially when MR was associated with dysmorphic features and/or major malformations and growth retardation. As shown here, the number of patients with a low body weight, malformation and a family history was proportionally similar to that reported by de Vries et al. . While the presence of these characteristics and the moderate to severe degree of MR should have increased the likelihood of identifying an STR , no STRs were in fact detected. The clinical scores of our patients ranged from four to seven yet no STRs were detected, even though the patients had been carefully selected and were suspected of having a chromosomal anomaly. A similar conclusion was reached by van Karnebeek et al. .
Joyce et al.  reported one case of STR using MT-FISH that, after a detailed karyotypic review, was found to have semi-cryptic structural anomalies involving a chromosomal imbalance. In the same study, two STRs were found in normal individuals, which suggested polymorphism in subtelomeric regions .
A 13q deletion with mosaicism was detected in one of our patients; however, there was no phenotypic correlation with the clinical manifestations  and FISH with a specific subtelomeric probe was negative. This case indicates the need to establish a clinical correlation with the cytogenetic findings when an STR is detected, especially if the parents or a specific probe are unavailable.
A non-subtelomeric rearrangement could also explain our negative MT-FISH results. In this case, other techniques such as array comparative genomic hybridization may be useful. Based on our findings, we cannot exclude unknown monogenic disease or multi-factorial conditions as at least part of the etiology of MR and developmental delay in these patients.