There are increasing numbers of geriatric hip fractures among the ageing population in Hong Kong.1 Patients who sustain an osteoporotic fracture are at increased risk of sustaining further osteoporotic fractures.2 3 The cumulative incidence of second hip fracture was 5.1% at 2 years and 8.6% at 8 years.4 This situation can be improved by implementing better guidelines for secondary drug prevention of fragility fractures. Appropriate treatment of patients with fragility fractures has been shown to reduce subsequent risk of fragility fracture by up to 50%.5 6 7

Many countries in the world have well-established guidelines to close this post-fracture care gap. However, this problem has been overlooked in Hong Kong and the situation is not improving. Diagnosis and treatment of osteoporosis differs among hospitals and specialties. There are no standardised guidelines for treating this particular group of elderly patients. The aim of the present study was to determine the current practice in Hong Kong regarding secondary drug prevention of fragility fractures after osteoporotic hip fracture, in order to make recommendations to implement better guidelines.

In Hong Kong, about 98% of all hospital admissions for hip fracture were admitted to public hospitals rather than private hospitals.8 Patient records from 2009 to 2012, including data on the dispensation of anti-osteoporosis medication to patients aged ≥65 years with new fragility hip fractures, were retrieved from the Hospital Authority Clinical Data Analysis and Reporting System. Patients with hip fracture were identified using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes 81.52, 81.51, 81.40, 79.15, 79.35, or 78.55 under subdivision Operation Theatre Management System–linked diagnosis. Patients who took anti-osteoporosis medication before the fracture and those with pathological fractures were excluded. For the remaining patients who were eligible for secondary drug prevention, we determined the intervention rate each year by determining the percentage of patients receiving anti-osteoporosis medication within 1 year after hip fracture. Version 4 of the strengthening the reporting of observational studies in epidemiology (STROBE) guidelines for cross-sectional studies was used in the preparation of this manuscript.

A total of 15 866 patients with osteoporotic hip fracture who met the criteria were included. From records on anti-osteoporosis medicine initiation, the intervention rate between 2009 and 2012 was found to be different each year, from as low as 9% in 2010 and as high as 15% in 2009 (Fig). The prescription rate for anti-osteoporosis medication was 14.78% in 2009, 25.03% in 2010, 9.24% in 2011, and 11.26% in 2012. Among the specialties prescribing anti-osteoporosis medication, orthopaedic surgeons initiated 63% of the prescriptions, whereas physicians initiated 37%. The anti-osteoporosis drugs prescribed in descending order were alendronic acid (76%), ibandronic acid (12%), strontium ranelate (5%), zoledronic acid (4%), risedronic acid (1%), teriparatide (1%), and denosumab (1%). The rate of anti-osteoporosis medication prescription was between 7% and 31% among the seven public acute hospitals with orthopaedic emergency admission included in the study.

A 2015 study of geriatric hip fractures showed that there had been a steady increase in the incidence of geriatric hip fracture in Hong Kong.1 The worldwide incidence of geriatric hip fractures is also projected to increase.9 We expect to see more patients with fragility fractures in our daily practice with the growing ageing population.

Patients with geriatric hip fracture carry a high mortality rate; the overall 30-day mortality is 3.01% and 1-year mortality is 18.56%.1 Older age and male sex are associated with an increase in mortality and a higher excess mortality rate following surgery.1 Patients with a second episode of hip fracture have been found to have an even higher mortality rate.4 By initiating anti-osteoporosis medication, those subsequent fragility fractures could be prevented.

The British Orthopaedic Association sets standards for surgeons to comply with in order to improve the quality and outcomes of care and also to reduce costs.2 Bone health management includes calcium and vitamin D supplement, osteoporosis treatment, and bone densitometry measurement. According to the American Society for Bone and Mineral Research Task Force 2012, patients with hip fracture should receive pharmacological treatment to prevent additional fractures, because they are clearly at risk for recurrent hip or other osteoporotic fractures, and initiation of bisphosphonate therapy after hip fracture has been shown to reduce the risk of a second hip fracture.10

The main limitation of the present study was that the data were mainly retrieved from a database of patient records. The accuracy of these records depends on the correct entry by clinicians of the diagnosis of hip fracture. Another limitation is that the government drug dispensation record does not included data from patients who choose to receive anti-osteoporosis medication in the private sector. This may create an underestimation of the treatment rate.

Although the treatment rate may have been underestimated in the present study, worldwide rates of osteoporosis treatment after hip fracture have been reported to be as low as 10% to 20% within 1 year.11 12 13 14 15 16 17 18 19 20 A recent study in Hong Kong showed that 33% of patients with hip fracture were prescribed medication for osteoporosis in the 6 months after discharge from the hospital.21 There are also wide discrepancies in drug prescription rates among different hospitals.

There are several potential reasons for these differences in drug prescription rates among hospitals. Firstly, different hospitals follow different working guidelines for the treatment of osteoporosis after hip fracture. Without standardisation of the guidelines, there can be a lack of clarity regarding the responsibility to undertake this care. Siris et al14 found that some physicians did not realise the significance of the initiation of anti-osteoporosis medication after fragility fractures, causing underdiagnosis and undertreatment of osteoporosis. Secondly, some clinicians refer patients to physicians for initiating osteoporosis treatment; especially in centres without geriatric support, these follow-up appointments with physicians can be up to 1 year after discharge from the hospital. Thirdly, many geriatric patients may have renal failure and may be contra-indicated for certain first-line anti-osteoporosis medication such as bisphosphonates. They may be unable to afford other more expensive self-financed anti-osteoporosis medication. Other factors that affect prescription rates include concerns about medication, and the available time and funds for diagnosis and treatment.13

The prevalence of femoral neck osteoporosis based on hip T-score of less than -2.5 was 47.8% in men and 59.1% in women in a Hong Kong study of 239 geriatric hip fractures.21 In the present study, the intervention rate each year was found to be only 9% to 15% across 2009 to 2012. There is obviously still a huge post-fracture gap in secondary prevention. Many patients with fragility fracture do not receive osteoporosis treatment for >1 year after hip fracture. Furthermore, there was little to no improvement in the prescription rates among the 4 years studied. Huge improvements could be achieved by raising the awareness of secondary drug prevention of osteoporosis and increasing the motivation of physicians.

Improvements can only be achieved with involvement of both the government and the individual specialties. The government should allocate more resources and implement a best practice framework for patients with hip fracture at high risk of secondary fracture. The government should also subsidise more anti-osteoporosis medications, so that better treatment can be provided in complicated and severe cases. Because the treatment of osteoporosis differs among hospitals and specialties, a fragility fracture committee or a fracture liaison service can coordinate and standardise patient care by setting up and implementing an easy-to-follow protocol. More education on the treatment of osteoporosis should be provided for orthopaedic and medical departments, to raise awareness and update the relevant knowledge in anti-osteoporosis medication advancement. In some complicated cases of osteoporosis, the involvement of different specialties is essential. The formation of geriatric-orthopaedic working groups and their early involvement in the perioperative and postoperative period can help ensure that optimal care is provided to all patients. Even with anti-osteoporosis medication, a good rehabilitation programme with fall prevention is required; this should be set up in collaboration with allied health professionals. With cooperation between the government and different hospital specialties, more secondary fragility fractures can be prevented. Patients will benefit from prevention of the morbidity and mortality associated with secondary fragility fracture.

Recently there has been debate on osteoporosis treatment and atypical femur fractures. Modi et al22 report that adherence to oral bisphosphonates is low, estimating that, of patients who are prescribed oral bisphosphonates, fewer than 40% are still taking them after 1 year. Although atypical femur fractures have been reported at very low frequencies, not only with bisphosphonate use but also following treatment with denosumab,23 patients are becoming increasingly reluctant to take anti-osteoporosis medication. An analysis of three randomised controlled trials of bisphosphonates concluded that treating 1000 women with osteoporosis for 3 years with a bisphosphonate will prevent approximately 100 vertebral or non-vertebral fractures (number needed to treat: 10).24 Importantly, for the 100 fractures prevented, bisphosphonates might cause 0.02 to 1.25 atypical femur fractures, assuming the relative risk ranges from 1.2 to 11.8 (number needed to harm: 800 to 43 300).25 Hence the beneficial effect of osteoporosis treatment still outweighs the risk for atypical femur fracture.

In Hong Kong, about 98% of all hospital admissions for hip fracture were admitted to public hospitals rather than private hospitals.8 Public hospitals in Hong Kong face a huge financial burden and lack of health care resources for providing optimal care to the ageing population. The cost associated with the prescription of anti-osteoporosis medication is of concern of the government. However, the tremendous hospital expenditure related to hip fracture care can be easily overlooked. In Hong Kong, the direct medical cost for each hip fracture was US$8831.9 in 2018, with the projected direct cost of US$84.7 million in total.26 In 2014, 84% of the drugs prescribed for osteoporosis were bisphosphonates.27 The annual cost of prescription of bisphosphonates per patient was approximately HK$174. Although multiple patients must be treated to prevent a single fracture, reducing the number of subsequent osteoporotic fractures can help the government to achieve significant cost-savings.

Despite the numerous benefits of anti-osteoporosis medication for patients with fragility fractures, the prescription rate remains low not only in Hong Kong, but also in the other parts of the world. Physicians should be aware of the benefits of anti-osteoporosis medication for patients with fragility fractures and guidelines for osteoporosis treatment should be developed and used more widely.

There is a large post-fracture care gap in secondary drug prevention for patients with osteoporotic hip fracture in Hong Kong. The majority of the patients are neither diagnosed nor tested for osteoporosis. Most remained untreated for 1 year after the osteoporotic hip fracture. The Hong Kong Hospital Authority needs to allocate more resources to implement a best practice framework for patients with hip fracture at high risk of secondary fracture, so that they receive appropriate anti-osteoporosis medication. By reducing the number of subsequent osteoporotic fractures, the Hospital Authority can realise substantial cost-savings.

Concept and design: All authors.
Acquisition of data: MY Cheung, AWH Ho.
Analysis or interpretation of data: MY Cheung, AWH Ho.
Drafting of the article: MY Cheung, AWH Ho.
Critical revision for important intellectual content: MY Cheung, AWH Ho.

All authors have disclosed no conflicts of interest. All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity. An earlier version of this paper was presented as a poster at the Annual Congress of the Hong Kong Orthopaedic Association, 6 to 8 November 2015, Hong Kong; at the World Congress on Osteoporosis, Osteoarthritis and Musculoskeletal Diseases, 26 to 29 March 2015, Milan, Italy; and at the 15th Regional Osteoporosis Conference, 24 to 25 May 2014, Hong Kong.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

1. Man LP, Ho AW, Wong SH. Excess mortality for operated geriatric hip fracture in Hong Kong. Hong Kong Med J 2016;22:6-10. Crossref

2. British Orthopaedic Association. The Care of Patients with Fragility Fracture. London: British Orthopaedic Association; 2007.

3. Johnell O, Kanis JA, Odén A, et al. Fracture risk following an osteoporotic fracture. Osteoporos Int 2004;15:175-9. Crossref

4. Lee YK, Ha YC, Yoon BH, Koo KH. Incidence of second hip fracture and compliant use of bisphosphonate. Osteoporos Int 2013;24:2099-104. Crossref

5. Lyles KW, Colón-Emeric CS, Magaziner JS, et al. Zoledronic acid and clinical fractures and mortality after hip fracture. N Engl J Med 2007;357:1799-809. Crossref

6. Harris ST, Watts NB, Genant HK, et al. Effects of risedronate treatment on vertebral and nonvertebral fractures in women with postmenopausal osteoporosis: a randomized controlled trial. Vertebral Efficacy with Risedronate Therapy (VERT) Study Group. JAMA 1999;282:1344-52. Crossref

7. Black DM, Cummings SR, Karpf DB, et al. Randomised trial of effect of alendronate on risk of fracture in women with existing vertebral fractures. Fracture Intervention Trial Research Group. Lancet 1996;348:1535-41. Crossref

8. Chau PH, Wong M, Lee A, Ling M, Woo J. Trends in hip fracture incidence and mortality in Chinese population from Hong Kong 2001-09. Age Ageing 2013;42:229-33. Crossref

9. Gullberg B, Johnell O, Kanis JA. World-wide projections for hip fracture. Osteoporos Int 1997;7:407-13. Crossref

10. Eisman JA, Bogoch ER, Dell R, et al. Making the first fracture the last fracture: ASBMR task force on secondary fracture prevention. J Bone Miner Res 2012;27:2039-46. Crossref

11. Andrade SE, Majumdar SR, Chan KA et al. Low frequency of treatment of osteoporosis among postmenopausal women following a fracture. Arch Intern Med 2003;163:2052-7. Crossref

12. Feldstein A, Elmer PJ, Orwoll E, Herson M, Hillier T. Bone mineral density measurement and treatment for osteoporosis in older individuals with fractures: a gap in evidence-based practice guideline implementation. Arch Intern Med 2003;163:2165-72. Crossref

13. Elliot-Gibson V, Bogoch ER, Jamal SA, Beaton DE. Practice patterns in the diagnosis and treatment of osteoporosis after a fragility fracture: a systematic review. Osteoporos Int 2004;15:767-78. Crossref

14. Siris ES, Bilezikian JP, Rubin MR, et al. Pins and plaster aren’t enough: a call for the evaluation and treatment of patients with osteoporotic fractures. J Clin Endocrinol Metab 2003;88:3482-6. Crossref

15. Kamel HK, Hussain MS, Tariq S, Perry HM, Morley JE. Failure to diagnose and treat osteoporosis in elderly patients hospitalized with hip fracture. Am J Med 2000;109:326-9. Crossref

16. Torgerson DJ, Dolan P. Prescribing by general practitioners after an osteoporotic fracture. Ann Rheum Dis 1998;57:378-9. Crossref

17. Follin SL, Black JN, McDermott MT. Lack of diagnosis and treatment of osteoporosis in men and women after hip fracture. Pharmacotherapy 2003;23:190-8. Crossref

18. Juby AG, De Geus-Wenceslau CM. Evaluation of osteoporosis treatment in seniors after hip fracture. Osteoporos Int 2002;13:205-10. Crossref

19. Harrington JT, Broy SB, Derosa AM, Licata AA, Shewmon DA. Hip fracture patients are not treated for osteoporosis: a call to action. Arthritis Rheum 2002;47:651-4. Crossref

20. Gardner MJ, Brophy RH, Demetrakopoulos D, et al. Interventions to improve osteoporosis treatment following hip fracture. A prospective randomized trial. J Bone Joint Surg Am 2005;87:3-7. Crossref

21. Ho AW, Lee MM, Chan EW, et al. Prevalence of pre-sarcopenia and sarcopenia in Hong Kong Chinese geriatric patients with hip fracture and its correlation with different factors. Hong Kong Med J 2016;22:23-9. Crossref

22. Modi A, Siris ES, Tang J, Sen S. Cost and consequences of noncompliance with osteoporosis treatment among women initiating therapy. Curr Med Res Opin 2015;31:757-65. Crossref

23. Selga J, Nuñez JH, Minguell J, Lalanca M, Garrido M. Simultaneous bilateral atypical femoral fracture in a patient receiving denosumab: case report and literature review. Osteoporos Int 2016;27:827-32. Crossref

24. Black DM, Kelly MP, Genant HK, et al. Bisphosphonates and fractures of the subtrochanteric or diaphyseal femur. N Engl J Med 2010;362:1761-71. Crossref

25. Black DM, Rosen CJ. Clinical practice. Postmenopausal osteoporosis. N Engl J Med 2016;374:254-62. Crossref

26. Cheung CL, Ang SB, Chadha M, et al. An updated hip fracture projection in Asia: The Asian Federation of Osteoporosis Societies study. Osteoporos Sarcopenia 2018;4:16-21. Crossref

27. Ho KC, Tsoi MF, Cheung TT, Cheung CL, Cheung BM. Increase in prescriptions for osteoporosis and reduction in hip fracture incidence in Hong Kong during 2005-2014. Hong Kong Med J 2016;22(Suppl 1):25S.

Congenital long QT syndrome (LQTS) is a life-threatening cardiac arrhythmia syndrome, which leads to sudden death in young people.1 Congenital LQTS is characterised by prolonged QT interval (QTc) on electrocardiogram (ECG) and occurrence of syncope or cardiac arrest. During the past two decades, there have been major advancements in the understanding of the genetic factors underlying the clinical manifestations, prognosis, and subtype-specific therapy of LQTS.1 2 3Seventeen disease-causing LQTS genes have been identified, each of which can lead to dysfunction in the potassium, calcium, or sodium cardiac ion channels.4 In this study, we review the clinical characteristics, genetic profile, management strategy, and outcome of our local LQTS paediatric patients.

Our study included all children, adolescents, and young adults diagnosed with congenital LQTS from January 1997 to December 2016 in the Department of Paediatric Cardiology, Queen Mary Hospital, which is the only tertiary paediatric cardiology referral centre in Hong Kong. All except three patients were Chinese. Diagnosis of congenital LQTS was reviewed with reference to the 2015 European Society of Cardiology guidelines for the management of patients with ventricular arrhythmias and the prevention of sudden cardiac death.5 Long QT syndrome was diagnosed with either corrected QTc ≥480 ms in repeated 12-lead ECG measurements or LQTS risk score >3 points, as proposed by Schwartz et al.6 The presence of a confirmed pathogenic LQTS mutation, irrespective of the QTc, was also used for diagnosis of LQTS. Secondary causes of LQTS were excluded.

Demographic data, clinical presentation, family history, QTc at presentation, and genetic tests were retrospectively reviewed. Clinically important presentation was defined as clinical symptoms or rhythm disturbances that warranted concern. Incidental findings of isolated prolonged QTc were not regarded as clinically important presentation.

We reviewed the treatment modalities, including beta blocker therapy, implantable cardioverter defibrillator (ICD), permanent pacemaker, and left cardiac sympathetic denervation. Clinical outcomes up to December 2016 were summarised. Syncope, seizure, aborted cardiac arrest, appropriate ICD shock, or sudden cardiac death after diagnosis of LQTS were considered as breakthrough cardiac events.

Genetic tests were offered to all patients after informed consent was provided by their parents or guardians. Blood samples were sent to the Molecular Genetics Laboratory of Victorian Clinical Genetic Services, Melbourne, Australia, for genetic testing. Before 2014, six common LQTS genes were tested (KCNQ1, KCNH2, SCN5A, KCNE1, KCNE2, KCNJ2) by sequencing of the entire coding region of all known transcripts of the genes. Multiplex ligation-dependent probe amplification analysis was performed on five genes (KCNQ1, KCNH2, SCN5A, KCNE1, KCNE2) to detect deletions or duplications. After 2014, next-generation sequencing was used to identify mutations in an arrhythmia gene panel to replace sequencing of the six LQTS genes (details of the arrhythmia full panel can be found at http://www.vcgs.org.au/tests/cardiac-gene-panels). Before referral to our unit, 13 patients had genetic tests performed by local or overseas genetic testing centres.

Mutations in the LQTS loci classified as pathogenic or likely pathogenic were considered as genotype positive in our cohort. Cascade testing was offered to first-degree relatives of patients identified as genotype positive.

Statistical analysis was performed with the SPSS (Window version 17.0; SPSS Inc, Chicago [IL], United States). Continuous variables were expressed as mean ± standard deviation, median and range. We used the standard t test for comparisons of continuous data. The P<0.05 were deemed statistically significant. Kaplan-Meier survival curves were created with censoring at first breakthrough cardiac event or last follow-up, and analysis was made using the log rank test.

During the study period, 59 patients (33 males) in 52 families were identified who fulfilled the diagnostic criteria as described. Nine individuals were diagnosed by ECG screening of our index cases including a 25-year-old young adult. The mean follow-up duration of the cohort was 5.33 ± 4.65 years. Four patients were lost to follow-up, but no death was reported in the territory-wide Hospital Authority electronic patient record. Five patients were under the care of adult cardiologists at the follow-up.

Table 1 shows the characteristics of our patients. The mean age at diagnosis was 8.12 ± 5.19 years (range, 0-25 years). For those patients who had clinically important presentation, the mean age at presentation was 8.75 ± 5.12 years (range, 0.00-16.95 years). Boys were younger than girls at presentation (6.39 ± 5.00 years vs 10.51 ± 4.55 years, P=0.016).

The mean QTc of the cohort was 504 ± 47 ms. The median Schwartz score was 4 points (range, 1-6 points). Index patients had longer mean QTc (512 ± 46 ms) when compared with screened family members (462 ± 25 ms, P=0.002).

Five (8.47%) patients had congenital heart defects: secundum atrial septal defect (n=1; genotype negative), ventricular septal defect (n=1; LQTS type 2 [LQT2]), tetralogy of Fallot (n=2; LQT2 and LQTS type 8 [LQT8]), and transposition of great arteries (n=1; genotype negative). One patient had bilateral sensorineural hearing loss and was subsequently confirmed to have LQTS type 1 (LQT1).

Figure 1 illustrates the mode of initial presentation of our patients. Forty-four patients had clinically important presentation at diagnosis. Syncope without convulsion was the most common mode of presentation (37.3%), among which around 40% of cases were stress-related. Convulsion was also a common symptom (12%). Aborted cardiac arrest occurred in six (10.2%) individuals.

Three patients presented with sinus bradycardia, one of whom was detected prenatally. Four patients, including three infants, had 2:1 atrioventricular (AV) block at diagnosis. A significant proportion of patients with LQTS (25.4%) did not have clinically important presentation at diagnosis; most of them had incidental ECG findings of prolonged QTc during medical check-ups or were identified by family cascade screening.

The clinical outcomes of patients with LQTS in our cohort with clinically important presentation are summarised in Figure 2.

All patients were offered beta blocker therapy. However, 14 patients refused beta blocker (11 patients had clinically important presentation). Seven patients were on beta blocker but stopped subsequently.

Metoprolol was the initial choice of beta blocker for most of our patients (n=27), whereas 10 patients had atenolol as initial choice. Propranolol was used in eight infants. Eight patients receiving metoprolol, atenolol, or propranolol later changed to nadolol to enhance compliance or for better control of breakthrough symptoms. Mexiletine was added as an adjuvant therapy for five patients with LQTS type 3 (LQT3) who were symptomatic. Among the 31 patients with an initial history of convulsion, syncope, or dizziness (mean follow-up duration, 4.81 ± 3.84 years), 21 became asymptomatic after medication and/or lifestyle modification. Two patients who had recurrent symptoms after initial beta blocker therapy became event-free after a change from metoprolol/atenolol to nadolol. Patients without clinically important presentation at diagnosis remained asymptomatic with or without beta blocker therapy, irrespective of presence of documented pathogenic mutation.

Implantable cardioverter defibrillator was implanted in 13 patients, whose mean QTc was 501 ± 42 ms. Six of these patients had initially presented with aborted ventricular tachycardia (VT)/ventricular fibrillation (VF) arrest. Five patients received ICD implantation because of recurrent symptom or subsequent VT/VF despite beta blocker therapy. Two of these five patients experienced appropriate shocks after ICD implantation. One patient with pathogenic KCNE1 mutation had syncope due to sinus arrest with long pauses; ICD was offered for primary prevention of sudden death in addition to pacing therapy. One patient who had AV block at birth developed subsequent unprovoked syncope at aged 6 years despite medical treatment and his pacing system was upgraded to ICD. In total, four patients experienced appropriate ICD shocks despite beta blocker treatment. There were no more ICD shocks after reinforcement of medication compliance, adjustment of dosage, and in one patient switching of metoprolol to nadolol.

Left cardiac sympathetic denervation was performed via video-assisted thoracoscopic approach in two patients, together with ICD therapy. Both of them were free of cardiac events on follow-up.

Pacemakers were implanted in five patients. Four of these five patients had functional AV block due to prolonged QTc. Normal AV node conduction recovered with time in these four children. The fifth patient had complete heart block after surgical repair of congenital heart condition (transposition of great arteries with ventricular septal defect).

The Kaplan-Meier survival curve is shown in Figure 3. Overall, the breakthrough cardiac event–free survival was 93.0% ± 0.034% at 1 year, 80.7% ± 0.065% at 5 years, and 72.6% ± 0.080% at 10 years for the entire cohort. Patients who had clinically important presentation at baseline had a higher risk of developing breakthrough cardiac events (P=0.048) compared with those who did not. There was no mortality during the study period.

All but four patients underwent genetic testing. Testing was not possible in two patients owing to loss to follow-up before genetic testing could be offered; these two patients were strong phenotypes of LQTS with Schwartz scores of 4 and 5 points, respectively. The other two patients were first-degree relatives of confirmed genotype-positive patients.

Seven patients were genotype negative. Four of them were tested before 2014. Nine patients had their mutated genes classified as variants of unknown significance. We also included three patients (LQT1, n=2; LQT2, n=1), reported by Mak et al,7 whose genetic tests were performed in a local laboratory.

Thirty-eight (69.1%) patients among those tested were confirmed to have pathogenic mutations for LQTS (Table 2). Eight mutations were novel (8/33, 24.2%). Most of the pathogenic mutations were missense mutation (30/33, 90.9%). Ten (16.9%) patients had LQT1 (KCNQ1). Twelve (20.3%) patients had LQT2 (KCNH2), whereas seven (11.9%) patients had LQT3 (SCN5A). Three patients had LQTS type 5 (KCNE1). There were five patients (three families) with LQT8 resulting from a rare pathogenic mutation in CACNA1C. LQT8 is linked to Timothy syndrome with multiple extracardiac manifestations. Only one of our patients with LQT8 had classical features of Timothy syndrome, with syndactyly, developmental delay, and congenital heart defect (tetralogy of Fallot). One patient had LQTS type 16 (CALM3). Figure 1 shows the details of the genotype information of LQTS genes identified.

Clinical characteristics of patients with LQT1, LQT2, or LQT3 are detailed in Figure 4. Syncope with stress was the predominant presentation in patients with LQT1 (60%). In contrast, syncope without stress was the main form of presentation in patients with LQT2 (41.7%). In patients with LQT3, around 30% had an initial presentation of aborted VF cardiac arrest. A high proportion (50%) of patients with LQT3 developed subsequent VT/VF despite medical therapy, compared with 14.3% in patients with LQT1 and 28.6% in patients with LQT2.

Nine patients were identified to have LQTS during family screening of index patients. Two patients had strong phenotypes but did not have genetic tests. Five patients had LQTS phenotypes with confirmation by genetic testing. Two asymptomatic children in the same family were referred to us and were identified by cascade screening of their father’s pathogenic mutation.

Long QT syndrome is a rare inherited disorder associated with an increased propensity to polymorphic VT/VF, syncope, and sudden cardiac death. In this report, we describe the clinical features and genetic profile of 59 LQTS paediatric patients managed in the only tertiary paediatric cardiology referral centre in Hong Kong over 20 years.

The prevalence of childhood LQTS was estimated to be 1:2000 in an Italian birth cohort.8 In a recent Japanese study, the estimated probability of diagnosing LQTS was 1:3300 in children aged 6 years and 1:1000 in those aged 12 years.9 Our results indicate that the prevalence of diagnosed LQTS in Hong Kong children was less than 1:10 000, suggesting an underdiagnosis of the condition. This is likely due to under-recognition of symptomatic LQTS in young patients who were treated for recurrent seizure and unexplained syncope. Without an ECG screening programme, many asymptomatic LQTS children also remain undiagnosed.

The lack of comprehensive screening for family members of adult LQTS probands is another reason for underdiagnosis, as only two children from a single family were referred to us from adult cardiologists over the 20-year study period. In addition, a 5-year-old child of a mother with known LQTS was not referred until he presented with convulsions.

Molecular autopsy for young victims of sudden cardiac death is not implemented in Hong Kong. In many developed countries, affected family members of LQTS sudden death victims are identified early through this pathway to prevent sudden cardiac death. We believe that the total number of symptomatic and asymptomatic young patients with LQTS in Hong Kong is much higher than what we have studied in our single tertiary referral centre.10 11

Congenital LQTS is usually diagnosed in patients presenting with syncope, unexplained seizure, and aborted cardiac arrest. The mean QTc of our patients was 504 ± 47 ms, and median Schwartz score of the entire cohort was 4 points (range, 1-6 points). This indicates that the patients that were referred to our centre were patients with more severe symptoms, resulting in a higher likelihood of a diagnosis of LQTS, based on clinical criteria.

Previous reports have shown that the risk of clinical events in boys (aged <15 years) with LQTS is significantly higher than that in girls with LQTS.12 In our cohort, we also confirmed that symptomatic boys were significantly younger than girl at diagnosis or presentation.

Sinus bradycardia and functional AV block are well reported in perinatal LQTS.13 14 15 The youngest patient in our group presented with fetal bradycardia. We also noted 2:1 AV block in three infants at presentation. Their AV conduction normalised with a significant decrease in QTc during follow-up. Paediatricians or family doctors should suspect LQTS in young infants with slow heart rates.

Few links between LQTS and congenital heart disease have been reported, apart from Timothy syndrome. A recent single-centre review of 49 LQTS genotype-positive patients identified 11 (22%) cases with concomitant conotruncal anomalies and/or aortic arch anomalies.16 In our cohort, five (8.5%) patients with LQTS had concomitant congenital heart disease. Two cases were diagnosed in the perioperative period. Prolonged QTc in the context of congenital heart disease can be confounded by several factors, including postoperative electromechanical factors, intrinsic, or postoperative QRS abnormalities. Therefore, the diagnosis of LQTS could be masked in patients with congenital heart disease. We suggest that ECG of patients with congenital heart disease should be evaluated carefully for QTc.

Throughout the world, 75% to 80% of patients with LQTS have identifiable genetic mutations, with LQT1, LQT2, or LQT3 accounting for 90% of cases. Pathogenic LQTS genetic mutations were identified in 69.1% of the patients in our cohort who were tested, which is comparable with other LQTS cohorts.17 Similarly, we had predominant genotypes of LQT1 (10/59, 16.9%), LQT2 (12/59, 20.3%), and LQT3 (7/59, 11.9%). In a recent single-centre study of LQTS in China, LQT2 was also the most common genotype.18 We also identified five (8.5%) patients with rare CACNA1C (LQT8) mutations, all of whom were Hong Kong Chinese. Without a study of a large Chinese population in the past 5 years for cross reference, we cannot be certain whether LQT8 is more prevalent in Chinese than in other ethnic groups.

Beta blockers are the mainstay of treatment for all LQTS genotypes. In a registry of 1530 patients with LQTS, all beta blockers seemed equally effective in reducing risk of a first cardiac event after beta blocker initiation. For patients with LQT1, no single type of beta blocker has been found superior, although nadolol was found to be superior for patients with LQT2.19 However, another study suggested that symptomatic LQT1 and LQT2 patients on metoprolol had a higher rate of recurrence of cardiac events.20 In the present study, 21 patients were prescribed metoprolol, two of whom required switching to another beta blocker due to recurrent symptoms. Because of the relatively short follow-up duration and small number of patients in our cohort, it is impossible to conclude on the efficacy of each beta blocker for patients with each genotype. Genotype-guided therapy is advocated in contemporary management in LQTS. Based on the available evidence, we are inclined to use nadolol as our first choice in symptomatic LQT2 patients. In symptomatic LQT3 patients, dual therapy using beta blockers and mexiletine are used. Mexiletine is a sodium channel blocker shown to shorten QTc in LQT3 patients.21 22

In addition to medical therapy with beta blockers, treatment of LQTS can also include lifestyle modifications, sympathetic denervation, and device therapy. With a multi-modality management strategy and genotype-guided therapy, outcome of LQTS have improved markedly over the past decades. The event-free survival was 96% at 1 year, 93% at 5 years, and 90% at 10 years, as reported recently in a large single-centre study which included 83% asymptomatic probands.23 In the present study, the breakthrough cardiac event-free survival was 93.0% at 1 year, 80.7% at 5 years, and 72.6% at 10 years (1). The event rate in the present study is likely higher than that of the abovementioned study because we have a higher proportion of probands (85%), of whom 88% were symptomatic at presentation. Breakthrough cardiac events were mainly related to non-compliance to our treatment advice.

Our study was based on single-hospital data and referral bias is expected. We may have received referral of patients with LQTS with more severe symptoms. In addition, a short duration of follow-up in our patients (mean 5.3 years) may have led to underestimation of clinical cardiac events and mortality.

Our study demonstrated the high yield of genetic testing and the importance of genetic information in predicting the prognosis of patients with LQTS and guiding their treatment. Early identification of affected family members through cascade screening of mutated gene was also demonstrated. We hope that public genetic services can continue to develop, to enable genetic testing to be offered to all patients with suspected channelopathies. We also advocate the establishment in Hong Kong of inherited arrhythmia clinics or cardiac genetic clinics in the public sector, as implemented in many other countries. Such clinics have proven effectiveness in reducing sudden cardiac death associated with inherited arrhythmia syndrome.24

Our study provides insight into the clinical and molecular profiles of young patients with LQTS in the only tertiary paediatric cardiology referral centre in Hong Kong. The LQT1, LQT2 and LQT3 genotypes are the most common in mutation-positive patients. Early identification of LQTS, administration of beta blocker therapy, device therapy, and lifestyle modifications can prevent sudden cardiac death. However, the breakthrough cardiac event survival was only 72.6% at 10 years. Further optimisation of the treatment strategy by genotype-guided therapy may reduce recurrent symptoms and improve prognosis.

Concept and design: APY Liu, BHY Chung, TC Yung.
Acquisition of data: SY Kwok, CYY Chan, TC Yung.
Analysis or interpretation of data: SY Kwok, CYY Chan, JLF Fung, CCY Mak.
Drafting of the article: SY Kwok, APY Liu.
Critical revision for important intellectual content: BHY Chung, KS Lun, TC Yung.

We would like to express our gratitude to all paediatricians and physicians who referred patients with LQTS to our department, and to the adult cardiologists who cared for our grown-up patients. We are also grateful for the contributions of the local university laboratories and the pathology departments of Hospital Authority hospitals in conducting genetic tests on some of our patients. We would like to express our gratitude to Ms Yo-yo WY Chu for her participation in providing genetic services to our patients.

All authors have disclosed no conflicts of interest. All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

Results of this study were presented in the following meetings: (1) World Congress of Pediatric Cardiology & Cardiac Surgery 2017, Barcelona, Spain, 16-21 Jul 2017; (2) The 13th Congress of Asian Society for Pediatric Research 2017, Hong Kong, 6-8 Oct 2017; (3) The 26th Annual Scientific Congress, Hong Kong College of Cardiology 2018, Hong Kong, 15-17 Jun 2018. Abstract of this study was published in the Journal of the Hong Kong College of Cardiology (Kwok SY, Liu AP, Lun KS, et al. Clinical and genetic profile of congenital long QT syndrome in Hong Kong—18-year experience in pediatrics. J HK Coll Cardiol 2018;26:60).

The expenses of the genetic analysis used in our study were sponsored by the Children’s Heart Foundation of Hong Kong.

This study received ethics approval from the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong Western Cluster.

1. Hobbs JB, Peterson DR, Moss AJ, et al. Risk of aborted cardiac arrest or sudden cardiac death during adolescence in the long-QT syndrome. JAMA 2006;296:1249-54. Crossref

2. Weintraub RG, Gow RM, Wilkinson JL. The congenital long QT syndromes in childhood. J Am Coll Cardiol 1990;16:674-80. Crossref

3. Mizusawa Y, Horie M, Wilde A. Genetic and clinical advances in congenital long qt syndrome. Circ J 2014;78:2827-33. Crossref

4. Giudicessi JR, Ackerman MJ. Calcium revisited: new insights into the molecular basis of long-QT syndrome. Circ Arrhythm Electrophysiol 2016;9:e002480. Crossref

5. Priori SG, Blomström-Lundqvist, Mazzanti A, et al. 2015 ESC Guidelines for the management of patients with ventricular arrhythmias and the prevention of sudden cardiac death: The Task Force for the Management of Patients with Ventricular Arrhythmias and the Prevention of Sudden Cardiac Death of the European Society of Cardiology (ESC). Endorsed by: Association for European Paediatric and Congenital Cardiology (AEPC). Eur Heart J 2015;36:2793-867. Crossref

6. Schwartz PJ, Crotti L. QTc behavior during exercise and genetic testing for the long-QT syndrome. Circulation 2011;124:2181-4. Crossref

7. Mak CM, Chen SP, Mok NS, et al. Genetic basis of channelopathies and cardiomyopathies in Hong Kong Chinese patients: a 10-year regional laboratory experience. Hong Kong Med J 2018;24:340-9. Crossref

8. Schwartz PJ, Stramba-Badiale M, Crotti L, et al. Prevalence of the congenital long-QT syndrome. Circulation 2009;120:1761-7. Crossref

9. Yoshinaga M, Kucho Y, Nishibatake M, Ogata H, Nomura Y. Probability of diagnosing long QT syndrome in children and adolescents according to the criteria of the HRS/EHRA/APHRS expert consensus statement. Eur Heart J 2016;37:2490-7. Crossref

10. Kwok SY, Pflaumer A, Pantaleo SJ, Date E, Jadhav M, Davis AM. Ten-year experience in atenolol use and exercise evaluation in children with genetically proven long QT syndrome. J Arrhythmia 2017;33:624-9. Crossref

11. Marcondes L, Crawford J, Earle N, et al. Long QT molecular autopsy in sudden unexplained death in the young (1-40 years old): lessons learnt from an eight year experience in New Zealand. PLoS One 2018;13:e0196078. Crossref

12. Goldenberg I, Moss AJ. Long QT syndrome. J Am Coll Cardiol 2008;51:2291-300. Crossref

13. Cuneo BF, Strasburger JF, Yu S, et al. In utero diagnosis of long QT syndrome by magnetocardiography. Circulation 2013;128:2183-91. Crossref

14. Mitchell JL, Cuneo BF, Etheridge SP, Horigome H, Weng HY, Benson DW. Fetal heart rate predictors of long QT syndrome. Circulation 2012;126:2688-95. Crossref

15. Horigome H, Nagashima M, Sumitomo N, et al. Clinical characteristics and genetic background of congenital long-QT syndrome diagnosed in fetal, neonatal, and infantile life a nationwide questionnaire survey in Japan. Circ Arrhythmia Electrophysiol 2010;3:10-7. Crossref

16. Ebrahim MA, Williams MR, Shepard S, Perry JC. Genotype positive long QT syndrome in patients with coexisting congenital heart disease. Am J Cardiol 2017;120:256-61. Crossref

17. Schwartz PJ, Priori SG, Spazzolini C, et al. Genotype-phenotype correlation in the long-QT syndrome: gene-specific triggers for life-threatening arrhythmias. Circulation 2001;103:89-95. Crossref

18. Gao Y, Liu W, Li C, et al. Common genotypes of long QT syndrome in China and the role of ECG prediction. Cardiology 2016;133:73-8. Crossref

19. Abu-Zeitone A, Peterson DR, Polonsky B, McNitt S, Moss AJ. Efficacy of different beta-blockers in the treatment of long QT syndrome. J Am Coll Cardiol 2014;64:1352-8. Crossref

20. Chockalingam P, Crotti L, Girardengo G, et al. Not all beta-blockers are equal in the management of long QT syndrome types 1 and 2 higher recurrence of events under metoprolol. J Am Coll Cardiol 2012;60:2092-9. Crossref

21. Wilde AA, Moss AJ, Kaufman ES, et al. Clinical aspects of type 3 long-QT syndrome: an international multicenter study. Circulation 2016;134:872-82. Crossref

22. Ruan Y, Liu N, Bloise R, Napolitano C, Priori SG. Gating properties of SCN5A mutations and the response to mexiletine in long-QT syndrome type 3 patients. Circulation 2007;116:1137-44. Crossref

23. Rohatgi RK, Sugrue A, Bos JM, et al. Contemporary outcomes in patients with long QT syndrome. J Am Coll Cardiol 2017;70:453-62. Crossref

24. Adler A, Sadek MM, Chan AY, et al. Patient outcomes from a specialized inherited arrhythmia clinic. Circ Arrhythm Electrophysiol 2016;9:e003440. Crossref

In 1970, Nagel et al1 first reported the transmission of single-lead electrocardiogram (ECG) data via radio system to hospital physicians in Florida, US, for diagnostic purposes. To date, prehospital 12-lead ECG programmes have been implemented in various countries. In 2015, the American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care recommended that prehospital 12-lead ECG should be acquired early for patients with possible acute coronary syndrome.2 The acquisition and tele-transmission of ECG data to the accident and emergency department (AED) allows rapid diagnosis of ST-segment elevation myocardial infarction (STEMI) by attending physicians. The on-call cardiologist and cardiac catheterisation laboratory (CCL) can then be notified early. This significantly shortens the door-to-balloon (D2B) time and improves patient outcome.3

The Hong Kong Fire Services Department (HKFSD) is the major emergency ambulance service provider in Hong Kong. Ambulances are deployed by the Fire Service Control Centre after calls to 999 and are manned by ambulance crew in accordance with pre-set protocols approved by the Medical Director of the HKFSD. Patients are transported to the nearest public hospital based on their geographical location. There is no choice with respect to the destination hospital, and there was no primary diversion for chest pain or STEMI in the project period.

Queen Mary Hospital (QMH) is a tertiary care hospital providing 24-hour emergency medical services in Hong Kong. Since November 2010, QMH has been the only public hospital in Hong Kong providing a 24-hour primary percutaneous coronary intervention (PPCI) service for all patients with STEMI.

In November 2015, the HKFSD and the QMH AED jointly launched a pilot project named ‘Prehospital Ambulance 12-Lead Electrocardiogram for Chest Pain Patients in Hong Kong West Cluster’. The present paper reports on the results of the first phase of that project. We hypothesised that prehospital ECG would shorten the D2B time and reduce mortality in patients with STEMI treated in AEDs in Hong Kong.

The present retrospective observational study analysed data from the ‘Prehospital Ambulance 12-Lead Electrocardiogram for Chest Pain Patients in Hong Kong West Cluster’ pilot project. That project began on 12 November 2015 and its first phase ended on 31 December 2016. All 122 ambulance crew members involved with the study received a half-day theory and hands-on training by clinical specialists of Zoll Medical Corporation (Chelmsford [MA], US) regarding the performance of prehospital 12-lead ECG. During the project period, there were 20 HKFSD ambulances operating in the catchment area of QMH; 10 were not operated during the night shift. Of these 20 ambulances, 15 were equipped with X Series® Monitor/Defibrillator (Zoll Medical Corporation) with ECG and tele-transmission functions. These 15 ambulances were deployed by the HKFSD to Aberdeen Ambulance Depot, Pok Fu Lam Ambulance Depot, Mount Davis Ambulance Depot, and Sheung Wan Fire Station.

In addition to the conventional management of chest pain or discomfort of cardiac origin, ambulance crews performed prehospital ECG (Fig). The ambulance crews were trained to use a standardised script in Chinese or English to explain the indication, procedure, benefit, and risk of performing prehospital ECG. Patients and accompanying relatives were provided with sufficient time for questions before informed consent was obtained. Patients were excluded if they were under 12 years of age; were in cardiac arrest; exhibited airway or breathing that could not be managed; had Glasgow Coma score ≤13; had first systolic blood pressure <90 mm Hg; had respiratory rate <10 or >29 breaths per minute; or were otherwise unable to consent.

The prehospital ECG was obtained in the ambulance compartment before its departure from the scene and was immediately tele-transmitted to the AED for interpretation by a physician. When a new prehospital ECG was transmitted, staff at the AED were notified via a dedicated laptop with alarm, fax, and email, as well as an alert phone call from the ambulance crew en route. The patient’s Hong Kong Identity Card number was also conveyed through the alert phone call from the ambulance. This allowed early identification of any new changes in the prehospital ECG, compared with previous ECG data stored in the hospital clinical management system record.

The most senior attending AED physician available was responsible for reading and interpreting the prehospital ECG. If the AED physician identified ST-segment elevation in two or more contiguous leads, the on-call cardiologist was paged and AED manpower and equipment were arranged immediately. Patients without such ST-segment elevation were triaged by experienced AED nurses upon arrival, in accordance with their overall clinical condition.

To compare the characteristics and outcomes of patients in this pilot project with those of other patients with chest pain (ie, those attending the AED by ambulance without prehospital ECG, or by self-arranged transport), data from the cardiac care unit of QMH were collected for comparison. Data of all patients attending the AED with emergency coronary angiography, with or without PPCI, were analysed. The D2B or door-to-catheter (D2C) time of different subgroups was determined and analysed as the primary outcome of this study. Secondary outcomes, including performance in daytime or night-time AED registration, triage accuracy, and 30-day mortality, were assessed in further subgroup analysis.

From 12 November 2015 to 31 December 2016, ambulance crews equipped with X Series® attended to 841 patients presenting with cardiac chest pain. Verbal consent to perform prehospital ECG was obtained from 731 (86.9%) patients and these were included for analysis. A mean 1.76 prehospital ECGs were successfully performed and transmitted per day during the pilot project period. Reasons for not performing prehospital ECG are shown in Table 1.

In total, 60% of patients were male. The patients’ mean age was 71 years, with a difference of 10 years between male (mean, 67 years) and female (mean, 77 years) patients. More patients registered at the AED during daytime hours (08:00-17:59; n=380 [52.0%]) than during night-time hours (18:00-07:59; n=351 [48.0%]). Most patients were triaged as category 3 in the AED (n=577) [78.9%]; those triaged as category 1 (critical), 2 (emergency), 4 (semi-urgent), and 5 (non-urgent) were 39 (5.3%), 57 (7.8%), 58 (7.9%), and 0, respectively.

In all, 93% of cases with cardiac chest pain were managed by AED physicians without further on-site consultation. Consultations with physicians from the cardiac care unit were initiated by the AED physician for 53 patients: 22 (41.5%) were performed on or before patient arrival at the AED, whereas the remaining 31 (58.5%) were conducted after patient arrival.

Of the 731 patients who received pre-hospital ECG, 534 (73.1%) patients were admitted to the medical ward, 96 (13.1%) to the emergency medical ward, 33 (4.5%) to the cardiac care unit, 13 (1.8%) to the surgical ward, three (0.4%) to the intensive care unit, and one (0.1%) to the neurosurgical ward for further management of their symptoms. Of the remaining patients, 39 (5.3%) were discharged from the AED, eight (1.1%) were discharged against medical advice, three (0.4%) disappeared, and one (0.1%) was certified in the AED.

Of the 731 patients who received pre-hospital ECG, 26 (3.6%) patients were clinically diagnosed with STEMI by AED physicians. Coronary angiogram, with or without PPCI, was arranged immediately for 25 of these patients. The remaining patient had a terminal malignancy and was offered non-invasive treatment after discussion with the patient’s family.

The mean D2B and D2C times were compared among 124 patients with clinically diagnosed STEMI: 25 patients with STEMI treated with prehospital ECGs; 58 patients with STEMI who were treated by ambulance crews without prehospital ECGs; and 41 self-transported patients with STEMI treated in the QMH AED during the pilot project period (Table 2). A statistically significant difference was found in the mean D2B time (P=0.003). Patients with prehospital ECGs had the shortest mean D2B (93 minutes) and D2C (71 minutes) times. Additionally, a greater percentage of patients with prehospital 12-lead ECG had D2B or D2C time ≤90 minutes.

Table 3 shows that there were significant differences in D2B or D2C times among the three groups of patients with STEMI attending the AED during daytime hours (08:00-17:59); there were no significant differences between the groups during night-time hours. The overall 30-day mortality of all patients with clinically diagnosed STEMI was 8% (10 of 124 patients) [Table 4]. Fisher’s exact test did not demonstrate any statistically significant difference between the intervention and control groups.

When clinical diagnoses of STEMI were used as a standard to verify diagnoses made by the X Series® Monitor/Defibrillator diagnostic algorithm, there were 14 true positive, three false positive, 702 true negative, and 12 false negative cases. This corresponded to sensitivity (53.8%), specificity (99.6%), positive predictive value (82.4%), negative predictive value (98.3%), and accuracy (97.9%).

In an AED without point-of-care cardiac biomarker testing capability, it is impossible to immediately detect a rise and/or fall in cardiac biomarkers, as described in the ‘Fourth universal definition of myocardial infarction’.4 Some patients in the AED with chest discomfort or other ischaemic symptoms, who developed ST elevation in two contiguous leads, were clinically diagnosed with STEMI before confirmation of typical biomarker changes, resulting in corresponding modification to medical management. This caused inaccuracy in calculation of the diagnostic performance of the diagnostic algorithm used in the X Series®, when the clinical diagnosis of STEMI in the AED was used as a standard for comparison.

The theoretical benefits of prehospital 12-lead ECG included early diagnosis by AED physicians, early activation of the Acute Myocardial Infarction Clinical Pathway with cardiologist input, and activation of the CCL. It worked well during office hours, when the full team of interventional cardiologists, cardiac care nurses, and radiographers were on-site. The shortest D2B time was 41 minutes. However, during non-office hours or night-time hours, this team had to return from their homes for PPCI. This was the time-limiting factor during night-time hours and there was no statistically significant difference in night-time D2B or D2C time in our series for ambulance patients with or without prehospital 12-lead ECG. Possible solutions might be the provision of sufficient manpower for a 24-hour on-site interventional cardiology team. Other measures, such as cardiologists receiving prehospital 12-lead ECG data through their mobile phones, or activation of the CCL by emergency physicians (with an increased risk of inappropriate activation) could be considered.

A greater percentage of patients with STEMI with prehospital 12-lead ECGs received reperfusion therapy with PPCI ≤90 minutes (Table 2). There was also a reduction in 30-day mortality to 0% in this subgroup (Table 4). However, whether a causeand- effect relationship existed is uncertain, due to our small sample size. Additionally, a statistically significant difference in 30-day mortality could not be demonstrated in our pilot project. Other cardiovascular outcome measurements, including re-infarction rate, major adverse cardiac events, and heart failure rate could be studied to provide insights regarding the benefit of prehospital 12-lead ECGs.

The D2B or D2C time could be improved through measures such as the performance of prehospital 12-lead ECGs on scene (rather than in-ambulance), or skipping AED consultation with direct cardiac care unit admission for patients suspected of STEMI on prehospital ECGs.5

Inappropriate activation of the CCL occurred when the interventional cardiologist provided an alternative diagnosis or considered the patient not to be a candidate for PPCI, with subsequent cancellation of the catheterisation procedure. In QMH, the CCL was activated by the on-call cardiologist, not the emergency physician, for appropriate patients with STEMI after on-site assessment. Therefore, inappropriate activation of the CCL by the emergency physician was 0% in our series. Instead, false positivity existed. False positivity was defined as patients with absence of thrombus causing obstruction in the culprit vessels and absence of a typical rise and fall of cardiac enzymes. In our pilot project, 17 of 124 patients with STEMI admitted through the AED had an emergency coronary angiogram performed without PPCI. Excluding one patient with triple vessel occlusion who failed PPCI and required emergent coronary artery bypass grafting, this yielded a false positive rate of 12.9% (16 of 124 patients). Prehospital 12-lead ECG extended the emergency assessment of patient condition to the prehospital phase. This allowed comparison of prehospital and AED ECGs. Dynamic ECG changes that occurred during the ambulance journey were more likely to be detected. Whether this might help to reduce the false positive activation of CCL or false negative discharge of patients from AED can be studied in the future.

In our pilot project, all ambulances equipped with X Series® Monitor/Defibrillator served the QMH catchment area without primary diversion. If primary diversion of patients with STEMI to a hospital with 24-hour PPCI service were to be implemented, on-scene ECG interpretation should be highly sensitive to avoid under-diversion. The ECG data could be interpreted by ambulance crews, machine diagnostic algorithm, or emergency physicians or cardiologists through tele-transmission. However, not all ambulance crews in Hong Kong are trained to interpret ECG data. Additionally, when compared with clinical diagnosis of STEMI in the AED, the sensitivity of the diagnostic algorithm used was 53.8% in our pilot project. Given this moderate sensitivity, immediate ECG interpretation by experienced emergency physicians or cardiologists is preferable. This diagnostic process is inevitably time-consuming. If ECG were performed on-scene instead of in the ambulance compartment, transport of the patient from scene to ambulance and the diagnostic process could happen simultaneously. Before ambulance departure from the scene, input from the emergency physician or cardiologist would be readily available to guide appropriate destination AED selection.

Triage by AED nurses must be both efficient and accurate, especially during busy periods. However, chest pain is a subjective perception, as are all types of pain. Before the availability of ECG, the vital signs and brief clinical contact between triage nurses and patients determined the assigned triage category for each patient. Because patients with STEMI belonged to category 1 (critical), it is inevitable that there would be increased under-triage of STEMI cases in the control group (Table 5). Prehospital ECG allowed more accurate triage of patients with STEMI.

In total, 61.3% (76/124) [Table 3] of emergency coronary angiograms with or without PPCI were performed during daytime hours (08:00-17:59). However, this total includes patients who registered during non-office hours on Saturday, Sunday, and public holidays; during some of these, activation of the CCL involved the interventional cardiology team returning from home, as would be necessary during night-time hours. Because night-time D2B or D2C times were consistently longer than those recorded during daytime hours (Table 3), daytime D2B or D2C times were most likely overestimated. Studies have shown that in units with no significant difference in reperfusion time during daytime and night-time hours, the mortality of patients with STEMI was not influenced by whether patients presented during standard working hours or outside of these hours.6 Measures could be implemented to shorten the night-time D2B or D2C time for improved patient outcomes.

The Acute Myocardial Infarction Clinical Pathway was first established in QMH in 2007; by 2011, it successfully reduced the 30-day mortality rate of acute myocardial infarction from 18.4% to 14.9%.7 The mortality in our pilot project was 8.2% in the control group with conventional management (Table 4). This reduction of mortality in the control group was likely a combined result of advances in cardiovascular medications and diagnostic or therapeutic technologies, as well as improvement of patient health awareness in recent years. With implementation of prehospital 12-lead ECG, further reduction of mortality was expected. In our subgroup with prehospital 12-lead ECG and PPCI, 30-day mortality of 0% was recorded. However, a larger study over longer period of time is needed to establish whether there is a statistically significant clinical impact of prehospital 12-lead ECG on long-term mortality.

Approximately one-third (41 of 124) of patients with AED diagnosis of STEMI arranged self-transport, rather than calling an ambulance. This underuse of the ambulance service is not unique to Hong Kong8 9 10and has been a well-documented cause of delayed hospital presentation8 10 and reperfusion time.10 In addition to the inability to perform prehospital ECG, patients with STEMI who self-transported to the AED were deprived of early assessment by the ambulance crew, administration of aspirin and/or sublingual nitrate, early notification of the AED for patients with unstable vitals, and immediate cardiopulmonary resuscitation (with defibrillation if necessary). Thus, wise use of the ambulance service should be advocated.

According to the HKFSD, 88% of the QMH catchment area was covered by the 15 ambulances with X Series® Monitor/Defibrillator installed. Owing to this incomplete catchment area coverage, 48 ambulance patients did not receive prehospital ECG before PPCI. A greater benefit might be observed in the future, if all ambulances in the territory were capable of transmitting prehospital ECG, together with primary diversion of patients with STEMI.

Consent to perform prehospital 12-lead ECG was not obtained from 110 of 841 (13.1%) patients with cardiac chest pain (71 female and 39 male patients). Refusal to consent may result in worse outcomes for patients with STEMI with delayed PPCI. Public education regarding the benefit of performing prehospital ECG may reduce refusal rates.

Of 124 patients with STEMI attending the AED, 20 (16.1%) consulted private physicians before attendance to the AED; a number of ECGs performed in these private clinics or hospitals documented STEMI. However, apart from handwritten referral letters, there is currently no formal direct communication channel between private physicians and AED physicians. More convenient means of prehospital communication with or without teletransmission of prehospital 12-lead ECG performed in private clinics or hospitals could be explored. The benefits of seamless communication between private physicians and the AED are not limited to STEMI alone and may affect many other medical conditions.

Prehospital 12-lead ECG is technologically feasible in Hong Kong and shortens the D2B time. However, shorter reperfusion time was recorded only during daytime hours. Promotion of prehospital 12-lead ECG and proper utilisation of ambulance services for patients with cardiac chest pain may allow additional patients with STEMI to benefit from prehospital ECG.

Concept or design: All authors.
Acquisition of data: KS Cheung, YC Siu, RHW Chan.
Analysis or interpretation of data: KS Cheung, LP Leung.
Drafting of the article: KS Cheung.
Critical revision for important intellectual content: All authors.

We would like to thank the staff of the following institutions: Hong Kong Fire Services Department, for performing prehospital electrocardiograms and providing prehospital data; Division of Cardiology, Department of Medicine, Queen Mary Hospital, for providing data on door-to-balloon and door-to-catheter time; Department of Accident and Emergency, Queen Mary Hospital, for collecting patient clinical data; Emergency Medical Unit, LKS Faculty of Medicine, the University of Hong Kong, for statistical analyses; and Zoll Medical Corporation (269 Mill Rd, Chelmsford, MA 01824-4105, US), for providing the machines, training, and technical support free of charge.

All authors have disclosed no conflicts of interest. All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Analysis of data from this pilot project was approved by the Institutional Review Board of the University of Hong Kong/ Hospital Authority Hong Kong West Cluster (UW17-318). Patient consent was waived.

1. Nagel EL, Hirshman JC, Nussenfeld SR, Rankin D, Lundblad E. Telemetry-medical command in coronary and other mobile emergency care systems. JAMA 1970;214:332-8. Crossref

2. O’Connor RE, Al Ali AS, Brady WJ, et al. Part 9: acute coronary syndromes: 2015 American Heart Association Guidelines Update for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation 2015;132:S483-500. Crossref

3. Le May MR, Davies RF, Dionne R, et al. Comparison of early mortality of paramedic-diagnosed ST-segment elevation myocardial infarction with immediate transport to a designated primary percutaneous coronary intervention center to that of similar patients transported to the nearest hospital. Am J Cardiol 2006;98:1329-33. Crossref

4. Thygesen K, Alpert JS, Jaffe AS, et al. Fourth universal definition of myocardial infarction (2018). Eur Heart J 2018 Aug 25. Epub ahead of print. Crossref

5. Anderson LL, French WJ, Peng SA, et al. Direct transfer from the referring hospitals to the catheterization laboratory to minimize reperfusion delays for primary percutaneous coronary intervention: insights from the National Cardiovascular Data Registry. Circ Cardiovasc Interv 2015;8:e002477. Crossref

6. Cockburn J, Karimi K, Hoo S, et al. Outcomes by day and night for patients bypassing the emergency department presenting with ST-segment elevation myocardial infarction identified with a pre-hospital electrocardiogram. J Interv Cardiol 2015;28:24-31. Crossref

7. Wong KL, Wong YT, Yung SY, et al. A single centre retrospective cohort study to evaluate the association between implementation of an acute myocardial infarction clinical pathway and clinical outcomes. Int J Cardiol 2015;182:82-4. Crossref

8. Hong CC, Sultana P, Wong AS, Chan KP, Pek PP, Ong ME. Prehospital delay in patients presenting with acute ST-elevation myocardial infarction. Eur J Emerg Med 2011;18:268-71. Crossref

9. Song L, Yan HB, Yang JG, et al. Factors associated with use of emergency medical service for acute myocardial infarction in Beijing [in Chinese]. Zhonghua Yi Xue Za Zhi 2010;90:834-8.

10. Mathews R, Peterson ED, Li S, et al. Use of emergency medical service transport among patients with ST-segment-elevation myocardial infarction: findings from the National Cardiovascular Data Registry Acute Coronary Treatment Intervention Outcomes Network Registry—Get With The Guidelines. Circulation 2011;124:154-63. Crossref

Recently, an increased incidence of suicide among students has triggered immense public concern regarding the mental health of adolescents and young adults. In 2014, there were 52 suicides in the age-group 15 to 24 years. The number of suicides in this age-group increased in subsequent years to 68 in 2015 and 75 in 2016.1 A report from the Centre for Health Protection in Hong Kong showed that, in 2012, the prevalences of mild, moderate, and severe depressive symptoms in adolescents and children were 36.4%, 14.7%, and 4.2%, respectively.2 The Population Health Survey 2003/2004 conducted collaboratively by the Department of Health and the Department of Community Medicine of the University of Hong Kong revealed that the median score for the State-Trait Anxiety Inventory was highest in those aged 25 to 34 years, whereas the median score for the Centre for Epidemiologic Studies Depression Scale was highest in those aged 15 to 34 years.3 A web-based survey targeting first-year students receiving tertiary education in Hong Kong concluded that, in 2006, the prevalence of depression was 20.9%, while that of anxiety was 41.2%.4 Another recent study, the Hong Kong Mental Morbidity Survey,5 revealed that the most common mental problem in Hong Kong was mixed anxiety and depressive disorder; moreover, there was a strong association between anxiety symptoms and depressive symptoms.

According to the Diagnostic and Statistical Manual of Mental Disorders, published by the American Psychiatric Association, major depressive disorder is defined as the presence of five or more of the listed symptoms for most of the days during the same 2-week period; at least one of the symptoms must be either depressed mood or loss of interest or pleasure. Notably, these symptoms should reflect a change from previous functioning. Generalised anxiety disorder (GAD) is characterised by excessive worries that cause distress and interfere with psychosocial functioning. These worries frequently happen without precipitants, exhibit longer durations, and are accompanied by three or more of the six listed additional symptoms.

Various factors associated with major depressive disorder and GAD have been identified, including alcohol use, illicit drug use, tobacco use, and level of physical activity.6 However, evidence for an association between academic pressure and these two psychiatric disorders remains unknown among university undergraduate students in Hong Kong. We aimed to provide an update regarding the prevalence of depressive and anxiety symptoms experienced by undergraduate students in Hong Kong, and to identify factors associated with these symptoms.

Full-time undergraduate students of the eight universities in Hong Kong that are funded by the University Grants Committee (UGC) were the target participants in this study. A cross-sectional study design was employed and the target number of students to be enrolled from each university was 150; the total target number of students to be recruited was 1200.

The inclusion criteria for this study were as follows: full-time undergraduate students in the eight UGC-funded universities who were ≥18 years, were able to understand the consent, provide a valid oral consent, comprehend the questionnaire, and had not previously participated in this research. Students who failed to fulfil all inclusion criteria were excluded from the study.

Convenience sampling was employed. Questionnaires were distributed at popular locations within all eight university campuses on school days in September and October 2016. Participants were asked to complete three sets of questionnaires, including the 9-item Patient Health Questionnaire (PHQ-9) for screening of depressive symptoms, the 7-item GAD scale (GAD-7) for screening of anxiety symptoms, and a socio-demographic questionnaire to identify factors associated with depressive and anxiety symptoms.7 8 When completed questionnaires were being returned, respondents were reminded that they would not be able to withdraw from the study after returning the questionnaires, as the questionnaires did not contain any personal identifying information to allow individual retrieval.

The PHQ-9 was adopted to screen for depressive symptoms in this study. Both Chinese and English versions were provided. The PHQ-9 has been shown to be a valid and reliable tool for assessing depressive symptoms in the Hong Kong general population.7 8 Another study showed that the PHQ-9 was a useful tool to detect both major depression and subthreshold depression.9 According to the PHQ-9, a score of <5 was defined as none-minimal depressive symptoms, score of 5 to 9 as mild, score of 10 to 14 as moderate, score of 15 to 19 as moderately severe, and score of ≥20 as severe.

The GAD-7 is a practical self-report anxiety questionnaire that is generally used in out-patient and primary care settings for referral to a psychiatrist to confirm the diagnosis of GAD. Both Chinese and English versions were available. The adoption of GAD-7 in this study was due to its good reliability, as well as its criteria, construct, and factorial and procedural validity.10 11 According to the GAD-7, a score of <5 was defined as none-minimal anxiety symptoms, score of 5 to 9 as mild, score of 10 to 14 as moderate, and score of ≥15 as severe.

Descriptive analysis was performed to summarise the demographics and statuses of depressive and anxiety symptoms of the respondents. A binary logistic regression was conducted to ascertain the effects of various covariates on the odds of exhibiting mild to severe depressive symptoms, on the basis of PHQ-9 results. A PHQ-9 score of <5 was defined as no depressive symptoms, while a score of ≥5 was defined as mild to severe depressive symptoms. Another binary logistic regression was used to ascertain the effects of multiple covariates on the odds that participants would exhibit mild to severe anxiety symptoms, on the basis of GAD-7 results. A GAD-7 score of <5 was defined as no anxiety symptoms, while a score of ≥5 was defined as mild to severe anxiety symptoms. A P value of <0.05 was regarded as a significant difference.

In total, 1480 undergraduate students from the eight UGC-funded universities were approached from 6 September 2016 to 3 October 2016. From these, 1200 completed questionnaires were collected, for a response rate of 81.1%. A total of 81 students did not meet at least one of the inclusion criteria and their responses were subsequently excluded from analysis (Fig). The remaining 1119 students fulfilled all inclusion criteria and their responses were analysed by SPSS (Mac version 24; IBM Corp, Armonk [NY], United States).

The mean (standard deviation) age of the respondents was 19.81 (1.48) years (range, 18-29 years). Among the respondents, 426 (38.1%) were male and 693 (61.9%) were female. Most respondents were Chinese (92.7%). The respondents’ demographics are shown in Table 1.

Among the 1119 valid questionnaires analysed, 767 (68.5%) were found to have mild to severe depressive symptoms (Table 2).

For screening of anxiety symptoms, 18 of the 1119 respondents were excluded because they did not complete the GAD-7 questionnaires; therefore, results are based on the 1101 valid responses. A total of 599 (54.4%) respondents were found to have mild to severe anxiety symptoms. A higher prevalence of anxiety was observed in females than males in all three categories of severity (Table 3).

The binary logistic regression model for the effects of multiple covariates on the odds of having mild to severe depressive symptoms, on the basis of the PHQ-9 results, was statistically significant with χ2=375.006, P<0.001, with 18 degrees of freedom (Table 4). The model explained 43.1% of the variance in depression severity, as shown by Nagelkerke R², and correctly classified 79.6% of the cases. Mild to severe anxiety symptoms, as screened by GAD-7, were significantly associated with mild to severe depressive symptoms (P<0.001). In contrast, several lifestyle and psychosocial variables, including regular exercise (P<0.001), self-confidence (P=0.01), satisfaction with academic performance (P=0.019), and optimism towards the future (P<0.001) were inversely related to mild to severe depressive symptoms.

Another binary logistic regression model for the effects of multiple covariates on the odds of having mild to severe anxiety symptoms, on the basis of the GAD-7 results, was statistically significant with χ²=374.842, P<0.001, with 20 degrees of freedom (Table 5). The model explained 41.3% of the variance in depression severity, as shown by Nagelkerke R², and correctly classified 76.9% of the cases. Mild to severe anxiety symptoms were associated with level of academic difficulty (P=0.028) and mild to severe depressive symptoms, as screened by PHQ-9 (P<0.001). Certain lifestyle and psychosocial variables, including satisfaction with friendship (P=0.001), sleep quality (P=0.003), and self-confidence (P=0.003) were inversely associated with mild to severe anxiety symptoms.

The results revealed that 68.5% of respondents had mild to severe depressive symptoms and 54.4% had mild to severe anxiety symptoms. These rates are higher than and comparable to the results of a similar survey conducted by the University of Hong Kong 10 years ago, respectively.4 This could be attributed to the increasing academic pressure after major reforms of the education system in Hong Kong,12 uncertainties in career prospects due to fluctuations in the socio-political environment, and the more prevalent use of social media.13

This study revealed that mild to severe depressive symptoms, as screened by PHQ-9, were associated with mild to severe anxiety symptoms, as screened by GAD-7. This is consistent with previous studies, including a study in patients with depression showing that 85% of those with major depression were also diagnosed with generalised anxiety.14 Sharing common risk factors and symptoms may explain their co-existence.15 16 Another study has shown that anxiety is more frequently an antecedent of depression, but the reverse relationship was not observed.17 This could be explained by the withdrawal and submissive techniques adopted by individuals with anxiety, in response to social exclusion.18 However, a definite causal relationship remains unclear.

Regular exercise could decrease the occurrence of depression via both physiological and psychological mechanisms. Exercise exerts an excitatory effect on the monoamine and endorphin neurotransmitter systems; notably, monoamines are depleted in patients with depression.19 Psychologically, exercise is reported to improve self-esteem and self-perception through self-actualisation and gaining pleasure from an expanded social circle.20

Psychosocial factors, including higher self-confidence, satisfaction with academic performance, and optimism towards the future are inversely related to depressive symptoms. These could be the presentation or consequences of depression; these factors may have protective effects against developing depression.21 Optimism towards the future, apart from personal factors, is also related to community factors. Risk factors include community disadvantage, safety, and discrimination; an important protective factor is community connectedness. These factors were associated with depressive symptoms.22 This implies that, in addition to promoting personal mental health, community-level risk and protective factors that affect individuals’ optimism towards future are sufficiently important to receive broader attention.

For the analysis of anxiety, the level of academic difficulty was associated with mild to severe anxiety symptoms. A higher level of academic difficulty may create greater stress and anxiety for students; subsequent unsatisfactory academic results may complete the vicious cycle. Several tips were suggested by the Anxiety and Depression Association of America (ADAA) regarding test anxiety reduction.23 Female gender was not associated with mild to severe anxiety symptoms, which conflicts with previous studies.15 No clear causality was identified, but some potential aetiologies include higher stress levels from academic and interpersonal issues.

Satisfaction with friendship, sleep quality, and self-confidence were three factors inversely associated with mild to severe anxiety symptoms. Having good interpersonal relationships with peers could be a protective factor against anxiety; however, symptoms of anxiety may also hinder friendship, thus explaining the inverse relationship. Insomnia is a symptom of anxiety. According to a survey conducted by ADAA, worrying about falling asleep at night was identified as a cause for an increased level of anxiety.24 Recommendations that could help in falling asleep and achieving better sleep quality include maintaining 7 to 9 hours of uninterrupted sleep, establishing a regular bedtime routine, and avoiding electronic gadgets, coffee, and tea, as well as designing a relaxing environment for sleeping.24 Lack of self-confidence could be a cause of anxiety, but anxiety symptoms could also diminish one’s self-confidence. A bilateral association is possible.

There were several limitations in our study. Firstly, as with all cross-sectional studies, it was not possible to establish causality between the identified factors and symptoms of depression and anxiety. Hence, the identified factors are regarded as associated factors, which could be either the causes or the results of depression or anxiety. To further explore the relationships between these factors and symptoms of depression and anxiety, we reviewed the available literature to provide supporting evidence. Secondly, convenience sampling was adopted, but there could be self-selection bias when inviting students to complete the questionnaires, limited by the locations and times at which questionnaire distribution was conducted. The questionnaires were distributed in the open areas of the eight UGC-funded universities in Hong Kong. Withdrawn university students affected by depression or anxiety might not have been reached or might have refused to participate in the study; hence, the results may be underestimates. Thirdly, we could not attain our target number of 150 students from each university, except from the Hong Kong Polytechnic University. Importantly, this study used screening questionnaires to assess the depressive and anxiety symptoms experienced by the students, but no clinical diagnoses were made by health care professionals. However, with limited resources, the use of validated screening questionnaires was considered a cost-effective approach to explore the situation in general and thus was used in this study. Possible directions for further studies include the implementation of stratified random sampling, expansion of the sample size of the study, and performance of a reliability test to reduce information bias. Including detailed backgrounds of students’ disciplines and faculties, as well as ensuring a diverse population to include both local and non-local students, as well as Chinese-speaking and non-Chinese-speaking students, could allow further subgroup analysis; students from different faculties may experience depressive symptoms or anxiety symptoms at different severities due to differences in curricula. In this research, participants who were screened to have depressive and/or anxiety symptoms were not notified because of the lack of identifiable personal information. Provision of the results to participants and recommendations regarding help-seeking information are suggested for future studies. This study result may not fully reflect the severity of depressive and anxiety symptoms among students, as the study method could not reach severely depressed individuals who had socially isolated themselves.

This study showed that over 50% of university students in the eight UGC-funded universities expressed some degree of depressive symptoms (68.5%) or anxiety symptoms (54.4%). Notably, 9% of these university students exhibited moderate to severe depressive symptoms and 5.8% of the studied students showed severe anxiety symptoms. Students with regular exercise, higher self-confidence, better satisfaction with academic performance, and more optimism towards the future experienced fewer depressive symptoms. Students with better satisfaction with friendship, better sleep quality, higher self-confidence, and lower levels of academic difficulty experienced fewer anxiety symptoms.

Concept or design of study: KWC Lun, CK Chan.
Acquisition of data: All authors.
Analysis or interpretation of data: All authors.
Drafting of the article: All authors.
Critical revision for important intellectual content: KWC Lun, CK Chan.

We thank Dr LM Ho, assistant IT director of the Division of Epidemiology and Biostatistics of School of Public Health of the University of Hong Kong, for his statistical support and advice in conducting this study. We also thank the School of Public Health of the University of Hong Kong, for its support in conducting this study.

All authors have disclosed no conflicts of interest. All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

The LKS Faculty of Medicine, University of Hong Kong reimbursed project expenses under the title of Health Research Project for HKD500.

This study was reviewed and approved by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster.

1. Hong Kong Jockey Club Centre for Suicide Research and Prevention, The University of Hong Kong. Statistics. Available from: https://csrp.hku.hk/statistics/. Accessed 18 Sep 2018.

2. Centre for Health Protection, Department of Health, Hong Kong SAR Government. Depression: beyond feeling blue. Non-Communicable Diseases Watch 2012. Available from: https://www.chp.gov.hk/files/pdf/ncd_watch_sep2012.pdf. Accessed 20 Apr 2016.

3. Centre for Health Protection, Department of Health, Hong Kong SAR Government. Population Health Survey 2003/2004. Collaborative project of Department of Health and Department of Community Medicine, University of Hong Kong, Available from: https://www.chp.gov.hk/files/pdf/report_on_population_health_survey_2003_2004_en.pdf. Accessed 20 Apr 2016.

4. Wong JG, Cheung EP, Chan KK, Ma KK, Tang SW. Web-based survey of depression, anxiety and stress in first-year tertiary education students in Hong Kong. Aust N Z J Psychiatry 2006;40:777-82. Crossref

5. Lam LC, Wong CS, Wang MJ, et al. Prevalence, psychosocial correlates and service utilization of depressive and anxiety disorders in Hong Kong: the Hong Kong Mental Morbidity Survey (HKMMS). Soc Psychiatry Psychiatr Epidemiol 2015;50:1379-88. Crossref

6. Cairns KE, Yap MB, Pilkington PD, Jorm AF. Risk and protective factors for depression that adolescents can modify: a systematic review and meta-analysis of longitudinal studies. J Affect Disord 2014;169:61-75. Crossref

7. Yu X, Tam WW, Wong PT, Lam TH, Stewart SM. The Patient Health Questionnaire-9 for measuring depressive symptoms among the general population in Hong Kong. Compr Psychiatry 2012;53:95-102. Crossref

8. Wang W, Bian Q, Zhao Y, et al. Reliability and validity of the Chinese version of the Patient Health Questionnaire (PHQ-9) in the general population. Gen Hosp Psychiatry 2014;36:539-44. Crossref

9. Martin A, Rief W, Klaiberg A, Braehler E. Validity of the Brief Patient Health Questionnaire Mood Scale (PHQ-9) in the general population. Gen Hosp Psychiatry 2006;28:71-7. Crossref

10. Spitzer RL, Kroenke K, Williams JB, Löwe B. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med 2006;166:1092-7. Crossref

11. Tong X, An D, McGonigal A, Park SP, Zhou D. Validation of the Generalized Anxiety Disorder-7 (GAD-7) among Chinese people with epilepsy. Epilepsy Res 2016;120:31-6. Crossref

12. Yeh YC, Yen CF, Lai CS, Huang CH, Liu KM, Huang IT. Correlations between academic achievement and anxiety and depression in medical students experiencing integrated curriculum reform. Kaohsiung J Med Sci 2007;23:379-86. Crossref

13. O’Keeffe GS, Clarke-Pearson K, Council on Communications and Media. The impact of social media on children, adolescents, and families. Pediatrics 2011;127:800-4. Crossref

14. Hall RC, Platt DE, Hall RC. Suicide risk assessment: a review of risk factors for suicide in 100 patients who made severe suicide attempts. Evaluation of suicide risk in a time of managed care. Psychosomatics 1999;40:18-27. Crossref

15. Blanco C, Rubio J, Wall M, Wang S, Jiu CJ, Kendler KS. Risk factors for anxiety disorders: common and specific effects in a national sample. Depress Anxiety 2014;31:756-64. Crossref

16. van Ameringen M. Comorbid anxiety and depression in adults: Epidemiology, clinical manifestations, and diagnosis. 15 Jun 2017. Available from: https://www.uptodate.com/contents/comorbid-anxiety-and-depression-in-adults-epidemiology-clinical-manifestations-and-diagnosis. Accessed 24 Jul 2017.

17. Muris P. Normal and Abnormal Fear and Anxiety in Children and Adolescents. London: Elsevier; 2007.

18. Richards CS, O’Hara MW, editors. The Oxford Handbook of Depression and Comorbidity. New York: Oxford University Press; 2014.Crossref

19. Buckworth J, Dishman RK. Exercise Psychology. Champaign: Human Kinetics; 2002.

20. Faulkner GE, Taylor AH, editors. Exercise, Health and Mental Health: Emerging Relationships. New York: Routledge; 2005. Crossref

21. Wang KT. Perfectionism, depression, and self-esteem: a comparison of Asian and Caucasian Americans from a collectivistic perspective [dissertation]. Pennsylvania: The Pennsylvania State University; 2007.

22. Stirling K, Toumbourou JW, Rowland B. Community factors influencing child and adolescent depression: A systematic review and meta-analysis. Aust N Z J Psychiatry 2015;49:869-86. Crossref

23. Anxiety and Depression Association of America. Test Anxiety. Available from: https://www.adaa.org/living-with-anxiety/children/test-anxiety. Accessed 23 Mar 2017.

24. Anxiety and Depression Association of America. Stress and Anxiety Interferes with sleep. Available from: https://www.adaa.org/understanding-anxiety/related-illnesses/other-related-conditions/stress/stress-and-anxietyinterfere. Accessed 23 Mar 2017.

Intellectual disability (ID) is estimated to affect 1% to 3% of the population in Western societies.1 It is almost two-fold greater in prevalence in low-and middle-income countries, compared with high-income countries. Importantly, the General Household Survey in 2014 showed the prevalence rate of ID to be approximately 1.0% to 1.4% in Hong Kong.2

Intellectual disability is defined as ‘significant limitations both in intellectual functioning and in adaptive behaviour, as expressed in conceptual, social, and practical adaptive skills’.3 These difficulties are evident above the age of 18; ID is indicated by an intelligence quotient (IQ) of approximately two standard deviations (SD) or more below the population mean (IQ ≤70) on the IQ test.

Developmental delay (DD) describes the developmental level of a child, typically <5 years old, who is substantially below the average standard of his peers. Global DD is defined as a significant delay in two or more domains: gross motor, fine motor, language, cognitive, social, or activities of daily living. Significant delay refers to scores >2 SD below the mean on norm-referenced age-appropriate developmental tests.4

In 2015 and 2016, more than 3000 children per year were diagnosed with DD by the Child Assessment Service (CAS). A study conducted in 2003 to 2004 showed that 80% of children with significant delay and 30% of children with borderline delay were later confirmed to exhibit ID at an older age.5 Overall, 30% of children with ID had a co-morbid diagnosis of autism spectrum disorder (ASD).

The aetiology of ID is complex. While milder forms of ID are suspected to typically result from the interplay of genetic and environmental factors,6 biological causes, particularly genetic causes, are often identified in children with significant cognitive delays (IQ <50).4 Rauch et al7 studied 670 subjects, generally <6 years of age, with ID: in 39.5%, ID was related to a genetic cause; in 1.3%, it was due to an acquired or environmental cause; and in 50% to 60%, it did not exhibit a known aetiology.

Chromosomal microarray (CMA) or array comparative genomic hybridisation, is recommended by many international professional organisations as a first-tier genetic investigation for children with unexplained DD, ID, or ASD.4 8 9 Compared with conventional karyotyping, CMA is able to detect copy number variants (CNVs) with much finer resolution and is not reliant on staining and visual resolution limits. In 2010, a review of 33 published studies—involving 21 698 patients with DD, congenital anomalies, or autism—found the diagnostic yield of CMA to be 15% to 20% across all studies, compared with 3% for the standard G-banded karyotype.9 In a group of 94 patients with no symptoms other than ID, and no clear dysmorphic features, the diagnostic yield was 6.4%.8 According to the American Academy of Neurology guideline in 2011, CMA testing was abnormal in approximately 7.8% of patients with global DD or ID. The yield was higher (10.6%) in those with syndromic features.8

Children with ASD who had co-morbid ID were more likely to yield molecular diagnoses.10 Approximately 10% of patients with ASD exhibit a de-novo CNV, as detected by CMA.11 Among ASD children without syndromic features, only 6% received a molecular diagnosis.

In Hong Kong, two studies have investigated the use of CMA in patients with DD, ID, ASD, or multiple congenital anomalies (MCA). Chong et al12 found clinically significant CMA results in 20 of 105 patients (19%). Tao et al13 found a diagnostic yield of 11% for pathogenic or likely pathogenic results in 327 children, ages 1 month to >20 years. Excluding patients with MCA, the diagnostic yield of CMA for DD, ID, or ASD was approximately 4.2%.13

Chromosomal microarray has high clinical utility. Firstly, it shortens the diagnostic odyssey and may avoid unnecessary investigations, which reduces both individual and societal costs associated with testing and medical care.8 14 Secondly, it may lead to a clinically actionable recommendation. The prognostic information related to diagnosis from CMA may alert other potential co-morbid conditions that cannot be predicted on the basis of physical examination alone. In a retrospective review of 1792 patients with DD, ID, ASD, or MCA who underwent CMA testing, individuals with a positive diagnosis had a higher rate of clinical actionable recommendations than those with an uncertain result (54% vs 34%, P=0.01).15 In Hong Kong, a detection rate of 8.6% was reported for clinically actionable CNVs,13 which was comparable to the reported rates of 3.6% to 7% in Western studies.15 16 17 Thirdly, it allows estimation of recurrence risk and informed decisions regarding reproductive options for the parents’ future pregnancies.

Children’s cognitive development at ≥5 years of age is more stable if the level of ID is known. Children with DD undergo assessment at CAS to facilitate their transition into primary school. Since 2012, the Clinical Genetic Service (CGS) of the Department of Health has provided CMA testing for DD, ID, or ASD. The presence of dysmorphic features, early onset of DD, increased severity of DD, and family history are common features that prompt a genetic referral. Collaboration between CAS and CGS can potentially narrow the service gap for children with DD, ID, or ASD by enabling early access to diagnostic genomic testing, thus facilitating shorter waiting time for genetic and genomic investigation(s) and a more client-friendly turnover time for results.

The aim of this study was to investigate the genetic profile and diagnostic yield of CMA in children with moderate, severe, and profound ID. The data obtained from this study are expected to be useful in future service planning for children with DD or ID. The diagnostic yield of CMA for children with more severe forms of ID is suspected to be higher than the generally quoted figures of CMA (10%) for investigation of ID. This study targeted children with more significant ID, which is more likely to be related to an underlying genetic aetiology.

This cross-sectional territory-wide study recruited children who attended CAS for developmental assessments, before Primary 1 entry, over a 12-month period from July 2016 to June 2017. All children were at least 5 years of age. Inclusion criteria were children with moderate, severe, or profound ID, with or without ASD. According to the International Statistical Classification of Diseases and Related Health Problems, tenth revision,18 moderate ID was defined as IQ 35 to 49; severe ID was defined as IQ 20 to 34; and profound IQ was defined as IQ <20. Exclusion criteria included known causes of ID: (i) antenatal causes such as congenital brain malformation or intrauterine infections; (ii) perinatal causes such as prematurity (<34 weeks), birth asphyxia, or hypoxic ischaemic encephalopathy; (iii) postnatal causes such as intracranial bleeding, intracranial infection, or brain trauma; and (iv) other biological causes such as inborn errors of metabolism, brain tumour, neuromuscular disorders, neurodegenerative disorders, or cerebral palsy.

Unexplained ID in this study was defined as children with no identifiable causes for ID who did not meet any of the exclusion criteria. These children were non-syndromic and non-dysmorphic. In addition, they had neither MCA nor family history of ID or ASD among first- and second-degree relatives. The presence of MCA was defined as the involvement of two or more organ systems.

Children were assessed by paediatricians with the Griffiths Mental Developmental Scales,19 or by clinical psychologists with the Wechsler Preschool and Primary Scale of Intelligence–Revised.20 The Diagnostic and Statistical Manual of Mental Disorders, 4th edition,21 ASD diagnostic criteria were used for assessment of ASD.

For children with genetic diagnoses or who were known clients of CGS, their medical files were retrieved from CAS and CGS for review. Children with syndromic or dysmorphic features, MCA, or significant family history, who had not been previously referred to CGS, were referred for a formal genetic consultation before genetic or genomic investigations were recommended by a clinical geneticist.

An expedited pathway was offered for children with unexplained ID. Pre-genetic counselling was provided by a paediatrician at CAS, followed by direct blood examination for CMA and Fragile X syndrome (FGX) testing at CGS. Consultation with a geneticist was arranged if either CMA or Fragile X testing yielded abnormal outcomes. Otherwise, clients did not consult a geneticist for further counselling.

For each patient, 3 mL of blood in ethylenediaminetetraacetic acid was sent to the laboratory at CGS. All samples were tested by PerkinElmer CGX^TM v2 60K arrays designed by Agilent SurePrint technology, in accordance with the manufacturer’s instructions. The coverage of the array demonstrated an average resolution of 140 kb across the genome, and ≤40 kb in regions of clinical relevance. It evaluated >245 known genetic syndromes and >980 gene regions of functional significance in human development. Data were analysed by Genoglyphix software (Signature Genomics, Spokane [WA], United States). Genomic coordinates were based on genome assembly hg19.

Detected CNVs were systematically evaluated for clinical significance by comparison with information in the proprietary Genoglyphix Chromosome Aberration Database (Signature Genomics), internal laboratory database at CGS and the Department of Health, and public databases (Database of Genomic Variants, International Standards for Cytogenomic Arrays Consortium, and Database of Chromosomal Imbalance and Phenotype in Humans using Ensembl Resources). Categorisation of CNVs was based on available phenotypes and comparison of phenotypes with genes in the region of copy gain or loss. This was performed by searching the following databases: Online Mendelian Inheritance in Man, PubMed, RefSeq, and the University of California Santa Cruz genome browser.22 Confirmatory fluorescence in situ hybridisation (FISH), multiplex ligation-dependent probe amplification (MLPA), or conventional karyotyping was performed as indicated. Parental testing was offered to aid further interpretation and classification. Copy number variants were classified as pathogenic, likely pathogenic, uncertain clinical significance, or benign, in accordance with the 2011 American College of Medical Genetics practice guidelines.23 Only pathogenic and likely pathogenic CNVs were regarded as clinically significant.

The number of subjects to be recruited was estimated based on the average number of children with moderate, severe, and profound ID in the CAS database. In 2013 to 2015, the average number of children with moderate ID or worse was approximately 270 children per year. With the assumption that 60% of cases were unexplained,7 potential cases eligible for CMA were estimated as 160 children per year. Literature showed that the diagnostic yield of CMA was 10% in identifying abnormal cases in similar settings.13 A 95% confidence interval was desired, with a reliability (d) of 0.05, in obtaining a diagnostic yield (ˆp) of 10% in this study. The sample size needed was determined following a previously published method1:

Hence, a target sample size of 138 was needed.24

The genetic profile of children in the study was described. The diagnostic yield from CMA was calculated according to the severity of ID. The Freeman-Halton test was used to test associations between the severity of (a) ID and CMA and (b) dysmorphism and CMA findings. The null hypothesis was that there was no association between the severity of ID or dysmorphism and CMA findings. P<0.05 indicated an association between the severity of ID or dysmorphism and CMA findings. The Freeman-Halton test was conducted by using SAS/STAT 9.22.

From July 2016 to June 2017, there were a total of 339 children diagnosed with more severe forms of ID: 241 (71%) children had moderate ID, 49 (14.5%) had severe ID, and 49 (14.5%) had profound ID. Eighty-three children were excluded for the following reasons: (1) they met predefined exclusion criteria; (2) their family could not participate due to geographical reasons (eg, family lived in China); (3) a language barrier affected their understanding of study details (eg, the children or their families spoke primarily Nepalese or Sri Lankan); or (4) their parents could not be contacted for consent. A total of 31 children opted not to participate in the study. In all, 225 (66.4%) of 339 children participated in the study.

Among the 225 children, 116 (51.6%) had a co-morbid diagnosis of ASD. Male (n=151) to female (n=74) ratio was 2:1. The age ranged from 5 to 10 years old with a mean age of 6.6 years old. In all, 71.5% of children had moderate ID, 14.7% had severe ID, and 13.8% had profound ID. Two hundred twenty-one (98%) children had Chinese parents. There were two pairs of consanguineous parents: one Indian couple and one Pakistani couple.

As shown in Table 1, 68 (30.2%) children were diagnosed with a genetic condition. The percentage of a positive genetic diagnosis increased with the severity of ID. Chromosomal abnormalities comprised 76% (n=52) of the total genetic diagnoses. The most common syndromic diagnosis was Down syndrome (n=22). There were two cases of FGX. Three children had chromosome 22 microdeletion syndromes—one exhibited the more common chromosome 22q11.2 microdeletion syndrome (DiGeorge syndrome), whereas the other two exhibited chromosome 22q13.3 deletion syndrome.

Of the 225 participating children, 138 underwent CMA testing; 53 (38%) children were referred to the Clinical Genetic Service by the expedited referral pathway. Table 2 shows that 16 (11.6%) children demonstrated clinically significant CNVs that explained their ID phenotype and 10 (7.2%) had variants of uncertain significance (VUS). The diagnostic yield of CMA increased with severity of ID: it was 8.7% in moderate ID, 17.6% in severe ID, and 23.5% in profound ID (P<0.05; Table 3). Among the 16 children with clinically significant CNVs, 11 demonstrated copy number loss (deletion), four demonstrated copy number gain (duplication), and one demonstrated an unbalanced translocation between chromosome 7q and 20p (Table 4). One case of Angelman syndrome was detected by CMA and later confirmed with MLPA. One case of Cri du chat syndrome was detected by CMA and later confirmed with FISH. In total, 69% of pathogenic or likely pathogenic CNVs were de novo. Ten children (7.2%) were incidentally identified as carriers of disease: six were alpha thalassemia heterozygous carriers, one was a heterozygous carrier of Joubert syndrome type 4, one was a heterozygous carrier of autosomal recessive disease Joubert syndrome and nephronophthisis, one was a heterozygous carrier of autosomal recessive deafness affecting the OTOA gene, and one was a carrier of Klinefelter syndrome.

The overall diagnostic yield of CMA among children with ID (11.6%) was consistent with studies performed in other regions of the world. The diagnostic yield of CMA increased with severity of ID and was much higher in children with dysmorphism (45.8% vs 4.4%, P<0.05).

Variants of uncertain significance are not uncommon. In all, 7.2% of children in this cohort had VUS (Table 5). Regular follow-up and reassessment by a clinical geneticist is necessary for these children, because VUS may eventually be re-classified as pathogenic or benign as clinical and genomic data accumulate in the literature.

The paradigm shift in the medical genetic and genomic field from the phenotype-first approach to the genotype-first approach is revolutionary. Traditionally, a phenotype-first approach was used to guide the investigation of possible genetic diagnoses, eg, karyotyping for Down syndrome, or specific assays, such as FISH, for DiGeorge syndrome. In the past decade, CMA has allowed more comprehensive unbiased discovery of microdeletion and microduplication syndromes throughout the human genome. Since the 1980s, it has been well-known that certain chromosomal microdeletion and microduplication syndromes are associated with recognisable forms of ID and DD. Classical examples include 15q11-q13 deletion, associated with Prader-Willi and Angelman syndromes, and 22q11.2 deletion, associated with DiGeorge syndrome (velocardiofacial syndrome). Thus far, approximately 50 to 60 recurrent microdeletion or duplication syndromes have been identified in children with DD or ID.

Although CMA is robust, it cannot replace a formal genetic consultation for children with clinically suspected genetic conditions. As an example, in Prader-Willi syndrome, 70% to 75% of cases can be detected by CMA, as they are due to a paternal 15q microdeletion subtype; 20% to 25% of cases require a more specific methodology for genetic confirmation. Therefore, clinical correlation and expert assessment remain necessary.

Males are more susceptible to ID than females; more than 100 X-linked genes are associated with ID.25 X-linked ID constitutes 5% to 10% of ID in males. One of the best-known causative genes for ID is FMR1; mutations of FMR1 result in FGX. The estimated incidence of FGX is approximately 1 in 4000 males and 1 in 5000 to 1 in 8000 females (approximately 0.5% of cases of ID) in Western countries. Peprah26 reported that the incidence of FGX in countries/regions with significant Asian populations, such as Canada, Estonia, Japan, and Taiwan, was significantly lower than in Western countries. In a study of 553 male children between the ages of 6 months and 18 years, Chen et al27 estimated the prevalence of FGX in mainland China to be approximately 0.93% among children with moderate to severe ID. Among the 225 children in our cohort, only two were diagnosed with FGX. Both exhibited ASD and severe ID. The typical physical characteristics of FGX, such as narrow face, protruding ears, and macro-orchidism, are often less obvious in early childhood; notably, they may become more prominent as the child approaches adolescence. This lack of early physical characteristics increases the diagnostic challenge for clinicians. Fragile X syndrome testing, regarded as first-tier genetic testing for DD and ASD in many international guidelines, has been a standard genetic investigation for ID or ASD in Hong Kong for many years.

Incomplete penetrance of a genomic condition within the same family is not uncommon. Notably, there were two such cases of 16p13.11 microduplication syndrome in this cohort. Patient 12 (Table 4) exhibited subtle dysmorphism comprising downslanting palpebral fissures, prominent ears, and mild right ptosis. Left undescended testes and umbilical hernia were operated in infancy. He exhibited global DD and was later diagnosed with moderate ID. His mother and two sisters had an identical chromosomal defect, but exhibited normal intelligence. Patient 13 demonstrated a more severe phenotype with hirsutism, bushy eyebrows, frontal bossing, hearing loss, visual problems, and profound ID. His mother and elder brother, both carrying the microduplication, exhibited normal intelligence. There likely exist unknown environmental or genetic modifiers to modulate susceptibility to ID caused by this microduplication. Thus, relying on family history to determine whether ID is hereditary can be misleading.

An important aspect with respect to obtaining a genetic diagnosis is patient prognosis. The 16p13.11 duplication syndrome is associated with an aortic root defect. In this study, the two affected children and their affected family members were referred for monitoring by echocardiogram. Similarly, Patient 2 exhibited 3q13.31 microdeletion syndrome, which is associated with diabetes mellitus and deafness; this patient was referred for audiological assessment and counselled on lifestyle management to minimise the risk of diabetes. In this study, three of 16 CMA-positive cases (18.8%) were clinically actionable.

Pre-test genetic counselling is as important as post-test counselling. Coincidental findings of genetic changes that either predict adult-onset conditions or reveal carrier status for recessive or X-linked conditions are common. In the present cohort, 10 children were identified as carriers of genetic conditions, including one child diagnosed with Klinefelter syndrome. He presented with moderate ID and ASD. Chromosomal microarray identified a copy number gain of the entire X chromosome. Klinefelter syndrome can be associated with learning disabilities, as well as delayed speech and language development. While a small, but significant, downward shift in mean overall IQ has been reported, general cognitive abilities of patients with Klinefelter syndrome are not typically in the ID range.28 An extra X chromosome may have contributed partially, but could not entirely explain the severity of ID. The major implications are that individuals with Klinefelter syndrome have a higher risk of endocrine dysfunction, fertility problems, male breast cancer, and autoimmune disease.

This study provided important information with respect to service planning for children with ID in Hong Kong. It allowed testing of an expedited referral mechanism between CAS and CGS, in which cases with unexplained ID benefitted through a significant reduction of waiting time for both pre-testing genetic counselling and investigation turnover time. This study included 38% of the 138 children who were not referred to CGS. The ideal future approach may be to extend the expedited mechanism for children with early-onset significant DD. It can avoid unnecessary investigations, thus lowering stress for both child and parent; importantly, it may reduce societal costs.

There were several limitations in this study. Firstly, a complete genetic profile of ID was not generated, as this cohort excluded mild ID. Secondly, clients from minority cultural groups in Hong Kong were underrepresented, because the language barrier affected recruitment. More effort must be expended to ensure equal opportunities for children from diverse cultural backgrounds. Thirdly, the duration of the study was of insufficient length for commentary on trends regarding the genetic profile of ID in Hong Kong.

The overall diagnostic yield (11.6%) of CMA is compatible with other international cohorts. Chromosomal microarray yield increases with the severity of ID. These data further support the use of CMA as a first-tier investigation for children with significant unexplained ID in Hong Kong.

All authors have made substantial contributions to the concept or design of the study, acquisition of data, analysis or interpretation of data, drafting of the article, and critical revision for important intellectual content.

The authors would like to thank Mr Morris Wu of Child Assessment Service, Department of Health for his advice on the statistical analysis of the data.

The authors have no conflicts of interest to disclose. All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Approval was obtained from the Ethics Committee of the Department of Health, Hong Kong Special Administrative Region. Informed consent was obtained from parents or legal guardians. Parents and legal guardians were counselled about the indication for CMA, benefits and limitations of test, methodology, reporting time, and possible outcomes upon recruitment.

1. Leonard H, Wen X. The epidemiology of mental retardation: challenges and opportunities in the new millennium. Ment Retard Dev Disabil Res Rev 2002;8:117-34. Crossref

2. Social Data Collected via the General Household Survey. Special Topics Report No 62. Persons with Disabilities and Chronic Diseases. Hong Kong: Census and Statistics Department, Hong Kong SAR Government; 2014.

3. American Association on Intellectual and Developmental Disabilities. Intellectual Disability: Definition, Classification, and Systems of Supports. 11th ed. Washington, DC: American Association on Intellectual and Developmental Disabilities;2010: xvi, 259.

4. Moeschler JB, Shevell M, Committee on Genetics. Comprehensive evaluation of the child with intellectual disability or global developmental delays. Pediatrics 2014;134:e903-18. Crossref

5. Tang KM, Chen TY, Lau VW, Wu MM. Clinical profile of young children with mental retardation and developmental delay in Hong Kong. Hong Kong Med J 2008;14:97-102.

6. Willemsen MH, Kleefstra T. Making headway with genetic diagnostics of intellectual disabilities. Clin Genet 2014;85:101-10. Crossref

7. Rauch A, Hoyer J, Guth S, et al. Diagnostic yield of various genetic approaches in patients with unexplained developmental delay or mental retardation. Am J Med Genet A 2006;140:2063-74. Crossref

8. Michelson DJ, Shevell MI, Sherr EH, Moeschler JB, Gropman AL, Ashwal S. Evidence report: Genetic and metabolic testing on children with global developmental delay: report of the Quality Standards Subcommittee of the American Academy of Neurology and the Practice Committee of the Child Neurology Society. Neurology 2011;77:1629-35. Crossref

9. Miller DT, Adam MP, Aradhya S, et al. Consensus statement: chromosomal microarray is a first-tier clinical diagnostic test for individuals with developmental disabilities or congenital anomalies. Am J Hum Genet 2010;86:749-64. Crossref

10. Tammimies K, Marshall CR, Walker S, et al. Molecular diagnostic yield of chromosomal microarray analysis and whole-exome sequencing in children with autism spectrum disorder. JAMA 2015;314:895-903. Crossref

11. Sebat J, Lakshmi B, Malhotra D, et al. Strong association of de novo copy number mutations with autism. Science 2007;316:445-9. Crossref

12. Chong WW, Lo IF, Lam ST, et al. Performance of chromosomal microarray for patients with intellectual disabilities/developmental delay, autism, and multiple congenital anomalies in a Chinese cohort. Mol Cytogenet 2014;7:34. Crossref

13. Tao VQ, Chan KY, Chu YW, et al. The clinical impact of chromosomal microarray on paediatric care in Hong Kong. PLoS One 2014;9:e109629. Crossref

14. Tirosh E, Jaffe M. Global developmental delay and mental retardation—a pediatric perspective. Dev Disabil Res Rev 2011;17:85-92. Crossref

15. Coulter ME, Miller DT, Harris DJ, et al. Chromosomal microarray testing influences medical management. Genet Med 2011;13:770-6. Crossref

16. Ellison JW, Ravnan JB, Rosenfeld JA, et al. Clinical utility of chromosomal microarray analysis. Pediatrics 2012;130:e1085-95. Crossref

17. Riggs ER, Wain KE, Riethmaier D, et al. Chromosomal microarray impacts clinical management. Clin Genet 2014;85:147-53. Crossref

18. World Health Organization. The ICD-10 classification of mental and behavioural disorders: diagnostic criteria for research: Geneva: World Health Organization; 1993.

19. Luiz D, Barnard A, Knosen N, Kotras N, Faragher B, Burns LE. Griffiths Mental Development Scales–Extended Revised. Two to Eight Years. Administration Manual. Oxford, UK: Hogrefe; 2006.

20. Wechsler D. Wechsler Preschool and Primary Scale of Intelligence–Revised. WPPSI-R: Psychological Corporation; 1989.

21. Diagnostic and Statistical Manual of Mental Disorders: DSM-IV. Washington, DC: American Psychiatric Association;1994: 535.

22. Kan AS, Lau ET, Tang WF, et al. Whole-genome array CGH evaluation for replacing prenatal karyotyping in Hong Kong. PLoS One 2014;9:e87988. Crossref

23. Kearney HM, Thorland EC, Brown KK, et al. American College of Medical Genetics standards and guidelines for interpretation and reporting of postnatal constitutional copy number variants. Genet Med 2011;13:680-5. Crossref

24. Boada R, Janusz J, Hutaff-Lee C, Tartaglia N. The cognitive phenotype in Klinefelter syndrome: a review of the literature including genetic and hormonal factors. Dev Disabil Res Rev 2009;15:284-94. Crossref

25. Lubs HA, Stevenson RE, Schwartz CE. Fragile X and Xlinked intellectual disability: four decades of discovery. Am J Hum Genet 2012;90:579-90. Crossref

26. Peprah E. Fragile X syndrome: the FMR1 CGG repeat distribution among world populations. Ann Hum Genet 2012;76:178-91. Crossref

27. Chen X, Wang J, Xie H, et al. Fragile X syndrome screening in Chinese children with unknown intellectual developmental disorder. BMC Pediatr 2015;15:77. Crossref

28. Daniel WW. Biostatistics: A Foundation for Analysis in the Health Sciences. 9th ed. Hoboken, NJ: J Wiley & Sons; 2009.