‘Consanguinity’ is a term derived from the Latin word ‘consanguineus’, meaning ‘of the same blood’. In medical genetics, consanguineous union is generally referred as a union between couples related as second cousins or closer.1 The prevalence of consanguinity varies significantly worldwide, depending on cultural background, religious belief, and geography. The highest rates are estimated in the Near and Middle East and in Northern Africa, where 20% to 50% of marriages are consanguineous.1 2 The prevalence in Southern Europe, South America, and Japan is about 1% to 5%, whereas Western European countries, North America, and Oceania have the lowest prevalence of <1%.1 2

Consanguineous union increases the risk of genetic disorders in offspring, especially for autosomal recessive diseases. However, recent studies suggest that parental consanguinity is also a risk factor for other adverse outcomes, even in developed multi-ethnic countries where the prevalence of consanguineous marriages is perceived as lower. For example, in Vienna where the background consanguinity rate was <1%, Posch et al3 reported that 39.7% of consanguineous couples had obstetric history of congenital malformations or genetic disorders. Becker et al4 reported that 6.1% of consanguineous couples were referred to a specialist centre in Germany for a history of major fetal anomalies. A 10-year retrospective analysis conducted in Australia, where the consanguinity rate is 5.5%, concluded that parental consanguinity was associated with higher rates of threatened premature labour, fetal congenital abnormality, stillbirth, and perinatal mortality.5 In that study, consanguinity was also found to be an independent risk factor of nearly 3-fold for stillbirth.

In Hong Kong, parental consanguinity is more frequent among non-Chinese ethnic minorities, which account for 8% of the total population.6 Internationally, healthcare workers lack knowledge on the risks of consanguinity.7 8 9 Inconsistencies in information provided during genetic counselling and screening has been observed.10 Consanguineous couples are often unaware of the potential health hazards in their offspring.11 12 13 The level of concern and awareness of the adverse effects of parental consanguinity among patients and physicians is low, and available data on consanguinity in Hong Kong are limited. Therefore, in the present study, we aimed to clarify the prevalence and characteristics of pregnancies from consanguineous unions in Hong Kong, and to assess the related effects on maternal, perinatal, and child health outcomes.

The Prenatal Diagnosis Clinic in Tuen Mun Hospital is responsible for counselling consanguineous couples. Dating ultrasound and counselling sessions for Down syndrome screening are arranged for all pregnant women who have their booking appointment in our locality. At the booking appointment, patients are also asked about consanguinity. Hospital-accredited interpreters are arranged for couples who are not fluent in Cantonese or English. Identification of consanguineous cases depends on self-reporting by couples. A pedigree chart is constructed for each case. Couples are counselled about the possible effects of parental consanguinity on pregnancy outcomes, and advised to attend antenatal care regularly.

A retrospective cohort study of all parental consanguinity cases over a 10-year period from 1 January 2007 to 31 December 2016 was conducted. The antenatal records of these cases were reviewed. Details were gathered about pregnancy loss, fetal congenital abnormalities, pregnancy and perinatal outcomes, and neonatal and childhood development in the preceding pregnancy. The family history of each case was also collected from patient records, including known genetic or congenital anomalies, or intellectual or developmental disabilities. A morphology scan was arranged for consanguineous cases. Each family pedigree was studied to determine the degree of parental consanguinity (Fig). Only couples fulfilling the definition of consanguineous unions (second cousins or closer) were included for analysis in the present study.

Socio-demographic characteristics were collected, including ethnicity, maternal and paternal age, religious beliefs, working status, education level, and occupation. Maternal antepartum and peripartum characteristics, and fetal and perinatal information were available. Information about the neonatal, infancy, and childhood outcomes of the offspring were retrieved from the public sector electronic record system.

The relationship between consanguinity and fetal, neonatal, infant, or childhood diseases that required long-term paediatric management was evaluated and categorised into one of three categories:

Category A—Improbable association with consanguinity: cases known to be caused by numerical or structural chromosomal abnormalities, or not to have an autosomal recessive mode of inheritance;

Category B—Probable association with consanguinity: cases known to have an autosomal recessive mode of inheritance, particularly when both parents were found to be the carriers of genetic disorders; and

Category C—Possible/unclear association with consanguinity: cases where the mode of inheritance was unclear, or when genetic testing was unremarkable.

The characteristics and outcomes of consanguineous cases were compared with a control group of non-consanguineous unions. The next record of a non-consanguineous case of the same ethnicity after that of a case of consanguineous union was selected as the control. This ensured the similar composition of ethnicity which might have socio-economic effects on the maternal and fetal outcomes within the study and control groups.14 As some consanguineous couples might have contributed more than one pregnancies in our database, only adverse past obstetric outcome in the immediately preceding pregnancy was counted in the analysis, and any positive family history reported by such couples was counted as one case only, in order to prevent duplicated entries for multigravida women. Most previous studies have not evaluated the effects of closer consanguinity that might increase risks of hereditary disorders.5 15 16 To evaluate the effect of degree of inbreeding, comparisons were made among ‘first cousin or closer’ (including first cousin and double first cousin), ‘beyond first cousin’ (including first cousin once removed and second cousin), and non-consanguineous relationships.

Approval of this study was granted by the research and ethics committee of the study hospital. Guidelines for reporting observational studies according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement were followed.

Statistical analysis was performed using SPSS (Windows version 22.0; IBM Corp, Armonk [NY], US). Cross-tabulation between degrees of consanguinity and the different variables was performed in order to evaluate the characteristics of the study population. Differences in continuous variables were compared using t test or one-way analysis of variance. Differences in categorical variables were analysed with Chi squared test or Fisher’s exact test. Linear regression was carried out to adjust the collinearity among variables. Multivariate logistic regression analysis was used to determine the risk of consanguinity for adverse pregnancy and perinatal outcomes, with adjustment of significant confounders. Adjusted odds ratio (OR) with 95% confidence interval (CI) were calculated. Statistical significance was established for P<0.05.

Of 56 657 fetuses, 334 (0.6%) were conceived by consanguineous parents; of these, the majority (85.0%, 284 of 334) were ethnic Pakistani (among whom the prevalence of consanguineous union is highest, at 30.5%), followed by Indian (6.2%), Nepalese (2.7%), Filipino (0.4%), and Chinese (0.04%) [Table 1]. Of all consanguineous unions, the majority were first cousin consanguineous unions (76.6%) and double first cousin unions (1.8%); together, these were categorised as first cousin or closer (≤1C) unions. The remainder were categorised as beyond first cousin (>1C) unions, and included first cousin once removed unions (5.1%), and second cousin unions (16.5%). Comparison of background variables including maternal and paternal age, education level, religion, length of stay in Hong Kong, marital status, working status, occupation, parity, and body mass index showed no significant differences between the consanguineous group and the non-consanguineous control group (Table 2).

Women in consanguineous unions were significantly more likely to have experienced congenital abnormality (10.5% vs 0.4%; P<0.001), unexplained intrauterine fetal demise (4.2% vs 0.4%; P=0.005) and neonatal death (4.6% vs 1.2%; P=0.024) in the preceding pregnancy, and family history of congenital abnormality (4.6% vs 0%; P<0.001) than were non-consanguineous controls (Table 3). Down syndrome screening was offered to all women, but the attendance was only about one-fifth for all groups.

In terms of major maternal and perinatal complications, there were no significant differences between the non-consanguineous control group and the overall consanguineous group or the subgroups, except that pregnancies of ≤1C unions were more often complicated with pre-eclampsia (4.2% vs 1.2%; P=0.02) than were those of the non-consanguineous control group (Table 4).

Altogether there were 58 fetuses and 14 fetuses having different abnormalities, from 55 consanguineous and 14 control couples respectively (Table 5). Offspring of consanguineous couples had a higher risk of having category C disorders (OR=4.60; 95% CI=2.35-9.00) or category B disorders (OR=8.70; 95% CI=1.06-71.36), compared with those of non-consanguineous couples. The overall prevalence of category C disorders (14.7%) was higher than that of category B disorders (2.4%). Compared with the non-consanguineous control group, the prevalence of category C disorders was significantly higher in the ≤1C subgroup (OR=5.59; 95% CI=2.83-11.06); it was lower in the >1C subgroup, but the difference was not significant.

The prevalence of structural malformations was higher in the consanguineous group than that in the non-consanguineous control group, especially for those abnormalities involving cardiovascular, musculoskeletal, and urological systems (Table 5). Parental consanguinity also significantly increased the risk of developmental delay in offspring of consanguineous couples (OR=6.72, 95% CI=1.48-30.63) and in those of ≤1C couples (OR=7.64, 95% CI=1.64-35.58). Autism spectrum disorder was more prevalent in offspring of consanguineous couples (2.1%) than in those of non-consanguineous couples (0%) [P=0.008]. The diseases recorded in the consanguineous group and in the control group are detailed in online supplementary Appendices 1 and 2, respectively.

To the best of our knowledge, this is the first comprehensive study in Hong Kong describing the prevalence of parental consanguinity. Our results support those of previous studies that revealed a higher prevalence of parental consanguinity in certain ethnic groups, and the higher prevalence of known genetic disorders (category B) among their offspring. In addition, our study has revealed that the prevalence of fetal structural abnormalities, developmental delay, and autism spectrum disorders (category C) are also high. This has implications for prenatal counselling and diagnosis, and related healthcare services.

Our comparison of maternal age and parity showed no significant difference between the consanguineous group and control group. This is in contrast to findings by Islam et al16 and Hosseini-Chavoshi et al,17 who found that women in consanguineous unions were younger and of higher parity in Iran and Oman, where the consanguinity rate was more than 30%. Studies in India and Pakistan populations also showed that mothers in consanguineous relationships were more likely to be socially and economically disadvantaged.11 18 The similarity in the socio-economic characteristics between the consanguineous and non-consanguineous unions of our study indicates that socio-economic factors are unlikely to be causes of the poorer fetal outcomes, both in the index pregnancy and the preceding pregnancy, found in our consanguineous group.

We identified eight offspring with autosomal recessive diseases in the consanguineous group, including three cases of beta-thalassaemia major and five cases of other rarer diseases (online supplementary Appendix 1). Although the carrier status of thalassaemia can be screened by low mean corpuscular volume of red blood cells, the carrier status of other recessive disorders can be more complex. For some disorders, comprehensive genetic carrier screening using exome sequencing is required.4 19 20 21 Our data provide useful information for preconception counselling for consanguineous couples. However, exome sequencing is expensive, and this screening test is not yet available in public hospitals. Health education and information about the availability of carrier screening should be provided to all pregnant women, regardless of cultural, religious, or socio-economic background. Once a consanguineous couple is diagnosed to be the carrier of a genetic disease, they should be encouraged to discuss carrier screening with their siblings, who may also carry the same recessive gene and be in consanguineous union. Access to obstetric care and genetic counselling services in prenatal diagnosis clinics allows couples to make informed choices. Knowledge on various cultural, religious, or socio-economic issues allows healthcare workers to provide appropriate support and to best advise patients.

Our results revealed that category C disorders are more prevalent among offspring of consanguineous couples, especially in the ≤1C subgroup. Fetal structural ultrasonographic examination should be offered to ≤1C couples, especially for the cardiovascular, urological, and skeletal systems.22 23 24 25 26 Detailed genetic counselling and investigation services must be offered to ≤1C couples if fetal abnormalities are detected.3 4

Our results revealed increased risk of developmental and behavioural disorders for offspring of consanguineous couples. However, disorders such as developmental delay and autism spectrum disorder are not diagnosable before birth. Preconception and prenatal counselling should be offered to consanguineous couples, who should also be reminded about regular postnatal follow-up examinations, in order to avoid any delay in diagnosing any developmental or behavioural disorders.27

Pakistani ethnicity accounted for only 1.6% of all fetuses but 85% of consanguineous couples in our study. According to the Hong Kong 2016 population by-census, 0.25% of the total Hong Kong population was of Pakistani ethnicity.6 However, the majority of this local Pakistani population is within potentially reproductive age-groups (15-24 years, 19.2%; 25-34 years, 14.9%; 35-44 years, 21.3%), and they tend to have more children per couple than do ethnic Chinese couples.6 It is essential to include information about the increased risks of parental consanguinity during the antenatal care and provide appropriate genetic counselling once a consanguineous couple is identified.

In addition to poor fetal outcomes, we also found a 3-fold increased risk of pre-eclampsia among women in ≤1C unions. Familial aggregation and possible genetic correlation of pre-eclampsia have been observed, but the exact effect of consanguinity remains controversial.28 29 Mumtaz et al15 suggested that parental consanguinity is a risk factor of 1.6-fold for preterm birth at less than 33 weeks of gestation. Low birth weight has also been associated with first-cousin relationships, but the risk increase was found to be marginal (OR=1.36)30. Our study did not confirm higher incidences of antepartum, peripartum, neonatal and perinatal complications in overall consanguinity. Findings on the effect of consanguinity on various complications are inconsistent, especially when these complications are multifactorial in pathogenesis.5 15 27 29 30

One limitation of our study is the retrospective nature that might have led to incompleteness of information for analysis, especially when previous pregnancies were not in Hong Kong. Another limitation is that some of the fetal abnormalities classified under category C may in fact be category B disorders, as some of them recurred in the same couples (online supplementary Appendix 1); the majority of category C disorders did not receive genetic investigations. However, there is a high dependence on public health service in our locality, and this facilitated data retrieval of postnatal, infancy, and childhood outcomes of the offspring. Different types of parental consanguinity were also included in our analysis to provide the stratified risks according to the degree of inbreeding. Collection of socio-economic characteristics was also comprehensive. The same composition of ethnicity in both the consanguineous and control groups further minimised the socio-economic confounding effects in the analysis. Another limitation is that the genetic data were often incomplete or not up-to-date for the studied cases, which were recorded from 2007 to 2016.

It is recommended that a territory-wide prospective study is conducted on consanguineous couples to further delineate their healthcare needs in Hong Kong.

Identification of consanguineous couples is essential to ensure appropriate referral for preconception or prenatal counselling and diagnosis. Our study showed the majority of consanguineous unions in Hong Kong are of Pakistani ethnicity. International studies have reported that in addition to recessive genetic diseases, offspring of consanguineous unions have higher incidences of non–genetically confirmed structural abnormalities, developmental delay, and autism spectrum disorders. The present study confirms this in the Hong Kong population. Information on the increased risks associated with parental consanguinity should be included in genetic counselling for consanguineous couples, so that they can make informed decisions.

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.

Concept or design of the study: All authors.
Acquisition of data: KH Siong.
Analysis or interpretation of data: KH Siong, TY Leung.
Drafting of the manuscript: KH Siong, TY Leung.
Critical revision for important intellectual content: All authors.

The authors have no conflicts of interest to disclose.

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

Ethics approval was obtained from New Territories West Cluster Clinical Research Ethics Committee (Ref NTWC/CREC/18012).

1. Bittles A. Consanguinity and its relevance to clinical genetics. Clin Genet 2001;60:89-98. Crossref

2. Consang.net. Global prevalence of consanguinity. Available from: http://www.consang.net/index.php/Global_prevalence. Accessed 1 Apr 2018.

3. Posch A, Springer S, Langer M, Blaicher W, Streubel B, Schmid M. Prenatal genetic counseling and consanguinity. Prenat Diagn 2012;32:1133-8. Crossref

4. Becker R, Keller T, Wegner RD, et al. Consanguinity and pregnancy outcomes in a multi-ethnic, metropolitan European population. Prenat Diagn 2015;35:81-9. Crossref

5. Kapurubandara S, Melov S, Shalou E, Alahakoon I. Consanguinity and associated perinatal outcomes, including stillbirth. Aust N Z J Obstet Gynaecol 2016;56:599-604. Crossref

6. Hong Kong SAR Government. Hong Kong 2016 Population By-census. Available from: https://www.bycensus2016.gov.hk/en/bc-articles.html. Accessed 1 Apr 2018.

7. Barrett P. A review of consanguinity in Ireland—estimation of frequency and approaches to mitigate risks. Ir J Med Sci 2016;185:17-28. Crossref

8. Teeuw ME, Hagelaar A, ten Kate LP, Cornel MC, Henneman L. Challenges in the care for consanguineous couples: an exploratory interview study among general practitioners and midwives. BMC Fam Pract 2012;13:105. Crossref

9. Aalfs CM, Smets EM, de Haes HC, Leschot NJ. Referral for genetic counselling during pregnancy: limited alertness and awareness about genetic risk factors among GPs. Fam Pract 2003;20:135-41. Crossref

10. Bennett RL, Hudgins L, Smith CO, Motulsky AG. Inconsistencies in genetic counseling and screening for consanguineous couples and their offspring: the need for practice guidelines. Genet Med 1999;1:286-92. Crossref

11. Joseph N, Pavan KK, Ganapathi K, Apoorva P, Sharma P, Jhamb JA. Health awareness and consequences of consanguineous marriages: a community-based study. J Prim Care Community Health 2015;6:121-7. Crossref

12. Jaber L, Romano O, Halpern GJ, Livne I, Green M, Shohat T. Awareness about problems associated with consanguineous marriages: survey among Israeli Arab adolescents. J Adolesc Health 2005;36:530. Crossref

13. Teeuw ME, Loukili G, Bartels EA, ten Kate LP, Cornel MC, Henneman L. Consanguineous marriage and reproductive risk: attitudes and understanding of ethnic groups practising consanguinity in Western society. Eur J Hum Genet 2014;22:452-7. Crossref

14. Khalil A, Rezende J, Akolekar R, Syngelaki A, Nicolaides KH. Maternal racial origin and adverse pregnancy outcome: a cohort study. Ultrasound Obstet Gynecol 2013;41:278-85. Crossref

15. Mumtaz G, Nassar AH, Mahfoud Z, et al. Consanguinity: a risk factor for preterm birth at less than 33 weeks’ gestation. Am J Epidemiol 2010;172:1424-30. Crossref

16. Islam MM. Effects of consanguineous marriage on reproductive behaviour, adverse pregnancy outcomes and offspring mortality in Oman. Ann Hum Biol 2013;40:243-55. Crossref

17. Hosseini-Chavoshi M, Abbasi-Shavazi MJ, Bittles AH. Consanguineous marriage, reproductive behaviour and postnatal mortality in contemporary Iran. Hum Hered 2014;77:16-25. Crossref

18. Bhopal RS, Petherick ES, Wright J, Small N. Potential social, economic and general health benefits of consanguineous marriage: results from the Born in Bradford cohort study. Eur J Public Health 2014;24:862-9. Crossref

19. Teebi AS, El-Shanti HI. Consanguinity: implications for practice, research, and policy. Lancet 2006;367:970-1. Crossref

20. Fareed M, Afzal M. Genetics of consanguinity and inbreeding in health and disease. Ann Hum Biol 2017;44:99-107. Crossref

21. Committee on Genetics. Committee Opinion No. 690: Carrier screening in the age of genomic medicine. Obstet Gynecol 2017;129:e35-40. Crossref

22. Stoll C, Alembik Y, Roth MP, Dott B. Parental consanguinity as a cause for increased incidence of births defects in a study of 238,942 consecutive births. Ann Genet 1999;42:133-9.

23. Sheridan E, Wright J, Small N, et al. Risk factors for congenital anomaly in a multiethnic birth cohort: an analysis of the born in Bradford study. Lancet 2013;382:1350-9. Crossref

24. Nabulsi MM, Tamim H, Sabbagh M, Obeid MY, Yunis KA, Bitar FF. Parental consanguinity and congenital heart malformations in a developing country. Am J Med Genet A 2003;116:342-7. Crossref

25. Becker SM, Al Halees Z, Molina C, Paterson RM. Consanguinity and congenital heart disease in Saudi Arabia. Am J Med Genet 2001;99:8-13. Crossref

26. Yunis K, Mumtaz G, Bitar F, et al. Consanguineous marriage and congenital heart defects: a case-control study in the neonatal period. Am J Med Genet A 2006;140:1524-30.

27. Abbas HA, Yunis K. The effect of consanguinity on neonatal outcomes and health. Hum Hered 2014;77:87-92. Crossref

28. Berends AL, Steegers EA, Isaacs A, et al. Familial aggregation of preeclampsia and intrauterine growth restriction in a genetically isolated population in The Netherlands. Eur J Hum Genet 2008;16:1437-42. Crossref

29. Sezik M, Ozkaya O, Sezik HT, Yapar EG, Kaya H. Does marriage between first cousins have any predictive value for maternal and perinatal outcomes in pre-eclampsia? J Obstet Gynaecol Res 2006;32:475-81.

30. Poorolajal J, Ameri P, Soltanian A, Bahrami M. Effect of consanguinity on low birth weight: a meta-analysis. Arch Iran Med 2017;20:178-84.

Clostridioides difficile infection (CDI) is a leading cause of healthcare-associated diarrhoea and is associated with significant morbidity and mortality.1 In the United States, C difficile has become the most lethal acute enteric pathogen, with the Centers for Disease Control designating it as an urgent antibiotic resistance threat in 2015, one of only three pathogens to earn this distinction (http://www.cdc.gov/drugresistance/biggest_threats.html).2 In Hong Kong, an annual increase of 26% in CDI was noted from 2006 to 2014.3 This is even more alarming for older adults aged ≥65 years, in whom the estimated crude incidence rate is 133 to 207 cases per 100 000 population.4 It is believed that CDI arises when the normal gut microbiota is disrupted. This allows for the colonisation of C difficile, whose production of cytotoxins leads to disease.5 A recent study showed that apart from microbiota dysbiosis, enteric virome dysbiosis may also play an important role.6 Risk factors include nursing home care, recent hospitalisation, antibiotics exposure, and proton-pump inhibitor use.4 First-line treatment for non-fulminant C difficile–associated diarrhoea includes the cessation of unnecessary medications and treatment with vancomycin or fidaxomicin. For more severe cases, a combination of vancomycin and intravenous metronidazole with early consideration of surgery may be required.7 However, it is well known that a significant proportion of cases relapse or recur despite first-line therapy.8 For recurrent cases, taper and pulse regimens of vancomycin can be offered, but response rates decline further with multiple recurrences.9 Fidaxomicin is another promising treatment option, but its usage is limited by its price and availability.10 These difficult-to-treat cases are associated with extended hospital stays and may result in widespread nosocomial outbreaks. In their seminal paper, van Nood et al11 demonstrated that faecal microbiota transplantation (FMT), in which infusion of faeces from healthy donors is given to individuals with disease, was efficacious for recurrent CDI. Since then, this finding has been confirmed by a systematic review and meta-analysis of >160 clinical studies.12 In view of the limited options available for treatment of recurrent or refractory cases, we decided to establish a pilot FMT service to address this treatment gap. Initially, screening of relatives meeting strict donor criteria and willing to donate fresh stools was performed; later on, with the establishment of the Healthy Donor Stool Biobank by the Faculty of Medicine, The Chinese University of Hong Kong in 2017, pre-screened frozen stools were made available. The objectives of this study, the first of its kind in Hong Kong, was to assess the safety, efficacy, and technical and logistical feasibility of performing FMT.

We conducted a single-centre, retrospective review of all consecutive cases of recurrent or refractory CDI with FMT performed at an academic hospital from 2013 to 2018. Data on patient demographics, co-morbidities, route of administration of FMT, treatment efficacy, adverse events, and other relevant clinical data were extracted from the Clinical Management System of the Hospital Authority, Hong Kong or from review of case notes under the auspices of the FMT registry. Stringent donor criteria based on guidance and protocols were applied to fresh and frozen donor stool to ensure patient safety.13 Screening for viral hepatitis, blood-borne viruses, pathogenic bacteria, enteric viruses, syphilis, multidrug resistant organisms, Helicobacter pylori, C difficile, and parasites was performed (online supplementary Appendix). The FMT product was delivered either via a feeding tube or through scope channels during oesophagogastroduodenoscopy (OGD) or colonoscopy. Feeding tube position was confirmed prior to FMT delivery. For the OGD route, the endoscopist performed a routine OGD to ensure there were no ulcerations or other contra-indications for FMT. Afterwards, the endoscopist tried to pass the scope as distally as possible (at least to the second or third parts of the duodenum) to minimise reflux of the FMT back into the stomach. During the procedure carbon dioxide for luminal insufflation was used to minimise gaseous distention, discomfort, and the urge to retch or vomit. Before FMT delivery, the patient is sat upright to minimise the risk of aspiration. For the colonoscopy route, the endoscopist performed a routine colonoscopy to ensure there are no large ulcers or other contra-indications for FMT. Afterwards, the endoscopist tried to pass the scope as proximally as possible, but if the patient’s tolerance was an issue, administering FMT into the left side of the colon or even the rectum was considered reasonable, depending on the premorbid state. The majority of FMT procedures were performed in the Endoscopy Centre with close monitoring, followed by further observation in the wards for any potential adverse events. The present report was compiled in accordance with the PROCESS statement.14

A total of 24 patients with recurrent or refractory CDI who received FMT were identified. The patients’ baseline characteristics are shown in the Table. Their median age was 70 years (interquartile range=45.0-78.3 years). More than two-thirds of the patients were male. More than half of the patients were in either a bedridden or chair-bound state, a surrogate of poor functional status. The majority (>80%) of patients had been hospitalised within the most recent 3 months or were long-term care facility residents. All patients had at least one co-morbidity. The most common co-morbidity was hypertension (n=10, 41.7%), followed by inflammatory bowel disease (IBD) [n=6, 25.0%] and stroke (n=5, 20.8%). The FMT was performed by experienced gastroenterologists in accordance with published protocols11 and the latest guidelines.15 All patients were given 50 g of FMT product unless otherwise stated. The FMT was delivered via feeding tube in 11 (45.8%), OGD in eight (33.3%), and colonoscopy in six (25.0%) of the patients (one patient had FMT performed via both OGD and colonoscopy during the same session). Resolution of diarrhoea without relapse within 8 weeks was achieved in 21 out of 24 patients (87.5%). Three patients (12.5%) did not show significant improvement in their symptoms after the therapy; therefore, a clinical decision was made to perform repeat FMT, after which all three patients showed resolution of diarrhoea. Two of those were confirmed C difficile–negative on repeat stool testing; the remaining patient was frail and did not save repeat stools, but no documented recurrence within 8 weeks was noted. No deaths occurred at 30 days. The procedure was generally well tolerated, with no serious adverse events attributable to FMT. The most common complication was abdominal pain (n=3, 12.5%) after FMT. Bloating was reported in one (4.2%) patient.

Because it replenishes colonic microbial diversity, FMT removes ecological niches that would otherwise be occupied by C difficile.16 A recent study has shown that the prevalence of CDI in Asia is similar to the high prevalence in North America and Europe.17 Furthermore, it is well known that the treatment efficacy of antibiotics declines with multiple recurrences,9 rendering them less effective in difficult-to-treat cases. We established Hong Kong’s first FMT service to address the treatment gap that currently exists. Our results suggest that the efficacy of FMT for recurrent or refractory CDI is comparable to that reported in the literature, with an excellent safety profile.11 12

According to the Census and Statistics Department, the population of Hong Kong is projected to increase from 6.8 million in mid-2003 to 8.38 million in mid-2033 with a continuous ageing trend.18 The proportion of those aged ≥65 years is projected to rise markedly, from 11.7% in 2003 to 27% in 2033. Life expectancy in Hong Kong has already increased to age 79.5 years, behind only Japan and Switzerland.18 Increasing numbers of elderly people living in residential care homes, together with the widespread use of antibiotics and proton-pump inhibitors4, mean that the incidence of CDI in elderly people will likely increase in Hong Kong, as it has in other developed countries.18

The incidence of IBD in Hong Kong has risen by almost 3-fold in the past two decades.19 Recently in the West, as the incidence and severity of CDI has increased in the general population, even greater increases have been described in patients with IBD.2 In the present study, a significant proportion of patients with CDI also had IBD, and it is likely that the incidence and severity of CDI in IBD patients in Hong Kong will increase further in future. Establishment of FMT services has been advocated for healthcare systems to effectively manage expected increases in refractory or recurrent CDI.20

A concerted effort by policymakers will be necessary to establish a dedicated centre to provide territory-wide FMT services to tackle the growing disease burden. Such a centre would require a multidisciplinary team involving gastroenterologists, microbiologists, infectious disease specialists, specialty nurses, and research personnel together with the cooperation of wards, the healthy donor stool biobank, and endoscopy centres. Such an FMT centre may receive territory-wide consultations and referrals. It could also provide training, knowledge transfer, and accreditation to healthcare professionals. Currently, the status of FMT as a therapeutic agent is still evolving. The United States Food and Drug Administration classifies human stool as a biological agent and asserts that its use in FMT should be regulated, although their draft guidance mentions that it intends to exercise enforcement discretion for its use in recurrent CDI that does not respond to standard therapies.21 This highlights the importance of proper oversight and governance to mitigate any potential risks that may arise. Data on long-term safety outcomes are also scarce, highlighting the importance and role of registries to provide more insight, another function a centralised and specialised FMT centre would be able to undertake.

Our study has several limitations. As a case series, the data presented are mainly descriptive and uncontrolled with the possibility of bias. Further, the sample size is relatively small. Controlled trials with a larger sample size are required to confirm these findings and to optimise the timing of FMT administration. Despite these limitations, resolution of diarrhoea was at a similar rate in the present study to that reported in the literature.11 Furthermore, as the only centre in Hong Kong that has provided FMT, the results reported here would be highly relevant for both clinicians and policymakers concerning its safety, efficacy, and feasibility.

Our results show that FMT is safe, efficacious, and feasible for treating patients with recurrent or refractory CDI in Hong Kong. Given the lack of effective alternatives for difficult-to-treat cases of CDI, and demographic trends that will likely lead to increased incidence of CDI, demand for FMT is expected to rise. A territory-wide FMT service should be established to address this expected increase in demand.

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.

Concept or design of the study: All authors.
Acquisition of data: RN Lui, LHS Lau, KCY Cheung.
Analysis or interpretation of data: RN Lui, LHS Lau, KCY Cheung.
Drafting of the manuscript: RN Lui, SH Wong, PKS Chan, SC Ng.
Critical revision for important intellectual content: All authors.

All authors have disclosed no conflicts of interest.

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

The present study was reviewed and approved by the Joint Chinese University of Hong Kong/New Territories East Cluster Clinical Research Ethics Committee (Ref 2017.260).

1. Lessa FC, Mu Y, Bamberg WM, et al. Burden of Clostridium difficile infection in the United States. N Engl J Med 2015;372:825-34. Crossref

2. Khanna S, Shin A, Kelly CP. Management of Clostridium difficile infection in inflammatory bowel disease: expert review from the clinical practice updates committee of the AGA institute. Clin Gastroenterol Hepatol 2017;15:166-74. Crossref

3. Ho J, Dai RZ, Kwong TN, et al. Disease burden of Clostridium difficile infections in adults, Hong Kong, China, 2006-2014. Emerg Infect Dis 2017;23:1671-9. Crossref

4. Wong SH, Ip M, Hawkey PM, et al. High morbidity and mortality of Clostridium difficile infection and its associations with ribotype 002 in Hong Kong. J Infect 2016;73:115-22. Crossref

5. Mylonakis E, Ryan ET, Calderwood SB. Clostridium difficile–associated diarrhea: a review. Arch Intern Med 2001;161:525-33. Crossref

6. Zuo T, Wong SH, Lam K, et al. Bacteriophage transfer during faecal microbiota transplantation in Clostridium difficile infection is associated with treatment outcome. Gut 2018;67:634-43.

7. McDonald LC, Gerding DN, Johnson S, et al. Clinical Practice Guidelines for Clostridium difficile infection in adults and children: 2017 update by the Infectious Diseases Society of America (IDSA) and Society for Healthcare Epidemiology of America (SHEA). Clin Infect Dis 2018;66:987-94. Crossref

8. Pépin J, Routhier S, Gagnon S, Brazeau I. Management and outcomes of a first recurrence of Clostridium difficile–associated disease in Quebec, Canada. Clin Infect Dis 2006;42:758-64. Crossref

9. Johnson S. Recurrent Clostridium difficile infection: a review of risk factors, treatments, and outcomes. J Infect 2009;58:403-10. Crossref

10. Cornely OA, Miller MA, Louie TJ, Crook DW, Gorbach SL. Treatment of first recurrence of Clostridium difficile infection: fidaxomicin versus vancomycin. Clin Infect Dis 2012;55 Suppl 2:154-61. Crossref

11. van Nood E, Vrieze A, Nieuwdorp M, et al. Duodenal infusion of donor feces for recurrent Clostridium difficile. N Engl J Med 2013;368:407-15. Crossref

12. Lai CY, Sung J, Cheng F, et al. Systematic review with meta-analysis: review of donor features, procedures and outcomes in 168 clinical studies of faecal microbiota transplantation. Aliment Pharmacol Ther 2019;49:354-63. Crossref

13. Woodworth MH, Neish EM, Miller NS, et al. Laboratory testing of donors and stool samples for fecal microbiota transplantation for recurrent Clostridium difficile infection. J Clin Microbiol 2017;55:1002-10. Crossref

14. Agha RA, Borrelli MR, Farwana R, et al. The PROCESS 2018 statement: updating consensus Preferred Reporting Of CasE Series in Surgery (PROCESS) guidelines. Int J Surg 2018;60:279-82. Crossref

15. Cammarota G, Ianiro G, Tilg H, et al. European consensus conference on faecal microbiota transplantation in clinical practice. Gut 2017;66:569-80. Crossref

16. Seekatz AM, Aas J, Gessert CE, et al. Recovery of the gut microbiome following fecal microbiota transplantation. MBio 2014;5:e00893-14. Crossref

17. Borren NZ, Ghadermarzi S, Hutfless S, Ananthakrishnan AN. The emergence of Clostridium difficile infection in Asia: a systematic review and meta-analysis of incidence and impact. PLoS One 2017;12:e0176797. Crossref

18. Census and Statistics Department, Hong Kong SAR Government. Hong Kong population projection 2004-2033, report of the task force on population policy. Available from: https://www.censtatd.gov.hk/media_workers_corner/pc_rm/hong_kong_population_projections_2004_2033__/index.jsp. Accessed 8 Mar 2019.

19. Lui RN, Ng SC. The same intestinal inflammatory disease despite different genetic risk factors in the East and West? Inflamm Intest Dis 2016;1:78-84. Crossref

20. Costello SP, Tucker EC, La Brooy J, Schoeman MN, Andrews JM. Establishing a fecal microbiota transplant service for the treatment of Clostridium difficile infection. Clin Infect Dis 2016;62:908-14. Crossref

21. Guidance for industry: enforcement policy regarding investigational new drug requirements for use of fecal microbiota for transplantation to treat Clostridium difficile infection not responsive to standard therapies. Food and Drug Administration, US Department of Health and Human Services, US Government; 2016.

Enhanced recovery after surgery (ERAS) is a multimodal pathway developed to improve recovery after major surgery. Since its formal introduction in the 1990s, ERAS has been adopted quickly because of the cost efficiency derived from its reduction in length of hospital stays, an important issue in the context of current rapidly increasing healthcare costs and the consequent need for optimisation.1 2 Application of ERAS integrates various medical interventions involving surgeons, anaesthetists, physiotherapists, dieticians, and nurses.3 The benefits of ERAS have been well proven in colectomy.4 5 6 7 Liver cancer is the fourth leading cause of cancer death in both sexes worldwide.8 Liver resection remains the mainstay of curative treatment for liver cancer. Liver resection is associated with a high rate of postoperative morbidity ranging from 15% to 48%9 10 and a postoperative hospital stay of 9 to 15 days.11 The high rates of complications lead to prolonged hospital stay and increase costs of hospitalisation.

An ERAS programme combines a number of elements that aim to enhance postoperative recovery, facilitate earlier discharge, and reduce surgical stress response.3 4 It mainly focuses on minimising the impact of surgery on patient homeostasis.12 The reduction of postoperative physiological stress by attenuation of the neurohormonal response to the surgical intervention not only provides the basis for a faster recovery but also diminishes the risk of organ dysfunction and complications.13 Programmes for ERAS consist of well-organised pathways of clinical interventions that begin with out-patient preoperative information, counselling, and physical optimisation; proceed to pre-, intra-, and post-operative protocol-driven actions; and end with patient discharge following pre-established criteria.14 The main pillars of ERAS are extensive preoperative counselling, no bowel preparation, no sedative premedication, no preoperative fasting, preoperative carbohydrate loading, tailored anaesthesiology, perioperative intravenous fluid restriction, non-opioid pain management, no routine use of drains and nasogastric tubes, early removal of the urinary catheter, and early postoperative feeding and mobilisation.15 16 Several major studies have suggested that ERAS is feasible and significantly reduces complications and the length of hospital stay for patients undergoing colonic resection.4 5 6 7 17 Furthermore, ERAS has been successfully applied to urological,18 cardiovascular,19 gynaecological,20 orthopaedic,21 and thoracic surgeries.22 However, the literature on ERAS after liver resection is limited. The aim of the present study was to evaluate the safety and efficacy of an ERAS programme for open or laparoscopic liver resection.

This was a prospective feasibility study carried out in a tertiary academic hospital. The inclusion criteria recruited all consecutive patients undergoing elective liver resection who were aged 18 to 70 years, with American Society of Anesthesiologists (ASA) grade I or II, with no severe physical disabilities, who required no assistance with activities of daily living, and with informed consent available. Patients undergoing emergency surgery, who had received preoperative portal vein embolisation, who were expected to receive concomitant procedures other than cholecystectomy, who were mentally incapable of written consent, and women who were pregnant were excluded.

During the same period, 42 patients who fulfilled the same inclusion criteria underwent liver resection and a conventional perioperative programme, as the On-Q Pain Buster system (I-Flow Corporation, Lake Forest [CA], US) was not available for financial reasons. None of the control patients were assigned to that group because they refused the ERAS programme. Propensity score matching analysis was used to minimise bias and confounding factors in patient selection, and 20 matched pairs of patients were generated for comparison.

The same team of hepatobiliary surgeons experienced in both laparoscopic and liver surgery performed all operations. Our open and laparoscopic techniques have been described previously.23 In brief, open hepatectomy was performed via right subcostal incisions with upward midline extensions and in some cases with left subcostal extensions. In most cases, the liver was mobilised in standard fashion before parenchymal transection, whereas in the rest, we adopted the anterior approach or the hanging technique. Liver transection was performed with a cavitron ultrasonic surgical aspirator (Valleylab, Boulder [CO], US) and TissueLink (TissueLink Medical Inc, Dover [DE], US). For laparoscopic hepatectomy, a combination of TissueLink and LigaSure (Valleylab) were used for liver transection. The Pringle manoeuvre was not routinely applied during liver resection. Endovascular staplers (Tyco Healthcare, Norwalk [CT], US) were used to divide larger vascular pedicles.

Details of the ERAS programme and the conventional perioperative programme are summarised in Table 1. The design of the ERAS programme was based on consensus between our surgeons, anaesthetists, physiotherapists, dieticians and nurses, who reviewed the relevant literature and made appropriate adjustments to suit the local situation. Patients who were to undergo elective hepatectomy were first screened in an out-patient clinic or in wards for eligibility for the ERAS programme. Patients who fulfilled the inclusion and exclusion criteria were interviewed by the principal investigator or co-investigators for the recruitment and preoperative counselling. A guided tour on surgical ward led by a trained nurse and an information booklet about the preoperative guidance were given to each patient. The booklet described the method used for respiratory rehabilitation, daily medical events after admission, daily mobilisation goals, and nutritional goals after the operation. The patient was seen at a preoperative anaesthesia clinic for preoperative assessment of risk adjustment and education about the fast-track anaesthetic and postoperative pain management, especially during mobilisation.

All patients received a 20-mL local infiltration of local anaesthesia (0.25% levobupivacaine) followed by continuous wound instillation at 4 mL/h for 72 h using the On-Q Pain Buster System balloon pump (I-Flow Corporation). Pain control was supplemented using opioid-sparing multimodal analgesia, including oral paracetamol and non-steroidal anti-inflammatory drugs. For minimally invasive liver resection, continuous infiltration of the wound with local anaesthetic agents was used and early mobilisation started on postoperative day 0. The principal investigator held regular audit meetings with the research team and medical/nursing staff to ensure protocol compliance.

Patients could be discharged if they fulfilled the discharge criteria, which consisted of (1) adequate pain control with oral analgesics, (2) absence of nausea, (3) ability to tolerate solid food, (4) liver function on an improving trend, (5) mobilisation and self-support as compared to the preoperative level, and (6) acceptance of discharge by the patient.

The primary outcome of the study was total postoperative hospital stay, including that of patients readmitted within 30 days after surgery. The secondary outcomes of the study included the readmission rate and morbidity and mortality within 30 days.

The clinical outcomes of patients who underwent liver resection and received the ERAS programme were compared with those who received a conventional perioperative programme in the same period. Propensity score matching analysis was performed to control for potential bias. Sex, age, number of co-morbidities, ASA grade, diagnosis, presence of cirrhosis, and type of resection were chosen as our baseline covariates to calculate each patient’s propensity score. The propensity scores were estimated by fitting a logistic regression model with the above covariates. The patients were then matched by their propensity scores using one-to-one nearest neighbour matching without replacement.

Statistical analyses and propensity score matching calculations were performed using SPSS (Windows version 20.0; IBM Corp, Armonk [NY], US). Chi squared tests (or Fisher’s exact tests, when appropriate) were used to compare categorical data. Mann-Whitney U tests were used to compare continuous, non-normally distributed outcomes between treatment groups. A two-sided P<0.05 was considered to be statistically significant.

A total of 20 patients who underwent liver resection at Prince of Wales Hospital, Hong Kong, from September 2015 to July 2016, were recruited into the ERAS programme. Their median age was 58 years (range, 33-77 years). The majority (n=19, 95%) of the patients were in ASA grade II. Hepatocellular carcinoma (n=13, 65%) and colorectal liver metastasis (n=5, 25%) were the main indications for operation. All patients had Child-Pugh score class A. Major and minor hepatectomy were performed in eight (40%) and 12 (60%) patients, respectively. Minimally invasive hepatectomy (laparoscopic or robotic) were performed in nine patients, and the remaining 11 (55%) patients received open hepatectomy. There were no major complications as defined by the Clavien-Dindo classification of surgical complications, and no patients required readmission.24 25 Only two (10%) patients developed minor complications, which were wound seroma (n=1, 5%) and urinary retention (n=1, 5%).

The demographics of patients in the ERAS and conventional perioperative programme groups were comparable (Table 2). Perioperative outcomes are summarised in Table 3. There were no significant differences in patient demographics, liver function, tumour characteristics, or surgical techniques between the two groups. When compared with the conventional perioperative programme, the ERAS programme was associated with a significantly shorter postoperative hospital stay (5 vs 6 days, P=0.033). There was no significant difference in rates of postoperative complications or readmission.

Results from the present study indicate that the ERAS programme is safe and feasible in both open and laparoscopic liver resections in Hong Kong. There was a significant reduction in the length of postoperative hospital stay in the ERAS group.

Although ERAS programmes are not new, their development in liver resection has been relatively slow because of the operation’s high complexity and the high frequency of underlying liver cirrhosis in this group of patients. Patients with liver cirrhosis who undergo liver resection have special concerns that require special attention. Because the ERAS principles for liver resection were adapted from colonic surgery, more evidence is needed to prove the benefits of ERAS in liver resection and to tailor the elements of ERAS to liver resection.

For example, most ERAS programmes in open liver surgery use thoracic epidural analgesia. However, patients who undergo liver surgery experience transient coagulopathy after the operation, which may increase the risk of spinal hematoma if epidural analgesia is used. One previous study indicated that epidural analgesia increases the risk of bleeding and prolongs prothrombin time after liver resection.26 Furthermore, the majority of patients with liver cancer in our locality also had co-existing liver cirrhosis. This group of patients is coagulopathic even before liver resection, and the risk of bleeding complications related to the epidural analgesic is a particular concern.27 In the present study, we used an infusion pump for continuous infiltration of the wound with local anaesthetic agents for pain control in patients who underwent open liver resection. The acute pain service provided regular ward rounds to review pain control. Other analgesics would be added if pain control was unsatisfactory. We have previously studied the analgesic efficacy of this infusion pump in open liver surgery and found that total morphine consumption was reduced in patients who received continuous wound instillation of local anaesthetic after open liver surgery. This technique also effectively reduced pain at rest and after spirometry.28 The small size of the device can facilitate early mobilisation during the postoperative period. Recent recommendations of ERAS guidelines for liver surgery suggest that routine thoracic epidural analgesia is not recommended and that a wound infusion catheter is a good alternative.29

Restrictive use of surgical site drains after operation is one of the key elements of most ERAS protocols to support early mobilisation and reduce postoperative pain and discomfort.30 Recent meta-analyses did not recommend routine abdominal drainage in elective uncomplicated hepatectomy.31 However, cirrhotic patients are at risk of developing ascites and bleeding after liver resection. Therefore, according to the ERAS society recommendations for perioperative care for liver surgery, the available evidence is inconclusive, and no recommendation can be given either for or against the use of prophylactic drainage after hepatectomy.29 Data from larger studies are needed to evaluate the role of intra-abdominal drains in this specific group of patients.

Nevertheless, ERAS protocols might still be beneficial to cirrhotic patients, particularly in regard to the omission of overnight fasting and carbohydrate loading. Cirrhotic patients have decreased hepatic glycogen storage and impaired gluconeogenesis: an overnight fast is equivalent to a fast of 2 to 3 days in a healthy person. Omission of overnight fasting and carbohydrate loading may lessen the nutritional stress for these patients.

Shorter hospital stays have been reported after minimally invasive liver resection.23 32 33 Whether a similar decrease in hospital stay can be achieved by open surgery with an optimised fast-track programme remains unclear. In the current series, length of hospital stay was reduced by 1 day in both the minimally invasive and open surgery groups. However, only the difference in the open surgery group reached statistical significance. The major limitation of our study is its small sample size. Therefore, it did not have enough power to demonstrate statistical significance in small differences. Early reports on ERAS in liver surgery have demonstrated a significant reduction in hospital stay by 2 to 6 days.34 35 36 Some might contend that it was careful selection of patients that resulted in the reduction of length of stay. However, diverse groups publishing on consecutive series with ERAS principles have shown consistent results.30 It is highly likely that the ERAS protocol can shorten hospital stays. However, whether it can lead to a reduction in healthcare costs will be the focus of future studies in this field. Another limitation of this study is the uncertainty of balance of characteristics between the two groups. Standardised mean differences showed imbalances of some demographics (eg, body mass index and extent of hepatectomy) between the treatment groups, but the P values did not reach statistical significance. Again this is caused by the small sample size, which yields a model that is not sensitive enough to detect small differences.

The ERAS programme for liver resection is safe and feasible. It resulted in a reduction in hospital stay without an increase in morbidity and mortality. Larger-scale studies are needed to optimise the programme’s elements and study its cost-effectiveness.

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.

Concept and design of study: CCN Chong, SKC Chan, PBS Lai.
Acquisition of data: WY Chung, YS Cheung, AKY Fung, AKW Fong, HT Lok.
Analysis or interpretation of data: CCN Chong.
Drafting of the article: CCN Chong.
Critical revision for important intellectual content: PBS Lai, SKC Chan, KF Lee, J Wong.

We would like to thank Mr Philip Ip for his statistical support in this project.

As an editor of the journal, PBS Lai was not involved in the peer review process. Other authors have no conflicts of interest to disclose.

The results of this project were presented in the 12th Biennial E-AHPBA Congress 2017 (23-26 May 2017, Mainz, Germany) and in the RCSEd/CSHK Conjoint Scientific Congress 2018 (15-16 September 2018, Hong Kong).

This project was supported by a Direct Grant from the Chinese University of Hong Kong (Ref No: MD14705).

The study was approved by the Joint Chinese University of Hong Kong–New Territories East Cluster Clinical Research Ethics Committee (CREC 2015.024).

1. Kehlet H, Wilmore DW. Evidence-based surgical care and the evolution of fast-track surgery. Ann Surg 2008;248:189-98. Crossref

2. Basse L, Raskov HH, Hjort Jakobsen D. Accelerated postoperative recovery programme after colonic resection improves physical performance, pulmonary function and body composition. Br J Surg 2002;89:446-53. Crossref

3. Kehlet H. Multimodal approach to postoperative recovery. Curr Opin Crit Care 2009;15:355-8. Crossref

4. Lemmens L, van Zelm R, Borel Rinkes I, van Hillegersberg R, Kerkkamp H. Clinical and organizational content of clinical pathways for digestive surgery: a systematic review. Dig Surg 2009;26:91-9. Crossref

5. Spanjersberg WR, Reurings J, Keus F, van Laarhoven CJ. Fast track surgery versus conventional recovery strategies for colorectal surgery. Cochrane Database Syst Rev 2011;(2):CD007635. Crossref

6. Khoo CK, Vickery CJ, Forsyth N, Vinall NS, Eyre-Brook IA. A prospective randomized controlled trial of multimodal perioperative management protocol in patients undergoing elective colorectal resection for cancer. Ann Surg 2007;245:867-72. Crossref

7. Gatt M, Anderson AD, Reddy BS, Hayward-Sampson P, Tring IC, MacFie J. Randomized clinical trial of multimodal optimization of surgical care in patients undergoing major colonic resection. Br J Surg 2005;92:1354-62. Crossref

8. Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global cancer statistics 2018: GLOBOCAN estimates of incidences and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2018;68:394-424. Crossref

9. Benzoni E, Molaro R, Cedolini C, et al. Liver resection for HCC: analysis of causes and risk factors linked to postoperative complications. Hepatogastroenterology 2007;54:186-9.

10. Karanjia ND, Lordan JT, Fawcett WJ, Quiney N, Worthington TR. Survival and recurrence after neo-adjuvant chemotherapy and liver resection for colorectal metastases: a ten year study. Eur J Surg Oncol 2009;35:838-43. Crossref

11. Jensen LS, Mortensen FV, Iversen MG, Jørgensen A, Kirkegaard P, Kehlet H. Liver surgery in Denmark 2002-2007 [in Danish]. Ugeskr Laeger 2009;171:1365-8.

12. Kehlet H. Fast-track surgery-an update on physiological care principles to enhance recovery. Langenbecks Arch Surg 2011;396:585-90. Crossref

13. Varadhan KK, Lobo DN, Ljungqvist O. Enhanced recovery after surgery: the future of improving surgical care. Crit Care Clin 2010;26:527-47,x. Crossref

14. Muller S, Zalunardo MP, Hubner M, Clavien PA, Demartines N; Zurich Fast Track Study Group. A fast-track program reduces complications and length of hospital stay after open colonic surgery. Gastroenterology 2009;136:842-7. Crossref

15. Wind J, Hofland J, Preckel B, et al. Perioperative strategy in colonic surgery; LAparoscopy and/or FAst track multimodal management versus standard care (LAFA trial). BMC Surg 2006;6:16. Crossref

16. Polle SW, Wind J, Fuhring JW, Hofland J, Gouma DJ, Bemelman WA. Implementation of a fast-track perioperative care program: what are the difficulties? Dig Surg 2007;24:441-9. Crossref

17. Kehlet H. Fast-track colorectal surgery. Lancet 2008;371:791-3. Crossref

18. Pruthi RS, Nielsen M, Smith A, Nix J, Schultz H, Wallen EM. Fast track program in patients undergoing radical cystectomy: results in 362 consecutive patients. J Am Coll Surg 2010;210:93-9. Crossref

19. Barletta JF, Miedema SL, Wiseman D, Heiser JC, McAllen KJ. Impact of dexmedetomidine on analgesic requirements in patients after cardiac surgery in a fast-track recovery room setting. Pharmacotherapy 2009;29:1427-32. Crossref

20. Hansen CT, Sørensen M, Møller C, Ottesen B, Kehlet H. Effect of laxatives on gastrointestinal functional recovery in fast-track hysterectomy: a double-blind, placebo-controlled randomized study. Am J Obstet Gynecol 2007;196:311.e1-7. Crossref

21. Andersen LØ, Gaarn-Larsen L, Kristensen BB, Husted H, Otte KS, Kehlet H. Subacute pain and function after fast-track hip and knee arthroplasty. Anaesthesia 2009;64:508-13. Crossref

22. Das-Neves-Pereira JC, Bagan P, Coimbra-Israel AP, et al. Fast-track rehabilitation for lung cancer lobectomy: a five-year experience. Eur J Cardiothoracic Surg 2009;36:383-91. Crossref

23. Lee KF, Chong CN, Wong J, Cheung YS, Wong J, Lai P. Long-term results of laparoscopic hepatectomy versus open hepatectomy for hepatocellular carcinoma: a case-matched analysis. World J Surg 2011;35:2268-74. Crossref

24. Clavien PA, Barkun J, de Oliveira ML, et al. The Clavien-Dindo classification of surgical complications: five-year experience. Ann Surg 2009;250:187-96. Crossref

25. Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004;240:205-13. Crossref

26. Sakowska M, Docherty E, Linscott D, Connor S. A change in practice from epidural to intrathecal morphine analgesia for hepato-pancreato-biliary surgery. World J Surg 2009;33:1802-8. Crossref

27. Ho AM, Lee A, Karmakar MK, et al. Hemostatic parameters after hepatectomy for cancer. Hepatogastroenterology 2007;54:1494-8.

28. Chan SK, Lai PB, Li PT, et al. The analgesic efficacy of continuous wound instillation with ropivacaine after open hepatic surgery. Anaesthesia 2010;65:1180-6. Crossref

29. Melloul E, Hübner M, Scott M, et al. Guidelines for perioperative care for liver surgery: enhanced recovery after surgery (ERAS) society recommendations. World J Surg 2016;40:2425-40. Crossref

30. Ljungqvist O, Scott M, Fearon KC. Enhanced recovery after surgery: a review. JAMA Surg 2017;152:292-8. Crossref

31. Gavriilidis P, Hidalgo E, de’Angelis N, Lodge P, Azoulay D. Re-appraisal of prophylactic drainage in uncomplicated liver resections: a systematic review and meta-analysis. HPB (Oxford) 2017;19:16-20. Crossref

32. Nguyen KT, Laurent A, Dagher I, et al. Minimally invasive liver resection for metastatic colorectal cancer: a multi-institutional, international report of safety, feasibility, and early outcomes. Ann Surg 2009;250:842-8. Crossref

33. Kazaryan AM, Pavlik Marangos I, Rosseland AR, et al. Laparoscopic liver resection for malignant and benign lesions: ten-year Norwegian single-center experience. Arch Surg 2010;145:34-40. Crossref

34. Lee A, Chiu CH, Cho MW, et al. Factors associated with failure of enhanced recovery protocol in patients undergoing major hepatobiliary and pancreatic surgery: a retrospective cohort study. BMJ Open 2014;4:e005330. Crossref

35. van Dam RM, Hendry PO, Coolsen MM, et al. Initial experience with a multimodal enhanced recovery programme in patients undergoing liver resection. Br J Surg 2008;95:969-75. Crossref

36. Stoot JH, van Dam RM, Busch OR, et al. The effect of a multimodal fast-track programme on outcomes in laparoscopic liver surgery: a multicentre pilot study. HPB (Oxford) 2009;11:140-4. Crossref

Many plants are poisonous to humans. Surveys of various toxicology centres in Australia,1 Germany,2 3 Morocco,4 New Zealand,5 Sweden,6 Thailand,7 the United Kingdom,8 and the United States9 10 showed that plant exposure was responsible for 1.8% to 8% of all inquiries. Most plant exposures did not result in significant toxicity, but severe and life-threatening poisonings have been reported. As reported by the Moroccan Poison Control Centre from 1980 to 2011, plants were the cause of approximately 5% of all fatal cases of intoxication encountered.4 Consumption of wild plants as “medicinal herbs” or food is not an uncommon practice in Hong Kong, and severe plant poisoning cases have been reported.11 12 13 14 15 16 17 18 19 However, local epidemiology data on plant poisoning are sparse and limited.

Owing to the variable clinical features and rare occurrence of plant poisonings, diagnosing and treating these patients remain a challenge to our frontline clinicians. Although history of plant exposure is important, it is often insufficient to pinpoint the incriminating toxic plant. Misidentification of plant is a common cause of poisoning in the first place. Morphological identification of a poisonous plant and biochemical confirmation are generally not available to guide immediate clinical management. Therefore, clinical features are critical for initiating supportive treatment, which is a common strategy for treating poisonings of an uncertain nature.

Classifications of clinically significant plant toxins have been proposed in order to aid rapid recognition and management of these patients.20 21 Plant toxins can be broadly classified into four groups: (1) cardiotoxic toxins, such as cardiac glycosides; (2) neurotoxic toxins, such as gelsemium alkaloids, grayanotoxins, solanaceous tropane (anticholinergic) alkaloids, strychnine, and brucine; (3) cytotoxic toxins, such as colchicine, mimosine, plumbagin, toxalbumins; vinblastine, and vincristine; and (4) gastrointestinal-hepatotoxic toxins, such as amaryllidaceous alkaloids, calcium oxalate raphide, cycasin, pentacyclic triterpenoids, phorbol esters, phytolaccatoxins, plumericin, pyrrolizidine alkaloids, saponins, steroidal alkaloids, tetrahydropalmatine, and teucvin.

The Hospital Authority Toxicology Reference Laboratory (HATRL), the only tertiary clinical toxicology laboratory in Hong Kong, provides support to all local hospitals in managing patients with complex poisoning problem, including cases of plant poisonings. Besides plant identification, HATRL provides specialised toxicology testing for certain plant toxins, especially for those with significant clinical toxicity and management impact.16 22 The aim of the present study was to identify the plant species most commonly involved in cases of plant poisoning in Hong Kong, in order to promote awareness among local clinicians and to provide a reference for diagnosing and managing plant poisonings.

All cases of suspected plant-related poisoning referred to the HATRL from January 2003 to December 2017 were retrospectively reviewed. Clinical data were collected by reviewing the laboratory database and patient medical records. Demographic characteristics, clinical presentation, drug history, laboratory and toxicological findings, progress, and outcomes of these patients were reviewed. The causal relationships between the clinical presentation and plant use were evaluated based on known adverse effects of the specified plant or toxins, the temporal sequence, exclusion of other underlying or concurrent diseases which could otherwise account for the clinical presentation, and progress after discontinuation of plant use. Severity of the poisoning cases was graded according to an established poisoning severity score as follows23: mild, transient, and spontaneously resolving symptoms (mild); pronounced or prolonged symptoms (moderate); severe or life-threatening symptoms (severe); and fatal poisoning (fatal). Descriptive statistics were used to present the results.

In collaboration with the Hong Kong Herbarium, Agriculture, Fisheries and Conservation Department, available plant specimens were sent for morphological identification. Mass spectrometry–based and microscopy-based tests for specific plant toxin(s) were developed, and performed on selected cases, guided by clinical presentation of the patients and types of the toxic plant(s) exposed. Mass spectrometry–based analytical platforms can identify the majority of the clinically significant plant toxins affecting the cardiovascular, neurological, gastrointestinal, hepatological, and renal systems in plant and biological specimens. The poisonous plants commonly encountered in Hong Kong and the corresponding plant toxins are summarised in Table 1.24

The ethics committee exempted the study group from obtaining patient consent because the presented data were anonymised, and the risk of identification was low. The STROBE guidelines were used to ensure the reporting of this study.25

A total of 62 cases of confirmed plant poisoning, involving 26 plant species, were identified within the study period. The patients were referred from 14 local hospitals administered by the Hospital Authority. Almost all patients (n=60, 97%) were Chinese, 29 (47%) were male, and 33 (53%) were female. The median age was 50 years (range, 1 month to 83 years), including three children (range, 1-23 months) and four adolescents (range, 15-18 years). The route of exposure was oral in 60 (97%) patients and topical in two (3%) patients. The majority of the patients (n=55, 89%) developed acute toxic symptoms after a single use of the plants, while the remaining patients (n=7, 11%) reported a history of prolonged exposure. Details of the plant poisoning cases, including the involved plant species and toxins, clinical presentation and outcome of these patients are summarised in Table 2.

Gastrointestinal toxicity was the most common clinical presentation (n=30, 48%). Of these 30 patients, 17 developed oromucosal irritation after ingestion of calcium oxalate raphide–containing plants. The other 13 patients presented with gastroenteritis-like symptoms, such as nausea, vomiting, abdominal pain and diarrhoea, after ingestion of plant containing gastrointestinal irritants (n=11) or cytotoxic toxins (n=2). Neurological toxicity was the second most common presenting symptoms (n=22, 35%), mainly caused by gelsemium alkaloids (n=12), Solanaceous tropane alkaloids (n=4), and grayanotoxins (n=3). Hepatotoxicity was encountered in six (10%) patients. Five patients had abnormal liver function test results owing to the use of teucvin (n=1), pyrrolizidine alkaloids (n=1), and other unknown hepatotoxin(s) (n=3); one patient developed cholestasis after exposure to pentacyclic triterpenoids. Cases of nephrotoxicity (n=2, 3%), cardiotoxicity (n=1, 2%), and dermatological toxicity (n=1, 2%) were also recorded.

Most of the patients experienced mild (n=49, 79%) or moderate toxicity (n=8, 13%). Of them, 55 patients recovered within days with supportive treatment. The other two patients with nephrotoxicity had residual renal impairment after discontinuation of plant use; in one of them a renal biopsy study showed tubulonephritic changes. The remaining five patients (8%) experienced severe toxicity after the use of Abrus precatorius (n=1), Gelsemium elegans (n=2), Rhododendron species (n=1), and Emilia sonchifolia (n=1), and all five required intensive care support.

The morphological identification and biochemical analysis process followed to identify the plants involved in the 62 cases of intoxication is shown in Figure 2. Most patients provided fresh plant specimens (n=53, 85%) or photograph of the plant (n=3, 5%) they had consumed. Among these specimens received by our laboratory, 43 could be identified morphologically with the aid of the Hong Kong Herbarium, Agriculture, Fisheries and Conservation Department. Biochemical analysis for specific plant toxin(s) were attempted in the plant or the biological specimens in 45 (73%) cases, with 41 yielding diagnostic information, including six cases with no plant specimens for identification and 13 cases with unsuccessful identification. There were 22 (35%) cases with both informative morphological identification and biochemical results. Among these 22 cases, the morphological identification and biochemical results were coherent in 20. For the remaining two cases with clinical gelsemium poisoning, gelsemium alkaloids were detected in both the urine and plant specimens. However, the plant specimen was morphologically identified as Cassytha filiformis. Cassytha filiformis is a non-toxic plant that has been shown to parasitise Gelsemium elegans and absorb gelsemium alkaloids, leading to poisoning.18 Patient accuracy in identification of plants was low. Although 56 (90%) patients had given common names of the plants they consumed, only 19 of them were correct.

Patients consumed the plants as medicinal herbs (n=31, 50%), as food (n=29, 47%), in an attempted suicide (n=1, 2%), or accidentally, in one paediatric patient (n=1, 2%). Causes of poisoning in patients using the plants as medicinal herbs or food included plant misidentification (n=34), unawareness/underestimation of the potential toxicity (n=20), or contamination of non-toxic plants with poisonous ones (n=6). The sources of the poisonous plants used included self-collecting from parks or the countryside (n=37, 60%), obtaining from friends or relatives (n=20, 32%), growing at home (n=3, 5%), and buying from wet markets (n=2, 3%). The plants were obtained in Hong Kong (n=49, 79%) or mainland China (n=12, 19%), with one (2%) case collected from the Philippines.

The 62 cases reported herein represent the largest series of plant poisoning cases in Hong Kong confirmed by either morphological identification or biochemical confirmation. Comparing our results with those of studies from other regions, the pattern of plant poisoning is considerably different. Plant poisonings reported in Western countries are predominantly accidental exposure in children,3 7 10 whereas in the present study, most patients were adults poisoned after intentional use of wild plants. The highly urbanised and industrialised nature of Hong Kong explains the low incidence of paediatric accidental poisoning. The relatively high rate of adult poisoning may be related to the long-standing tradition and Chinese culture of using wild plant as “medicinal herbs” and “vegetables”. Recreational abuse of toxic plants, such as Datura species,26 27 which are responsible for a significant number of poisoning cases reported in other regions, was not observed locally.

The most common type of poisoning in the current study was the use of raphide-containing plants, such as Alocasia macrorrhizos, accounting for more than 25% of cases. The roots of these plants are frequently mistaken as edible taro.28 Although all cases in this study were mild, severe outcomes including oropharyngeal oedema, upper airway obstruction, or systemic toxicity are possible.29 30

Similar to previous reports of plant poisoning,3 5 7 most patients in the present study developed mild transient symptoms and recovered with supportive management and discontinuation of exposure. However, there were five cases of severe poisoning requiring intensive care support, involving the use of Gelsemium elegans, Rhododendron species, Abrus precatorius, and Emilia sonchifolia (Fig 1).

Gelsemium elegans is one of the most notorious poisonous plants in Hong Kong and South-East Asian countries and has been used for homicidal and suicidal purposes.31 32 Severe gelsemium toxicity can result in respiratory depression and even death. In our case series, two clusters of “hidden” gelsemium poisoning were identified where the patients consumed non-toxic parasitic Cassytha filiformis which absorbed gelsemium alkaloids from Gelsemium elegans on which it grew.18 Similar cases of hidden gelsemium poisoning from contaminated dried herbs Ficus hirta have been reported in Hong Kong.16

Rhododendron species and other plants in the Ericaceae family contain grayanotoxins.33 Severe grayanotoxin poisoning may result in respiratory depression and arrhythmias.34 “Mad honey” containing nectar of Ericaceae plants is known to be a source of exposure in Hong Kong,35 but poisonings due to direct consumption of Rhododendron flowers are not uncommon.12

The seed of Abrus precatorius contains the protein abrin, an extremely poisonous toxin similar to ricin that can result in multi-organ failure.36 37 These seeds are sometimes used in beaded jewellery, in addition to being collected, providing another potential source of exposure.37 Cases of abrin poisoning due to suicidal attempt or accidental ingestion have been reported worldwide,38 39 40 but such poisoning is rare in Hong Kong.

The toxic constituents in Emilia sonchifolia are pyrrolizidine alkaloids.41 Massive acute ingestion can lead to hepatotoxicity and coagulopathy, while chronic low-dose exposure can result in liver cirrhosis and hepatic veno-occlusive disease.42 43 44 Certain types of traditional Chinese medicine, such as Flos farfarae and Herba senecionis scandentis, are potential sources of pyrrolizidine alkaloids exposure in Hong Kong,45 46 but toxicity due to wild plant consumption is relatively rare.14

Timely diagnosis of plant poisonings is very difficult. Local epidemiology data are scarce, and reports published in other parts of the world are of limited use because plant species are geographically specific. Patients might volunteer a history of plant exposure, but the information they provide may be uninformative, misleading, or incorrect. In our case series, among those who provided a common name of the plant, only 34% were correct.

Our results demonstrate the importance of adopting a complementary approach, incorporating clinical toxidrome, morphological identification of plant specimens, and biochemical analysis of plant toxins, in order to achieve maximal diagnostic efficacy. Clinical history may not be particularly informative. The presenting toxidrome is more objective and useful for identifying the plant responsible for the poisoning, and this diagnostic approach is well supported in the literature.20 21

Table 1 summarises and categorises commonly encountered local poisonous plants by toxidromes. This may serve as quick reference for clinicians to allow rapid identification of the plant and prompt management of the poisoning. However, clinicians should also be aware that the toxicity of some plants may not be well characterised, and that some plants may give rise to non-specific toxidrome, rendering this approach less useful in certain cases.

Specialised toxicological investigation may provide additional evidence to confirm or refute the provisional diagnosis. In most situations, morphological identification of the plant is sufficient to achieve a diagnosis, if a well-preserved plant specimen is provided. However, a plant specimen is not always available; even if available, it may not be representative of the plants consumed by the patient, or may have been deformed due to prior processing. An equally important tool is target testing for specific plant toxin(s) in the plant or biological specimens based on clinical suspicion. This is a powerful tool for confirming the diagnosis, especially when plant specimen is not available; and for identifying cases of contamination with an unclear or hidden source.18 However, these investigations cannot aid immediate management in the emergency room.

Diagnosis of plant poisoning requires a proactive rather than a reactive approach. A knowledge database comprising the clinical toxicity, morphology, and biochemical analysis of local toxic plants is an important tool to prepare clinicians and healthcare professionals, such as that provided by the Hospital Authority.24 This public resource can also be used to educate the general public on the dangers of wild plants.

As a retrospective study, this study has several limitations. This study is subjected to selection bias as the data collection is a passive procedure depending on test or consultation requests by clinicians, and there might be incomplete or missing data. Moreover, for those with history of exposure or mild clinical symptoms, toxicology testing or consultation may not be requested, our data might therefore include only cases with more severe outcome in the overall continuum of plant poisoning. Inevitably, some plant exposures were not reported to our laboratory, and thus our findings may underestimate the actual number of cases.

Diagnosing plant poisoning is challenging and the epidemiology of plant poisonings is geographically specific. Clinicians should consider a complementary approach with consideration of the clinical toxidromes, local epidemiological data, botanical and toxicological findings to help recognition of the plants involved in cases of intoxication. Plant specimens and biological specimens should be saved whenever possible for toxicological analysis. Clinicians should be aware of local poisonous plants and their toxicities. Although most plant exposure resulted in a self-limiting disease, severe poisonings were encountered. The public should be educated about the potential hazards of consuming plants obtained from the wild and discouraged from engaging in this practice.

All authors had full access to the data, contributed to the study (including concept or design, acquisition of data, analysis or interpretation of data, drafting of the manuscript, and critical revision for important intellectual content), approved the final version for publication, and take responsibility for its accuracy and integrity.

All authors have disclosed no conflicts of interest.

Some of the cases have been published by the authors in electronic form24 and in book form (Mak WL, Lam YH, Ching CK, Chan SS, Chong YK, Ng WY, editors. Atlas of Poisonous Plants in Hong Kong—A Clinical Toxicology Perspective. Hong Kong: Hospital Authority Toxicology Reference Laboratory; 2016), and presented at the Hospital Authority Toxicology Services Scientific Conference 2016 (“Poisonous plants in Hong Kong—a clinical perspective”). Some of the cases have been previously published as case reports by the authors and other units.12 13 14 15 16 17 18 19

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

The study was approved by the Hong Kong Hospital Authority Kowloon West Cluster Research Ethics Committee (Ref. KW/EX-16-036[96-18(TCM)]).

1. Children’s Health Queensland Hospital and Health Service, Queensland Government. Queensland Poisons Information Centre Annual report 2014. Available from: https://www.childrens.health.qld.gov.au/wp-content/uploads/PDF/poisons/annual-reports/2014.pdf. Accessed 30 May 2018.

2. Müller D, Desel H. Common causes of poisoning: etiology, diagnosis and treatment. Dtsch Arztebl Int 2013;110:690-9.

3. Plenert B, Prasa D, Hentschel H, Deters M. Plant exposures reported to the Poisons Information Centre Erfurt from 2001-2010. Planta Med 2012;78:401-8. Crossref

4. Rhalem N, Chaoui H, Lahcen O, Soulaymani A, Bencheikh RS. Toxic deaths: data from the Poison Control Centre of Morocco (CAPM). XXXV International Congress of the European Association of Poisons Centres and Clinical Toxicologists (EAPCCT) 26-29 May 2015, St Julian’s, Malta. Clin Toxicol 2015;53:265.

5. Slaughter RJ, Beasley DM, Lambie BS, Wilkins GT, Schep LJ. Poisonous plants in New Zealand: a review of those that are most commonly enquired about to the National Poisons Centre. N Z Med J 2012;125:87-118.

6. Swedish Poisons Information Centre. Swedish Poisons Information Centre. Annual Report 2015. Available from: https://giftinformation.se/globalassets/publikationer/annual-report-2015.pdf. Accessed 18 May 2018.

7. Sriapha C, Tongpoo A, Wongvisavakorn S, et al. Plant poisoning in Thailand: a 10-year analysis from Ramathibodi Poison Center. Southeast Asian J Trop Med Public Health 2015;46:1063-76.

8. Public Health England. National Poisons Information Service Report 2016/17. Available from: http://www.edinburghclinicaltoxicology.org/s/NPIS-2016-17-report.pdf. Accessed 30 May 2018.

9. Gummin DD, Mowry JB, Spyker DA, Brooks DE, Fraser MO, Banner W. 2016 annual report of the American Association of Poison Control Centers’ National Poison Data System (NPDS): 34th annual report. Clinical Toxicol (Phila) 2017;55:1072-252. Crossref

10. Enfield B, Brooks DE, Welch S, et al. Human plant exposures reported to a regional (Southwestern) Poison Control Center over 8 years. J Med Toxicol 2018;14:74-8. Crossref

11. Chan TY, Chan LY, Tam LS, Critchley JA. Neurotoxicity following the ingestion of a Chinese medicinal plant, Alocasia macrorrhiza. Hum Exp Toxicol 1995;14:727-8. Crossref

12. Poon WT, Ho CH, Yip KL, et al. Grayanotoxin poisoning from Rhododendron simsii in an infant. Hong Kong Med J 2008;14:405-7.

13. Poon WT, Chau TL, Lai CK, et al. Hepatitis induced by Teucrium viscidum. Clin Toxicol (Phila) 2008;46:819-22. Crossref

14. Wu JS, Poon WT, Ma CK, et al. Budd-Chiari syndrome secondary to toxic pyrrolizidine alkaloid exposure. Hong Kong Med J 2013;19:553-5. Crossref

15. Pang CT, Ng HW, Lau FL. Oral mucosal irritating plant ingestion in Hong Kong: epidemiology and its clinical presentation. Hong Kong J Emerg Med 2010;17:477-81. Crossref

16. Lai CK, Chan YW. Confirmation of gelsemium poisoning by targeted analysis of toxic gelsemium alkaloids in urine. J Anal Toxicol 2009;33:56-61. Crossref

17. Fung HT, Lam KK, Lam SK, Wong OF, Kam CW. Two cases of Gelsemium elegans Benth. poisoning. Hong Kong J Emerg Med 2007;14:221-4. Crossref

18. Cheung WL, Law CY, Lee HC, et al. Gelsemium poisoning mediated by the non-toxic plant Cassytha filiformis parasitizing Gelsemium elegans. Toxicon 2018;154:42-9. Crossref

19. Cheng KL, Chan YC, Mak TW, Tse ML, Lau FL. Chinese herbal medicine-induced anticholinergic poisoning in Hong Kong. Hong Kong Med J 2013;19:38-41.

20. Diaz JH. Poisoning by herbs or plants: rapid toxidromic classification and diagnosis. Wilderness Environ Med 2016;27:136-52. Crossref

21. Lin TJ, Nelson LS, Tsai JL, et al. Common toxidromes of plant poisonings in Taiwan. Clin Toxicol (Phila) 2009;47:161-8. Crossref

22. Ng SW, Ching CK, Chan AY, Mak TW. Simultaneous detection of 22 toxic plant alkaloids (aconitum alkaloids, solanaceous tropane alkaloids, sophora alkaloids, strychnos alkaloids and colchicine) in human urine and herbal samples using liquid chromatography-tandem mass spectrometry. J Chromatogr B Analyt Technol Biomed Life Sci 2013;942-943:63-9. Crossref

23. Persson HE, Sjöberg GK, Haines JA, Pronczuk de Garbino J. Poisoning severity score. Grading of acute poisoning. J Toxicol Clin Toxicol 1998;36:205-13. Crossref

24. Hospital Authority Toxicology Reference Laboratory. Atlas of poisonous plants in Hong Kong—a clinical toxicology perspective. 2016. Available from: http://www3.ha.org.hk/toxicplant. Accessed 30 May 2018.

25. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP; STROBE Initiative. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol 2008;61:344-9. Crossref

26. Centers for Disease Control and Prevention (CDC). Suspected moonflower intoxication-Ohio, 2002. MMWR Morb Mortal Wkly Rep 2003;52:788-91.

27. Boumba VA, Mitselou A, Vougiouklakis T. Fatal poisoning from ingestion of Datura stramonium seeds. Vet Hum Toxicol 2004;46:81-2.

28. Lin TJ, Hung DZ, Hu WH, Yang DY, Wu TC, Deng JF. Calcium oxalate is the main toxic component in clinical presentations of Alocasis macrorrhiza (L) Schott and Endl poisonings. Vet Hum Toxicol 1998;40:93-5.

29. Farré M, Xirgu J, Salgado A, Peracaula R, Reig R, Sanz P. Fatal oxalic acid poisoning from sorrel soup. Lancet 1989;2:1524. Crossref

30. Kalliala H, Kauste O. Ingestion of rhubarb leaves as cause of oxalic acid poisoning. Ann Paediatr Fenn 1964;10:228-31.

31. Xiang H, Zhou YJ, Huang PL, et al. Lethal poisoning with Gelsemium elegans in Guizhou, China. Public Health 2016;136:185-7. Crossref

32. Zhou Z, Wu L, Zhong Y, et al. Gelsemium elegans poisoning: a case with 8 months of follow-up and review of the literature. Front Neurol 2017;8:204. Crossref

33. Lampe KF. Rhododendrons, mountain laurel, and mad honey. JAMA 1988;259:2009. Crossref

34. Lennerz C, Jilek C, Semmler V, Deisenhofer I, Kolb C. Sinus arrest from mad honey disease. Ann Intern Med 2012;157:755-6. Crossref

35. Chen SP, Lam YH, Ng VC, et al. Mad honey poisoning mimicking acute myocardial infarction. Hong Kong Med J 2013;19:354-6. Crossref

36. Dickers KJ, Bradberry SM, Rice P, Griffiths GD, Vale JA. Abrin poisoning. Toxicol Rev 2003;22:137-42. Crossref

37. Fernando C. Poisoning due to Abrus precatorius (jequirty bean). Anaesthesia 2001;56:1178-80. Crossref

38. Wooten JV, Pittman CT, Blake TA, et al. A case of abrin toxin poisoning, confirmed via quantitation of L-abrine (N-methyl-L-tryptophan) biomarker. J Med Toxicol 2014;10:392-4. Crossref

39. Jang DH, Hoffman RS, Nelson LS. Attempted suicide, by mail order: Abrus precatorius. J Med Toxicol 2010;6:427-30. Crossref

40. Subrahmanyan D, Mathew J, Raj M. An unusual manifestation of Abrus precatorius poisoning: a report of two cases. Clin Toxicol (Phila) 2008;46:173-5. Crossref

41. Cheng D, Röder E. Pyrrolizidine alkaloids from Emilia sonchifolia [in German]. Planta Med 1986;6:484-6. Crossref

42. Wiedenfeld H. Plants containing pyrrolizidine alkaloids: toxicity and problems. Food Addit Contam Part A Chem Anal Control Expo Risk Assess 2011;28:282-92. Crossref

43. Chen Z, Huo JR. Hepatic veno-occlusive disease associated with toxicity of pyrrolizidine alkaloids in herbal preparations. Neth J Med 2010;68:252-60.

44. Kumana CR, Ng M, Lin HJ, Ko W, Wu PC, Todd D. Herbal tea induced hepatic veno-occlusive disease: quantification of toxic alkaloid exposure in adults. Gut 1985;26:101-4. Crossref

45. Centre for Food Safety, Food and Environmental Hygiene Department, Hong Kong SAR Government. Risk assessment studies report no. 56. Chemical hazard evaluation. Pyrrolizidine alkaloids in food. 2017. Available from: https://www.cfs.gov.hk/english/programme/programme_rafs/files/Pyrrolizidine_Alkaloids_in_Food_e.pdf. Accessed 10 Nov 2018.

46. Gao H, Ruan JQ, Chen J, et al. Blood pyrrole-protein adducts as a diagnostic and prognostic index in pyrrolizidine alkaloid-hepatic sinusoidal obstruction syndrome. Drug Des Devel Ther 2015;9:4861-8. Crossref

Sudden death is defined as death occurring within 1 hour of the onset of symptoms or within 24 hours of the victim being seen alive.1 Sudden death due to an underlying heart disease is known as sudden cardiac death (SCD). The worldwide annual incidence of SCD is about 4 to 5 million cases per year.2 A tragic and devastating complication of a number of heart diseases, SCD is most often unexpected and has major implications for the surviving family and the community. The majority of SCD cases in middle-aged and older individuals are caused by coronary artery disease; however, SCD is rare in young people and the causes are more diversified.3 4 Autopsy studies have shown no structural heart disease was found in up to 30% of young SCD victims.5 6 7 8

Molecular autopsy was first described by Ackerman et al9 in 1999, to determine the cause of death in uncertain cases after conventional autopsy by genetic analysis. Post-mortem genetic studies have shown that SCD in these victims can be caused by fatal arrhythmias secondary to a group of inheritable cardiac electrical disorders collectively known as sudden arrhythmia death syndrome (SADS).10 11 12 13 14 These include Brugada syndrome (BrS), long QT syndrome (LQTS), short QT syndrome, catecholaminergic polymorphic ventricular tachycardia (CPVT), arrhythmogenic right ventricular cardiomyopathy (ARVC), hypertrophic cardiomyopathy (HCM), and other cardiomyopathies.

Because SADS-related conditions are genetic diseases, there are two different approaches to identify SADS among young sudden unexplained death (SUD) victims. The first approach involves detailed clinical and targeted genetic examination of the surviving relatives of SCD victims. Studies using this approach suggest that SADS may account for approximately 40% of autopsy-negative sudden death in young people.10 15 However, this approach may not be able to identify subjects with concealed form of SADS due to incomplete penetrance and variable expressivity of the pathological mutations. The second approach is to perform molecular autopsy on SCD victims, which involves post-mortem genetic testing for SADS. A landmark study on molecular autopsy by the Mayo Clinic showed over one third of SCD cases hosted a presumably pathogenic mutations of cardiac ion channel diseases.8 16 Thus a combined approach using both cardiac evaluation of surviving relatives and molecular autopsy on SUD victims should give a higher yield on elucidating the underlying causes of SUD.

In Hong Kong, there are scarce data on the prevalence and types of SADS underlying SCD or SUD in young people. The present study aimed to investigate the prevalence and types of SADS in an East Asian population, and to perform a pilot study on incorporating cardiac evaluation of surviving relatives and molecular autopsy by next-generation sequencing (NGS) into conventional forensic investigations.

In the present study, first, we carried out a 5-year retrospective review of the records of all autopsies for young sudden death victims performed in public mortuaries in Hong Kong. Second, we conducted a 2-year study to determine the prevalence and types of SADS as the underlying causes of SCD among local young victims through conventional and molecular autopsy, and evaluated their first-degree relatives.

The Forensic Pathology Service, Department of Health, provides all public autopsy services for over 7 million people in Hong Kong. We performed this retrospective review of the records of all autopsies in public mortuaries for young sudden death victims (aged 5-40 years) between 1 January 2008 and 31 December 2012. Sudden death was defined as death occurring within 1 hour of the onset of symptoms or within 24 hours of the victim being seen alive.1 Sudden death victims were recruited into the study for a detailed review of their autopsy records. Data including age, height, weight, sex, circumstances of death, clinical history of cardiac disease, and pathologic findings at autopsy were collected and analysed. Sudden death victims whose deaths were caused by trauma, accidents, drowning, and drug toxicity, and those whose autopsy records were either incomplete or not accessible for retrospective review were excluded.

In this prospective study, young SCD victims (aged 5-40 years) were identified and recruited into the study by forensic pathologists after a finding of either an inheritable arrhythmogenic cardiomyopathy or no anatomical cause of death (including other structural heart disease) on autopsy, and a negative toxicology screening. Clinical history, including personal history of arrhythmic events and history surrounding the sudden death event of the SCD victim was collected during identification interviews with next-of-kin as far as possible by forensic pathologists. DNA-friendly blood samples were collected for molecular autopsy. Written informed consent for a molecular autopsy was obtained from the next-of-kin of each victim. The first-degree relatives of the victims were referred by forensic pathologists to the study centre for genetic counselling and recruitment into the study. Clinical history, including personal history of arrhythmic events and history surrounding the sudden death event of the SCD victim was collected from family members. All recruited first-degree relatives underwent clinical evaluation including 12-lead electrocardiogram (ECG), signal-averaged ECG, echocardiogram, 24-hour Holter analysis and treadmill exercise testing. Additional investigations were used only as required, such as flecainide provocation testing (if BrS is suspected in subjects >18 years) and targeted genetic screening (if positive molecular autopsy findings in the index SUD victim). All first-degree relatives provided written informed consent for these procedures and for publication of the results.

A panel of 35 genes implicated in SADS for BrS, LQTS, short QT syndrome, CPVT, ARVC, and HCM was tested using NGS (Table 1). The NGS was performed using targeted gene capture technique on a MiSeq Sequencing System (Illumina, Inc, San Diego [CA], United States). Target regions of interest were restricted to the coding regions and the 10-bp flanking regions. Target rates indicated percentage of bases with a minimum depth of coverage of 20×. Alignments to the February 2009 (GRCh37/hg19) human genome assembly and variant calls were generated using dual pipelines, SoftGenetics NextGENe (v2.4.1) and an in-house one where sequencing reads were aligned by Burrows-Wheeler Aligner (v0.7.5a-r405) to hg19 genome and processed with picard-tools (v1.114). Local realignment for indels and base quality recalibration were performed with Genome Analysis Toolkit (GATK, v3.2). Variant calls were made with UnifiedGenotyper (GATK v3.2). Only those genes included in the requested gene panel were processed for variant calling. Variants identified were annotated and analysed with VariantStudio (v2.2.1; Illumina, Inc); in general, variants with an allele frequency of <0.1% for dominant disorders or <1.0% for recessive disorders were reported. Pathogenic and likely pathogenic variants were confirmed by Sanger sequencing; benign and likely benign variants were not reported.

The genetic variant pathogenicity was established according to the Practice Guidelines for the Interpretation and Reporting of Unclassified Variants in Clinical Molecular Genetics by the Clinical Molecular Genetics Society (http:// www.acgs.uk.com/media/774853/evaluation_and_reporting_of_sequence_variants_bpgs_june_2013_-_finalpdf.pdf). Major criteria include degree of conservation, population allele frequencies, co-segregation pattern, literature data, functional studies and in silico prediction. The pathogenicity of novel missense variants was analysed by PolyPhen-2, SIFT, MutationTaster and Assessing Pathogenicity Probability in Arrhythmia by Integrating Statistical Evidence (https://cardiodb.org/APPRAISE/) and that of novel splicing variants was by Splice Site Finder-like, MaxEntScan, GeneSplicer and Human Splicing Finder, wherever appropriate. Splicing variants were considered to be damaging if the score was more than 10% lower than the wild-type prediction. Allele frequencies among populations were as reported in the Genome Aggregation Database (gnomAD, http://gnomad.broadinstitute.org/).

A diagnosis of SADS was established when molecular autopsy in SCD victims identified a positive genetic variant implicated in SADS or generally accepted clinical criteria for a particular disease were fulfilled in the SCD victims or their first-degree relatives. The descriptive statistics were analysed using Excel 2016 (Microsoft Corp, Redmond [WA], United States).

There were 17 187 autopsies performed during the study period and 2748 (16%) deaths were aged 5 to 40 years. There were 420 sudden death victims. Among them, 289 (69%) were SCD (male:female ratio=9:2). The median age was 32 years (range, 7-40 years). Coronary artery diseases accounted for 35% of the causes of death; other causes of death were aortic dissection (6%), myocarditis (6%), left ventricular hypertrophy (4%), dilated cardiomyopathy (9%), other structural heart diseases (9%), HCM (4%), ARVC (2%), and unexplained (25%).

There were 32 SCD victims aged 5 to 40 years between 1 July 2014 and 30 June 2016 with a finding of either an inheritable arrhythmogenic cardiomyopathy or no anatomical cause of death (including other structural heart disease) on autopsy and negative toxicology results. Eleven individuals were excluded: two who had no Hong Kong identity card and nine whose families refused consent. Finally, 21 SCD victims (18 male, 3 female) were recruited into the study. Table 2 shows the clinical and forensic data of the 21 SCD victims. The median age was 31 years (range, 14-39 years). From the conventional autopsy, 18 (86%) were unrevealing; two (10%) SCD victims had ARVC and one (5%) had HCM. The majority died during resting (48%) or sleeping (24%). Only three SCD victims (14%) died during exercise. Three (14%) SCD victims had family history of SCD and seven (33%) had a history of syncope. Many (71%) of the SCD victims had unremarkable past health. Genetic analysis showed 29% with positive heterozygous genetic variants (Table 3). Seven variants were identified in seven genes, six variants of which were novel. The genetic background was heterogeneous without any common mutations found. Combining conventional and molecular autopsy findings, the cause of death in 12 (57%) still remains unknown; cause of death was confirmed in four (19%) as ARVC, in two (10%) as BrS, and one (5%) each in HCM, LQTS, and CPVT.

Overall, more than half of first-degree relatives (6 of 11 individuals) who underwent genetic testing carried the positive genetic variants. Among them, SADS-related clinical features were detected in three first-degree relatives from two families (cases 14 and 16). The true clinical phenotype can sometimes be detected in surviving relatives upon appropriate clinical evaluation, which in turn significantly aids the interpretation of genetic findings. Cases 14 and 16 illustrate the importance of this (Fig 1). The first-degree relatives of these two cases showed SADS-related clinical features after cardiologist’s assessment.

Case 14 was a 17-year-old male victim who died of sudden nocturnal death. He had an episode of syncope few months prior to his death but he did not seek medical attention. Family screening by clinical evaluation found no structural heart disease but ECG revealed prolonged QTc interval in his asymptomatic father and sister (530 ms and 480 ms, respectively) suggesting a diagnosis of LQTS (Fig 1). Molecular autopsy found heterozygous NM_199460.2(CACNA1C):c.2276C>G NP_955630.2:p.(Ala759Gly). This novel variant is predicted to be damaging and absent in the gnomAD. CACNA1C is the gene encoding for the alpha-1c subunit of the type 1 voltage-dependent calcium channel involved in the cardiac myocyte membrane polarisation. Its mutations have been reported in BrS and LQTS.

Case 16 was a 19-year-old male victim who also died of sudden nocturnal death. There was no known syncope or preceding symptoms ahead of the collapse. He had history of abnormal heart beat, although ECG was not done. His family history was unremarkable. Autopsy revealed pulmonary hypertension. Molecular autopsy detected heterozygous NM_198056.2(SCN5A):c.2893C>T (p.Arg965Cys) which has been reported as a disease causing mutation of BrS.17 18 19 The same variant has also been reported in patients with LQTS (with digenic mutation in KCNH2 in one case).20 21 The allele frequency of this variant (rs199473180) is reported to be 0.058% in East Asians (gnomAD). In silico analyses by SIFT, MutationTaster, and PolyPhen-2 have revealed this variant to be damaging, disease causing, and probably damaging, respectively. Functional study showed that the p.(Arg965Cys) mutant led to slower recovery from inactivation as a result of channels with a more negative potential in steady state inactivation.18 Family screening found his father carrying the same SCN5A mutation and a type 2 Brugada ECG pattern (Fig 1). However, no type 1 Brugada ECG feature was revealed by a flecainide provocation test. The clinical phenotype in this case supports the pathogenicity of this novel genetic variant.

To elucidate the cause of SCD, a comprehensive post-mortem evaluation is required. Figure 2 shows the combined approach using both cardiac evaluation of surviving relatives and molecular autopsy on SUD victims which would give a higher yield on elucidating the underlying causes of SUD. Any history of syncope, cardiac symptoms especially related to exertion, emotion and stress, previous ECG, circumstances of SCD (activity at the time of death), any family history of cardiac disease, premature or sudden death, near-arrest attack, and epilepsy should be investigated. Patients with SADS can present with or be (incorrectly) labelled as having epilepsy.22 23 24 25 Guidelines on post-mortem examination of SUD in the young have been published by the Royal College of Pathologists of Australasia (https://www.rcpa.edu.au/getattachment/89884c69-f066-411d-a3d1- 39460444db13/Guidelines-on-Autopsy-Practice.aspx). In addition to traditional autopsy, some reports have used whole-body computed tomography and magnetic resonance imaging to identify structural heart abnormalities such as ARVC and HCM.26 27 However, these imaging modalities were not used in the present study.

Technologies applied in molecular autopsy have changed from Sanger sequencing targeting a few major SUD-related genes to NGS targeting an expanded gene panel of up to 200 genes.28 The former approach achieved diagnostic yields of around 15% to 20%.8 16 29 30 31 32 With increasing throughput capacities at more affordable costs, the large gene panel or exome approach by NGS is becoming more appealing in molecular autopsy. A proof-of-principle exome-wide study was carried out among 50 SUD subjects, with likely pathogenic variants identified in 32%.33 The diagnostic yield from various NGS studies, including our own, is up to 35%, despite the different lengths of gene lists.13 34 35 36 However, the spectrum of diseases might be different and rarer causes might be identified, because substantial clinical suspicion is typically lacking among young SCD victims with unremarkable premorbid history. However, a negative genetic result does not necessarily exclude the possibility of a genetic basis for the disease.

The major difficulty of molecular autopsy is establishing the causation between the death and the genetic variants. Applying molecular autopsy to the investigation of death causes involves probabilistic rather than binary “yes or no” answers. Interpretation of variant pathogenicity relies heavily on the characteristics of the genetic variant, allele rarity in population frequencies, in silico predictions, co-segregation patterns among affected family members, previous literature data on reports of similar cases, and available functional data. This approach follows the usual practice by which pathologists deal with genetic reporting.

There are two main advantages of molecular autopsy in SCD. First, molecular autopsy enables a correct diagnosis of SCD to be established, which can bring some level of closure to the family. Findings of a SADS-related condition can differentiate a natural cause from an unnatural cause of death, such as the possibility of drowning precipitated by an inherited arrhythmia during swimming. Second, the ultimate goal of diagnosing SCD is to prevent another tragedy among family members. The majority of SADS-related disorders are autosomal dominant inherited, and family members are at 50% risk of inheriting the mutation, with variable penetrance. Family pedigree for at least three generations should be included in the extended clinical workup, and genetic counselling should be considered for second- or higher-degree relatives wherever appropriate.

There are limitations to our study. First, SCD victims with autopsy conducted in a public hospital were not recruited in this study. Hence, the sample numbers may not be representative for the whole territory. Second, no premorbid ECG findings of the victims were available to correlate with the genetic results. Third, despite multiple efforts, no genetic mutations were found in over two thirds of the SCD victims in our study. Expanding the gene list may be able to increase the yield. The associated phenotypes and pathogenesis of some genes are still yet to be further elucidated.

Sudden arrhythmia death syndrome is a significant cause of SCD in the young. This study is first of its kind in an East Asian population and provides important data on the prevalence and types of SADS among young SCD victims. This is also the first local feasibility study on incorporating cardiac evaluation of surviving relatives and NGS molecular autopsy into the conventional forensic investigations. The interpretation of genetic variants in the context of SCD is complicated and we recommend its analysis and reporting by qualified pathologists. This model may be considered to cover all age-groups of SCD victims, as well as other potential applications such as sudden unexpected death in epilepsy, or sudden infant death syndrome. This local pilot study should be considered an important advance in diagnosing young SCD victims in East Asian populations.

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 manuscript, and critical revision for important intellectual content. 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.

All authors have disclosed no conflicts of interest.

The 5-year review was presented at the Asia Pacific Heart Rhythm Society Meeting on 3 to 6 October 2013, in Hong Kong. Part of the SADS HK Study results was presented at CardioRhythm on 24 to 26 February 2017, in Hong Kong and was published (J HK Coll Cardiol 2016;24:34).

This work was partly funded by SADS HK Foundation Limited, Hong Kong.

The study was approved by local ethics committees (Ethic Committee of the Department of Health (L/M 601/2013) and Kowloon West Cluster Research Ethics Committee (KWC-REC Reference KW/FR-13-023-67-05)). Consent was obtained from the next-of-kin of the SCD victims and from the first-degree relatives themselves under study.

1. Puranik R, Chow CK, Duflou JA, Kilborn MJ, McGuire MA. Sudden death in the young. Heart Rhythm 2005;2:1277-82. Crossref

2. Chugh SS, Reinier K, Teodorescu C, et al. Epidemiology of sudden cardiac death: clinical and research implications. Prog Cardiovasc Dis 2008;51:213-28. Crossref

3. Basso C, Corrado D, Thiene G. Cardiovascular causes of sudden death in young individuals including athletes. Cardiol Rev 1999;7:127-35. Crossref

4. Link MS. Sudden cardiac death in the young: epidemiology and overview. Congenit Heart Dis 2017;12:597-9. Crossref

5. Chugh SS, Kelly KL, Titus JL. Sudden cardiac death with apparently normal heart. Circulation 2000;102:649-54. Crossref

6. Corrado D, Basso C, Thiene G. Sudden cardiac death in young people with apparently normal heart. Cardiovasc Res 2001;50:399-408. Crossref

7. de Noronha SV, Sharma S, Papadakis M, Desai S, Whyte G, Sheppard MN. Aetiology of sudden cardiac death in athletes in the United Kingdom: a pathological study. Heart 2009;95:1409-14. Crossref

8. Tester DJ, Medeiros-Domingo A, Will ML, Haglund CM, Ackerman MJ. Cardiac channel molecular autopsy: insights from 173 consecutive cases of autopsy-negative sudden unexplained death referred for postmortem genetic testing. Mayo Clin Proc 2012;87:524-39. Crossref

9. Ackerman MJ, Tester DJ, Driscoll DJ. Molecular autopsy of sudden unexplained death in the young. Am J Forensic Med Pathol 2001;22:105-11. Crossref

10. Tester DJ, Ackerman MJ. The role of molecular autopsy in unexplained sudden cardiac death. Curr Opin Cardiol 2006;21:166-72. Crossref

11. Wever-Pinzon OE, Myerson M, Sherrid MV. Sudden cardiac death in young competitive athletes due to genetic cardiac abnormalities. Anadolu Kardiyol Derg 2009;9 Suppl 2:17-23.

12. Hellenthal N, Gaertner-Rommel A, Klauke B, et al. Molecular autopsy of sudden unexplained deaths reveals genetic predispositions for cardiac diseases among young forensic cases. Europace 2017;19:1881-90. Crossref

13. Hertz CL, Christiansen SL, Ferrero-Miliani L, et al. Next-generation sequencing of 100 candidate genes in young victims of suspected sudden cardiac death with structural abnormalities of the heart. Int J Legal Med 2016;130:91-102. Crossref

14. Christiansen SL, Hertz CL, Ferrero-Miliani L, et al. Genetic investigation of 100 heart genes in sudden unexplained death victims in a forensic setting. Eur J Hum Genet 2016;24:1797-802. Crossref

15. Semsarian C, Hamilton RM. Key role of the molecular autopsy in sudden unexpected death. Heart Rhythm 2012;9:145-50. Crossref

16. Tester DJ, Spoon DB, Valdivia HH, Makielski JC, Ackerman MJ. Targeted mutational analysis of the RyR2-encoded cardiac ryanodine receptor in sudden unexplained death: a molecular autopsy of 49 medical examiner/coroner’s cases. Mayo Clin Proc 2004;79:1380-4. Crossref

17. Priori SG, Napolitano C, Gasparini M, et al. Natural history of Brugada syndrome: insights for risk stratification and management. Circulation 2002;105:1342-7. Crossref

18. Hsueh CH, Chen WP, Lin JL, et al. Distinct functional defect of three novel Brugada syndrome related cardiac sodium channel mutations. J Biomed Sci 2009;16:23. Crossref

19. Yu JH, Hu JZ, Zhou H, et al. SCN5A mutation in patients with Brugada electrocardiographic pattern induced by fever [in Chinese]. Zhonghua Xin Xue Guan Bing Za Zhi 2013;41:1010-4.

20. Jimmy JJ, Chen CY, Yeh HM, et al. Clinical characteristics of patients with congenital long QT syndrome and bigenic mutations. Chin Med J (Engl) 2014;127:1482-6.

21. Lieve KV, Williams L, Daly A, et al. Results of genetic testing in 855 consecutive unrelated patients referred for long QT syndrome in a clinical laboratory. Genet Test Mol Biomarkers 2013;17:553-61. Crossref

22. Medford BA, Bos JM, Ackerman MJ. Epilepsy misdiagnosed as long QT syndrome: it can go both ways. Congenit Heart Dis 2014;9:E135-9. Crossref

23. Partemi S, Cestèle S, Pezzella M, et al. Loss-of-function KCNH2 mutation in a family with long QT syndrome, epilepsy, and sudden death. Epilepsia 2013;54:e112-6. Crossref

24. Goyal JP, Sethi A, Shah VB. Jervell and Lange-Nielson Syndrome masquerading as intractable epilepsy. Ann Indian Acad Neurol 2012;15:145-7. Crossref

25. Tu E, Bagnall RD, Duflou J, Semsarian C. Post-mortem review and genetic analysis of sudden unexpected death in epilepsy (SUDEP) cases. Brain Pathol 2011;21:201-8. Crossref

26. Roberts IS, Benamore RE, Benbow EW, et al. Post-mortem imaging as an alternative to autopsy in the diagnosis of adult deaths: a validation study. Lancet 2012;379:136-42. Crossref

27. Puranik R, Gray B, Lackey H, et al. Comparison of conventional autopsy and magnetic resonance imaging in determining the cause of sudden death in the young. J Cardiovasc Magn Reson 2014;16:44. Crossref

28. Semsarian C, Ingles J. Molecular autopsy in victims of inherited arrhythmias. J Arrhythm 2016;32:359-65. Crossref

29. Liu C, Zhao Q, Su T, et al. Postmortem molecular analysis of KCNQ1, KCNH2, KCNE1 and KCNE2 genes in sudden unexplained nocturnal death syndrome in the Chinese Han population. Forensic Sci Int 2013;231:82-7. Crossref

30. Doolan A, Langlois N, Chiu C, Ingles J, Lind JM, Semsarian C. Postmortem molecular analysis of KCNQ1 and SCN5A genes in sudden unexplained death in young Australians. Int J Cardiol 2008;127:138-41. Crossref

31. Skinner JR, Crawford J, Smith W, et al. Prospective, population-based long QT molecular autopsy study of postmortem negative sudden death in 1 to 40 year olds. Heart Rhythm 2011;8:412-9. Crossref

32. Hofman N, Tan HL, Alders M, et al. Yield of molecular and clinical testing for arrhythmia syndromes: report of 15 years’ experience. Circulation 2013;128:1513-21. Crossref

33. Bagnall RD, Das K J, Duflou J, Semsarian C. Exome analysis-based molecular autopsy in cases of sudden unexplained death in the young. Heart Rhythm 2014;11:655-62. Crossref

34. Bagnall RD, Weintraub RG, Ingles J, et al. A prospective study of sudden cardiac death among children and young adults. N Engl J Med 2016;374:2441-52. Crossref

35. Farrugia A, Keyser C, Hollard C, Raul JS, Muller J, Ludes B. Targeted next generation sequencing application in cardiac channelopathies: analysis of a cohort of autopsy-negative sudden unexplained deaths. Forensic Sci Int 2015;254:5-11. Crossref

36. Tester DJ, Ackerman MJ. Postmortem long QT syndrome genetic testing for sudden unexplained death in the young. J Am Coll Cardiol 2007;49:240-6. Crossref