Approximately 10.6 persons in Hong Kong die of coronary heart diseases each day.1 Coronary heart diseases represent a spectrum of disorders, from common atherosclerosis to life-threatening ST-segment elevation myocardial infarction (STEMI). The European Society of Cardiology has divided the total ischaemic time in patients with STEMI into patient and system delays.2 System delays are further subdivided into emergency medical system and non-emergency medical system (hospital) delays. Minimising delays of all types would be most beneficial for patient outcome. To the best of our knowledge, there have been no studies of the components of ischaemic time in patients with STEMI in Hong Kong.

The Hong Kong Fire Services Department (HKFSD) serves as the primary emergency ambulance provider in Hong Kong. Ambulance services are activated by the Fire Service Control Centre upon calls to 999; the ambulances are staffed in accordance with protocols approved by the HKFSD Medical Director.3 Patients are transported to the nearest public hospital, based on their geographical location. There is no opportunity for patients to choose the destination hospital, and no primary diversions are used for chest pain/STEMI.

To shorten the total ischaemic time in patients with STEMI, the HKFSD and the Department of Accident and Emergency (AED) of Queen Mary Hospital (QMH) jointly launched a pilot project named ‘Prehospital Ambulance 12-lead Electrocardiogram for Chest Pain Patients in Hong Kong West Cluster’. Our group previously published an article using data from the first phase of this pilot project (12 November 2015 to 31 December 2016), demonstrating the impact of prehospital 12-lead electrocardiogram (ECG) on door-to-balloon time (D2B) in patients with STEMI who received primary percutaneous coronary intervention (PPCI).3 However, the majority of D2B time constituted the post-admission period, whereas myocardial injury likely began with the onset of symptoms before hospital admission. We hypothesised that prehospital ECG would also shorten the ischaemic time before hospital admission, including the first medical contact-to-first ECG time, AED door-to-triage and AED door-to-consultation time, and length of stay in the AED. In the present analysis, pre-admission data from our pilot project were compared with those of patients with STEMI who received the standard management. By assessing the patient flow from the onset of symptoms until admission, patient and system delays could be identified, and potential improvements could be implemented accordingly.

This was a retrospective observational study of data from the first and second phases of the ‘Prehospital Ambulance 12-lead Electrocardiogram for Chest Pain Patients in Hong Kong West Cluster’ pilot project (12 November 2015 to 5 November 2017). Data were provided by the QMH AED, QMH Cardiac Care Unit (CCU), and HKFSD.

The QMH CCU provided the list of all patients admitted through the AED with emergency coronary angiography, with or without PPCI. This included patients who travelled to hospital by self-arranged transport, as well as those who travelled by ambulance with and without prehospital 12-lead ECG. Patients were excluded if they had STEMI that developed during in-patient stay, if they had post-arrest malignant arrhythmia, or if they had extracorporeal-membrane-oxygenation-assisted cardiopulmonary resuscitation (ECMO-CPR). Patients secondarily transported from St John Hospital on an outlying island were also excluded. The HKFSD and QMH AED provided data regarding the first and second phases of our pilot project, in which 15 HKFSD ambulances in Hong Kong West Cluster were equipped with X Series Defibrillator/Monitor (Zoll Medical Corporation, Chelmsford [MA], US). Prehospital 12-lead ECGs of chest pain patients were tele-transmitted to the AED for immediate assessment by an emergency physician. Physicians from CCU received alerts regarding patients with suspected STEMI in ECG. This pilot project began on 12 November 2015; the first phase ended on 31 December 2016. In the first phase, all prehospital ECGs were obtained in ambulance compartments before departure from the scene. In the second phase (1 January 2017 to 5 November 2017), prehospital ECGs were performed either upon ambulance on scene or in the ambulance compartment, as determined by the ambulance crew based on the feasibility of carrying the X Series Defibrillation/Monitor to the scene.

The primary objective of this study was to investigate whether there were significant differences in patient and system delays between the use of prehospital 12-lead ECG and standard care. The primary outcomes were the patient and system delay times in minutes. Patient delay was measured from the time of symptom onset to the time of emergency response system activation. The times of emergency response system activation were the time of the 999 call for ambulance patients and the time of AED registration for patients with self-arranged transport. System delay was measured from the time of emergency response system activation to the time of admission through the AED. System delay was further stratified as the time from ambulance on scene to AED registration; ambulance on scene to first ECG (for patients travelling by ambulance); AED door-to-triage time, AED door-to-first AED ECG time, AED door-to-physician consultation time, and length of stay in the AED (for all patients). The secondary objective of this study was to assess accuracy of triage, benefit of performing prehospital ECG when ambulance on scene compared with in the ambulance compartment, ability to call CCU on/before patient arrival, and adherence to HKFSD performance pledge. Triage accuracy was measured by proportion of patients correctly triaged as category 1. Benefit of performing prehospital ECG when ambulance on scene compared with in the ambulance compartment was measured by shortened time in minutes to undergo ECG when the former is performed. Ability to call CCU on/before patient arrival was measured by the proportion of patients with CCU calls on/before AED registration. Adherence to HKFSD performance pledge was measured as the time from 999 call to the time of ambulance arrival at the street address.

Data were analysed by Excel 2010 (Microsoft Corp, Redmond [WA], US) and SPSS (Windows version 25.0; IBM Corp, Armonk [NY], US). The non-parametric Mann-Whitney U test was used to analyse statistical differences between groups.

In all, 245 patients were admitted through the AED to the CCU with emergent coronary angiogram, with or without PPCI. In total, 43 patients were excluded because they developed STEMI during in-patient stay, exhibited post-arrest malignant arrhythmia, or required ECMO-CPR. In addition, five patients who were secondarily transported from St John Hospital were excluded. Data were analysed for a total of 197 patients from CCU and our pilot project. In the first phase, 118 patients were included; all underwent prehospital ECGs in the ambulance compartment. In the second phase, 79 patients were included; 36 and eight underwent prehospital ECGs in the ambulance compartment and on scene, respectively. The remaining 35 patients were transported by ambulances without prehospital ECG capabilities (Fig).

The median patient delay time was 90 minutes; 12% of patients experienced delays of >12 hours. There was significant difference in patient delay between patients travelling by ambulance and those who used self-arranged transport (Table 1).

There were significant differences between patients with and without prehospital ECG, in terms of median time from ambulance on scene to AED registration (Table 2), and ambulance on scene to first ECG (Table 3). Prehospital ECG was available 5 minutes earlier if performed upon ambulance on scene compared with that in ambulance compartment. This 5-minute difference also gave us an estimate of time required to transfer the patient to the ambulance from the scene where the patient experienced chest pain. Prehospital ECG performed at the time of ambulance on scene was available 35 minutes earlier than the first ECG (in the AED) for patients who used self-arranged transport.

After excluding six patients with unknown time of 999 call or unknown time of ambulance arrival on scene, the median time from 999 call to ambulance on scene was 8 minutes (interquartile range, 6-10 minutes). Overall, 119 of 123 (96.7%) patients had times within 17 minutes (12 minutes to street address performance pledge by HKFSD, plus 5 minutes travel time between ambulance and scene). Forty-three patients with STEMI underwent prehospital ECG and had known CCU call times; of these, ambulance crews in 7 of 8 with ECG performed on scene (88%), and 25 of 35 with ECG performed in the ambulance compartment (71%), were able to contact the CCU physician on or before patient arrival in AED. With prehospital ECG, AED door-to-triage time, AED door-to-first AED ECG time, AED door-to-physician consultation time, and length of stay in the AED were all significantly shortened. In addition, more patients with STEMI were correctly triaged as category 1 if they had undergone prehospital ECG (Table 4).

The MEDEA study showed that female sex and age >65 years were factors associated with delayed presentation in patients with acute myocardial infarction.4 Perceived barriers in care-seeking and symptom congruence were also associated with prehospital delays in Hong Kong Chinese patients with acute myocardial infarction.5 Although current guidelines recommend PPCI by an experienced team within 12 hours of symptom onset as a reperfusion strategy in patients with STEMI,2 12% of patients in our study called 999 or arrived at the AED >12 hours after symptom onset (Table 1). These delays might have worsened their outcomes. Moreover, 15 of 197 (7.6%) were evaluated by another physician before they presented to the AED, and those other physicians might have performed ECG in their clinics. Regardless of the availability of ECG in the other clinics, there is no formal communication channel between the AED and private physicians, with the exception of referral letters. These factors may also have contributed to delays.

A key factor for reducing prehospital delay is education of the public, especially high-risk patients, regarding the prudent use of emergency medical services and the typical symptoms of myocardial infarction. Online symptom checkers have been devised for patient self-diagnosis. However, an audit study found only moderate accuracy in these algorithms.6 Because patients with STEMI often have other medical co-morbidities that require regular follow-up by family physicians or general out-patient clinics, education regarding typical symptoms of myocardial infarction and the availability of prehospital ECG assessments could be provided during these regular consultations. Notably, prehospital ECG can be performed only after a patient has activated the emergency response system; therefore, it cannot shorten the patient delay unless it is coupled with public education regarding STEMI symptoms and public awareness of prehospital ECG availability.

Approximately one-third of patients with STEMI in our pilot project used self-arranged transport for travelling to the hospital. This phenomenon is not unique to Hong Kong.7 8 9 Compared with patients travelling by ambulance, a greater proportion of patients travelling by self-arranged transport had patient delay time of >12 hours (Table 1). Self-arranged transport patients did not undergo prehospital ECG. The delayed activation of the emergency response system deprived such patients of early assessment by an ambulance crew, as well as early notifications to the AED and CCU physicians of unstable vitals and a potential need for life-saving treatment. It is important to educate the public on appropriate use of ambulance services. Mass media campaigns regarding use of ambulances for chest pain have been shown to increase emergency medical system use by people with chest pain and suspected acute coronary syndromes.10 In this study, there was a significant difference in the median times for ambulance patients to make 999 calls and for self-arranged transport patients to register in the AED (Table 1). This may be related to differences in awareness of STEMI symptoms, different patient attitudes regarding management of their own symptoms, and different treatment expectations regarding ambulance services. Education should emphasise early 999 calls for patients with STEMI symptoms, instead of the use of self-arranged transport.

For emergency ambulance calls, the HKFSD has a performance pledge of within 12 minutes for 92.5% of patients, from the time of the call to the arrival of an ambulance at the designated street address.11 For chest pain patients with suspected myocardial infarction, even when 5 additional minutes are added to include travel time from the ambulance to the scene (as estimated in Table 3), this pledge remains more stringent than those of ambulance services in other countries. For example, the average response time by the London Ambulance Service for corresponding category 2 calls is 18 minutes.12 In our 123 patients with STEMI who were transported by ambulances, the median response time was 8 minutes; 96.7% of patients received a response within the target of within 17 minutes during the pilot project. By enhancing the mobilising system and commissioning new fire stations/ambulance depots at various locations, the response time for chest pain patients is expected to decrease further.

Before this pilot project was implemented, there were concerns regarding delayed transport to the hospital due to the performance of prehospital ECG. In contrast to our expectations, the median scene-to-AED registration time was 3 minutes shorter in patients who underwent prehospital ECG, compared with patients who did not undergo prehospital ECG (Table 2). However, every ambulance journey was unique and involved a different travel distance. Because the actual distances travelled and the proportions of time used for performing prehospital ECG in each ambulance journey were not available, a larger data analysis is needed to clarify these findings.

Among the 44 patients with STEMI who underwent prehospital ECG, the CCU physician was called before patient arrival in 7 of 8 (88%) patients for whom ECG was performed on scene and in 25 of 35 (71%) patients for whom ECG was performed in the ambulance compartment. The CCU call time was unknown for the remaining one patient. Therefore, performing prehospital ECG on scene allowed earlier CCU consultations. It has been recommended that 12-lead ECG recording and interpretation should be performed as soon as possible upon initial medical contact, with a target maximum delay of 10 minutes.13 This target was achieved in our pilot project in patients for whom prehospital ECG was performed on scene (median time: 6 minutes) [Table 3]. We were near this target in patients for whom ECG was performed in the ambulance compartment (median time: 11 minutes). Therefore, ECG performance on scene, rather than in the ambulance compartment, is recommended whenever feasible.

Preliminary history and vital signs collected during ambulance transport were immediately given to the triage nurse upon patient arrival. This shortened the triage time to 0 minutes (door-to-triage time) in patients for whom prehospital ECG was performed. Moreover, triage accuracy was improved: more patients with STEMI (40/44) who underwent prehospital ECGs were correctly triaged as category 1 (critical) and therefore received immediate treatment. Shortened times for door-to-first AED ECG and door-to-consultation were evident. Patients with STEMI who were transported by ambulance could provide notification to the AED before their arrival, thereby enabling pre-arrival preparation of an acute care room. In addition, an immediate AED physician assessment with repeat ECG could be performed upon patient arrival. These changes resulted in an overall reduction in the length of AED stay (nearly by half), compared with the length of AED stay for patients who did not undergo prehospital ECG. Shortened length of AED stay is beneficial for percutaneous coronary intervention centres where patients with STEMI cannot yet bypass the AED with direct catheterisation laboratory admission.

Delayed activation of the emergency response system and choice of transportation contributed to patient delay. Prehospital 12-lead ECG, preferably performed on scene, can shorten system delay and total ischaemic time in STEMI management.

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

All authors have disclosed no conflicts of interest.

We would like to thank the staff of the following institutions: (1) Hong Kong Fire Services Department, for performing prehospital electrocardiograms and providing prehospital data; (2) Division of Cardiology, Department of Medicine, Queen Mary Hospital, for providing data on door-to-balloon and door-to-catheter time; (3) Department of Accident and Emergency, Queen Mary Hospital, for collecting patient clinical data; (4) Mr Fan Min of Emergency Medicine Unit, Li Ka Shing Faculty of Medicine, The University of Hong Kong, for statistical analyses; and (5) Zoll Medical Corporation (Chelmsford [MA], US), for providing the machines, training, and technical support free of charge.

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

Analysis of data from this pilot project was approved by the Institutional Review Board of The University of Hong Kong/ Hospital Authority Hong Kong West Cluster (UW19-184). The requirement for patient consent was waived.

1. HealthyHK, Department of Health, Hong Kong SAR Government. Coronary heart diseases. Available from: https://www.healthyhk.gov.hk/phisweb/en/healthy_facts/disease_burden/major_causes_death/coronary_heart_disease/. Accessed 21 Apr 2019.

2. Ibanez B, James S, Agewall S, et al. 2017 ESC Guidelines for the management of acute myocardial infarction in patients presenting with ST-segment elevation: The Task Force for the management of acute myocardial infarction in patients presenting with ST-segment elevation of the European Society of Cardiology (ESC). Eur Heart J 2018;39:119-77. Crossref

3. Cheung KS, Leung LP, Siu YC, et al. Prehospital 12-lead electrocardiogram for patients with chest pain: a pilot study. Hong Kong Med J 2018;24:484-91. Crossref

4. Ladwig KH, Fang X, Wolf K, et al. Comparison of delay times between symptom onset of an acute ST-elevation myocardial infarction and hospital arrival in men and women <65 years versus ≥65 years of age: findings from the Multicenter Munich Examination of Delay in Patients Experiencing Acute Myocardial Infarction (MEDEA) Study. Am J Cardiol 2017;120:2128-34. Crossref

5. Li PW, Yu DS. Predictors of pre-hospital delay in Hong Kong Chinese patients with acute myocardial infarction. Eur J Cardiovasc Nurs 2018;17:75-84. Crossref

6. Semigran HL, Linder JA, Gidengil C, Mehrotra A. Evaluation of symptom checkers for self diagnosis and triage: audit study. BMJ 2015;351:h3480. Crossref

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

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

9. Lavery T, Greenslade JH, Parsonage WA, et al. Factors influencing choice of pre-hospital transportation of patients with potential acute coronary syndrome: an observational study. Emerg Med Australas 2017;29:210-6. Crossref

10. Nehme Z, Cameron PA, Akram M, et al. Effect of a mass media campaign on ambulance use for chest pain. Med J Aust 2017;206:30-5. Crossref

11. Fire Services Department, Hong Kong SAR Government. Performance pledge. Available from: https://www.hkfsd.gov.hk/eng/aboutus/performance.html. Accessed 21 Apr 2019.

12. NHS London Ambulance Service. New ambulance response categories. Available from: https://www.londonambulance.nhs.uk/calling-us/17086-2/. Accessed 21 Apr 2019.

13. Diercks DB, Peacock WF, Hiestand BC, et al. Frequency and consequences of recording an electrocardiogram >10 minutes after arrival in an emergency room in non-ST-segment elevation acute coronary syndromes (from the CRUSADE Initiative). Am J Cardiol 2006;97:437-42. Crossref

Respiratory syncytial virus (RSV) infection poses a heavy disease burden in children worldwide.1 2 Haemodynamically significant congenital heart disease (hs-CHD) has been mainly studied and identified as a risk factor for severe RSV infection.3 4National immunoprophylaxis policies for RSV in children have been adopted worldwide.5 6 7 8 9 10 11 In Hong Kong, under the Paediatric Coordinating Committee of the Hospital Authority, guidelines and government funding for RSV immunoprophylaxis for children with bronchopulmonary dysplasia of prematurity were established in 2012.12 However, no consensus has been reached regarding guidelines for RSV immunoprophylaxis for children with congenital or acquired heart disease (HD), because the epidemiology and impact of RSV infection on these patients have not been delineated in Hong Kong.13 14 15 16 It has been suggested that the high morbidity and mortality rates of RSV infection in children with hs-CHD observed during the pre-palivizumab era3 are no longer applicable, owing to advances in healthcare. To widen the scope of the study regarding HD there is a need to conduct an updated local study regarding the epidemiology and impact of RSV infection on children with all types of HD, with the aim of providing evidence-based recommendations for RSV immunoprophylaxis.

Hong Kong has a population of 7.48 million, including a paediatric population (aged ≤18 years) of approximately 1.1 million. Over 90% of in-patient service and nearly all tertiary service is provided by the public health system. There are 12 public hospitals providing approximately 1500 acute paediatric beds, of which 45 are paediatric intensive care beds; the total annual discharges and deaths are approximately 88 500. The present study was a multicentre retrospective case-control study of RSV infection in children in four hospitals with large paediatric departments, including 22 paediatric intensive care beds and approximately 630 acute paediatric beds, from 1 January 2013 to 31 December 2015.

Paediatric patients were recruited using the Clinical Data Analysis and Reporting System electronic database of the Hong Kong Hospital Authority. The inclusion criteria were: (1) any discharge diagnosis of RSV infection, including bronchiolitis and pneumonia [International Classification of Diseases, Ninth Revision, codes: 079.6 (0), 466.0 (9), 466.11, 480.1]; and (2) laboratory confirmation of RSV infection from patients’ nasopharyngeal or endotracheal secretions, either by immunofluorescent antigen staining (D³ Ultra 8 DFA; Diagnostic Hybrids Inc, Athens [OH], US) or RNA detection by reverse transcriptase polymerase chain reaction (Xpert Xpress Flu/RSV; Cepheid, Sunnyvale [CA], US).

Information was collected regarding epidemiologic characteristics, demographics, and clinical information (eg, laboratory and pharmacy data, pre-existing co-morbidities, complications, reinfection, and intensive care unit [ICU] and ventilator requirements). Complications were defined as new secondary diagnosis related to RSV infection in a specific episode. Heart disease was defined as the presence of any active HD, including structural defects and cardiomyopathies. Patients who showed complete resolution of HD were considered normal. Diagnosis was confirmed by echocardiography, with or without cardiac catheterisation. Details of HD were retrieved, such as haemodynamic status, need for medication, and need for operation. One designated paediatrician from each hospital verified and retrieved information from the patients’ paper records, if needed. The numbers of hospital admissions for acute respiratory infections in the same period were also retrieved from each hospital. Meteorological information was obtained from the electronic database of the Hong Kong Observatory.

First, the epidemiology of RSV infection and its association with meteorological conditions in the entire cohort was studied. Then patients were excluded if they had significant co-morbidities other than HD, including chronic lung disease, neuromuscular problems, and immunocompromised status, which are expected to increase the severity of RSV infection.17 18 19 Patients with social problems awaiting placement in the hospital causing undue prolonged hospital stay were also excluded. The remaining patients were then divided into heart disease (HD) and control (without any co-morbidities) groups to compare the severity of RSV infection. Cardiac patients alone were analysed to identify cardiac predictors of severe outcome from RSV infection (Fig 1).

Statistical analysis was conducted using SPSS (Windows version 23.0; IBM Corp, Armonk [NY], US). Continuous data were tested for normality. There was a large difference in sample size between the HD and control groups; therefore, non-parametric tests were used to compare these groups: the Mann-Whitney U test was used for univariate analysis of continuous data, and the Chi squared test or Fisher’s exact test were used for categorical variables, where appropriate. The same statistical analyses were repeated using a smaller control group (1/10 of the original size), which was obtained by generating an age- and sex-matched random sample from the original control group. This repeat analysis yielded similar results, which showed that the statistical tests used were acceptable despite the large discrepancy in sample size. Univariate and multivariate regression analyses were also performed. A backward variable selection method was used for model building. Results were considered statistically significant when P≤0.05.

There were 3538 RSV-related hospital admissions in the four paediatric departments during the 3-year period, which constituted 11.8% of acute respiratory infection admissions to the four departments and 43% of all RSV admissions to public hospitals in Hong Kong. The age distribution of the admitted patients is shown in Table 1. The RSV infections primarily affected children aged <5 years: 96.2% in the entire cohort and 97.3% in the HD group. Furthermore, RSV infections were most common in children aged <1 year: 44.6% in the entire cohort and 56.3% in the HD group. There was a mild male sex preponderance in the entire cohort, as well as in the HD group: male-to-female ratios of 1.32 and 1.29). The mortality rates were 0.14% (n=5) in the entire cohort and 0% in the HD group. All patients who died had an underlying chronic illness, such as congenital hereditary muscular dystrophy, spinal muscular atrophy, acute leukaemia, or asthma.

The aggregated monthly trend of RSV infection is shown in Figure 2. Most infections (87.7% of those in the entire cohort and 91.1% of those in the HD group) occurred during the period from January to September, peaking from March to August (the spring and summer months). With respect to meteorological factors, bivariate correlation analysis of the entire cohort showed that the aggregated monthly RSV incidence had a statistically significant positive correlation with the monthly mean relative humidity (r=0.71, P=0.010) and statistically significant negative correlations with the monthly mean wind speed (r=-0.80, P=0.002) and monthly mean atmospheric pressure (r=-0.62, P=0.032) [Table 2]. Backward linear regression revealed that only monthly mean wind speed was significantly and independently associated with monthly RSV incidence (B=-31.716, 95% confidence interval [CI]=-48.61 to -14.82; P=0.002).

Overall, 225 episodes were excluded because they involved patients who had co-morbidities other than HD. The HD group had 112 episodes involving 105 patients, while the control group had 3201 episodes involving 3155 patients. The demographic and clinical details of both groups are depicted in Table 3. The HD and control groups had similar age distribution (median [interquartile range]=0.9 years [0.46-1.97 years] vs 1.19 years [0.45-2.4 years]; P=0.068) and sex proportion (male preponderance of 56.3% vs 56.9%, P=0.899). There was no statistically significant difference in mortality rate (0% in HD group vs 0.03% in control group; P=1). Patient-based analysis revealed that the HD group had a higher RSV re-infection rate (>1 episode in the study period) than the rate in the control group (7.6% vs 1.7%, P<0.001). Respiratory syncytial virus infection was statistically significantly more severe in the HD group than in the control group in terms of the hospital length of stay (median=4 days [3-6 days] vs 2 days [2-4 days]; P<0.001), respiratory failure requiring respiratory support (17.0% vs 5.1%; P<0.001), requirement of ICU care (11.6% vs 1.4%, P<0.001), requirement of invasive or non-invasive mechanical ventilation (3.6% vs 0.4%, P=0.003), and occurrence of complications (28.6% vs 9.8%, P<0.001). Respiratory failure and dehydration were common complications in both groups. However, heart failure exacerbation (n=13) and arrhythmia (n=3) only occurred in the HD group, while febrile convulsion (n=96), acute myocarditis (n=2), and Kawasaki disease (n=8) were only observed in the control group. There was no statistically significant difference between the two groups in terms of co-infection rate (8.9% vs 7.8%; P=0.665). More than half of co-infections (55.4%) were due to respiratory organisms: rhinovirus was most common, followed by adenovirus, Haemophilus influenzae, and Pneumococcus bacteria (Table 4).

In the HD group, 26 patients had a history of partial surgical repair of heart lesions, while nine others required urgent cardiac surgery after recovery from RSV infection. Four cardiac risk factors were identified: (1) heart failure requiring medications (n=26); (2) pulmonary hypertension (n=7); (3) severe airway abnormalities associated with congenital heart disease (CHD) such as pulmonary artery sling (n=6); and (4) cyanotic CHD (n=1). Each patient could have multiple risk factors; 75 patients had no risk factors. Regression analyses involving age, sex, and the four risk factors were conducted to identify variables that could predict the severity of RSV infection. Backward multivariate linear regression showed that total hospital length of stay was positively and independently associated with heart failure (B=4.65, 95% CI=2.55-6.74, P<0.001), pulmonary hypertension (B=5.52, 95% CI=1.89-9.15, P=0.003), and airway abnormalities (B=10.35, 95% CI=6.44-14.25, P<0.001). The ICU length of stay was positively and independently associated with airway abnormalities (B=5.90, 95% CI=3.73-8.07, P<0.001). Backward binary logistic regression revealed that the requirement of ICU care was positively associated with pulmonary hypertension (adjusted odds ratio [aOR]=20.67, 95% CI=3.74-114.21, P=0.001) and airway abnormalities (aOR=15.50, 95% CI=2.56-93.84, P=0.003). Similarly, respiratory failure was positively associated with both pulmonary hypertension (aOR=9.27, 95% CI=1.41-61.05, P=0.021) and airway abnormalities (aOR=11.21, 95% CI =1.84-68.33, P=0.009) [Table 5].

To the best of our knowledge, this is the first study reviewing the epidemiology and impact of RSV infection on children with HD in Hong Kong. The representative sample in this study included 43% of all RSV-related hospitalisations in the public sector. Respiratory infection due to RSV was common in children in Hong Kong, causing 11.8% of hospital admissions for acute respiratory infection; it was most common in children aged <5 years, similar to the findings of another study in Hong Kong (96.2% vs 98.4%).13 However, the incidence of respiratory infection due to RSV in patients aged <1 year in the present study was lower than that in Western studies (44.6% vs 75%-90%).2 This is potentially because babies in Hong Kong are typically cared for at home with the help of grandparents or domestic helpers, rather than attending nursery facilities. Coupled with the small number of children in most Hong Kong families, this may have reduced the rate of cross-infection. 20 Male sex was identified as a risk factor for RSV-related hospitalisation, as in prior studies, but the male preponderance in this study was slightly lower than that of other studies (1.32:1 vs 1.44-1.65:1).2 13 20 21 22 23 The mortality rate of RSV-related hospitalisation was extremely low (0.14%), which was similar to the rates in local and worldwide studies (0.15%-1%).2 13 15 Most deaths occurred in patients who had underlying chronic illnesses.

Although the seasonality of RSV infection in Hong Kong, a subtropical area, is not well-defined in the manner observed in temperate regions, there is a degree of seasonal variation. As in other studies from Hong Kong and Singapore,13 14 15 16 24 this study showed that RSV infection peaked in the humid spring and summer months, but was less common from October to December during the dry and windy season. Indeed, analysis of meteorological data showed that RSV infection was positively correlated with relative humidity and negatively correlated with both wind speed and atmospheric pressure. Relative humidity has been consistently associated with RSV infection rate in other studies.13 16 25 Notably, RSV in large particle aerosol form is more stable at higher humidity and may thus favour transmission.25 While temperature, rainfall, and sunlight were associated with RSV infection in previous studies,13 16 25 this study showed that wind speed was negatively associated with RSV infection. We presume that strong wind disturbs the stability of RSV in aerosol. We recommend that individuals who care for young children maintain a dry and well-ventilated environment to decrease the likelihood of cross-infection. In addition, this study highlighted the differences in RSV seasonality in Hong Kong, relative to other parts of Asia. Taiwan has two biennial peaks (spring and autumn), while Malaysia has a single peak infection period (September to December).22 23 Thus, local epidemiological and climatic data are pivotal in determining the appropriate timing for RSV immunoprophylaxis.

In this study, we included all types of HD, rather than CHD alone; notably, acquired HD, such as cardiomyopathy, can increase vulnerability to RSV infection.26 The characteristics of RSV infection in cardiac patients have changed. First, the percentage of cardiac patients in this study was lower than that in prior studies (3.2% vs 8%-16.4%).3 22 27 28 29 Second, the outcome of RSV infection in this study was less severe than that of prior studies with respect to the requirement of ICU care (11.6% vs 30.4%-63%), requirement of mechanical ventilation (3.6% vs 19%-24%), and rate of mortality (0% vs 2.5%-37%).3 22 27 28 29 These differences may be related to the ongoing tendencies for economic and healthcare improvement; they may also be related to good antenatal ultrasound service, termination of pregnancies involving severe CHD, and easy access to medical service in Hong Kong. However, as in other studies, children with HD in Hong Kong exhibit significantly more severe outcomes from RSV infection than do children without co-morbidity.15 22 26 29 30 31

In the present study, myocarditis and Kawasaki disease were complications of RSV infection in the non-cardiac group. These associations have not been extensively reported in prior studies, and further investigation is needed. The lack of these complications in the HD group may be explained by the much smaller sample size in that group (the HD group had 112 episodes involving 105 patients, while the control group had 3201 episodes involving 3155 patients). The rate of co-infection with other respiratory viruses in this study was low compared with that reported in a meta-analysis (4.3% vs ≤50%)2; however, the findings were similar in that dual infection led to a more severe disease course and that rhinovirus was the most common co-infecting agent. Notably, adenovirus was the second most common agent in this study, whereas it was human bocavirus in the meta-analysis; the presence of this virus is not routinely assessed in Hong Kong.

Within the HD group, we identified three severity predictors for RSV infection, namely heart failure requiring medications, pulmonary hypertension, and severe airway abnormalities associated with CHD. The first two have been well described3 22 26 27 and were included in the American Academy of Pediatrics guideline for RSV prophylaxis11; severe airway abnormality associated with CHD has not been extensively investigated in previous studies of RSV infection. Pulmonary artery sling with tracheal stenosis is not easily identified during prenatal examinations and some affected patients remain undiagnosed for a few months after birth. The disease may first be identified during respiratory infection, such as RSV, and most patients exhibit severe courses of disease. In contrast to the findings of the Taiwanese study,32 we did not find cyanotic CHD to be a severity predictor. This was potentially because the number of such patients was very small in our series (n=1). Indeed, the number of live births of patients with severe cyanotic CHD is decreasing in Hong Kong due to the high rate of termination of pregnancies involving the disease. The number of surviving patients with severe untreated cyanotic CHD is also small, as they either receive some form of surgical treatment or do not survive early infancy.

Early studies stressed the importance of hygiene and infection control for the prevention of RSV infection.20 The IMpact-RSV study in the late 1990s led to the approval of palivizumab (a monoclonal antibody) for passive immunisation in the US in 1998 and in Europe in 1999 for children with chronic lung disease or prematurity.5 6 33 In 2003, the use of palivizumab in the US was extended to children with hs-CHD (heart failure, pulmonary hypertension, and cyanotic CHD) who were aged ≤2 years.7 34 Comparable national guidelines have been gradually adopted globally, including in Asian countries (Japan in 2006 and Korea in 2009).8 9 In temperate regions with a distinct RSV season length of approximately 6 months, the typical palivizumab regimen is monthly intramuscular injection at 15 mg/kg/dose during the RSV season, with a maximum of five doses. However, in subtropical regions without a clear RSV season, monthly injection may be necessary for the entire year, which is costly. In 2011, a study in Taiwan showed that a protocol of six consecutive monthly doses in at-risk children, beginning when the risk was initially detected in the first year of life, was also effective.10 Currently, the American Academy of Pediatrics limits routine RSV immunoprophylaxis to children who have acyanotic CHD with heart failure or pulmonary hypertension in the first year of life.11 In the United Kingdom, children who have cyanotic or acyanotic CHD with significant co-morbidities are recommended to receive RSV immunoprophylaxis until age 2 years.35 In 2017, however, an international steering committee broadened the indications to children with unoperated hs-CHD or symptomatic airway abnormalities who were aged ≤2 years, as well as children with cardiomyopathy requiring treatment who were aged ≤1 year.36 A Canadian-Italian group questioned whether children with hs-CHD need higher doses of immunoprophylaxis because of their increased inherent risks.37 Numerous studies have shown the safety and efficacy of RSV immunoprophylaxis and its long-term benefit in children with HD.6 30 38 39 40 It is therefore imperative to establish similar guidelines in Hong Kong. In accordance with local experience regarding bronchopulmonary dysplasia in premature babies,12 an RSV immunoprophylaxis scheme administered monthly for five doses, beginning in the first year of life, should be considered in children with HD who exhibit severity predictors such as heart failure, pulmonary hypertension, and/or severe airway abnormality related to CHD. The optimal timing for immunoprophylaxis may be during the local peak of infection, from March to August.

Because this was a retrospective hospital-based study, the incidence of RSV respiratory infection may have been underestimated; however, we presume that only patients with relatively mild disease were potentially omitted. As in all retrospective reviews, this study did not use a pre-defined protocol for management and clinical documentation. This limitation was partially addressed by analysing objective clinical details which are relatively standard and easily retrievable through electronic means. We also collected out-patient follow-up records when necessary to provide additional information regarding co-morbidities. However, instances of re-infection before and after the study period were not assessed due to technical limitations, which is not ideal.

Respiratory syncytial virus infections are common in Hong Kong, and the incidence peaks from March to August; prevalence is greatest in children aged <1 year, and there is a mild male preponderance. Infection is favoured by high relative humidity, low wind speed, and low atmospheric pressure. Heart disease, both congenital and acquired, is a distinct risk factor for severe RSV infection in terms of hospital length of stay, reinfection, complication, respiratory failure, and the requirements for ICU care and mechanical ventilation. In Hong Kong, an RSV immunoprophylaxis scheme administered monthly for five doses, beginning during spring and summer in the first year of life, should be considered in children with HD who exhibit any of the following severity predictors: heart failure, pulmonary hypertension, and severe airway abnormalities associated with CHD.

All authors contributed to the concept of study, acquisition and analysis of data. SH Lee and KL Hon wrote the article and had 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.

We thank Ms Kathy YC Tsang of Prince of Wales Hospital who provided expert and continuous statistical support for this study.

All authors have disclosed no conflicts of interest.

The results from this research have been presented, in part, at the following conferences:
1. The first phase of this research involving two hospitals on the impact of respiratory syncytial virus (RSV) infection in children with heart disease was presented in the 2nd Asian RSV Expert Forum, Singapore, 24 October 2016.
2. The preliminary data of this research involving four hospitals on the epidemiology and impact of RSV infection in children with heart disease were presented in: (a) "RSV infection in Hong Kong Children" forum organised by the Hong Kong Society of Paediatric Cardiology, 17 January 2017 and (b) the Working Group on Use of Palivizumab for RSV Infection Prophylaxis in Children under Hospital Authority Paediatric COC, 30 August 2017.
3. A modified abstract was selected as poster presentation in the Joint Annual Scientific Meeting of Hong Kong College of Paediatricians, 28 September 2019 (with prior permission by Editor-in-Chief of the Hong Kong Medical Journal).

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

This study was approved by the Research Ethics Committees of the Hospital Authority of Hong Kong (Ref KC/KE-16-0122/ER-2). There is no ethical concern and waiver for obtaining consent was approved by the Cluster Research Ethics Committees of the Hospital Authority of Hong Kong. All procedures performed in this study involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

1. Nair H, Nokes DJ, Gessner BD, et al. Global burden of acute lower respiratory infections due to respiratory syncytial virus in young children: a systematic review and meta-analysis. Lancet 2010;375:1545-55. Crossref

2. Bont L, Checchia PA, Fauroux B, et al. Defining the epidemiology and burden of severe respiratory syncytial virus infection among infants and children in Western countries. Infect Dis Ther 2016;5:271-98. Crossref

3. MacDonald NE, Hall CB, Suffin SC, Alexson C, Harris PJ, Manning JA. Respiratory syncytial viral infection in infants with congenital heart disease. N Engl J Med 1982;307:397-400. Crossref

4. Welliver RC. Review of epidemiology and clinical risk factors for severe respiratory syncytial virus (RSV) infection. J Pediatr 2003;143(5 Suppl):S112-7. Crossref

5. American Academy of Pediatrics: Committee on Infectious Diseases and Committee on Fetus and Newborn. Prevention of respiratory syncytial virus infections: indications for the use of palivizumab and update on the use of RSV-IGIV. Pediatrics 1998;102:1211-6. Crossref

6. Simoes EA. Immunoprophylaxis of respiratory syncytial virus: global experience. Respir Res 2002;3 Suppl 1:S26-33. Crossref

7. American Academy of Pediatrics: Committee on Infectious Diseases and Committee on Fetus and Newborn. Policy Statement: Revised indications for the use of palivizumab and respiratory syncytial virus immune globulin intravenous for the prevention of respiratory syncytial virus infections. Pediatrics 2003;112(6 Pt 1):1442-6. Crossref

8. Nakazawa M, Saji T, Ichida F, Oyama K, Harada K, Kusuda S. Guidelines for the use of palivizumab in infants and young children with congenital heart disease. Pediatr Int 2006;48:190-3. Crossref

9. Kim NK, Choi JY. Respiratory syncytial virus prevention in children with congenital heart disease: who and how? Korean J Pediatr 2011;54:197-200. Crossref

10. Chi H, Hsu CH, Chang JH, et al. A novel six consecutive monthly doses of palivizumab prophylaxis protocol for the prevention of respiratory syncytial virus infection in high-risk preterm infants in Taiwan. PLoS One 2014;9:e100981. Crossref

11. American Academy of Pediatrics: Committee on Infectious Diseases; American Academy of Pediatrics Bronchiolitis Guidelines Committee. Updated guidance for palivizumab prophylaxis among infants and young children at increased risk of hospitalization for respiratory syncytial virus infection. Pediatrics 2014;134:415-20. Crossref

12. Lee SR, Kwok KL, Ng DK, Hon KL. Palivizumab for infants <29 weeks in Hong Kong without a clear-cut season for respiratory syncytial virus infection—a cost-effectiveness analysis. J Trop Pediatr 2018;64:418-25. Crossref

13. Chan PK, Sung RY, Fung KS, et al. Epidemiology of respiratory syncytial virus infection among paediatric patients in Hong Kong: seasonality and disease impact. Epidemiol Infect 1999;123:257-62. Crossref

14. Hon KL, Leung TF, Cheng WY, et al. Respiratory syncytial virus morbidity, premorbid factors, seasonality, and implications for prophylaxis. J Crit Care 2012;27:464-8. Crossref

15. Leung TF, Lam DS, Miu TY, et al. Epidemiology and risk factors for severe respiratory syncytial virus infections requiring pediatric intensive care admission in Hong Kong children. Infection 2014;42:343-50. Crossref

16. Chan PK, Tam WW, Lee TC, et al. Hospitalization incidence, mortality, and seasonality of common respiratory viruses over a period of 15 years in a developed subtropical city. Medicine (Baltimore) 2015;94:e2024. Crossref

17. Hall CB, Powell KR, MacDonald NE, et al. Respiratory syncytial viral infection in children with compromised immune function. N Engl J Med 1986;315:77-81. Crossref

18. Groothuis JR, Gutierrez KM, Lauer BA. Respiratory syncytial virus infection in children with bronchopulmonary dysplasia. Pediatrics 1988;82:199-203.

19. Wilkesmann A, Ammann RA, Schildgen O, et al. Hospitalized children with respiratory syncytial virus infection and neuromuscular impairment face an increased risk of a complicated course. Pediatr Infect Dis J 2007;26:485-91. Crossref

20. Simoes EA. Environmental and demographic risk factors for respiratory syncytial virus lower respiratory tract disease. J Pediatr 2003;143(5 Suppl):S118-26. Crossref

21. Purcell K, Fergie J. Driscoll Children’s Hospital respiratory syncytial virus database: risk factors, treatment and hospital course in 3308 infants and young children, 1991 to 2002. Pediatr Infect Dis J 2004;23:418-23. Crossref

22. Lee JT, Chang LY, Wang LC, et al. Epidemiology of respiratory syncytial virus infection in northern Taiwan, 2001-2005—seasonality, clinical characteristics, and disease burden. J Microbiol Immunol Infect 2007;40:293-301.

23. Khor CS, Sam IC, Hooi PS, Quek KF, Chan YF. Epidemiology and seasonality of respiratory viral infections in hospitalized children in Kuala Lumpur, Malaysia: a retrospective study of 27 years. BMC Pediatr 2012;12:32. Crossref

24. Chew FT, Doraisingham S, Ling AE, Kumarasinghe G, Lee BW. Seasonal trends of viral respiratory tract infections in the tropics. Epidemiol Infect 1998;121:121-8. Crossref

25. Welliver R. The relationship of meteorological conditions to the epidemic activity of respiratory syncytial virus. Paediatr Respir Rev 2009;10 Suppl 1:6-8. Crossref

26. Kristensen K, Stensballe LG, Bjerre J, et al. Risk factors for respiratory syncytial virus hospitalisation in children with heart disease. Arch Dis Child 2009;94:785-9. Crossref

27. Moler FW, Khan AS, Meliones JN, Custer JR, Palmisano J, Shope TC. Respiratory syncytial virus morbidity and mortality estimates in congenital heart disease patients: a recent experience. Crit Care Med 1992;20:1406-13. Crossref

28. Wang EE, Law BJ, Stephens D. Pediatric Investigators Collaborative Network on Infections in Canada (PICNIC): prospective study of risk factors and outcomes in patients hospitalized with respiratory syncytial viral lower respiratory tract infection. J Pediatr 1995;126:212-9. Crossref

29. Jung JW. Respiratory syncytial virus infection in children with congenital heart disease: global data and interim results of Korean RSV-CHD survey. Korean J Pediatr 2011;54:192-6. Crossref

30. Chang RK, Chen AY. Impact of palivizumab on RSV hospitalizations for children with hemodynamically significant congenital heart disease. Pediatr Cardiol 2010;31:90-5. Crossref

31. Chu PY, Hornik CP, Li JS, Campbell MJ, Hill KD. Respiratory syncytial virus hospitalisation trends in children with haemodynamically significant heart disease, 1997-2012. Cardiol Young 2017;27:16-25. Crossref

32. Chiu SN, Shao PL, Chen HC, et al. Risk of respiratory syncytial virus infection in cyanotic congenital heart disease in a subtropical area. J Pediatr 2016;171:25-30.e1. Crossref

33. The IMpact-RSV Study Group. Palivizumab, a humanized respiratory syncytial virus monoclonal antibody, reduces hospitalization from respiratory syncytial virus infection in high-risk infants. Pediatrics 1998;102(3 Pt 1):531-7. Crossref

34. Feltes TF, Cabalka AK, Meissner HC, et al. Palivizumab prophylaxis reduces hospitalization due to respiratory syncytial virus in young children with hemodynamically significant congenital heart disease. J Pediatr 2003;143:532-40. Crossref

35. UK Government. The Green Book. Immunisation against infectious disease. Available from https://webarchive. nationalarchives.gov.uk/20130104181824/https://www.wp.dh.gov.uk/immunisation/files/2012/09/Green-Book-updated-040113.pdf. Accessed 17 Jul 2019.

36. Tulloh RM, Medrano-Lopez C, Checchia PA, et al. CHD and respiratory syncytial virus: global expert exchange recommendations. Cardiol Young 2017;27:1504-21. Crossref

37. Manzoni P, Paes B, Lanctôt KL, et al. Outcomes of infants receiving palivizumab prophylaxis for respiratory syncytial virus in Canada and Italy: An international, prospective cohort study. Pediatr Infect Dis J 2017;36:2-8. Crossref

38. Cohen SA, Zanni R, Cohen A, et al. Palivizumab use in subjects with congenital heart disease: results from the 2000-2004 Palivizumab Outcomes Registry. Pediatr Cardiol 2008;29:382-7. Crossref

39. Resch B, Michel-Behnke I. Respiratory syncytial virus infections in infants and children with congenital heart disease: update on the evidence of prevention with palivizumab. Curr Opin Cardiol 2013;28:85-91. Crossref

40. Wegzyn C, Toh LK, Notario G, et al. Safety and effectiveness of palivizumab in children at high risk of serious disease due to respiratory syncytial virus infection: a systematic review. Infect Dis Ther 2014;3:133-58. Crossref

Reference intervals (RIs) are an indispensable tool for clinical decision making in the interpretation of numerical pathology results. Simple yet elegant comparisons with reference subjects empower the interpreter with objective judgements and aid clinicians in diagnosis, treatment, monitoring, prognostication, and screening.1

Reference intervals are commonly defined as limiting values, usually upper and lower limits, between which a prespecified percentage (usually 95%) of results would fall.2 3 In daily practice, for most tests, there exists some degree of laboratory-specific bias related to differences in pre-analytical and analytical factors, such as the choices of analytical platform, methodology, and reagent. Therefore, it is desirable for laboratories to provide sets of laboratory-specific RIs following Clinical and Laboratory Standards Institute guideline C28-A3c. A laboratory may establish a new set of RIs by conducting an RI study with at least 120 reference individuals from each subgroup stratified by sex, age, and other parameters as appropriate.2 Conducting an RI study is challenging, as enormous efforts of human and financial resources are needed. As the list of analytes is long, it is almost impossible for every laboratory to repeat an RI study to accommodate future changes in methodology or analytical platforms.2 4 Alternatively, a laboratory may adopt the RIs established by other sources such as manufacturers or the literature and validate them with at least 20 reference individuals’ results. An additional option is for the laboratory to transfer previously established RIs according to mathematical formulas to account for differences in analytical factors.2 These methods ensure that each laboratory provides a set of clinically meaningful intervals to clinicians, aiding their management.

Therefore, for the same analyte, it is not uncommon to see different RIs across laboratories. This inter-laboratory coefficient of variation was reported by Ceriotti et al3 to be as high as 15% to 20% for the RIs of urea and creatinine. This could be reasonable for hormonal tests that are not optimally standardised, as demonstrated by the marked variations in RIs for thyroid hormones between four analytical platforms shown by a recent study in the UK.5 For analytes that are generally well standardised across platforms, such as plasma electrolytes, one would expect results generated by different laboratories to be comparable. Logically, with insignificant methodological bias, the RIs should be same for the specified homogenous population.

In 2007, the UK Pathology Harmony Group showed that laboratories were using different sets of Rls with no sound scientific basis despite using the same analytical platform and reagents.6 7 The same problem was later also revealed by a survey on RIs in Australasia.8 The differences in RIs were concluded to be unnecessary and would have created unneeded confusion during interpretation, which might lead to inappropriate investigations or treatments.9 10 Common RIs were offered as a solution to unite laboratory practices.4

In Hong Kong, we have observed a general trend of variation in RIs that resembles those in the UK and Australasia, with various RIs adopted for most tests, including general chemistry laboratory tests. Hence, we conducted the first local study to explore the situation with a territory-wide survey on RIs. The aim was to scientifically review the analytical variation of general chemistry laboratory tests between local laboratories and to examine the evidence for such variations.

Fourteen blood general chemistry analytes were included in this study, namely albumin, alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), total bilirubin, calcium, chloride, creatinine, gamma-glutamyl transferase (GGT), phosphate, potassium, sodium, total protein, and urea. We conducted a territory-wide survey involving 10 chemical pathology laboratories. All laboratories provided routine services to assess the 14 analytes, except for AST, chloride, and GGT, which were not evaluated in three laboratories. The instruments were Abbott Architect (labs 1-3), Beckman Coulter AU (labs 4-5), Roche Cobas (labs 6-9), and Siemens Dimension EXL (lab 10). Table 1 summarises the analytical platforms and methodologies.

The laboratories participated in the Condensed General Chemistry Programme provided by the Royal College of Pathologists of Australasia Quality Assurance Programs. In each cycle of the external quality assurance programme (EQAP), identical sets of QAP materials were analysed by each individual laboratory for the aforementioned blood general chemistry analytes using their own analytical platform. The use of QAP materials, which were commutable samples with the same properties as routinely analysed clinical samples, minimises the matrix effect. In routine clinical practice, EQAP safeguards laboratory performance by comparison with peers and reference methods. In the present study, we retrospectively review these readily available EQAP data from local laboratories for bias assessments. The participants provided their responses by email to the following items: (1) historical EQAP results of six cycles (105-11, 105-12, 105-15, 105-16, 106-03, and 106-04); (2) adult RIs in use for clinical service, and (3) analytical specification of assays.

Laboratory performance bias was assessed by percentage differences of EQAP results. Percentage difference was defined as the laboratory result minus the target value divided by the target value. The feasibility of applying common RIs among the laboratories was determined by the degree of agreement between the percentage differences and the corresponding allowable limits of performance.11 Data analyses were performed using Microsoft Excel 2016.

Figure 1 shows that half of the 14 analytes showed agreement across all laboratories. The inter-laboratory differences are within the corresponding target allowable limit of error (ALE) for AST (-3% to +9%; target ALE ±12%), chloride (-1% to +2%; ±3%), phosphate (-1% to 4%; ±8%), potassium (-2% to 3%; ±5%), sodium (-1% to 2%; ±2%). Three other analytes (albumin, ALT, and GGT) also showed agreement across nine laboratories with the Abbott, Beckman, and Roche platforms, except Siemens which was only used by one laboratory.

Figure 2 shows the inter-laboratory comparison of RIs for the 14 general chemistry analytes. Laboratories using the same platform generally adopted the same RIs, except for one laboratory using the Roche platform.

Notably, for the seven analytes mentioned above that showed agreement within the target ALE, the RIs differed substantially across the 10 laboratories. Particularly, the upper RI limit ranged from 34 to 50 U/L (coefficient of variation [CV]: 11%): in male samples and 30 to 40 U/L (9%) in female samples in AST; 107 to 109 mmol/L (0.9%) in chloride; 1.39 to 1.52 mmol/L (2.7%) in phosphate; 4.4 to 5.2 mmol/L (6.7%) in potassium; 144 to 148 mmol/L (0.7%) in sodium; 79 to 87 g/L (2.2%) in total protein; and 6.3 to 8.1 mmol/L (8.1%) in urea. The lower RIs ranged from 98 to 102 mmol/L (1.7%) in chloride; 0.72 to 0.88 mmol/L (6.2%) in phosphate; 3.4 to 3.6 mmol/L (2.6%) in potassium; 136 to 137 mmol/L (0.2%) in sodium; 63 to 68 g/L (2.2%) in total protein; and 2.5 to 3.1 mmol/L (7.4%) in urea.

The remaining analytes (albumin, ALT, ALP, calcium, creatinine, GGT, and total bilirubin) demonstrated substantial platform-specific bias exceeding the target ALE. High bias exceeding the ALE was observed for ALT (+12% to +20%; target ALE ±12%) and GGT (+11% to +14%; ±12%), with negative bias exceeding the ALE in albumin (-5.3% to -7.1%; ±6%), ALP (-11.4% to -15.3%; ±12%), and calcium (-5.6% to -7.1%; ±4%) present on the Siemens platform. Negative bias exceeding the ALE in ALP (-12.2% to -14.8%; ±12%) was also detected on the Roche platform. For calcium, negative bias exceeding the ALE (-4% to -6%; ±4%) was also detected at one laboratory using the Beckman platform. For creatinine, all laboratories were in agreement about concentrations ranging from 152 to 349 μmol/L. However, significant negative bias (-13% to -22%; ±12%) was observed for creatinine levels at the target value of 67 μmol/L on the Abbott, Siemens, and Roche instruments. For total bilirubin, half of the laboratories showed agreement within the ALE, while the remaining laboratories had significant negative bias (-14% to 17%; ±12%).

The investigated laboratories used different RIs despite employing the same analytical platforms, methods and reagents, for 11 out of the 14 analytes among those using the Abbott platform (labs 1-3), 11 out of 14 of analytes among those using Roche platforms (labs 6-9), and three out of the 14 of analytes among those using the Beckman platforms (labs 4-5).

Sex-specific RIs were not consistently provided for eight analytes (ALP, ALT, AST, phosphate, potassium, total bilirubin, total protein, and urea). For instance, sex-specific RIs were not provided by two laboratories for ALP, two for ALT, two for AST, five for potassium, five for urea, seven for total protein, eight for phosphate, and nine for total bilirubin.

Reference intervals are provided by laboratories as interpretative tools to aid clinical decision making. Theoretically, RIs could be affected by patient factors (eg, sex, age, ethnicity, biological variability), pre-analytical and analytical factors (eg, choice of method, reagents, platform, calibration), and statistical methodology.12 Therefore, for the same population, the RIs used for a test are inevitably influenced by the bias of the laboratory assays. In other words, RIs should theoretically be the same if the above-listed factors do not introduce significant bias.

In local practice, 10 analytes surveyed demonstrated sufficient agreement within the ALE between different analytical platforms across laboratories (Fig 1: AST, chloride, phosphate, potassium, sodium, total protein, and urea for all four platforms; albumin, ALT and GGT for Abbott, Beckman, and Roche platforms) [Fig 1]. These results confirmed the previous findings of bias assessment by the Australasian Association of Clinical Biochemists, which concluded that chloride, phosphate, potassium, sodium, total protein, and urea measurements showed sufficient similarity across analytical platforms and laboratories and that common RIs could be adopted.10 The same study found method-specific bias in AST levels averaging +22% for assays using pyridoxal-5-phosphate as an activator compared with those not using pyridoxal-5-phosphate.10 Our results showed a lesser degree of pyridoxal-5-phosphate–related bias (+7%), so this issue would not prevent the use of common RIs in the local scenario.

For analytes with demonstrated agreement across platforms and laboratories, the RIs are theoretically expected to be the same if obtained from the same group of reference (ie, ‘healthy’) individuals. In actual practice, for the seven analytes mentioned above, all of the adult RIs varied across laboratories, with the CV of the upper and lower limits of the RIs up to 11% and 7.4%, respectively. The inter-laboratory percentage differences of upper RI limits were up to 47% for AST (absolute difference: 16 U/L), 29% for urea (1.8 mmol/L), and 18% for potassium (0.8 mmol/L), and those of the lower RI limits were up to 24% for urea (0.6 mmol/L), 22% for phosphate (0.16 mmol/L), and 8% for total protein (5 g/L). We can compare the CVs of these analytes’ RIs against the corresponding between-subject biological variation (CV-G) values published by Ricos et al.13 The CV of the upper RI limits of potassium and sodium were 1.2 and 1.0 times those of CV-G, respectively while that of the lower RI limits of sodium and phosphate were 1.1 and 0.6 times those of CV-G, respectively. These RI variations generate significant additional bias during interpretation, which is often overlooked and unchecked. Furthermore, laboratories were using different RIs despite using the same analytical platforms and methodologies for these analytes. For example, among users of the Abbott platform, the potassium RIs of labs 1 and 2 were 3.6 to 5.2 mmol/L for samples of both sexes, while that of lab 3 was 3.5 to 4.5 mmol/L for male and 3.4 to 4.4 mmol/L for female samples. These variations were unnecessary, as supported by the sufficient agreement across analytical platforms and laboratories. Application of different RIs in various circumstances could lead to confusion among interpreters and hinder data management in the era of electronic health records.4 14 Similar trends of unexplained RI variations were previously observed in the UK for sodium, potassium, and other analytes, and this eventually lead to the Pathology Harmony group’s recommendation of harmonised RIs.7 At present, local laboratories often spend substantial human resources on decisions and maintenance regarding the appropriate RIs for large numbers of analytes. The use of common RIs for these seven analytes would unite local laboratory practices, facilitate electronic communications between laboratory information and electronic patient record systems, and streamline the maintenance of RIs.

For creatinine, low bias was noted for seven laboratories using the Jaffe methods, but this tendency spared the laboratories that used the enzymatic method on the Beckman platform. This bias was likely related to the high variability of the Jaffe method at low creatinine concentrations, which has been reported to be up to 30% on some platforms.15 While the remarkably good analytical agreement shown for the remaining five higher concentrations of creatinine support the use of common RIs, this should be cautiously reviewed, as the lowest concentration of creatinine (67 μmol/L) is very close to the lower RI limit. Further study of bias may be warranted for creatinine.

Substantial bias exceeding the ALE was demonstrated for the remaining six analytes, with high bias for ALT and GGT and low bias for albumin, ALP, and calcium on the Siemens platform; low bias for ALP on the Roche platform; low bias for calcium at one laboratory using the Beckman platform; and low bias for total bilirubin in labs 1 to 3, 7, and 8. Positive bias averaging 8% for albumin was observed for laboratories using the bromocresol green method compared with the bromocresol purple method, a pattern similar to the findings of Koerbin et al.10 16 Bias in ALT measurement could be attributed to differences in assay design,10 with an average of +7% bias shown for the assay using pyridoxal-5-phosphate over the assay that does not use it. Bias for calcium and total bilirubin could be related to methodological differences between platforms, while bias for ALP and GGT were likely to be specific to the analytical platform. While the feasibility of local common RIs for these six analytes was not confirmed by this study, our findings indicate that common RIs could still be considered for albumin, ALT, and GGT in laboratories using the Abbott, Beckman, and Roche platforms, which all laboratories except one use.

Variable adoptions of sex-specific RIs were another key finding of the survey. Heterogeneous and inconsistent practices of sex partitioning for RIs were noted in eight analytes (ALP, ALT, AST, phosphate, potassium, total bilirubin, total protein, and urea). Moreover, sex-specific RIs were sometimes different even within the same platform. For example, the upper RI limit of GGT in male samples differed by 35 U/L among users of the Roche platform, and the upper RI limit of ALP differed by 40 U/L and 7 U/L in male and female samples, respectively, among users of the Abbott platform. Common RIs with united practice of sex partitioning could be the solution to converge these practices.

Historically, heterogeneous and sometimes incomparable results of the same measurands could be obtained with different assays because of suboptimal standardisations in pre-analytical and analytical factors. Laboratory-specific RIs were advocated to compensate and allow for sound interpretations of laboratory results in clinical settings.17 Realising the need for assay standardisation, an enormous global effort has been taken in the past 60 years to study biological variability, standardise pre-analytical conditions and analytical methods, improve quality control, establish traceability of reference materials and methods, and implement EQAPs for various kinds of assays, led by the International Federation of Clinical Chemistry (IFCC) and other international/national organisations.18 Major successes have been realised for a large number of measurands, as listed on the website of the International Consortium for Harmonization of Clinical Laboratory Results.19

The concept of common RIs emerged in the early 2000s and has gained huge popularity over the past decade.4 The theory is simple: if the measured results of different assays are comparable, ie with adequate assay standardisation, the same RIs should be adopted given that the tests are performed on the same reference population.17 Redundant variations of RIs merely impair interpretation.

Presently, there are two types of common RIs: ‘objective’ and ‘subjective’ ones.20 Subjective common RIs were generally defined by scientific surveys and expert guidance with the harmonisation approach. Examples include the “agreed Pathology Harmony clinical biochemistry reference intervals for adults” for 15 general chemistry analytes recommended by the UK Pathology Harmony Group in 201121 and the “adult harmonised reference intervals” for 18 general chemistry analytes recommended by the Australasian Association of Clinical Biochemists and endorsed by the Royal College of Pathologists of Australasia in 2016.16 22 23 The two groups have since continued their work on harmonisation of various aspects of pathology in the past decade, with the UK Pathology Harmony Group working on the Pathology Harmony bookmark for tumour markers and requesting guidance for non-specialists, and the Australasian Association of Clinical Biochemists working on harmonisation of paediatric common RIs, serum protein electrophoresis reporting, lipid reporting, management and communication of high-risk lab results, arterial and venous blood gas RIs, and reporting of dynamic endocrine testing for adults and paediatric patients.6 7 8 16 24 25 26

Objective common RIs refer to those defined by well-conducted, multicentre reference studies, such as the Nordic Trueness Project, which was conducted with well-standardised pre-analytical and analytical handlings and the use of five control materials. The project involved 102 Nordic routine clinical biochemistry laboratories and more than 2500 carefully selected healthy reference individuals.27 The Nordic Reference Interval Project RIs for 25 general chemistry analytes were established and published in 2002 and implemented throughout Nordic countries in 2004 with the help of the Scandinavian Society of Clinical Chemistry.27 28 29 30 Among Asian countries, the Japan Society of Clinical Chemistry has recently published their nationwide common RIs for 40 laboratory tests determined by three multicentre RI studies.31 Table 2 8 21 23 28 31 summarises the common RIs published in different parts of world for the general chemistry analytes surveyed and the common RIs proposed by our study.

In 2017, the IFCC Committee on Reference Intervals and Decision Limits (C-RIDL) published two landmark papers on the results of their global multicentre study on reference values of 25 chemistry analytes in 13 386 healthy adults recruited from 12 countries, including China,32 with the use of a specially designed serum panel.33 34 The study explored the regionality and ethnicity of these reference values globally and provided invaluable information for the possibility of future derivation and transference of the established RIs through use of the C-RIDL serum panel.34

The relatively small number and choice of QAP specimens for retrospective methodological comparisons represent a major limitation of our survey. Artificial materials used in QAP specimens generally gave rise to more variable and method-dependent results due to matrix effects.9 Despite this, our survey demonstrated that methodological bias would not prevent the use of common RIs for seven general chemistry analytes. For the remaining analytes, we speculate that the degree of methodological bias may be exaggerated by the matrix effect of the QAP, ie, the actual analytical difference is likely to be smaller when tested with a patient sample. Our findings should be verified with a formal prospective bias study with a standardised protocol and the use of another set of blood specimens, preferably unadulterated human samples, with pre-assigned reference values to ensure commutability.

This survey compared the adult RIs of 14 general chemistry analytes among 10 chemical pathology laboratories using four different analytical platforms. Bias assessments and comparisons of RIs revealed that different and variable RIs were provided by the laboratories despite sufficient inter-laboratory and inter-platform agreement regarding the RIs of 10 general chemistry analytes. The use of common RIs was found to be feasible and is recommended for these 10 analytes. Such use would unify and improve local standards of clinical laboratory practice. A well-designed implementation plan for common RIs with support from stakeholders including clinicians, pathologists, and scientists would be vital for the success of such a substantial project. Figure 3 shows our proposed implementation plan for the introduction of common RIs in Hong Kong, modified from the plan suggested by Tate et al8 for the harmonisation of adult and paediatric RIs in Australasia. Furthermore, the concept of common RIs could be expanded to cover more general chemistry analytes, eg, creatine kinase and magnesium; special chemical tests, eg, therapeutic drug monitoring and hormones; other clinical laboratory specialties, such as haematology and immunology; and paediatric RIs.6 7 8

All authors contributed to the concept or design, drafting of the article, 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.

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

The present survey is a retrospective observational review of local laboratory practice and external quality assurance program data with no patient participation, no access to private or sensitive patient data, no collection or analysis of human body fluid or tissue. The quality assurance program materials used for the data collection in the survey are processed samples designed by the external quality assurance program organiser to mimic the properties of clinical sample. Therefore, ethics approval was not applicable for this study.

1. Plebani M, Lippi G. Reference values and the journal: why the past is now present. Clin Chem Lab Med 2012;50:761-3. Crossref

2. Defining, Establishing, and Verifying Reference Intervals in the Clinical Laboratory; Approved Guideline—Third Edition. Clinical and Laboratory Standards Institute; 2010.

3. Ceriotti F, Hinzmann R, Panteghini M. Reference intervals: the way forward. Ann Clin Biochem 2009;46:8-17. Crossref

4. Jones GR, Barker A, Tate J, Lim CF, Robertson K. The case for common reference intervals. Clin Biochem Rev 2004;25:99-104.

5. Barth JH, Luvai A, Jassam N, et al. Comparison of method-related reference intervals for thyroid hormones: studies from a prospective reference population and a literature review. Ann Clin Biochem 2018;55:107-12. Crossref

6. Berg J. The approach to pathology harmony in the UK. Clin Biochem Rev 2012;33:89-93.

7. Berg J. The UK Pathology Harmony initiative; The foundation of a global model. Clin Chim Acta 2014;432:22-6. Crossref

8. Tate JR, Sikaris KA, Jones GR, et al. Harmonising adult and paediatric reference intervals in Australia and New Zealand: an evidence-based approach for establishing a first panel of chemistry analytes. Clin Biochem Rev 2014;35:213-35.

9. Jones G, Barker A. Standardisation of reference intervals: an Australasian view. Clin Biochem Rev 2007;28:169-73.

10. Koerbin G, Tate JR, Ryan J, et al. Bias assessment of general chemistry analytes using commutable samples. Clin Biochem Rev 2014;35:203-11.

11. Jones GR, Sikaris K, Gill J. ‘Allowable limits of performance’ for external quality assurance programs—an approach to application of the Stockholm Criteria by the RCPA Quality Assurance Programs. Clin Biochem Rev 2012;33:133-9.

12. Ozarda Y. Reference intervals: current status, recent developments and future considerations. Biochem Med (Zagreb) 2016;26:5-16. Crossref

13. Ricós C, Alvarez V, Cava F, et al. Current databases on biological variation: pros, cons and progress. Scand J Clin Lab Invest 1999;59:491-500. Crossref

14. Koerbin G, Sikaris KA, Jones GR, et al. Evidence-based approach to harmonised reference intervals. Clin Chim Acta 2014;432:99-107. Crossref

15. Hoste L, Deiteren K, Pottel H, Callewaert N, Martens F. Routine serum creatinine measurements: how well do we perform? BMC Nephrol 2015;16:21. Crossref

16. Koerbin G, Sikaris K, Jones GR, et al. An update report on the harmonization of adult reference intervals in Australasia. Clin Chem Lab Med 2018;57:38-41. Crossref

17. Ceriotti F. Prerequisites for use of common reference intervals. Clin Biochem Rev 2007;28:115-21.

18. Siest G, Henny J, Gräsbeck R, et al. The theory of reference values: an unfinished symphony. Clin Chem Lab Med 2013;51:47-64. Crossref

19. International Consortium for Harmonization of Clinical Laboratory Results. Summary of measurand harmonization activities. Available from: www.harmonization.net/measurands/. Accessed 27 Mar 2019.

20. Ozarda Y, Sikaris K, Streichert T, Macri J; IFCC Committee on Reference Intervals and Decision Limits (C-RIDL). Distinguishing reference intervals and clinical decision limits—a review by the IFCC Committee on Reference Intervals and Decision Limits. Crit Rev Clin Lab Sci 2018;55:420-31. Crossref

21. Berg J, Lane V. Pathology Harmony; a pragmatic and scientific approach to unfounded variation in the clinical laboratory. Ann Clin Biochem 2011;48(Pt 3):195-7. Crossref

22. Jones GR, Koetsier S. Uptake of recommended common reference intervals for chemical pathology in Australia. Ann Clin Biochem 2017;54:395-7. Crossref

23. Koerbin G, Tate JR. Harmonising adult reference intervals in Australia and New Zealand—the continuing story. Clin Biochem Rev 2016;37:121-9.

24. Campbell C, Caldwell G, Coates P, et al. Consensus statement for the management and communication of high risk laboratory results. Clin Biochem Rev 2015;36:97-105.

25. Tate J, Caldwell G, Daly J, et al. Recommendations for standardized reporting of protein electrophoresis in Australia and New Zealand. Ann Clin Biochem 2012;49:242-56. Crossref

26. Australasian Association of Clinical Biochemists. Harmonisation: guideline, publications, history, committee. Available from: www.aacb.asn.au/aboutus/harmonisation. Accessed 27 Mar 2019.

27. Pedersen MM, Ornemark U, Rustad P, et al. The Nordic Trueness Project 2002: use of reference measurement procedure values in a general clinical chemistry survey. Scand J Clin Lab Invest 2004;64:309-20. Crossref

28. Rustad P, Felding P, Franzson L, et al. The Nordic Reference Interval Project 2000: recommended reference intervals for 25 common biochemical properties. Scand J Clin Lab Invest 2004;64:271-84. Crossref

29. Strømme JH, Rustad P, Steensland H, Theodorsen L, Urdal P. Reference intervals for eight enzymes in blood of adult females and males measured in accordance with the International Federation of Clinical Chemistry reference system at 37 degrees C: part of the Nordic Reference Interval Project. Scand J Clin Lab Invest 2004;64:371-84. Crossref

30. Rustad P, Felding P, Lahti A, Hyltoft Petersen P. Descriptive analytical data and consequences for calculation of common reference intervals in the Nordic Reference Interval Project 2000. Scand J Clin Lab Invest 2004;64:343-70. Crossref

31. Ichihara K, Yomamoto Y, Hotta T, et al. Collaborative derivation of reference intervals for major clinical laboratory tests in Japan. Ann Clin Biochem 2016;53:347-56. Crossref

32. Xia L, Chen M, Liu M, et al. Nationwide multicenter reference interval study for 28 common biochemical analytes in China. Medicine (Baltimore) 2016;95:e2915. Crossref

33. Ichihara K, Ozarda Y, Barth JH, et al. A global multicenter study on reference values: 1. Assessment of methods for derivation and comparison of reference intervals. Clin Chim Acta 2017;467:70-82. Crossref

34. Ichihara K, Ozarda Y, Barth JH, et al. A global multicenter study on reference values: 2. Exploration of sources of variation across the countries. Clin Chim Acta 2017;467:83-97. Crossref

Obstetric anal sphincter injuries (OASIS) is a serious complication of vaginal delivery that is associated with an increased risk of anal incontinence (complaint of involuntary loss of faeces or flatus).1 The incidence of OASIS is reportedly much lower in Hong Kong (0.32%) than in other countries, such as the United Kingdom, Norway, and Sweden (2.9%-4.2%).2 3 4 5 This could be affected by a number of factors. First, delivery practices in Hong Kong are quite different from elsewhere in the world, such that they include the use of a hands-on approach to protect the perineum and liberal use of episiotomy.6 The episiotomy rates are reportedly high in Hong Kong: 83.7% for primiparous women and 54.8% for multiparous women.5 Moreover, in Hong Kong, a left mediolateral episiotomy is used, whereas midline episiotomy or right mediolateral episiotomy are used in many other parts of the world.7 Second, there may be ethnic differences in pelvic floor biometry. In particular, Chinese women have a smaller hiatal dimension and reduced pelvic organ mobility.8 It is unclear how these differences in practice and pelvic floor biometry influence the incidence of OASIS.

Importantly, it is also possible that the reduced incidence of OASIS in Hong Kong is a result of underdetection. In a recent local prospective observational study, women were assessed by a single experienced clinician via rectal examination after either normal or instrumental vaginal delivery; the results of that study showed that the incidence of OASIS in primiparous Asian women in Hong Kong was 10%,6 which suggests that the OASIS rate might be higher than previously published. Obstetric anal sphincter injuries that are identified after an extended interval (such as during postnatal follow-up) is regarded as occult OASIS. There is limited information in the literature regarding occult OASIS; thus far, studies have been conducted in the United Kingdom and Australia.9 10

The use of endoanal ultrasound (US) may facilitate identification of OASIS.11 Endoanal US comprises a non-invasive assessment modality and is regarded as the gold standard in studies of anal sphincter injury.9 11 Moreover, all cases of clinically identified OASIS can also be identified on endoanal US.9 The aim of this study was to determine the prevalence of OASIS in primiparous women after normal vaginal delivery or instrumental delivery using endoanal US during postnatal follow-up. Understanding the prevalence and detection rates of OASIS can help inform training policies for midwives and doctors on the awareness and detection of OASIS.

This was a retrospective analysis of archived US volumes from two previously published studies that were performed at a tertiary university hospital in Hong Kong. The initial study recruited 442 nulliparous women in the first trimester, during the period from August 2009 to September 2010.12 13 The second study recruited 292 primiparous women at 1 to 3 days after instrumental delivery, during the period from September 2011 to May 2012. None of the women in either study reported symptoms of pelvic floor disorders, including faecal incontinence to solid or loose stool, before pregnancy.14 Details of deliveries, including any occurrence of perineal tearing, were recorded after each delivery. Ethics approval was obtained from The Joint Chinese University of Hong Kong–New Territories East Cluster Clinical Research Ethics Committee (Ref CRE-2013.332). The STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines were followed in the preparation of this report.15

Generally, each woman underwent perineal examination by the attending midwife or doctor who conducted the delivery, immediately after vaginal delivery. This information was immediately recorded in the medical record. Third- or fourth-degree tears were assessed and repaired by a trained obstetrician. The anorectal mucosa was repaired by continuous or interrupted sutures with 3-O Vicryl. Internal anal sphincter tears were repaired separately by interrupted end-to-end sutures with 2-O Vicryl. External anal sphincter (EAS) tears were repaired by overlapping or end-to-end sutures with 2-O Vicryl. Perineal muscles and the vagina were repaired with 2-O Vicryl. The diagnosis and operative record of each woman were immediately entered into the electronic medical record. The degree of perineal tear was defined using Sultan’s classification of perineal trauma.16

During postnatal follow-up (6-12 months after delivery), the urinary, bowel, and prolapse symptoms of each woman, as well as their quality of life, were assessed using the Chinese Pelvic Floor Distress Inventory (PFDI) and Pelvic Floor Impact Questionnaire (PFIQ).17 Assessment of the anal sphincter was performed with endoanal US using a 10-MHz 360-degree rotating probe (Focus 400, BK Medical; Gentofte, Denmark) with the woman in the lithotomy position. Automatic image acquisition was performed with two volumes stored for each woman.

Offline analysis of the endoanal US volumes was performed in 2018 by two experienced obstetricians (OYKW, SSCC) who were blinded to the clinical diagnosis and questionnaire information. An anal sphincter defect was defined as a discontinuity of >30 degrees in endosonographic images of the internal (hypoechoic ring) and/or external (mixed echogenic ring) sphincters.18 A partial-thickness EAS injury was defined as a defect of <50% thickness of the EAS, whereas a defect of >50% of the EAS was regarded as a full-thickness injury. We considered any EAS and/or internal anal sphincter injury to be OASIS. This follows the clinical classification of OASIS by Sultan.16 Each researcher reviewed all endoanal US volumes independently. Any discrepancies were resolved by consensus review of the relevant US volumes.

The PFDI and PFIQ are comprehensive validated instruments which assess the symptoms and impact of pelvic floor disorders.17 In this study, faecal incontinence was defined as an affirmative response to either item 38 (“Do you usually lose stool beyond your control if your stool is well formed?”) or item 39 (“Do you lose stool beyond your control if your stool is loose or liquid?”) of the PFDI. Flatal incontinence was defined as an affirmative response to item 40 (“Do you usually lose gas from the rectum beyond your control?”) of the PFDI.

Data were analysed by SPSS (Window version 22.0; IBM Corp, Armonk [NY], United States). Descriptive analyses were used to study the prevalence of OASIS on endoanal US. Means were compared between groups using the independent-samples t test. Comparisons of frequencies were made using the Chi squared test or Fisher’s exact test, where appropriate. Univariate analysis was performed to evaluate the influence of potential risk factors on OASIS. Differences with P<0.05 were considered to be statistically significant. Power calculations were not performed with regard to this specific research question, as this study comprised a subanalysis of two prior projects, as described earlier in this paper.

A total of 544 women who had vaginal delivery were enrolled in this study; 207 had normal vaginal delivery and 337 had instrumental delivery (285 vacuum extraction, 52 forceps). Ultrasound images were suboptimal for two women who had normal vaginal delivery; these women were excluded from the analysis.

The demographic data and delivery information are shown in Table 1. Left mediolateral episiotomy was performed in 187 (91.2%) women in the normal vaginal delivery and 336 (99.7%) women in the instrumental delivery group. The duration of active second stage was longer in the instrumental delivery group than in the normal vaginal delivery group (62.7 ± 40.9 min vs 27.9 ± 22.4 min, P<0.005), as a prolonged second stage was the most common indication for instrumental delivery in this cohort (48.4%). More women had epidural analgesia in the instrumental delivery group than in the normal vaginal delivery group (15.7% vs 8.8%, P=0.028). There was no significant difference between the normal vaginal delivery and instrumental delivery groups regarding the timing of endoanal US assessment (P=0.22).

The Figure shows endoanal US images of intact anal sphincters, as well as sphincters with different degrees of OASIS. There were discrepancies or uncertainties in the endoanal US analysis of 16 women with respect to the diagnosis of OASIS. The two researchers determined the diagnoses of these women by consensus review; six were diagnosed with OASIS and 10 were regarded as normal.

The prevalence of clinically detected OASIS was 0% in the normal vaginal delivery group and 1.8% (n=6) in the instrumental delivery group. Table 2 shows that the prevalence of OASIS detected by endoanal US was 7.8% (n=16; 95% confidence interval [CI]=4.1%-11.5%) in the normal vaginal delivery group and 5.6% (n=19; 95% CI=3.1%-8.1%) in the instrumental delivery group (P=0.415). Twenty-nine (82.9%) women had OASIS, as detected by endoanal US, that was not diagnosed during clinical assessment immediately after delivery. Therefore, the occult OASIS rate was 7.8% (95% CI=4.1%-11.5%) in the normal vaginal delivery group and 3.8% (95% CI=1.8%-5.8%) in the instrumental delivery group. In addition, 63.6% (n=21) of occult EAS injuries comprised partial-thickness EAS injuries, whereas 36.4% (n=12) comprised full-thickness EAS injuries. When women with OASIS were compared to those without OASIS, increased birth weight was the only delivery factor associated with an increased risk of OASIS (odds ratio [OR]=3.1, 95% CI=1.3%-7.6%, P=0.012) [Table 3].

Overall, nine (1.7%) and 29 (5.4%) women reported faecal incontinence to solid and loose stool, whereas 97 (17.9%) women reported flatal incontinence (Table 4). All affected women reported mild symptoms. Among the women with OASIS, only one (2.9%) with a repaired third degree (3a) tear reported symptoms of both (faecal incontinence to loose stool and flatal incontinence). Three women (10.3%) who had occult injury reported flatal incontinence. There were no associations between the presence of OASIS and faecal incontinence (P=0.71) or between the presence of OASIS and flatal incontinence (P=0.37).

Primiparity has been associated with increased risks of OASIS (ORs of 2.39 and 8.34) in large retrospective studies.19 20 In the present study, which included large number of primiparous women, the findings on endoanal US were compared with women’s reported symptoms of faecal and flatal incontinence. Importantly, there were no associations between faecal or flatal incontinence and the presence of OASIS.

After assessment by endoanal US, the prevalence of OASIS in the normal vaginal delivery group increased from 0% to 7.8% and that in the instrumental delivery group increased from 1.8% to 5.6%. Overall, 82.9% of women with OASIS detected by endoanal US had not been diagnosed with OASIS during clinical assessment immediately after delivery. This finding is consistent with the results of the study by Andrews et al.9 In that study, the prevalence of OASIS markedly increased from 11% to 24.5% when women were re-examined by an experienced research fellow; 87% of OASIS diagnoses were missed by midwives and 28% were missed by junior doctors.9 In our study, normal vaginal deliveries were primarily attended by midwives, whereas instrumental deliveries were performed by residents. The higher rate of occult OASIS in the normal vaginal delivery group suggests that midwives currently receive inadequate training for clinical identification of OASIS. Thus, to improve the detection of OASIS, midwives and doctors should be trained to recognise OASIS by performing a standardised vaginal and rectal examination after delivery.

Compared with previous studies, the rate of OASIS determined by endoanal US in our study (6.5%) was lower than the rate of 10% determined by a single examiner in a prospective observational study conducted in the same unit.6 This could be a result of the small sample size (70 subjects) in the prior study. Furthermore, most patients with OASIS (5/7) in that study were reported to have small 3a tears. There were no 3c or fourth-degree tears in that study. Following the same delivery practices, clinically detected small 3a tears may therefore appear normal in endoanal US. Furthermore, these tears might not result in long-term consequences.6 21

The finding of an overall lower OASIS rate in Hong Kong, compared with that in Asian women who deliver in Caucasian countries, is not new.6 Asian women who deliver in locations with more restrictive policies regarding episiotomy have shown higher rates of OASIS.22 23 24 In a study conducted in the United States, OASIS was found significantly more frequently in Asian women than in women of other ethnicities.23 In Australia, nulliparous women born in South Asia and South-East Asia were 2.6-fold and 2.1-fold more likely to exhibit OASIS than women born in Australia or New Zealand women.24 It is uncertain whether the increased rate of episiotomy might protect against OASIS in Asian women and contribute to the relative reduction in the rate of OASIS in Hong Kong. Thus, our unit is currently conducting a randomised controlled trial to compare restrictive and routine episiotomy. In addition to episiotomy, the delivery technique and hands-on approach might contribute to the relative reduction in the rate of OASIS. All deliveries in our study were conducted with women in a lithotomy position, with their feet on footplates or in stirrups. All midwives and doctors conducting the deliveries used hands-on techniques to protect the perineum in each woman. Either firm pressure or pressure with squeezing of the perineum, also known as the modified Ritgen manoeuvre, was used.6 Warm compresses were not commonly used by midwives and doctors in our study.

The OASIS rate in the normal vaginal delivery group was higher than that in the the instrumental delivery group, but this difference was not statistically significant. The majority of deliveries by women in the instrumental delivery group were performed using vacuum extraction. The rate of OASIS in these women could be similar to that of women in the normal vaginal delivery group. The OASIS rates were similar in women who delivered with the aid of vacuum extraction or with forceps, whereas previous studies showed that forceps delivery was associated with an increased risk of OASIS.19 20 25 The small number of forceps deliveries in this study might have led to insufficient statistical power to detect a difference between the two types of instrumental deliveries. Furthermore, the use of forceps was primarily restricted to patients who were low risk, and mostly comprised outlet/low-cavity forceps deliveries. Previous studies reported that macrosomia, higher birth weight (OR=1.14, 95% CI=1.0-1.3, P=0.039), and shorter perineal length were risk factors for OASIS.6 19 20 The present study had similar findings, in that higher birth weight was a risk factor for OASIS (OR=3.1, 95% CI=1.3-7.6, P=0.012). However, perineal length was not assessed, which is an important limitation of this study.

Flatal incontinence was present in 17.9% of women after delivery, which is comparable to the rate reported in previous studies.26 27 In addition to OASIS, irritable bowel syndrome, high body mass index, and mode of delivery constitute factors associated with flatal incontinence.20 21 Overall, 5.5% of women reported faecal incontinence; most of these women reported faecal incontinence to loose stool and mild symptoms only. Most obstetric anal sphincter injuries were not detected during clinical examination. Shortly after delivery, the presence of OASIS was not associated with symptoms of faecal or flatal incontinence, but a longer-term study is needed to confirm these findings. However, we previously found that only antenatal faecal incontinence symptoms increased the likelihood of faecal incontinence at 12 months after delivery (OR=6.1, 95% CI=1.8-21.5, P=0.005), whereas maternal characteristics, mode of delivery, and the presence of OASIS did not.28 In longer-term follow-up (3-5 years after delivery), 2.1% and 5.9% of women who had one vaginal delivery reported faecal incontinence to solid and loose stool, respectively.29

To the best of our knowledge, there have been no randomised controlled trials regarding the optimal timing for the use of endoanal US to assess OASIS after vaginal delivery. One randomised controlled trial has been conducted to compare clinical examination alone (control group) and clinical examination with additional endoanal US immediately after delivery (intervention group).30 31 The results of that study showed that US performed immediately after delivery—before repair—might detect more cases of OASIS: 5.6% of women were found to have full-thickness OASIS that was not recognised during clinical examination alone.31 However, the study also showed that five of 21 women underwent unnecessary intervention, as the sonographic defect could not be clinically located, despite surgical exploration.31 Therefore, the use of endoanal US immediately after delivery and before repair was not recommended.

Women with OASIS should undergo follow-up after delivery to assess symptoms of faecal incontinence. Currently, there is no consensus regarding the optimal mode of delivery for these women in subsequent pregnancies. Scheer et al32 and Karmarkar et al33 assessed women who had OASIS in subsequent pregnancies using a questionnaire, endoanal US, and manometry. Vaginal delivery was recommended for asymptomatic women with normal findings. Women were reassessed after subsequent deliveries. There were no statistically significant differences in anal manometry findings, anal symptoms, or quality of life following subsequent vaginal delivery or caesarean section.32 33 In the study by Scheer et al,32 new OASIS occurred in only one woman after a vaginal delivery. Therefore, decisions regarding the mode of delivery for subsequent pregnancies after OASIS should be based on clinical symptoms, anal manometry, and endoanal US. This would help to preserve anal sphincter function and avoid unnecessary caesarean sections. Currently, the value of the above assessments is limited in Hong Kong. The significance of an incidental finding of occult anal sphincter defect remains uncertain.

The prevalence of OASIS determined by endoanal US was higher than the rate determined by clinical practice. This may indicate that additional training for midwives and doctors may be required to improve the detection of OASIS. At 6 to 12 months after delivery, OASIS was not associated with symptoms of faecal or flatal incontinence, but a longer-term study is needed to confirm these findings.

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: RYK Cheung, SSC Chan.
Acquisition of data: OYK Wan, RYK Cheung, LL Lee, SSC Chan.
Analysis or interpretation of data: SPK Kwok, SSC Chan.
Drafting of the article: All authors.
Critical revision for important intellectual content: SPK Kwok, OYK Wan, RYK Cheung, SSC Chan.

The results from this research have been presented, in part, at the following conferences:
1. Wan OYK, Cheung RYK, Chan SSC. 6th Annual Meeting of the Asia-Pacific Urogynecology Association and 13th Japanese Society of Pelvic Organ Prolapse Surgery Joint Conference–Young Doctors Session. Okinawa, Japan, 22-24 March 2019 (oral abstract presentation).
2. Wan OYK, Kwok SPK, Cheung RYK, Chan SSC. Hospital Authority Convention 2019, Hong Kong, 14-15 May 2019 (e-poster presentation).
3. Kwok SPK, Wan OYK, Cheung RYK, Lee LL, Chung JPW, Chan SSC. Obstetrical and Gynaecological Society of Hong Kong Annual Scientific Meeting 2019, Hong Kong, 1-2 June 2019 (oral presentation).

As an editor of the journal, JPW Chung was not involved in the peer review process. Other 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.

Ethics approval was obtained from local institute, The Joint Chinese University of Hong Kong–New Territories East Cluster Clinical Research Ethics Committee (Ref CRE-2013.332). Written informed consent was obtained from all participants.

1. Fenner DE, Genberg B, Brahma P, Marek L, DeLancey JO. Fecal and urinary incontinence after vaginal delivery with anal sphincter disruption in an obstetrics unit in the United States. Am J Obstet Gynecol 2003;189:1543-50. Crossref

2. Thiagamoorthy G, Johnson A, Thakar R, Sultan AH. National survey of perineal trauma and its subsequent management in the United Kingdom. Int Urogynecol J 2014;25:1621-7. Crossref

3. Baghestan E, Irgens LM, Børdahl PE, Rasmussen S. Trends in risk factors for obstetric anal sphincter injuries in Norway. Obstet Gynecol 2010;116:25-34. Crossref

4. Ekéus C, Nilsson E, Gottvall K. Increasing incidence of anal sphincter tears among primiparas in Sweden: a population-based register study. Acta Obstet Gynecol Scand 2008;87:564-73. Crossref

5. Hong Kong College of Obstetricians and Gynaecologists. Territory-wide audit in Obstetrics & Gynaecology; 2009.

6. Bates LJ, Melon J, Turner R, Chan SS, Karantanis E. Prospective comparison of obstetric anal sphincter injury incidence between an Asian and Western hospital. Int Urogynecol J 2019;30:429-37. Crossref

7. Carroli G, Mignini L. Episiotomy for vaginal birth. Cochrane Database Syst Rev 2009;(1):CD000081. Crossref

8. Cheung RY, Shek KL, Chan SS, Chung TK, Dietz HP. Pelvic floor muscle biometry and pelvic organ mobility in East Asian and Caucasian nulliparae. Ultrasound Obstet Gynecol 2015;45:599-604. Crossref

9. Andrews V, Sultan AH, Thakar R, Jones PW. Occult anal sphincter injuries—myth or reality? BJOG 2006;113:195-200. Crossref

10. Guzmán Rojas RA, Shek KL, Langer SM, Dietz HP. Prevalence of anal sphincter injury in primiparous women. Ultrasound Obstet Gynecol 2013;42:461-6. Crossref

11. Sultan AH, Kamm MA, Hudson CN, Thomas JM, Bartram CI. Anal-sphincter disruption during vaginal delivery. N Engl J Med 1993;329:1905-11. Crossref

12. Chan SS, Cheung RY, Yiu KW, Lee LL, Leung TY, Chung TK. Pelvic floor biometry during first singleton pregnancy and the relationship with symptoms of pelvic floor disorders: a prospective observational study. BJOG 2014;121:121-9. Crossref

13. Chan SS, Cheung RY, Yiu KW, Lee LL, Chung TK. Pelvic floor biometry in Chinese primiparous women 1 year after delivery: a prospective observational study. Ultrasound Obstet Gynecol 2014;43:466-74. Crossref

14. Chung MY, Wan OY, Cheung RY, Chung TK, Chan SS. Prevalence of levator ani muscle injury and health-related quality of life in primiparous Chinese women after instrumental delivery. Ultrasound Obstet Gynecol 2015;45:728-33. Crossref

15. 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

16. Sultan AH. Obstetric perineal injury and anal incontinence. Clinical Risk 1999;5:193-6. Crossref

17. Chan SS, Cheung RY, Yiu AK, et al. Chinese validation of Pelvic Floor Distress Inventory and Pelvic Floor Impact Questionnaire. Int Urogynecol J 2011;22:1305-12. Crossref

18. Roos AM, Thakar R, Sultan A. Outcome of primary repair of obstetric anal sphincter injuries (OASIS): does the grade of tear matter? Ultrasound Obstet Gynecol 2010;36:368-74. Crossref

19. de Leeuw JW, Struijk PC, Vierhout ME, Wallenburg HC. Risk factors for third degree perineal ruptures during delivery. BJOG 2001;108:383-7. Crossref

20. Aukee P, Sundström H, Kairaluoma MV. The role of mediolateral episiotomy during labour: analysis of risk factors for obstetric anal sphincter tears. Acta Obstet Gynecol Scand 2006;85:856-60. Crossref

21. Ramalingam K, Monga AK. Outcomes and follow-up after obstetric anal sphincter injuries. Int Urogynecol J 2013;24:1495-500. Crossref

22. Hauck YL, Lewis L, Nathan EA, White C, Doherty DA. Risk factors for severe perineal trauma during vaginal childbirth: a Western Australian retrospective cohort study. Women Birth 2015;28:16-20. Crossref

23. Grobman WA, Bailit JL, Rice MM, et al. Racial and ethnic disparities in maternal morbidity and obstetric care. Obstet Gynecol 2015;125:1460-7. Crossref

24. Davies-Tuck M, Biro MA, Mockler J, Stewart L, Wallace EM, East C. Maternal Asian ethnicity and the risk of anal sphincter injury. Acta Obstet Gynecol Scand 2015;94:308-15.Crossref

25. Stedenfeldt M, Øian P, Gissler M, Blix E, Pirhonen J. Risk factors for obstetric anal sphincter injury after a successful multicentre interventional programme. BJOG 2014;121:83-91. Crossref

26. Boreham MK, Richter HE, Kenton KS, et al. Anal incontinence in women presenting for gynecologic care: prevalence, risk factors, and impact upon quality of life. Am J Obstet Gynecol 2005;192:1637-42. Crossref

27. Melville JL, Fan MY, Newton K, Fenner D. Fecal incontinence in US women: a population-based study. Am J Obstet Gynecol 2005;193:2071-6. Crossref

28. Chan SS, Cheung RY, Yiu KW, Lee LL, Chung TK. Prevalence of urinary and fecal incontinence in Chinese women during and after first pregnancy. Int Urogynecol J 2013;24:1473-9. Crossref

29. Ng K, Cheung RY, Lee LL, Chung TK, Chan SS. An observational follow-up study on pelvic floor disorders to 3-5 years after delivery. Int Urogynecol J 2017;28:1393-9. Crossref

30. Walsh KA, Grivell RM. Use of endoanal ultrasound for reducing the risk of complications related to anal sphincter injury after vaginal birth. Cochrane Database Syst Rev 2015;(10):CD010826. Crossref

31. Faltin DL, Boulvain M, Floris LA, Irion O. Diagnosis of anal sphincter tears to prevent fecal incontinence: a randomized controlled trial. Obstet Gynecol 2005;106:6-13. Crossref

32. Scheer I, Thakar R, Sultan AH. Mode of delivery after previous obstetric anal sphincter injuries (OASIS)—a reappraisal? Int Urogynecol J Pelvic Floor Dysfunct 2009;20:1095-101. Crossref

33. Karmarkar R, Bhide A, Digesu A, Khullar V, Fernando R. Mode of delivery after obstetric anal sphincter injury. Eur J Obstet Gynecol Reprod Biol 2015;194:7-10. Crossref

According to the Hong Kong Cancer Registry,1 prostate cancer is the third most common cancer in men. As in other cancers, biopsy is needed for histological confirmation of the diagnosis of prostate cancer before treatment is initiated. With the increasing age of the population, the incidence rate of this cancer is expected to increase; the frequency of prostate biopsy will therefore also increase. Hodge et al2 introduced the systematic sextant biopsy protocol under transrectal ultrasound guidance. Transrectal ultrasound-guided prostate biopsy (TRUSPB) has since become a widely accepted and routinely performed technique to detect prostate cancer.3 In Hong Kong, most urologists use TRUSPB to confirm the diagnosis of prostate cancer, particularly in patients with elevated prostate-specific antigen (PSA) or abnormal digital rectal examination; TRUSPB is also used in patients undergoing active surveillance of prostate cancer. However, there are complications associated with the use of TRUSPB. Most notably, because the procedure is performed via the rectum, there is a risk of postprocedural sepsis; the incidence of sepsis ranged was 2% to 4% in contemporary series,4 5 and sepsis-related mortality has also been reported.6

An increasing number of studies have demonstrated success in cancer diagnosis with extended biopsy using transperineal ultrasound-guided prostate biopsy (TPUSPB). In an early report, Kojima et al7 retrospectively assessed the usefulness of TPUSPB, which differs from TRUSPB in terms of patient position, puncture route, puncture site, and ultrasound probe.8 Most importantly, the TPUSPB enables urologists to thoroughly prepare the perineum with a disinfectant solution to eliminate the possibility of skin flora contamination of the puncture site.9 In addition, this procedure involves puncture of perineal skin under the guidance of a side-fire ultrasound probe without penetration of rectal mucosa, thereby avoiding the possibility that intestinal flora are transferred to the blood stream. An Australian study group showed that TPUSPB, in combination with antibiotic prophylaxis, could almost entirely prevent sepsis complications.10 Some authors have suggested that TPUSPB may be performed without antibiotic prophylaxis, thus reducing the risk of generating antibiotic resistance.4 Based on these potential benefits, our centre introduced TPUSPB beginning in January 2018. Subsequently, we have completely replaced TRUSPB with TPUSPB. In this study, we aimed to compare the outcomes of our initial series of patients who underwent TPUSPB with those of our previous cohort of patients who underwent TRUSPB.

In this retrospective cohort study, we compared 100 patients who underwent TPUSPB with 100 patients who underwent TRUSPB in our centre. This study was approved by our institutional ethics committee. All 100 patients who underwent TPUSPB from January 2018 to May 2018 (TPUSPB group) were included; 100 patients who underwent TRUSPB from January 2016 to April 2016 were also included. The indications for biopsy for both groups were serum PSA >4 ng/dL, abnormal digital rectal examination, and surveillance biopsy for patients under active surveillance.

The following data were retrieved from hospital records and compared between the two groups: age, serum PSA level, prostate size, PSA density, prostate cancer detection rate, and complications (eg, admission due to acute retention of urine, rectal bleeding, haematuria, fever, and sepsis). We used the Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3) as an acute change in total sequential organ failure assessment score ≥2 points due to the infection11: (1) respiratory rate ≥22/min, (2) altered mental activity, and (3) systolic blood pressure ≤100 mm Hg.

For both groups, the indications for prostatic biopsies were serum PSA level >4 ng/dL, abnormal digital rectal examination, or follow-up biopsy for patients under active surveillance. All patients underwent pre-procedure blood tests and urine tests to ensure there was no bleeding tendency or positive urine culture. Patients using antiplatelet or anticoagulant treatment were required to discontinue drugs prior to undergoing biopsy. All patients used a sodium phosphate rectal enema in the morning of the procedure and took oral prophylactic antibiotics (1 g amoxicillin-clavulanate and 500 mg ciprofloxacin) 2 hours before the procedure. For TPUSPB, numbing cream (2.5% lidocaine and 2.5% prilocaine) was applied over the perineal region 1 hour before the procedure and 1% lidocaine (20 mL) was injected into the perineum as local anaesthesia (LA) immediately prior to prostate biopsy, at a 45-degree angle from the midline and approximately 15 mm above the anus on either side. Details of the two procedures are described below. After either procedure, all patients were given an additional 1-day course of oral antibiotics (1 g amoxicillin-clavulanate and 500 mg ciprofloxacin).

When undergoing TRUSPB, patients assumed the left lateral position, as shown in Figure 1. Prostate size was measured using a transrectal biplanar ultrasound probe. Subsequently, 10 core biopsies were taken: five cores were taken from each side of the prostate at the base, mid, apex, upper lateral, and lower lateral regions. Each procedure was 5 to 10 minutes in duration.

When undergoing TPUSPB, patients assumed the Lloyd-Davies position prior to injection of lidocaine for LA (described above). After lidocaine injection, a biplanar ultrasound probe was inserted through the anus. The prostate size was measured, and 14-gauge angiocatheters were then inserted at the sites previously used for LA injection, as shown in Figure 2. Ten core biopsies were obtained in a manner similar to that of TRUSPB. Because of the different orientation of the biopsy needle, the apical biopsy was targeted towards the anterior fibromuscular layer. The biopsy needle was maintained parallel to the probe to ensure clear visualisation of the targeted area, which was possible when the whole needle was completely visualised on ultrasound (Fig 3). Each procedure was 10 to 15 minutes in duration. We also assessed the pain experienced during the TPUSPB at three time points, namely during probe insertion into the anus, LA injection, and biopsy procedures, by verbal analogue scale (0-10) during TPUSPB.

Statistical analyses were performed using SPSS (Windows version 24.0; IBM Corp, Chicago [IL], United States). For continuous variables, age was compared by independent t test, while PSA, prostate size, and PSA density were compared by the Mann-Whitney U test as they did not exhibit normal distributions. Normality was assessed by normal QQ plots and the Shapiro-Wilk test. When comparing categorical variables, including cancer detection rates and complication rates, the Chi squared test was used if the expected count in each cell was >5; otherwise, Fisher’s exact test was used. In addition, the Cochran-Mantel-Haenszel test was performed to assess whether there was an association between the biopsy method and cancer detection rate according to clinical stage. Differences with a two-sided P value of <0.05 were considered to be statistically significant.

The patient characteristics, prostate cancer detection rates, and complications in patients who underwent TPUSPB, compared with those who underwent TRUSPB, are summarised in Table 1. Age did not significantly differ between the two groups. The median serum PSA in the TPUSPB group was higher than that in the TRUSPB group (12.0 ng/dL vs 9.5 ng/dL, P=0.047). Moreover, the median prostate size in the TPUSPB was smaller than that in the TRUSPB group (46.2 mL vs 56.8 mL, P=0.003). Therefore, the PSA density of TPUSPB group was higher than that in the TRUSPB group (0.27 vs 0.16, P=0.001).

Stratified prostate cancer detection rates in patients who underwent TPUSPB, compared with those who underwent TRUSPB, are listed in Table 2. There was no statistically significant difference in overall prostate cancer detection rate between the two groups. In subgroup analysis stratified by serum PSA level, the TPUSPB group had a higher prostate cancer detection rate than the TRUSPB group among patients with 20 to 100 ng/mL PSA (50% vs 15%, P=0.036). However, there were no statistically significant differences in prostate cancer detection rates among other subgroups according to PSA levels. There were also no statistically significant differences in prostate cancer detection rates between the two groups upon stratification according to PSA density or clinical staging.

In analysis of 35 patients with positive cores in the TPUSPB group, 16 (45.7%) patients had at least one positive core in the anterior fibromuscular stroma. Among these 16 patients, 14 were diagnosed with high-risk prostate cancers, with multiple positive cores in each patient. The relatively high proportion of high-risk prostate cancer in the TPUSPB cohort might explain the relatively high number of positive cores in the anterior fibromuscular stroma.

In the TPUSPB group, 19 patients had previous negative findings in TRUSPB; three of these 19 (15.7%) were diagnosed with prostate cancer based on the findings of TPUSPB. Two of the three tumours were detected in the anterior fibromuscular layer, and the remaining tumour was found in the apical zone. In the TRUSPB group, 36 patients had previous negative findings in TRUSPB; six (16.7%) of these were diagnosed with prostate cancer based on the findings of the current TRUSPB.

Concerning about the pain experienced during TPUSPB, the pain scores reported by patients during probe insertion into the anus, LA injection, and biopsy procedures were 1-2, 1-2, and 2-4, respectively.

With respect to complications, we initially planned to use the Sepsis-3 described above, but no patients in this study developed clinical signs of infection that met the criteria for sepsis. However, four (4%) patients in the TRUSPB group developed fever requiring hospital admission compared with none in the TPUSPB group (P=0.121) [Table 3]. At least 1 week of intravenous antibiotic treatment was prescribed for all four of those patients with fever. Three patients in the TPUSPB group and one in the TRUSPB group developed acute retention of urine (P=0.621). No patients in the TPUSPB group and one in the TRUSPB group had rectal bleeding (P=1.000). There were no admissions due to haematuria in either group.

Since Hodge et al2 introduced the systematic sextant biopsy protocol, TRUSPB has become the main approach to detect prostate cancer worldwide. In recent years, many urologists have described increased risks of infection and sepsis associated with TRUSPB.12 Fluoroquinolone was previously thought to provide effective antibiotic prophylaxis, thus preventing infections associated with TRUSPB; however, fluoroquinolone-resistant bacteria and extended-spectrum beta-lactamase–producing bacteria are present within the intestinal flora of 40.4% and 41.0% of Chinese patients, respectively.13 Accordingly, the rate of post-TRUSPB sepsis is rising both in Hong Kong6 and worldwide.14 To reduce the rates of infection, many strategies have been attempted, including augmented prophylactic antibiotic protocols. However, no strategies have prevented the development of sepsis due to the transfer of faecal bacteria into the blood stream through TRUSPB puncture sites.10 Transperineal ultrasound-guided prostate biopsy has been suggested as a potentially safer alternative. Notably, the indications, workups, medications, and numbers of cores are identical between TRUSPB and TPUSPB. However, the techniques differ with respect to multiple aspects, including patients’ position and puncture route,15 as described in the Methods section of this paper.

From the pain scores recorded during TPUSPB, most patients tolerated the procedure well. Because the entire procedure was performed under LA, all patients could be discharged on the same day without the need for general anaesthesia or monitored anaesthesia care. In Asian nations, many patients must travel a considerable distance to reach the hospital; therefore, 1-day out-patient procedures are preferable for patients and their relatives. Moreover, some patients are high-risk or unfit for general anaesthesia or monitored anaesthesia care; procedures performed under LA are therefore much safer and more practical for them.

Transrectal ultrasound-guided prostate biopsy neglects prostate cancer located in the anterior fibromuscular stroma, whereas the TPUSPB does not.16 In our study, a significant proportion of positive prostatic cores were found in the anterior fibromuscular stroma among patients in the TPUSPB group. Moreover, some patients with prior negative findings in TRUSPB were diagnosed with prostate cancer in the anterior fibromuscular stroma based on the results of TPUSPB. Therefore, TPUSPB may have an advantage over TRUSPB in patients with previous negative biopsy findings.

This study had some limitations. First, a consistent number of cores was biopsied in all patients in the TPUSPB group, irrespective of prostate size. To improve the rate of prostate cancer detection, some experts have advocated for the use of different numbers of prostate biopsies, based on prostate size16—more biopsies should be taken in patients with larger prostates. Second, TPUSPB required more time than TRUSPB. However, as each step is standardised, the duration of the procedure may decrease. Third, there was no documentation of pain scores in the TRUSPB group; thus, a comparison could not be performed. Future studies should address these limitations. In particular, a larger sample size is needed to confirm whether TPUSPB is superior with respect to the rate of prostate cancer detection.

In summary, TPUSPB avoids penetration of the rectal mucosa and possible transfer of intestinal flora to the blood stream during the procedure. This contributed to the lack of infections in the present study. With respect to prostate cancer detection rate, TPUSPB is at least comparable to TRUSPB. Therefore, an increasing number of urologists may adopt this technique in the future.

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 study: KL Lo, KL Chui, CF Ng.
Acquisition of data: KL Lo, K Lim, JKM Li, J Wong, SK Mak.
Analysis or interpretation of data: KL Lo, CF Ng, SCH Leung, SF Ma.
Drafting of the manuscript: KL Lo, CF Ng.
Critical revision for important intellectual content: All authors.

As an editor of the journal, CF Ng was not involved in the peer review process. Other authors have disclosed no conflicts of interest.

We would like to thank for the support of the nursing staff in the Integrated Ambulatory Care Centre, North District Hospital for their support to the procedures.

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

This study was approved by the Joint Chinese University of Hong Kong–New Territories East Cluster Clinical Research Ethics Committee (Ref CREC 2018.323).

1. Hong Kong Cancer Registry, Hospital Authority, Hong Kong SAR Government. Prostate cancer in 2016. Available from: http://www3.ha.org.hk/cancereg/pdf/top10/rank_2016.pdf. Accessed 17 May 2019.

2. Hodge KK, McNeal JE, Terris MK, Stamey TA. Random systematic versus directed ultrasound guided transrectal core biopsies of the prostate. J Urol 1989;142:71-4. Crossref

3. Heidenreich A, Bastian PJ, Bellmunt J, et al. EAU guidelines on prostate cancer. Part 1: screening, diagnosis, and local treatment with curative intent-update 2013. Eur Urol 2014;65:124-37. Crossref

4. Toner L, Bolton DM, Lawrentschuk N. Prevention of sepsis prior to prostate biopsy. Investig Clin Urol 2016;57:94-9. Crossref

5. Chan ES, Lo KL, Ng CF, Hou SM, Yip SK. Randomized controlled trial of antibiotic prophylaxis regimens for transrectal ultrasound-guided prostate biopsy. Chin Med J (Engl) 2012;125:2432-5.

6. Ng CF, Chan SY. Re: the incidence of fluoroquinolone resistant infections after prostate biopsy—are fluoroquinolones still effective prophylaxis? J Urol 2008;180:1570-1. Crossref

7. Kojima M, Hayakawa T, Saito T, Mitsuya H, Hayase Y. Transperineal 12-core systematic biopsy in the detection of prostate cancer. Int J Urol 2001;8:301-7. Crossref

8. Xue J, Qin Z, Cai H, et al. Comparison between transrectal and transperineal prostate biopsy for detection of prostate cancer: a meta-analysis and trial sequential analysis. Oncotarget 2017;8:23322-36. Crossref

9. Chang DT, Challacombe B, Lawrentschuk N. Transperineal biopsy of the prostate—is this the future? Nat Rev Urol 2013;10:690-702. Crossref

10. Grummet JP, Weerakoon M, Huang S, et al. Sepsis and ‘superbugs’: should we favour the transperineal over the transrectal approach for prostate biopsy? BJU Int 2014;114:384-8. Crossref

11. Singer M, Deutschman CS, Seymour CW, et al. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA 2016;315:801-10. Crossref

12. Steensels D, Slabbaert K, De Wever L, Vermeersch P, Van Poppel H, Verhaegen J. Fluoroquinolone-resistant E. coli in intestinal flora of patients undergoing transrectal ultrasound-guided prostate biopsy—should we reassess our practices for antibiotic prophylaxis? Clin Microbiol Infect 2012;18:575-81. Crossref

13. Tsu JH, Ma WK, Chan WK, et al. Prevalence and predictive factors of harboring fluoroquinolone-resistant and extended-spectrum β-lactamase-producing rectal flora in Hong Kong Chinese men undergoing transrectal ultrasound-guided prostate biopsy. Urology 2015;85:15-21. Crossref

14. Williamson DA, Barrett LK, Rogers BA, Freeman JT, Hadway P, Paterson DL. Infectious complications following transrectal ultrasound-guided prostate biopsy: new challenges in the era of multidrug-resistant Escherichia coli. Clin Infect Dis. 2013;57:267-74. Crossref

15. Emiliozzi P, Corsetti A, Tassi B, Federico G, Martini M, Pansadoro V. Best approach for prostate cancer detection: a prospective study on transperineal versus transrectal six-core prostate biopsy. Urology 2003;61:961-6. Crossref

16. Ong WL, Weerakoon M, Huang S, et al. Transperineal biopsy prostate cancer detection in first biopsy and repeat biopsy after negative transrectal ultrasound-guided biopsy: the Victorian Transperineal Biopsy Collaboration experience. BJU Int 2015;116:568-76. Crossref