The Health Commission of Hubei province, China, first announced a cluster of patients with atypical pneumonia of unidentified pathogenic cause on 31 December 2019.1 The virus was isolated; its genome was then sequenced by a number of Chinese scientists who confirmed it to be a type of coronavirus. The virus was named severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), and the resulting disease was termed coronavirus disease 2019 (COVID-19) by the World Health Organization.2 The infectious disease centre at Princess Margaret Hospital, Hong Kong, is the designated local tertiary referral centre that provides treatment for patients diagnosed with COVID-19. Case reports available at the time of writing describe ground-glass lung changes in isolated patients with COVID-19.3 4 5 Herein, we evaluate the radiological features in the earliest group of patients with confirmed COVID-19 in Hong Kong. The aim of the present study was to provide insights into radiological identification and assessment of COVID-19 pneumonia.

Patients in this study were confirmed to have SARS-CoV-2 infection on the basis of a positive nasopharyngeal aspirate reverse transcription polymerase chain reaction result and/or a positive serological testing result. The first 20 confirmed patients from all hospitals in Hong Kong were sent to our infectious disease unit for quarantine and treatment.

The radiological images reviewed in this study were obtained with high-resolution computed tomography (CT) or conventional thoracic CT. The examinations were performed with a multi-slice 16-head detector scanner (LightSpeed; GE Medical Systems, Waukesha [WI], United States) in the infectious disease centre, which is equipped with a negative pressure ventilating system.

The following parameters were used for high-resolution CT of the thorax: voltage, 120 kVp; current, 30-300 mA (smart mA); field of view, 32-40 mm; and gantry rotation, 1.0 s. Conventional CT was performed with the following parameters: voltage, 120 kVp; current, 100-500 mA (smart mA); diagonal field of view, 40 mm; and gantry rotation, 0.5 s.

Radiographers who performed CT scans of patients with confirmed or suspected COVID-19 were required to wear full-body protective garments, in accordance with guidelines from infection control specialists. All radiographers wore disposable fluid-resistant gowns, gloves, face shields, face masks with a rating of at least N95 (3M; Aberdeen [SD], United States), disposable shoe wraps, and protective eyewear. Patients were also required to wear masks with a rating of at least N95. All surfaces in contact with or within 1 m of the patients were cleaned with antiviral agents after completion of scanning. Cleaning procedures were performed twice; subsequently, the CT suite was not used for at least 30 minutes to allow for several air exchanges prior to the entry of the next patient.

The radiological images were reviewed and interpreted by consensus; the reviewers were two consultant radiologists who were registered specialist radiologists under Hong Kong Medical Council, Fellows of the Royal College of Radiologists, and Fellows of the Hong Kong College of Radiologists with 20 years of experience each in body CT.

From 22 January 2020 to 29 February 2020, our hospital received 20 patients aged 25 to 80 years all with confirmed COVID-19 (Table). Chest radiographs were performed for all patients on admission; the most common finding was bilateral non-specific pulmonary infiltrates (Fig 1). Shortly after admission, 19 patients (11 men and eight women) underwent high-resolution CT or conventional plain CT thorax. One patient was asymptomatic and exhibited normal chest radiographs throughout the hospital stay; thus, no CT scans were performed for further evaluation. The median interval from confirmation of diagnosis to CT scanning was 3 days.

As indicated in the Table, ground-glass opacities (GGO) with peripheral subpleural distribution were observed in all patients (100%) [Figs 2, 3a, b]. Furthermore, 57.9% of the patients exhibited diffuse involvement of both upper and basal zones, 15.8% demonstrated upper zone predominance, and 26.3% demonstrated basal predominance. All lobes of the lungs were involved in 16 (84.2%) patients (Fig 2), subsegmental consolidative changes were present in 14 (73.7%) patients (Fig 3a, c), interlobular septal thickening was present in 12 (63.2%) patients (Fig 3a), bronchial wall thickening or dilation was present in 10 (52.6%) patients (Fig 3b), and crazy-paving patterns were present in six (31.6%) patients (Fig 3). Mediastinal lymph node enlargement (ie, short axis >1 cm), centrilobular nodule, and pleural effusion were not detected in any of the patients.

In summary, peripheral subpleural GGO without zonal predominance in the absence of centrilobular nodules, pleural effusion, and lymph node enlargement were consistent findings. Other common findings included septal thickening, consolidations, bronchial dilatation/wall thickening, and crazy-paving patterns.

The most common respiratory pathogens are viruses. The imaging findings of viral pneumonia are diverse and often overlap with the findings of other non-viral pneumonias and inflammatory conditions. Imaging findings have been described in recent outbreaks associated with emerging pathogens, including severe acute respiratory syndrome (SARS) coronavirus and Middle East respiratory syndrome (MERS) coronavirus.6 7 Although a definite diagnosis cannot be reached based on imaging features alone, recognition of viral pneumonia patterns can aid in identification of potentially infected patients, especially during a specific viral outbreak.

Peripheral subpleural GGO without zonal predominance in the absence of pleural effusion and lymph node enlargement were consistent findings in initial thoracic CT scans of patients with COVID-19. These findings coincide with recent reports of single patients in which the major findings comprised multifocal patchy GGO, most evident around the periphery.3 4 5 Several other findings including bronchial wall thickening, bronchial dilatation, septal thickening, and crazy-paving patterns were also observed in a subset of patients.

Similar to our findings, diseases caused by other β-coronaviruses (eg, SARS, MERS, and other endemic human β-coronaviruses including OC43 and HKU1) are also characterised by multifocal peripheral GGOs. Moreover, patients infected with those viruses rarely exhibit cavitation, lymphadenopathy, or pleural effusions,6 similar to the findings in the present study. However, our study showed that lung changes in patients with COVID-19 have no zonal predominance, which contrasts with the findings in patients with SARS or MERS, which predominantly affect basal zones.6 Air-space opacities are less frequent in patients with COVID-19 pneumonia, compared with patients with SARS or MERS, which suggests that the course of COVID-19 pneumonia may be less aggressive. Nonetheless, conclusions should not be drawn prematurely as this study only involved the initial radiological assessment. More insights into the temporal changes regarding radiological findings during the progression of disease will become available as these patients undergo follow-up scans. Further studies that include the clinical course of COVID-19 in these patients will be performed in the future.

Coronavirus disease 2019 is a highly contagious disease that requires high vigilance and rapid detection. Knowledge of common radiological patterns on CT thorax can help discern the extent of pulmonary involvement and potentially facilitate identification of patients with pneumonia in Hong Kong during the COVID-19 outbreak.

Concept or design: All authors.
Acquisition of data: All authors.
Analysis or interpretation of data: All authors.
Drafting of the manuscript: SK Li, FH Ng.
Critical revision of the manuscript for important intellectual content: SK Li, FH Ng.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

The authors have disclosed no conflicts of interest.

We would like to express our gratitude to the Infectious Disease Team and “dirty team” physicians of Princess Margaret Hospital, Hong Kong, for their professional patient care and invaluable contribution to the understanding of a novel disease.

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

This study was carried out with approval from the Kowloon West Cluster Ethics Committee (Ref KW/EX-20-032(144-20)). The requirement for patient consent was waived by the committee.

1. Centre for Health Protection, Hong Kong SAR Government. CHP closely monitors cluster of pneumonia cases on Mainland. 31 December 2019. Available from: https://www.info.gov.hk/gia/general/201912/31/P2019123100667.htm. Accessed 1 Feb 2020.

2. World Health Organization. Clinical management of severe acute respiratory infection when COVID-19 is suspected. Interim guidance. 12 January 2020. Available from: https://www.who.int/publications-detail/clinical-management-of-severe-acute-respiratory-infection-when-novel-coronavirus-(ncov)-infection-is-suspected. Accessed 1 Feb 2020.

3. Chan JF, Yuan S, Kok KH, et al. A familial cluster of pneumonia associated with the 2019 novel coronavirus indicating person-to-person transmission: a study of a family cluster. Lancet 2020;395:514-23. Crossref

4. Medlinkcn.com. 武漢19-nCoV 肺炎影像學表現初探 [in Chinese]. Available from: http://www.medlinkcn. com/?id=138. Accessed 1 Feb 2020.

5. Lei J, Li J, Li X, Qi X. CT imaging of the 2019 novel coronavirus (2019-nCoV) pneumonia. Radiology 2020;295:18. Crossref

6. Koo HJ, Lim S, Choe J, Choi SH, Sung H, Do KH. Radiographic and CT features of viral pneumonia. Radiographics 2018;38:719-39. Crossref

7. Franquet T. Imaging of pulmonary viral pneumonia. Radiology 2011;260:18-39. Crossref

The World Health Organization aims to eradicate hepatitis B virus (HBV) by 2030 and prevention of vertical transmission is a key element of its eradication efforts.1 The risk of chronic infection depends on the timing of infection acquisition and is highest during the perinatal period, such that chronic infection occurs in approximately 90% of newborns from HBV-infected mothers.2 The risk is dramatically reduced by administration of hepatitis B immunoglobulin to newborns at birth, in combination with a complete course of hepatitis B vaccination.3 Despite a 75% to 90% reduction in the carrier rate with these measures, they have not resulted in complete eradication of HBV infections. Among the maternal and obstetric factors examined, a high maternal HBV DNA level during pregnancy is the strongest risk factor leading to immunoprophylaxis failure.4 5 6 7

The immunoprophylaxis failure rate in Hong Kong is reportedly 1.1%, according to the results of a local prospective multicentre observational study.5 Immunoprophylaxis failure occurs only in those women with high viral load (ie, ≥6 log₁₀ copies/mL [≥171 821 IU/mL]; immunoprophylaxis failure rate 4.2%) or hepatitis B e-antigen (HBeAg)–positive status (immunoprophylaxis failure rate 4.5%). The use of antiviral treatment during the third trimester in highly viraemic mothers to suppress viral load has been advocated to reduce the risk of chronic HBV infection in newborns.8 9 To achieve this, it is essential to establish a management strategy that includes HBV DNA assessment for identification of high-risk patients, as well as initiation of prompt antiviral treatment in the antenatal period.

An enhanced service model for pregnant women who had positive screening results for hepatitis B surface antigen (HBsAg) was established on 15 May 2017 at Queen Mary Hospital in Hong Kong (Fig 1). All women were offered blood tests for HBV DNA, performed by the Department of Medicine, The University of Hong Kong, during their first antenatal visits at Tsan Yuk Hospital or Queen Mary Hospital. The cost of HK$400 per test was borne by the patients; the laboratory results were reviewed by hepatologists.

Pregnant women with HBV DNA levels of ≥200 000 IU/mL were triaged by hepatologists for an early clinic appointment, generally before 33 weeks of gestation, to receive counselling regarding potential antiviral treatment. Tenofovir disoproxil fumarate (TDF) 300 mg daily was chosen for its potent efficacy and risk of pre-existing or emergent resistant mutants from previous lamivudine and telbivudine treatments.10 11 12 Drug compliance and HBV DNA level were monitored regularly. All other pregnant women with viral loads of <200 000 IU/mL were also scheduled for an elective long-term follow-up appointment. Irrespective of viral load, all neonates born from pregnant women with hepatitis B were administered HBV vaccine and hepatitis B immunoglobulin within 12 hours of birth. The present study aimed to evaluate the patients’ acceptance and outcome of this enhanced service model for management of pregnant women carrying hepatitis B.

This retrospective review included all women who attended the antenatal clinic from 15 May 2017 to 31 December 2019. Information regarding hepatitis B carrier status, HBV DNA blood test acceptance, viral load, patient triage, and TDF acceptance were retrieved from the antenatal record system and clinical management system under the Hong Kong Hospital Authority. The HBV DNA assays were performed in Department of Medicine, The University of Hong Kong.

Each patient’s HBV carrier status was determined by an HBsAg test performed during pregnancy. Hepatitis B virus DNA level was considered high for viral loads of ≥200 000 IU/mL and low for viral loads of <200 000 IU/mL. The rate of antenatal acceptance of TDF was defined as the proportion of women taking TDF in the group with high viral loads who had been counselled by hepatologists. Descriptive statistics are reported.

Of 13 082 women who attended the antenatal clinic from 15 May 2017 to 31 December 2019, 375 pregnant women had positive HBsAg screening results; the carrier rate was 2.9%. In total, 102 (27.2%) women had undergone HBV DNA testing or received regular hepatological follow-up before pregnancy. Blood tests for HBV DNA and hepatological reviews were offered to 273 pregnant women. Two women refused further assessment and four women did not attend the blood test visit. Reasons for refusal or non-attendance were not documented. Of the four women who did not attend the blood test visit, two were reminded of the need for a blood test at subsequent antenatal visits, but did not complete the tests. Overall, the acceptance rate for hepatological review was 97.8% (267/273). Among the 267 women who accepted hepatological reviews, blood tests were not performed because of pregnancy termination due to trisomy 21 (n=1) and miscarriage (n=1). Thus, the final cohort comprised 265 pregnant women who had HBV viral load results available for triage assessment. The median gestational age at the time of HBV testing was 17 weeks.

Sixty (22.6%) pregnant women were HBV carriers with viral loads of ≥200 000 IU/mL; highest level was 688 000 000 IU/mL. The median age of women with high viral loads was not substantially lower than that of women with low viral loads (33 years vs 35 years; Wilcoxon rank sum test; not significant).

Fifty eight of the 60 patients with high viral loads were scheduled for hepatological review before the expected date of delivery. First hepatological review appointments were scheduled for 55 (91.7%) women and 44 (73.3%) women at or before 32 and 28 weeks of gestation, respectively. Among the three women who scheduled their first hepatological review appointment after 32 weeks of gestation, two delayed the referral submission and one had attended the antenatal clinic at an advanced stage of gestation (Fig 2).

One patient did not attend a hepatological review appointment before 28 weeks of gestation. The remaining 57 women with high viral loads received antenatal counselling regarding TDF and 56 women agreed to take the drug, thereby constituting an antenatal acceptance rate of 96.6% (56/58). The acceptance rates of TDF among women with high viral loads are shown in the Table.

Treatment with TDF commenced at a median gestational age of 26 weeks (range, 20-38 weeks). Overall, 44 (78.6%) women received 300 mg daily TDF at or before 28 weeks of gestation. Among the 12 women who began TDF treatment after 28 weeks of gestation, four had deferred blood tests for assessment of viral load, six had late antenatal clinic appointments, one did not attend the initial appointment at 28 weeks, and one attended the clinic at 33 weeks of gestation. Overall, 52 (92.9%) women commenced TDF treatment at or before 32 weeks of gestation.

Screening for HBV carrier status is a universal antenatal test in Hong Kong. Women who have positive screening results are counselled regarding the risk of vertical transmission. In 1983, a neonatal HBV vaccination programme was introduced in Hong Kong to cover vaccination for first newborns of carrier mothers. This became universal in November 1988; hepatitis B immunoglobulin and hepatitis B vaccine have since been administered to all babies born to mothers carrying HBV. These measures focus mainly on postnatal neonatal immunoprophylaxis; however, they lack a robust system for actively reducing the antenatal risk of vertical transmission, as well as a referral system that ensures long-term hepatological follow-up for carrier mothers.

The present study demonstrated an effective and acceptable approach involving HBV DNA testing during triage of obstetric patients for prevention of perinatal HBV transmission. Data from the Centre of Health Protection have shown that the HBsAg prevalence in pregnant women is decreasing, from >10% in the early 1990s to 5.0% in 2017.13 The HBV carrier rate in the present study (2.9%) was lower than that previously reported in Hong Kong. Our cohort included women with HBsAg who were tested from May 2017 to December 2019; the low carrier rate in the present study might reflect a continuous reduction in overall HBsAg prevalence, due to universal neonatal vaccination.14 However, more than 70% of the pregnant women in our cohort did not have HBV viral load testing or regular hepatological surveillance before pregnancy. This is an important public health concern, because HBV carriers with high viral loads are at higher risk of mother-to-child transmission of HBV, as well as development of liver cirrhosis and hepatocellular carcinoma. With proper antenatal education and general awareness, nearly 98% of the obstetric patients in our population were willing to undergo self-financed HBV DNA testing.

Viral load is a key factor that influences immunoprophylaxis failure4; a higher risk of immunoprophylaxis failure has been demonstrated in women with viral loads of ≥200 000 IU/mL. Positive HBeAg screening results, maternal age <35 years, and body mass index ≤21 kg/m² have been associated with a higher mean viral load.15 Our study showed that 22.6% of women had viral loads of ≥200 000 IU/mL, which was comparable to previous findings. Although age was not a statistically significant factor in the present study, a previous study showed that women with high viral loads were younger than women with low viral loads.5 The prevention of perinatal transmission is of considerable importance in achieving complete eradication of HBV. Incorporation of HBV DNA testing during pregnancy is a key element that can facilitate identification of at-risk pregnant women who may benefit from antenatal antiviral prophylaxis. Ideally, both liver function test and HBeAg should be assessed in pregnant women to determine their HBV disease status; antiviral treatment may be initiated for maternal indications. Although the presence of HBeAg suggests a high risk of immunoprophylaxis failure, HBeAg was not routinely assessed during triage in the present cohort because it was not regarded as an indicator of the need for antiviral treatment to prevent vertical transmission.9 16 17 Additionally, HBV DNA quantification provides a continuous assessment of risk according to viral load, compared to the dichotomous result of HBeAg screening. Therefore, HBV DNA level should be used to identify women who should receive antiviral treatment.18

Tenofovir disoproxil fumarate is the drug of choice for antenatal prophylaxis because of its potent effect and the possibility of mutants resistant to lamivudine and telbivudine, due to prior treatment with those drugs.11 12 19 Breastfeeding is not contra-indicated for women who are taking TDF. A highly promising rate of antenatal acceptance of TDF (96.6%) was observed among women who had undergone antenatal hepatological review. This indicates the need for surveillance and the usefulness of patient education during the antenatal period.

Although the optimal timing of antiviral treatment remains controversial, randomised controlled trials show that initiation of TDF during the period between 28 and 32 weeks of gestation is effective in reduction of immunoprophylaxis failure. Earlier initiation of antiviral treatment is unnecessary, because the immunoprophylaxis failure rate is not appreciably reduced. Postponement of treatment to a point later than 32 weeks of gestation may result in an insufficient duration of treatment and suboptimal viral load reduction at delivery.10 20 Guidelines from the American Association for the Study of Liver Diseases and the Asian Pacific Association for the Study of the Liver recommend initiation of treatment at 28 to 32 weeks of gestation.9 21 The present study demonstrated a feasible framework for triage of nearly all pregnant women with high viral loads before their dates of delivery. Nearly 93% were able to initiate antiviral prophylaxis at or before 32 weeks of gestation; this could only be attained with an appropriate hepatological review appointment during pregnancy. The arrangement could be further improved if women could attend antenatal visits earlier during pregnancy, blood test logistics could be simplified, and resources could be allocated more effectively.

To the best of our knowledge, the multidisciplinary efforts of obstetricians and hepatologists have enabled Queen Mary Hospital to become the first public hospital in Hong Kong with enhanced antenatal management for pregnant women carrying hepatitis B. The proportion of women with high viral loads was comparable to the proportions in previous studies. Our results indicated the usefulness of HBV DNA blood tests in pregnant women and high acceptance of antenatal antiviral treatment. Triage according to HBV DNA level allowed early hepatological review and commencement of antiviral medication, thereby reducing the viral load at the time of delivery and minimising the risk of vertical transmission. This service model was adopted as a framework for implementation of a fully funded enhanced antenatal service to prevent mother-to-child transmission of HBV in public maternity units, commencing 1 January 2020.

Concept or design: PW Hui.
Acquisition of data: C Ng, PW Hui.
Analysis or interpretation of data: C Ng, PW Hui.
Drafting of the manuscript: PW Hui.
Critical revision of the manuscript for important intellectual content: KW Cheung, CL Lai.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

All authors have disclosed no conflicts of interest.

The authors thank Mr John Yuen for performing HBV DNA analysis.

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

This research was approved by the Institutional Review Board of the University of Hong Kong / Hospital Authority West Cluster (Ref UW 20-092).

1. World Health Organization. Global Health Sector Strategy on viral hepatitis 2016-2021, towards ending viral hepatitis. 2016. Available from: http://apps.who.int/iris/bitstream/handle/10665/246177/WHO-HIV-2016.06-eng.pdf?sequence=1. Accessed 2 Feb 2020.

2. Edmunds WJ, Medley GF, Nokes DJ, Hall AJ, Whittle HC. The influence of age on the development of the hepatitis B carrier state. Proc Biol Sci 1993;253:197-201. Crossref

3. Lee C, Gong Y, Brok J, Boxall EH, Gluud C. Effect of hepatitis B immunisation in newborn infants of mothers positive for hepatitis B surface antigen: systematic review and meta-analysis. BMJ 2006;332:328-36. Crossref

4. Cheung KW, Seto MT, Wong SF. Towards complete eradication of hepatitis B infection from perinatal transmission: review of the mechanisms of in utero infection and the use of antiviral treatment during pregnancy. Eur J Obstet Gynecol Reprod Biol 2013;169:17-23. Crossref

5. Cheung KW, Seto MT, Kan AS, et al. Immunoprophylaxis failure of infants born to hepatitis B carrier mothers following routine vaccination. Clin Gastroenterol Hepatol 2018;16:144-5. Crossref

6. Wen WH, Chang MH, Zhao LL, et al. Mother-to-infant transmission of hepatitis B virus infection: significance of maternal viral load and strategies for intervention. J Hepatol 2013;59:24-30. Crossref

7. Zou H, Chen Y, Duan Z, Zhang H, Pan C. Virologic factors associated with failure to passive-active immunoprophylaxis in infants born to HBsAg-positive mothers. J Viral Hepat 2012;19:e18-25. Crossref

8. European Association for the Study of the Liver. EASL 2017 Clinical Practice Guidelines on the management of hepatitis B virus infection. J Hepatol 2017;67:370-98.

9. Terrault NA, Lok AS, McMahon BJ, et al. Update on prevention, diagnosis, and treatment of chronic hepatitis B: AASLD 2018 hepatitis B guidance. Hepatology 2018;67:1560-99. Crossref

10. Pan CQ, Duan Z, Dai E, et al. Tenofovir to prevent hepatitis B transmission in mothers with high viral load. N Engl J Med 2016;374:2324-34. Crossref

11. Hyun MH, Lee YS, Kim JH, et al. Systematic review with meta-analysis: the efficacy and safety of tenofovir to prevent mother-to-child transmission of hepatitis B virus. Aliment Pharmacol Ther 2017;45:1493-505. Crossref

12. Hu YH, Liu M, Yi W, Cao YJ, Cai HD. Tenofovir rescue therapy in pregnant females with chronic hepatitis B. World J Gastroenterol 2015;21:2504-9. Crossref

13. Viral Hepatitis Control Office, Special Preventive Programme, Centre for Health Protection, Department of Health, Hong Kong SAR Government. Surveillance of Viral Hepatitis in Hong Kong—2017 Update Report. Available from: https://www.chp.gov.hk/files/pdf/viral_hepatitis_report.pdf. Accessed 2 Feb 2020.

14. Lao TT, Sahota DS, Chan PK. Three decades of neonatal vaccination has greatly reduced antenatal prevalence of hepatitis B virus infection among gravidae covered by the program. J Infect 2018;76:543-9. Crossref

15. Cheung KW, Seto MTY, So PL, et al. Optimal timing of hepatitis B virus DNA quantification and clinical predictors for higher viral load during pregnancy. Acta Obstet Gynecol Scand 2019;98:1301-6. Crossref

16. Hu Y, Xu C, Xu B, et al. Safety and efficacy of telbivudine in late pregnancy to prevent mother-to-child transmission of hepatitis B virus: a multicenter prospective cohort study. J Viral Hepat 2018;25:429-37. Crossref

17. Jourdain G, Ngo-Giang-Huong N, Harrison L, et al. Tenofovir versus placebo to prevent perinatal transmission of hepatitis B. N Engl J Med 2018;378:911-23. Crossref

18. Cheung KW, Lao TT. Hepatitis B—Vertical transmission and the prevention of mother-to-child transmission [in press]. Best Pract Res Clin Obstet Gynaecol. In press.

19. Cheung KW, Seto MT, Lao TT. Prevention of perinatal hepatitis B virus transmission. Arch Gynecol Obstet 2019;300:251-9. Crossref

20. Yang X, Zhong X, Liao H, Lai Y. Efficacy of antiviral therapy during the second or the third trimester for preventing mother-to-child hepatitis B virus transmission: a systematic review and meta-analysis. Rev Inst Med Trop Sao Paulo 2020;62:e13. Crossref

21. Sarin SK, Kumar M, Lau GK, et al. Asian-Pacific clinical practice guidelines on the management of hepatitis B: a 2015 update. Hepatol Int 2016;10:1-98. Crossref

Total knee arthroplasty (TKA) is the most effective and efficacious surgical method to improve pain and function for patients with end-stage osteoarthritis; however, TKA has been associated with substantial blood loss. In addition to visible blood loss from the surgical field and wound drainage, hidden blood loss occurred in patients undergoing TKA, which resulted in mean blood loss of 1.5 L.1 Therefore, perioperative blood transfusion was needed in up to 38% of patients undergoing TKA.2

Blood transfusion is not risk-free. Often, no adverse effects are encountered by patients who undergo blood transfusion. However, adverse effects occasionally occur, ranging from minor allergic reactions to blood-borne infection and potentially fatal acute immune haemolytic reaction. With the implementation of best international practices to warrant blood transfusion safety by the Blood Transfusion Service in Hong Kong, the transfusion risk has significantly decreased in the past two decades.3 However, absolute safety in transfusion cannot be achieved because of the window period for detecting infections, possibility of emerging infections, and potential human errors related to the process of transferring collected blood from donors to transfusion recipients. Notably, researchers in Hong Kong reported the first two cases (worldwide) of transmission of Japanese encephalitis virus, via blood transfusion to immunocompromised hosts, in 2018.4 In addition to the general risks associated with transfusion, blood transfusion has been independently associated with poor surgical outcome. Specifically, patients who underwent transfusion exhibited an eight-fold to 10-fold excess risk of adverse outcomes, defined as postoperative complications in the American College of Surgeons National Surgical Quality Improvement Project.5 With respect to total hip or knee arthroplasty, a dose-dependent relationship between transfusion and risk of surgical site infection was observed.6

With increasing understanding regarding the benefits and risks of blood transfusion, as well as alternative approaches for patients who experience blood loss, the concept of patient blood management (PBM) was developed. The World Health Organization defines PBM as ‘a patient-focused, evidence-based and systematic approach to optimise the management of patients and transfusion of blood products for quality and effective patient care. It is designed to improve patient outcomes through the safe and rational use of blood and blood products and by minimising unnecessary exposure to blood products…’.7 The three major components of PBM are as follows: (1) optimisation of the patient’s own blood mass; (2) minimisation of blood loss; and (3) optimisation of physiological tolerance to anaemia.8 This new standard of care is now well-established in some centres in the US, Austria, and Western Australia, as well as nationally in the Netherlands. However, PBM remains an uncommon practice in Asia.

We introduced the PBM programme for patients undergoing TKA in our institution, beginning in 2014. Various components were introduced gradually (in phases) from 2014 to 2018. The key measures of PBM in preoperative, intra-operative, and postoperative periods were fully implemented in 2018. To the best of our knowledge, our PBM programme is a pioneer PBM programme in Hong Kong. The aim of the present study was to assess the effectiveness of the multimodal PBM approach in our university-based centre.

This single-centre retrospective study was approved by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (Ref UW 19-600). The requirement for patient consent was waived by the review board. We retrospectively collected blood transfusion data regarding patients who underwent unilateral primary TKA in our centre from 1 January 2013 to 31 December 2018. Patients who underwent one-stage bilateral primary TKA or revision TKA were excluded from our study. Patients with acquired or congenital coagulopathy, as well as those currently taking anticoagulants, were included in our study; notably, these patients were at greater risk of perioperative blood loss and transfusion.

The primary outcome measure was the mean yearly transfusion rate, which was defined as the number of patients who received transfusion after TKA (during the same hospitalisation episode) divided by the number of patients who underwent TKA during the period from 1 January to 31 December; this result was multiplied by 100. The mean units of blood given per transfusion episode in 1 year was defined as the cumulative total number of blood units transfused to patients after TKA in 1 year divided by the number of transfusion episodes in that year. Transfusion data were retrieved from the Clinical Data Analysis and Reporting System, a database developed by the Hong Kong Hospital Authority for research purposes; this database contains medical information recorded by the Hong Kong Hospital Authority since 1993.

Secondary outcome measures were mean length of hospital stay after surgery, the rate of unexpected readmission through the Emergency Department after discharge from the hospital, the proportion of patients who had early prosthetic joint infection requiring revision surgery within 90 days after index surgery, and the 90-day mortality rate. These data and other patient demographic data (eg, age and sex), perioperative data (eg, American Society of Anesthesiologists [ASA] physical status), and preoperative haemoglobin level were retrieved from our patient records, as well as a local joint replacement registry.

Patient blood management was a relatively new concept in Hong Kong when it was first implemented in our institution in 2014. Initially, this programme was a standalone surgeon-initiated programme without external support; it was implemented in sequential stages based on PBM guidelines provided by the National Blood Authority in Australia.9 The sequential stages of PBM implementation in our centre are described in Table 1. The PBM strategies included modern surgical, anaesthetic, and perioperative care. In 2014, PBM was initiated with instructions regarding proper indications for transfusion, single-unit transfusion policy,10 and restrictive transfusion policy with transfusion triggered at haemoglobin ≤8 g/dL in healthy individuals.11 In 2015, the traditional practice of routine placement of a surgical drain during TKA (associated with a higher transfusion rate12) was stopped; the use of topical tranexamic acid (injection of 1 g tranexamic acid into knee joint at the end of the surgical procedure, shown to reduce postoperative blood loss and transfusion rate13) was implemented to reduce perioperative blood loss. In 2016, preoperative anaemia screening and optimisation were initiated via collaboration with haematologists. Patients with preoperative haemoglobin <11 g/dL were examined for causes of anaemia, in accordance with Network for Advancement of Transfusion Alternatives guidelines14; their erythropoiesis and preoperative haemoglobin characteristics were optimised before TKA was performed. For example, patients with iron-deficiency anaemia were checked for gastrointestinal blood loss and prescribed iron supplementation; their haemoglobin levels were rechecked after supplementation to confirm achievement of ≥11 g/dL before TKA. To further reduce intra-operative blood loss, combined intravenous tranexamic acid (15 mg/kg administered intravenously at the induction of anaesthesia) and topical tranexamic acid were implemented; these are reportedly effective for reduction of blood loss.15 In 2017, a more stringent restrictive transfusion policy was adopted with transfusion triggered at haemoglobin ≤7 g/dL in healthy individuals. Moreover, active warming (a modern anaesthetic technique used during intra-operative care) was implemented to avoid intra-operative hypothermia16; this hypothermia has been associated with greater volume of blood loss and the need for transfusion.17 18 By the beginning of 2018, all above PBM strategies were fully implemented.

In addition to PBM strategies, other changes in patient selection and perioperative management were implemented between 2013 and 2018. First, the degree of medical co-morbidities may have differed because of the establishment of another joint replacement centre in July 2016; this new centre provided joint replacement surgeries for patients with improved medical fitness, among patients with end-stage osteoarthritis in our hospital. Therefore, the medical co-morbidities of the patients in 2013 and 2018 were compared on the basis of ASA status. Second, because of technological advances in the design of TKA prostheses over time, there were differences in the numbers of modern-design TKA prostheses between 2013 and 2018; these modern-design TKA prostheses aimed to improve knee kinematics, rather than reduce transfusion rate. However, these prostheses were unlikely to bias our transfusion rate results, according to a prior assessment of factors predictive of transfusion rate in patients undergoing TKA.19

In 2013, no PBM strategies had been implemented in our institution, whereas all strategies had been fully implemented by 2018. The Chi squared test was used to compare the transfusion rate between patients who underwent TKA in 2013 and those who underwent TKA in 2018; differences with P<0.05 were considered statistically significant. As mentioned above, medical co-morbidities of the patients in 2013 and 2018 were compared on the basis of ASA status.

In total, 262 and 215 patients underwent primary TKA in our centre in 2013 and 2018, respectively (Table 2). There were no significant differences in mean age (2013: 72.17±9.76 years; 2018: 72.49±9.27 years, P=0.71) or sex (male:female) ratio (2013, 61:201; 2018, 63:152, P=0.14) between the groups. The preoperative haemoglobin level was also similar between the groups (2013: 12.77±1.42 g/dL; 2018: 12.89±1.42 g/dL, P=0.35). However, there was a significant difference in ASA distribution between the groups (P=0.03). There was a comparatively greater proportion of patients with ASA Grade II status in 2018 (2013: 57.6%; 2018: 68.8%).

The primary outcome was the mean transfusion rate in 1 year. There was a significant difference in the mean transfusion rate after primary TKA between 2013 and 2018 (2013: 31.3%; 2018: 1.9%, P<0.001); however, there was no significant difference in the mean units of blood transfused per transfusion episode (2013: 1.62±0.78; 2018: 1.00±0.00, P=0.12). Moreover, the mean annual transfusion rate after primary TKA exhibited stepwise reduction as PBM strategies were implemented during the period from 2014 to 2018 (Fig).

Regarding secondary outcomes, the mean length of hospitalisation was significantly lower in 2018 (2013: 14.49±8.10 days; 2018: 8.77±10.14 days, P<0.001). However, there was no difference in the unexpected readmission rate through the Emergency Department (2013: 3.8%; 2018: 3.7%, P=0.96), the proportion of patients who exhibited early prosthetic joint infection within 90 days after index surgery (2013: 0.4%; 2018: 0%, P=0.36), or the proportion of patients with 90-day mortality (2013: 0%; 2018: 0.5%, P=0.27).

Blood transfusion is a life-saving therapy, but is a limited resource. In recent years, there have been recurrent blood shortages in Hong Kong, and the Hong Kong Red Cross has issued an urgent appeal for blood donors on several occasions.20 21 22 23 The amount of blood stored in blood banks is determined by supply and demand. To increase blood supply, additional blood donors are needed. The Annual Report of Hong Kong Red Cross in 2018/2019 revealed that 4% of blood donors were aged >60 years, and the largest group of blood donors were aged 41-50 years (23.7% of donors).24 With the increasing number of older people in Hong Kong, more blood donors are needed from older age-groups. In addition, healthcare professionals should be judicious in prescribing transfusion, and should consider methods to minimise transfusion. Our study demonstrated the effectiveness of implementing PBM in our centre. In addition to comparing the mean transfusion rate in 2013 and 2018, this study included an audit of the mean annual transfusion rate after primary TKA from 2014 to 2018 (Fig).

Globally, PBM is not a new concept. In May 2010, the World Health Organization formally recognised the importance of PBM and recommended its use to the 193 member states.25 Subsequently, PBM has been successfully implemented in Western countries, especially Australia26 and the US.27 The Asia-Pacific PBM Expert Consensus Meeting Working Group assessed the status of PBM in Asia.28 In Singapore, PBM was implemented nationally, beginning in 2013. The Ministry of Health and Blood Services Group actively promoted PBM at public hospitals in the first few years; regular national audits of PBM-related efforts have been performed since 2017 to promote appropriate use of red blood cell transfusion and implementation of preoperative anaemia screening for elective surgeries. Patient blood management programmes were successfully implemented in National University Hospital and Singapore General Hospital.28 29 In Korea, PBM was implemented through a professional initiative by the Korean Research Society of Transfusion Alternatives of the Republic of Korea in 2006; the Korean Patient Blood Management Research Group was formed to promote greater PBM use in 2016. The efforts of the Korean Patient Blood Management Research Group resulted in achievement of several PBM milestones in 2016.28 Notably, PBM was included in the Korean Transfusion Guidelines; a new steering committee was also formed, comprising leading physicians from various specialties. Patient blood management was successfully implemented in a number of Korean hospitals, which led to a reduction in transfusion rate.30 31 32 In Malaysia, PBM has been promoted at the local hospital level.28 In the Department of Maternal and Fetal Medicine at the Sultan Haji Ahmad Shah Hospital, women at high risk of anaemia are screened for iron deficiency anaemia in early pregnancy; iron-deficient women are provided oral or intravenous iron supplementation.

At our institution, PBM was first implemented in 2014 as a surgeons’ initiative in the Division of Joint Replacement Surgery. In addition to good surgical techniques, good perioperative care is an important determinant of surgical outcomes. Patient blood management is a component of our overall perioperative management protocol in the modern enhanced recovery after surgery programme. With implementation of PBM strategies and measures in the enhanced recovery after surgery programme, the length of hospitalisation was shortened in 2018, compared with 2013. Despite the shorter length of stay, there were no differences in the unexpected readmission rate through the Emergency Department, the proportion of patients who had early prosthetic joint infection within 90 days after index surgery, or the proportion of patients who had 90-day mortality.

To the best of our knowledge, there have been few studies regarding PBM in patients undergoing TKA in Hong Kong. In 2015, Lee et al33 reported their pioneering experience with implementation of PBM in patients undergoing TKA; their PBM protocols included typing and screening only for patients with preoperative haemoglobin of <11 g/dL, and restrictive transfusion triggered at haemoglobin 8 g/dL. When they compared outcomes before and after introduction of the PBM programme, the transfusion rate (before: 10.3%; after: 3.1%, P=0.046) and cross-match rate (before: 100%; after: 3.1%, P<0.001) both decreased. We implemented PBM in our institution, beginning in 2014. Modern surgical, anaesthetic, and perioperative techniques in PBM were gradually introduced from 2014 to 2018. Our PBM protocols are more comprehensive than those of Lee et al, because our protocols were designed in accordance with the PBM guidelines provided by the National Blood Authority in Australia.9 Therefore, our transfusion rate in TKA in 2018 decreased to 1.9%.

Prosthetic joint infection is a severe complication in arthroplasty; affected patients often require revision surgery. Notably, patients who underwent treatment for prosthetic joint infection exhibited a significant, independent risk of increased mortality, due to the direct adverse effect of infection and the indirect effect of poor underlying health condition.34 In a recent meta-analysis involving 21 770 patients who underwent total hip and knee arthroplasty, patients who received allogeneic blood transfusion had a significantly greater risk of surgical-site infection (pooled odds ratio: 1.71, P=0.02).35 Recently, a dose-dependent relationship was observed between allogeneic transfusion and surgical site infection after total hip or knee arthroplasty.6 Therefore, PBM was expected to reduce the incidence of surgical site infection in our centre. However, because of the relatively small sample size in our study and the relatively low incidence of prosthetic joint infection (approximately 1%), a difference in the incidence of prosthetic joint infection between 2013 and 2018 could not be identified. When PBM was fully implemented in our centre (2018), the transfusion rate after primary TKA was 1.9%; this was comparable to international reported values. Specifically, when PBM strategies were implemented in the US, the transfusion rates were approximately 4.5%.36 When PBM is implemented by high-volume surgeons with an eight-step checklist to reduce bleeding, the transfusion rate after TKA could be as low as 0.0044%.37 Therefore, a transfusion rate of 0% is achievable.

As the transfusion rate decreased in patients undergoing TKA, there were also benefits to the healthcare system. Blood transfusion involves many costs associated with blood transfer from donors to recipients (eg, collection, screening, storage, transportation, and prescription of donated blood). We do not have data regarding the cost of packed red blood cells in Hong Kong; however, the cost was estimated to be approximately 1130 USD/unit in a study conducted in the US.36 Therefore, reduced transfusions through implementation of PBM can result in lower healthcare expenditures, which are of considerable importance because of the increasing demand for TKA among the aging population in Hong Kong.

There were some limitations in this study. First, it was a retrospective study; thus, compliance with PBM strategies could not be fully verified. However, as each strategy was introduced throughout the course of the study, there was gradual reduction in the transfusion rate. Therefore, compliance with the strategies was presumably optimal. Second, because different strategies were implemented successively, the strategies with the greatest contribution to the reduced transfusion rate could not be identified. Third, because this was not a prospective randomised placebo-controlled interventional trial, a causal relationship between PBM strategies and reduction in transfusion rate could not be established. However, our study provided an assessment of real-world implementation of PBM strategies within a large hospital; thus, it comprises pioneering research in Hong Kong. Fourth, some potential cofounding factors may not have been identified or controlled in the present analysis. For example, the type of prosthesis used was not analysed as a separate factor. However, preoperative haemoglobin levels (the most significant predictor of blood transfusion19) were compared between both groups. To the best of our knowledge, there remains minimal relevant literature regarding the effect of TKA prosthesis on the transfusion rate.

In conclusion, our results demonstrated the effectiveness of PBM implementation on transfusion rate in patients undergoing TKA. From 2014 to 2018, there was a stepwise reduction in transfusion rate after TKA; this was similar to findings in previously published research. This is one of the few studies in Hong Kong to review PBM in surgical practice. Although we focused on patients undergoing TKA, the principles of PBM could be useful for other medical or surgical specialties.

All authors contributed to the concept or design of the study, acquisition of data, analysis or interpretation of data, drafting of the manuscript, and critical revision of the manuscript 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.

We thank Dr CK Lee, Chief Executive and Medical Director, Hong Kong Red Cross Blood Transfusion Service, Hong Kong, for providing advice on the patient blood management programme. We also acknowledge and express our gratitude to the following departments of our hospital for the support in this multidisciplinary project: Nursing Division, Department of Orthopaedics and Traumatology; Operation Theatre Services; and Blood Transfusion Committee.

This research has been presented in part as an oral presentation at the Hong Kong Hospital Authority Convention 2018.

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 Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (Ref UW 19-600). The requirement for patient consent was waived by the review board.

1. Sehat KR, Evans R, Newman JH. How much blood is really lost in total knee arthroplasty? Correct blood loss management should take hidden loss into account. Knee 2000;7:151-5. Crossref

2. Cushner FD, Friedman RJ. Blood loss in total knee arthroplasty. Clin Orthop Relat Res 1991;(269):98-101. Crossref

3. Lee CK. Risk minimization in transfusion transmitted infection in Hong Kong [thesis]. University of Hong Kong; 2017.

4. Cheng VC, Sridhar S, Wong SC, et al. Japanese encephalitis virus transmitted via blood transfusion, Hong Kong, China. Emerg Infect Dis 2018;24:49-57. Crossref

5. Ferraris VA, Hochstetler M, Martin JT, Mahan A, Saha SP. Blood transfusion and adverse surgical outcomes: the good and the bad. Surgery 2015;158:608-17. Crossref

6. Everhart JS, Sojka JH, Mayerson JL, Glassman AH, Scharschmidt TJ. Perioperative allogeneic red blood-cell transfusion associated with surgical site infection after total hip and knee arthroplasty. J Bone Joint Surg Am 2018;100:288-94. Crossref

7. World Health Organization. WHO Global Forum for Blood Safety: patient blood management. Available from: https://www.who.int/bloodsafety/events/gfbs_01_pbm/en/ Accessed 5 Jan 2020.

8. Isbister JP. The three-pillar matrix of patient blood management-an overview. Best Pract Res Clin Anaesthesiol 2013;27:69-84. Crossref

9. National Blood Authority Australia. Patient Blood Management Guidelines. Available from: https://www.blood.gov.au/pbm-guidelines. Accessed 9 Jan 2020.

10. National Blood Authority Australia. Single Unit Transfusion Guide. Available from: https://www.blood.gov.au/single-unit-transfusion. Accessed 9 Jan 2020.

11. Hardy JF. Current status of transfusion triggers for red blood cell concentrates. Transfus Apher Sci 2004;31:55-66. Crossref

12. Zhang Q, Liu L, Sun W, et al. Are closed suction drains necessary for primary total knee arthroplasty? A systematic review and meta-analysis. Medicine (Baltimore) 2018;97:e11290. Crossref

13. Panteli M, Papakostidis C, Dahabreh Z, Giannoudis PV. Topical tranexamic acid in total knee replacement: a systematic review and meta-analysis. Knee 2013;20:300-9. Crossref

14. Goodnough LT, Maniatis A, Earnshaw P, et al. Detection, evaluation, and management of preoperative anaemia in the elective orthopaedic surgical patient: NATA guidelines. Br J Anaesth 2011;106:13-22. Crossref

15. Shang J, Wang H, Zheng B, Rui M, Wang Y. Combined intravenous and topical tranexamic acid versus intravenous use alone in primary total knee and hip arthroplasty: A meta-analysis of randomized controlled trials. Int J Surg 2016;36:324-9. Crossref

16. Benson EE, McMillan DE, Ong B. The effects of active warming on patient temperature and pain after total knee arthroplasty. Am J Nurs 2012;112:26-33. Crossref

17. Rajagopalan S, Mascha E, Na J, Sessler DI. The effects of mild perioperative hypothermia on blood loss and transfusion requirement. Anesthesiology 2008;108:71-7. Crossref

18. Schmied H, Kurz A, Sessler DI, Kozek S, Reiter A. Mild hypothermia increases blood loss and transfusion requirements during total hip arthroplasty. Lancet 1996;347:289-92. Crossref

19. Boutsiadis A, Reynolds RJ, Saffarini M, Panisset JC. Factors that influence blood loss and need for transfusion following total knee arthroplasty. Ann Transl Med 2017;5:418. Crossref

20. Ernest K. Hong Kong Red Cross issues urgent appeal for blood donors as supplies dwindle. South China Morning Post [newspaper on the internet]. 2017 May 4: Health & Environment. Available from: https://www.scmp.com/news/hong-kong/health-environment/article/2092982/hong-kong-red-cross-issues-urgent-appeal-blood. Accessed 10 Mar 2020.

21. City urged to donate blood amid severe shortage. The Standard [newspaper on the internet]. 2017 Aug 11: Local. Available from: https://www.thestandard.com.hk/breaking-news/section/3/94995/City-urged-to-donate-blood-amid-severe-shortage. Accessed 10 Mar 2020.

22. The Government of the Hong Kong Special Administrative Region. Urgent appeal for blood donation [press release]. 2018 Jan 9. Available from: https://www.info.gov.hk/gia/general/201801/09/P2018010900651.htm. Accessed 10 Mar 2020.

23. Coronavirus fears drain Hong Kong's blood banks. Radio Television Hong Kong [newspaper on the internet]. 2020 Feb 12. Available from: https://news.rthk.hk/rthk/en/component/k2/1508174-20200212.htm. Accessed 10 Mar 2020.

24. The Annual Report of Hong Kong Red Cross 2018/2019. Available from: https://www.redcross.org.hk/en/publications/annual_reports.html. Accessed 15 Jan 2020.

25. World Health Assembly. Resolution WHA63.12: Availability, safety and quality of blood products. Available from: https://apps.who.int/medicinedocs/documents/s19998en/s19998en.pdf. Accessed 10 Jan 2020.

26. Farmer SL, Towler SC, Leahy MF, Hofmann A. Drivers for change: Western Australia Patient Blood Management Program (WA PBMP), World Health Assembly (WHA) and Advisory Committee on Blood Safety and Availability (ACBSA). Best Pract Res Clin Anaesthesiol 2013;27:43-58. Crossref

27. Whitaker B, Rajbhandary S, Kleinman S, Harris A, Kamani N. Trends in United States blood collection and transfusion: results from the 2013 AABB Blood Collection, Utilization, and Patient Blood Management Survey. Transfusion 2016;56:2173-83. Crossref

28. Abdullah HR, Ang AL, Froessler B, et al. Getting patient blood management Pillar 1 right in the Asia-Pacific: a call for action. Singapore Med J 2019 May 2. Epub ahead of print. Crossref

29. D E H O, Hadi F, Stevens V. Health economic evaluation comparing iv iron ferric carboxymaltose, iron sucrose and blood transfusion for treatment of patients with iron deficiency anemia (Ida) in Singapore. Value Health 2014;17:A784. Crossref

30. Lee ES, Kim MJ, Park BR, et al. Avoiding unnecessary blood transfusions in women with profound anaemia. Aust N Z J Obstet Gynaecol 2015;55:262-7. Crossref

31. Yoon HM, Kim YW, Nam BH, et al. Intravenous ironsupplementation may be superior to observation in acute isovolemic anemia after gastrectomy for cancer. World J Gastroenterol 2014;20:1852-7. Crossref

32. Na HS, Shin SY, Hwang JY, Jeon YT, Kim CS, Do SH. Effects of intravenous iron combined with low-dose recombinant human erythropoietin on transfusion requirements in iron-deficient patients undergoing bilateral total knee replacement arthroplasty. Transfusion 2011;51:118-24. Crossref

33. Lee QJ, Mak WP, Yeung ST, Wong YC, Wai YL. Blood management protocol for total knee arthroplasty to reduce blood wastage and unnecessary transfusion. J Orthop Surg (Hong Kong) 2015;23:66-70. Crossref

34. Zmistowski B, Karam JA, Durinka JB, Casper DS, Parvizi J. Periprosthetic joint infection increases the risk of one-year mortality. J Bone Joint Surg Am 2013;95:2177-84. Crossref

35. Kim JL, Park JH, Han SB, Cho IY, Jang KM. Allogeneic blood transfusion is a significant risk factor for surgical-site infection following total hip and knee arthroplasty: a meta-analysis. J Arthroplasty 2017;32:320-5. Crossref

36. Moskal JT, Harris RN, Capps SG. Transfusion cost savings with tranexamic acid in primary total knee arthroplasty from 2009 to 2012. J Arthroplasty 2015;30:365-8. Crossref

37. Lindman IS, Carlsson LV. Extremely low transfusion rates: contemporary primary total hip and knee arthroplasties. J Arthroplasty 2018;33:51-4. Crossref

Chemotherapy, radiotherapy, certain medications for cancer, and some rheumatological and haematological diseases are gonadotoxic, which can jeopardise patients’ fertility, particularly that of young cancer survivors.1 With advancements in treatment, the 5-year survival rate of patients with cancer in childhood and adolescence has increased to over 80%.2 Thus, improving their quality of life and reducing their risk of infertility is an important aspect of their management plan.

International clinical guidelines, including the American Society of Clinical Oncology (ASCO),3 European Society for Medical Oncology (ESMO),4 and Royal College of Radiologists guidelines,5 suggest discussion of fertility preservation (FP) with patients of childhood and reproductive age during the course of cancer therapy. However, a study showed that fewer than half of oncologists routinely refer their patients to reproductive endocrinologists, and even fewer oncologists follow the guidelines, despite their willingness to discuss infertility in relation to cancer therapy.6 Moreover, a lack of awareness and related training about FP among clinicians may cause underutilisation of FP services worldwide, including in Canada, the US, and Hong Kong.7 8 9 In Hong Kong, clinicians and patients alike may consider FP as an expensive, privatised option without subsidisation.

In a previous cross-sectional study that evaluated the awareness, attitudes, and knowledge of FP among clinicians across different specialties in Hong Kong, only 45.6% of clinicians were familiar with FP.10 As current medical students will become our future clinicians, it is important to assess their level of understanding and awareness of fertility and FP, as this may greatly influence their future practice and consideration of appropriate interventions to improve affected patients’ outcomes. Cross-sectional studies have been conducted to assess college students’ awareness and knowledge of fertility in Canada,11 Serbia,12 and the US.13 These studies in Western populations showed knowledge inadequacy about age-related fertility decline and FP. However, to the best of our knowledge, there are no studies in a Chinese population that have aimed to investigate the awareness, knowledge gaps, and attitudes of medical students regarding fertility and FP. Therefore, the aims of this study were to evaluate the awareness of, attitudes towards, and knowledge regarding FP among Chinese medical students in Hong Kong.

This was a cross-sectional survey conducted from November 2018 to June 2019. The study population consisted of Chinese medical students from The Chinese University of Hong Kong. Chinese undergraduate students aged ≥18 years studying in the Medicine programme and capable of communicating in English were included. Those who were non-Chinese, under age 18 years, incapable of communicating in English, and those who refused to join the study were excluded.

Eligible participants were invited to complete a self-administered online survey. The online survey was sent to participants by internal mass email and social networking applications. Snowball sampling was done by encouraging medical students to send the online survey to their classmates to boost the response rate.

The online survey was developed on an electronic form (MyCUform). The self-administered survey included a brief explanation and was comprised of four parts: (1) baseline demographic data (Table 1); (2) awareness of FP; (3) knowledge about FP (Table 2); multiple choice questions consisting of five or six options on knowledge about fertility and FP; and (4) attitudes towards FP (Tables 3 and Table 4). It consisted of 38 questions and took approximately 15 minutes to finish. The survey was developed after reviewing the literature.14 15 16 It was assessed for logical validity by three physicians in the Department of Obstetrics and Gynaecology, who reviewed the accuracy of the contents. The survey was also piloted on a small number of doctors and medical students for content clarity and modified accordingly to incorporate the pilot participants’ feedback. The final version was then administered to the full study’s participants.

Statistical analysis was performed using SPSS (Windows version 24.0; IBM Corp, Armonk [NY], US). Continuous data were described as means, standard deviations, and percentiles. Categorical data were summarised as frequencies and percentiles. Subgroup analyses were performed between gender and year groups, with the medical students split into junior (year 1-4) and senior (year 5-6) year groups. Junior medical students had not completed the clinical attachments of the obstetrics and gynaecology module rotation (OB-GYN), whereas senior medical students had finished the OB-GYN module in year 5. The categorical data were tested by Pearson’s Chi squared test or Fisher’s exact test to check for significant differences between groups. Results with P values of <0.05 were considered statistically significant.

Approximately 700 Chinese medical students were approached through social networking applications and email, and 243 completed the online survey (response rate: 34.7%). Table 1 summarises the participants’ demographic data. In terms of age distribution, there was no significant difference (P=0.597) between the two groups.

Overall, 71.2% (n=173) of respondents were aware of FP strategies. Despite this, 77.8% (n=189) of respondents were not familiar with any clinics or specialists who provide FP services. Senior students showed better awareness than junior students of the above items (P<0.001). The majority (86.8%, n=211) had not heard of any regulations related to FP. Gamete and embryo freezing were the most well-known FP methods, with female students being significantly more aware than male students of those methods (P<0.003).

The majority (71.2%, n=173) responded that they would discuss the option of FP with their patients as future doctors, even if the treatment had a <30% chance of causing infertility. If a treatment had a ≥70% chance of causing infertility, nearly all (95.5%, n=232) students would discuss FP.

Table 2 shows the results of the questions that address knowledge about fertility and FP. Overall, there were no gender differences besides the response regarding the age range of a woman’s significant fertility decline: proportionally more female students answered that item correctly (45.7% vs 38.8%; P<0.05).

Knowledge regarding FP was mainly acquired from electronic media (58.4%, n=142), medical school (57.6%, n=140), and medical professionals (38.7%, n=94). More senior students than junior students acquired fertility knowledge through medical school education (81.4% vs 30.7%; P<0.0001) and medical professionals (54.3% vs 21.1%; P<0.0001).

Most of the responding students (86.4%, n=210) wished to know more about FP, with 80.7% (n=196) of the students agreeing that there is a need to incorporate FP material into the medical curriculum.

Regarding attitudes, 94.2% (n=229) of medical students agreed that establishing one or two dedicated clinics or centres for FP counselling is necessary. Subgroup analysis indicated that more male medical students would like to have two dedicated centres (75.0% vs 68.2%, P=0.021). More senior than junior students agreed that FP should be available solely as a public service (89.1% vs 71.9%; P=0.001). Overall, 97.9% (n=238) of participants thought that practice guidelines for FP should be required. More than half of respondents (59.7%, n=145) agreed that there should be an age limit for FP. More female than male students agreed to set an age limit for FP (65.7% vs 51.5%, P=0.025).

Among various factors considered by medical students to determine whether to recommend FP to patients, the desire to have children (51.0%, n=124), the prognosis of cancer or a medical condition (23.5%, n=57), and time available before gonadotoxic treatment (7.4%, n=18) were the most likely considerations.

Most participants (80.2%, n=195) responded that the government should subsidise FP in patients undergoing gonadotoxic treatment, with senior students expressing stronger support for subsidisation than junior students (86.0% vs 73.7%; P<0.02). More than half of the responding students agreed that the government should subsidise 30% to 50% of the cost of FP procedures including sperm (79.8%, n=194) and egg freezing (80.2%, n=195), and in vitro fertilisation (IVF) [67.9%, n=165]. More senior than junior students thought that the government should subsidise >70% of the cost of IVF (34.1% vs 20.1%; P<0.02).

Table 3 shows the respondents’ attitudes towards elective and medical gamete and embryo freezing. More male students than female students expressed agreement with FP provision to men because of having no suitable partner (34.0% vs 22.1%; P<0.05). Table 4 illustrates attitudes towards family planning among the responding medical students. More male than female students were determined not to delay their family planning (16.5% vs 7.9%, P=0.037).

With the advancement of technology, FP has become increasingly effective at enabling patients who have undergone gonadotoxic treatment to raise families. However, as shown by many previous studies on medical students’ understanding of FP, this study reflects an overall worldwide tendency towards a lack of awareness and knowledge about fertility and FP.11 12 13 To the best of our knowledge, this is the first study that has aimed to investigate the awareness, knowledge, and attitudes of Chinese medical students regarding fertility and FP, as well as identifying their knowledge gaps in the subject.

The majority of Chinese medical students in Hong Kong have heard of at least one FP strategy. Senior students had significantly greater awareness than junior students of FP, reproductive techniques, and the availability of specialty clinics, likely reflecting the knowledge and exposure gained during the clinical OB-GYN module in the fifth year of study. This suggests that appropriate and timely education can improve medical students’ awareness of FP.

Previous studies have shown that even brief educational interventions about FP to medical students and house staff have potential benefits.13 To provide quality service, ASCO and ESMO guidelines suggest that patients with cancer be informed of their potential fertility decline and referred to FP services after treatment.3 4 Most students responded that they would refer patients to FP services even if treatment had only a low risk of infertility, but they were not familiar with the actual practice of FP, including relevant regulations and referral methods. Downloadable fact sheets on the effects of cancer treatment on fertility, available options for FP, and a list of service providers with reference costs stated should be available and accessible for proper patient education and counselling.

There are several misconceptions among medical students regarding knowledge about fertility and FP. The responding medical students tended to overestimate the age of the female fertility peak and the success rate of IVF and underestimate the risk of infertility in women at age 45 years (Table 2). Overestimation of female fertility has also been observed in overseas studies.11 12 13 17 18 Medical students have better knowledge about female fertility than male fertility. This could be explained by the fact that there have been many more studies about the concern of female fertility decline with age. As sex education programmes at the secondary and university levels mainly emphasise pregnancy prevention education rather than infertility awareness, there may be a lack of knowledge about the impact of ageing on fertility among women.11 Fertility-related knowledge should be included in the undergraduate medical curriculum and ideally be integrated in high school education as well to enhance public education on this topic. Beyond patient care, medical students’ inadequate knowledge about this topic also has great implications for their future careers and personal lives. They may delay their own family planning for career reasons without sufficient consideration of their impending fertility decline or may have a false sense of security regarding the success rate of IVF. Therefore, more education about fertility is required, which was also supported by the respondents of our study.

Electronic media play a significant role in the promotion of FP, especially among junior students. In contrast, medical school was the main source of FP knowledge for senior students, followed by medical professionals and electronic media. This is consistent with findings in American medical students and house staff.13 Media reports of female celebrities undergoing FP procedures, particularly egg freezing, could explain the high prevalence of student familiarity with this procedure and their lack of knowledge about other, less popular options.19 Knowledge acquisition through both the medical curriculum and electronic media have proven to be significant, particularly in people who have not received formal education about FP. This can be applied to the general public, although it poses a risk of promoting misinformation: people could be misled into having a false sense of security regarding successful childbearing late in one’s reproductive life through the use of FP methods.11 Therefore, caution should be used to ensure that materials released through the media are accurate.

Most of the responding medical students would first consider the patient’s desire when referring patients for FP, while 41.3% of physicians would consider the patient’s prognosis first.10 This finding could be explained by the medical students’ lack of medical practice experience and doctors’ awareness of resource limitations, as it has been shown that junior doctors make judgements mostly based on their own assumptions, compared with the experience-based judgements made by senior doctors.20

The private sector is the only current provider of FP services in Hong Kong. More male than female medical students who responded to our survey had positive attitudes towards the establishment of related facilities. The cost of freezing gametes is at least US$7800 (US$1 to HK$7.8),21 and the median monthly household income of a 1-person family in Hong Kong is US$1282.22 There are currently no gamete freezing subsidy programmes available. Such high costs are unaffordable to many patients, and especially patients with cancer are already financially burdened by their current treatment. In this regard, most medical students agreed that the government should subsidise FP services to patients undergoing gonadotoxic treatment, with more than half agreeing that 30% to 50% is a reasonable subsidy proportion.

There has been an increasing trend towards elective FP in recent years, which has opened up a debate about the ethicality of FP for non-medical indications and whether FP should only be provided for medical reasons. Our study reflected overwhelming support for FP for medical reasons. However, the responding medical students’ opinions were mixed regarding elective freezing for non-medical indications. Only approximately half of the respondents agreed that FP should be provided to people because they have not found a suitable partner, or because they delay family planning for the sake of career development. Despite the mixed responses regarding elective freezing, Chinese medical students from Hong Kong were more supportive of elective freezing than undergraduates and medical students from the US.23 This finding may be driven by the higher cost of FP in the US compared with that in Hong Kong.23 Indeed, elective freezing has gained popularity in recent years. Still, the greater acceptance of FP for medical reasons echoes the healthcare-related perceptions and expectations of patients in Hong Kong: fertility is not a necessity, and FP healthcare is considered a luxury as opposed to a necessity for immediate physical well-being.

This study’s participants were more supportive of elective freezing for women than men. This is likely because of their awareness that women’s fertility declines relatively earlier than that of men. Moreover, women in Hong Kong are expected to have a predominantly domestic and childrearing role within their families,16 largely because of deeply rooted traditional Chinese familial constructs in which women tend to take on homemaking roles. However, the male medical students who responded were also concerned about their own family planning. This study’s results show that more male than female medical students were determined not to delay their family planning or stated their intent to have gametes frozen because they had no suitable partner.

Most medical students agree that their future fertility is important to them. Over half of the responding medical students stated their intent to delay family planning for career development, and the majority plan to have their first child between age 30 and 34 years (the average age of residency completion is 29 years). However, few responded that they would consider undergoing FP treatments. This paradoxical response is consistent with the results of another study on Hong Kong university students that showed low inclination to seek help in the event of fertility problems compared with Western counterparts.16 The taboo of childlessness in Chinese culture may be another reason why planning and conversation are discouraged in the event of infertility.16 Thus, students may view FP techniques as drastic and unconventional, preferring natural conception. As discussed earlier, the lack of correct knowledge about age-related female fertility decline could also lead to such results. Medical training is long, potentially delaying doctors from starting families during their most fertile years. Education is beneficial not only to future patients, but also to current and future physicians’ quality of life.

As a form of convenience sampling, the online survey method was chosen to efficiently distribute the survey throughout the large population of medical students. However, this inherently came with limitations, including low response rate, small sample size, selection bias, and the inability to characterise non-respondents. Ideally, students of other medical schools in Hong Kong should be included to make our study more comprehensive; however, this would be logistically difficult. Because of constraints on advertising, the study consisted of slightly more senior than junior students and more female than male respondents. Moreover, our study did not specifically ask about the educational background of students that may have influenced their knowledge. However, there was no significant difference in the age distribution between the junior and senior student groups. Information bias may have also affected the responses from students who had not completed the OB-GYN module.

In conclusion, this study revealed important aspects of FP from the perspective of Chinese medical students in Hong Kong. In particular, we highlighted that awareness and knowledge of FP are limited among medical students. There is a strong worldwide need to increase education about and exposure of FP in the curriculum to improve medical students’ knowledge. More involvement from the public sector and enhancement of facilities in terms of service provision and financial support for FP services are also needed. Given the variety of perspectives on childbearing among different cultures, this should be taken into account when doctors consider FP options for their patients and themselves. Only by increasing awareness and knowledge can more accepting attitudes towards FP arise among our doctors, allowing for better clinical outcomes and quality of life for future patients.

Concept or design: JPW Chung, EYL Ng.
Acquisition of data: All authors.
Analysis or interpretation of data: EYL Ng.
Drafting of the article: All authors.
Critical revision for important intellectual content: JPW Chung.

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.

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

The authors would like to thank the medical students at The Chinese University of Hong Kong for their kind participation in the study.

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

Ethical approval for the study was obtained from the Survey and Behavioural Research Ethics Committee (Ref SBRE-18-168).

1. Husson O, Huijgens PC, van der Graaf WTA. Psychosocial challenges and health-related quality of life of adolescents and young adults with hematologic malignancies. Blood 2018;132:385-92. Crossref

2. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2019. CA Cancer J Clin 2019;69:7-34. Crossref

3. Oktay K, Harvey BE, Partridge AH, et al. Fertility preservation in patients with cancer: ASCO clinical practice guideline update. J Clin Oncol 2018;36:1994-2001. Crossref

4. Peccatori FA, Azim HA Jr, Orecchia R, et al. Cancer, pregnancy and fertility: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol 2013;24 Suppl 6:vi160-70. Crossref

5. The Royal College of Physicians, The Royal College of Radiologists, The Royal College of Obstetrics and Gynaecology. The effects of cancer treatment on reproductive functions: guidance on management 2007. Available from: https://www.rcr.ac.uk/system/files/ publication/field_publication_files/Cancer_fertility_ effects_Jan08.pdf. Accessed 3 Jan 2020.

6. Köhler TS, Kondapalli LA, Shah A, Chan S, Woodruff TK, Brannigan RE. Results from the Survey for Preservation of Adolescent Reproduction (SPARE) study: gender disparity in delivery of fertility preservation message to adolescents with cancer. J Assist Reprod Genet 2011;28:269-77. Crossref

7. Chung JP, Haines CJ, Kong GW. Sperm cryopreservation for Chinese male cancer patients: a 17-year retrospective analysis in an assisted reproductive unit in Hong Kong. Hong Kong Med J 2013;19:525-30. Crossref

8. Flink DM, Sheeder J, Kondapalli LA. A review of the oncology patient’s challenges for utilizing fertility preservation services. J Adolesc Young Adult Oncol 2017;6:31-44. Crossref

9. Liu SS, Chan KY, Leung RC, et al. Prevalence and risk factors of human papillomavirus (HPV) infection in Southern Chinese women—a population-based study. PLoS One 2011;6:1-7. Crossref

10. Chung JP, Lao TT, Li TC. Evaluation of the awareness of, attitude to, and knowledge about fertility preservation in cancer patients among clinical practitioners in Hong Kong. Hong Kong Med J 2017;23:556-61. Crossref

11. Bretherick KL, Fairbrother N, Avila L, Harbord SH, Robinson WP. Fertility and aging: do reproductive-aged Canadian women know what they need to know? Fertil Steril 2010;93:2162-8. Crossref

12. Vujčić I, Radičević T, Dubljanin E, Maksimović N, Grujičić S. Serbian medical students’ fertility awareness and attitudes towards future parenthood. Eur J Contracept Reprod Health Care 2017;22:291-7. Crossref

13. Anspach Will E, Maslow BS, Kaye L, Nulsen J. Increasing awareness of age-related fertility and elective fertility preservation among medical students and house staff: a pre- and post-intervention analysis. Fertil Steril 2017;107:1200-5.e1. Crossref

14. Hickman LC, Fortin C, Goodman L, Liu X, Flyckt R. Fertility and fertility preservation: knowledge, awareness and attitudes of female graduate students. Eur J Contracept Reprod Health Care 2018;23:130-8. Crossref

15. Schwartz D, Mayaux MJ, Spira A, et al. Semen characteristics as a function of age in 833 fertile men. Fertil Steril 1983;39:530-5. Crossref

16. Chan CH, Chan TH, Peterson BD, Lampic C, Tam MY. Intentions and attitudes towards parenthood and fertility awareness among Chinese university students in Hong Kong: a comparison with Western samples. Hum Reprod 2015;30:364-72.Crossref

17. Virtala A, Vilska S, Huttunen T, Kunttu K. Childbearing, the desire to have children, and awareness about the impact of age on female fertility among Finnish university students. Eur J Contracept Reprod Health Care 2011;16:108-15. Crossref

18. Yu L, Peterson B, Inhorn MC, Boehm JK, Patrizio P. Knowledge, attitudes, and intentions toward fertility awareness and oocyte cryopreservation among obstetrics and gynecology resident physicians. Hum Reprod 2016;31:403-11. Crossref

19. Moss R. Egg freezing is the fastest growing fertility treatment—here’s what you need to know. 2019. Available from: https://www.huffingtonpost.co.uk/ entry/egg-freezing-is-the-fastest-growing-fertilitytreatment- heres-what-you-need-to-know-about-it_ uk_5cd1b93ae4b04e275d50f3e0. Accessed 3 Jan 2020.

20. Nilsson MS, Pilhammar E. Professional approaches in clinical judgements among senior and junior doctors: implications for medical education. BMC Med Educ 2009;9:25. Crossref

21. Healthy Matters. Your Complete guide to egg freezing in Hong Kong. Jun 2019. Available from: https://www. healthymatters.com.hk/guide-egg-freezing-hong-kongwomen- need-know-preserving-fertility/. Accessed 3 Jan 2020.

22. Census and Statistics Department, Hong Kong SAR Government. Table E032: domestic households by household size and monthly household income (excluding foreign domestic helpers) [Table 9.4A in Quarterly Report on General Household Survey]. Population estimates. 2019. Available from: https://www.censtatd.gov.hk/hkstat/ sub/sp150.jsp?productCode=D5250036. Accessed 3 Jan 2020.

23. Mahesan A, Mundt S, Smith L, Stadtmauer L. Knowledge and attitudes regarding elective oocyte cryopreservation in medical students and undergraduates. Fertil Steril 2017;108:e109-10. Crossref

Despite advances in paediatric critical care medicine, death remains inevitable in some instances, due to various aetiologies.1 2 In paediatric critical care medicine settings, patients who would have otherwise died may be kept ‘alive’ by advanced cardiovascular and ventilatory support. Some patients on cardiopulmonary support may experience irreversible cessation of all brain function, which is regarded as brain death (BD).3 4 5 6 Because BD or cardiopulmonary death is equivalent to death, there is no obligation for the physician to provide further futile treatment.1 2 3 4 7 Nevertheless, miraculous survivals have been reported in lay media involving patients who were previously declared BD or dead, which has created a mis-informed understanding of BD.8 In this retrospective study, all patients who underwent BD assessment over a 16-year period were evaluated to determine whether survival occurred following BD assessment; the demographics of patients diagnosed with BD were compared with those of all other patients diagnosed with cardiopulmonary death at a paediatric intensive care unit (PICU). The null hypothesis was that there would be no demographic or clinical differences between patients diagnosed with BD and those diagnosed with cardiopulmonary death.

All children admitted to the PICU of a university-affiliated teaching hospital and trauma centre (Prince of Wales Hospital) between October 2002 and October 2018 were included in the study. The Prince of Wales Hospital provides tertiary PICU service for children, from birth to age 16 years, in the Eastern New Territories of Hong Kong. The institutional ethics committee approved this review and waived the requirement for patient consent.

The demographics and clinical characteristics of deceased children were collected from the principal author’s database (KLH), in which every PICU admission was registered; data were also collected retrospectively from the Clinical Management System of the hospital. All deaths were reviewed, including those of patients with clinical evidence of BD who underwent BD assessment. Brain death was defined as irreversible loss of all functions of the brain, including the brainstem. The presence of coma, absence of brainstem reflexes, and positive apnoea test were essential findings for diagnosis of BD. The diagnosis of BD was mainly clinical and was made in accordance with the hospital’s standard protocol for paediatric patients.4 9 Patients were classified either as BD or cardiopulmonary death.

The demographics and clinical characteristics of these two groups of patients were summarised as median (interquartile range [IQR]) or as number (percentage), and were compared using the Chi squared test, Fisher’s exact test, or Mann-Whitney U test, as appropriate. Patient characteristics included age, sex, length of PICU stay (time from PICU admission to withdrawal of ventilator support), and diagnoses associated with PICU admissions. The GraphPad Prism 6 software (GraphPad Software, La Jolla [CA], US) and SPSS (Windows version 19.0; IBM Corp, Armonk [NY], US) were used for statistical analysis. All comparisons were two-tailed, and P values <0.05 were considered statistically significant.

Of the 2784 children admitted to the PICU, 127 (4.6%) died in the PICU (Table 1). All but seven children were of Chinese ethnicity. There were 73 boys (57.5%) and 54 girls (42.5%); the median age was 3.2 years (IQR: 0.94-7.34 years). Most patients had not previously been admitted to the PICU (n=103, 81%), and most patients were aged >1 year (74.8%). Of the 127 patients who died, BD assessments were performed for 22 (17.3%) patients who had clinical evidence of BD; all 22 patients were diagnosed with BD. The remaining 105 (82.7%) patients were diagnosed with cardiopulmonary death.

Comparison of the two groups showed that length of PICU stay was significantly longer for patients with BD (8.5 days; IQR: 4.75-14 days) than for patients with cardiopulmonary death (2 days; IQR: 1-10 days; P=0.004). The two groups shared similar demographics. The most common diagnoses associated with death in the PICU were infections (29.1% of patients), oncological diagnoses (13.0%), and cardiovascular diagnoses (13.8%) [Table 1]. Comparison of the two groups showed that trauma (P=0.003) and intracranial events (P=0.041) were more common in patients with BD, whereas respiratory diagnoses (P=0.033) were more common in patients with cardiopulmonary death. With respect to the cause of injury, the combined proportion of accident and trauma-related injury was greater in patients with BD than in patients with cardiopulmonary death (27.3% vs 3.8%, P<0.001). Among patients with BD, the most common mechanisms of brain injury were hypoxic-ischaemic injury (eg, cardiac arrest, shock, and/or respiratory failure), central nervous system infection, and traumatic brain injury (Table 2). Organ donation was approved by the families of four of the 22 patients with BD (18.2%) and was performed successfully in three of these four patients.

Our local guideline for BD determination does not include patients aged <2 years. Nevertheless, we found no difference in the proportion of patients aged <2 years between the BD (n=7) and cardiopulmonary death groups (n=42) [31.8% and 40%, P=0.47]. There was a non-significant trend towards greater use of ancillary tests (eg, radionuclide cerebral perfusion scan or electroencephalography) for BD determination in patients aged <2 years, compared with patients aged >2 years (85.7% and 53.3%, P=0.19). The United Kingdom guidelines recommend that ancillary tests are not required in infants from gestational age of 37 weeks to 2 months after birth.10 None of the patients were within this age range in our study.

Family acceptance of the diagnosis of BD may have influenced the length of PICU stay in our study. Among patients with documented family acceptance of the diagnosis of BD, the time interval from BD to withdrawal of ventilator support was 0.5 days (range, 0-1.5 days; n=10). This interval was prolonged among patients with documented family resistance of the diagnosis of BD (median, 8 days; range, 5-16 days; n=5, P=0.005); three of the five patients’ families eventually accepted withdrawal of ventilator support, whereas the remaining two patients remained on ventilator support and lapsed into cardiac arrest after 16 days and 66 days.

Over this 16-year period, BD assessment was only performed in 22 (17.3%) patients who had clinical signs of BD; all 22 patients were confirmed to have BD. Notably, patients with BD had longer length of PICU stay and a greater combined proportion of accident and trauma-related injury, while patients with cardiopulmonary death had a greater frequency of respiratory diagnoses. In the present study, the percentage of patients with BD in the PICU was comparable to the numbers of patients with BD in two large reports (one from the US and the other from Canada; Table 3).5 6 Accident and trauma-related injury led to one in four diagnoses of BD in our study, whereas the proportions of accident and trauma-related injury, as well as traumatic brain injury, were higher in the US and Canada.

Guidelines for BD assessment vary in terms of the numbers of examinations, numbers and types of physicians, time intervals between examinations, and use of ancillary tests.11 12 In general, if BD is suspected, two physicians (neither of whom would be involved in organ harvesting from the patient) should perform two sets of brainstem examinations, at least 6 hours apart to ensure sufficient observation time. A single apnoea test should also be performed. If the results of these tests are positive, the patient can then be declared legally and clinically BD.1 2 3 4 Before these examinations, conditions that may confound the clinical diagnosis of BD should be excluded.1 2 3 4 11 12 Absence of the pupillary reflex to direct and consensual light, as well as the absence of corneal, cough, and gag reflexes, support the clinical diagnosis of BD. The calorie test can aid in determining the integrity of the oculovestibular reflex. A positive result consists of the absence of eye deviation when ice water is irrigated into an external auditory canal. The apnoea test is performed after the second examination of brainstem reflexes; only a single apnoea test is needed. Before the apnoea test is performed, the physician must confirm that the patient is not hypothermic, is euvolemic, and has normal arterial pressure of carbon dioxide and pressure of oxygen levels. The patient should then be connected to a pulse oximeter and the ventilator should be disconnected. Concurrently, 100% O₂ is delivered into the trachea at 6 L/min. A patient with BD may exhibit systolic blood pressure <90 mm Hg, significant oxygen desaturation, or cardiac arrhythmia. If respiratory movements are absent and the arterial pressure of carbon dioxide is ≥60 mm Hg, the apnoea test result is considered positive. If the patient is very unstable and an apnoea test might not be tolerated, or if the results of the apnoea test are inconclusive, physicians may opt for other neuro-diagnostic options (eg, four-vessel cerebral angiography, radionuclide cerebral perfusion scan, and/or electroencephalography). A lack of blood perfusion to the brain and lack of electrical activity would support a diagnosis of BD.

The length of PICU stay was longer for patients with BD than for patients with cardiopulmonary death; this differed from the trends observed in the US and Canada (Table 3).5 6 As noted in the Results, family acceptance of the diagnosis of BD may have influenced the length of PICU stay in our study. Unfortunately, not all stakeholders understand or accept the implications of a diagnosis of BD. In our experience, the reasons for the family’s resistance might be two-fold. First, it might be emotionally difficult to accept the death of a loved one, when the child is apparently ‘breathing’ and appears physically ‘well’ when ventilatory support is provided. Second, the family might have confused persistent vegetative state with BD3 4; notably, patients with persistent vegetative state have intact brainstem function, while patients with BD have an irreversible loss of brainstem function. In such instances of confusion, families may wish to wait for the patient’s ‘miraculous revival’.1 13 14 15 16 While the acknowledgement of BD as biological death may be counterintuitive to the public, there is a need to emphasise and accept that BD is legal death.17 Thus, public education is necessary to resolve common misconceptions regarding the biological and legal statuses of patients with BD.18 Notably, among university students in Hong Kong, improved knowledge has been shown to promote acceptance of the withdrawal of ventilator support following BD.19 From a physician’s perspective, withdrawal of ventilator support for patients with BD should not be regarded as withdrawal of life support; in addition, continued use of ventilator support that allows a patient to lapse into cardiac arrest is not a suitable option. Prolonged and unnecessary treatment in the PICU prevents other critically ill patients from using the PICU service; it also constitutes ineffective use of scarce medical resources. Abuse of PICU beds is undesirable because medical resources in the public sector are extremely competitive and limited.20 Further studies regarding physician counselling skills and family acceptance of the diagnosis of BD may improve resource utilisation.

Medical professionals should closely monitor aetiologies that can lead to BD and consider discussions with affected patients’ families at an early stage of medical treatment. These early discussions would allow more time for families to comprehend the implications of a BD assessment and potential positive test results. In a previous study, we found that prolonged length of PICU stay was associated with a Do-Not-Attempt-Resuscitation order, which was placed in nearly half of our PICU deaths; this finding implied that patients’ families often need considerable time to accept the end-of-life decision when futility of medical treatment becomes evident.1 Family acceptance of the diagnosis of BD is critical for successful management of such situations. If family acceptance is not achieved, physicians may become involved in a conflict with the family, which results in an ethical dilemma regarding the need to continue treatment for a patient with BD. For example, one patient in the present study remained in the PICU for 66 days due to this difficult situation. Communication to identify common values and establish options based on objective criteria may resolve potential disputes and allow physicians and families to reach agreements before, during, and after BD assessment.21

A practical aspect of BD assessment involves its implications for the organ donation process. Four of 22 patients’ families opted for organ donation; notably, all four families also accepted the diagnosis of BD. Successful donations of liver or kidneys were made from three patients. Acceptance of the diagnosis of BD may be associated with acceptance of organ donation and withdrawal of ventilator support.19 Support for organ donation, which was initiated by the organ transplant coordinator, had avoided potential instances of conflict. Acceptance of the diagnosis of BD could be a factor, in combination with other cultural and religious beliefs, for the lower organ donation rate than that observed in Western countries.12

In this study, one in five PICU deaths were BD. Acute hypoxic-ischaemic injury was the most common mechanism of brain injury; moreover, accident and trauma-related injuries were the cause of injury in one quarter of patients with BD. Diagnosis of BD was associated with significantly longer PICU stay. Notably, the organ donation rate was suboptimal. These findings emphasise the importance of injury prevention in childhood, as well as the need for education of the public regarding acceptance of BD and support for organ donation.

Concept or design: KL Hon.
Acquisition of data: KL Hon, CC Au, TT Tse, WS Lin, TC Leung, TC Chow.
Analysis or interpretation of data: KL Hon, CC Au, TT Tse, WS Lin, TC Leung, TC Chow, AKC Leung.
Drafting of manuscript: KL Hon, CC Au, TT Tse, WS Lin, TC Leung, TC Chow.
Critical revision of the manuscript for important intellectual content: KL Hon, CC Au, CK Li, HM Cheung, SY Qian, AKC Leung.

All authors have full access to the data, contribute to the study, approve the final version for publication, and take responsibility for its accuracy and integrity.

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

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

The Joint Chinese University of Hong Kong–New Territories East Cluster Clinical Research Ethics Committee approved this review (CREC Ref. No. 2016.116).

1. Hon KL, Poon TC, Wong W, et al. Prolonged non-survival in PICU: does a do-not-attempt-resuscitation order matter. BMC Anesthesiol 2013;13:43. Crossref

2. Hon KL, Luk MP, Fung WM, et al. Mortality, length of stay, bloodstream and respiratory viral infections in a pediatric intensive care unit. J Crit Care 2017;38:57-61. Crossref

3. Goila AK, Pawar M. The diagnosis of brain death. Indian J Crit Care Med 2009;13:7-11. Crossref

4. Sarbey B. Definitions of death: brain death and what matters in a person. J Law Biosci 2016;3:743-52. Crossref

5. Kirschen MP, Francoeur C, Murphy M, et al. Epidemiology of brain death in pediatric intensive care units in the United States. JAMA Pediatr 2019;173:469-76. Crossref

6. Joffe AR, Shemie SD, Farrell C, Hutchison J, McCarthy-Tamblyn L. Brain death in Canadian PICUs: demographics, timing, and irreversibility. Pediatr Crit Care Med 2013;14:1-9. Crossref

7. Citerio G, Murphy PG. Brain death: the European perspective. Semin Neurol 2015;35:139-44. Crossref

8. Daoust A, Racine E. Depictions of “brain death” in the media: medical and ethical implications. J Med Ethics 2014;40:253-9. Crossref

9. Verheijde JL, Rady MY, Potts M. Neuroscience and brain death controversies: the elephant in the room. J Relig Health 2018;57:1745-63. Crossref

10. Marikar D. The diagnosis of death by neurological criteria in infants less than 2 months old: RCPCH guideline 2015. Arch Dis Child Educ Pract Ed 2016;101:186. Crossref

11. Greer DM, Varelas PN, Haque S, Wijdicks EF. Variability of brain death determination guidelines in leading US neurologic institutions. Neurology 2008;70:284-9. Crossref

12. Chua HC, Kwek TK, Morihara H, Gao D. Brain death: the Asian perspective. Semin Neurol 2015;35:152-61. Crossref

13. Al-Shammri S, Nelson RF, Madavan R, Subramaniam TA, Swaminathan TR. Survival of cardiac function after brain death in patients in Kuwait. Eur Neurol 2003;49:90-3. Crossref

14. Burkle CM, Sharp RR, Wijdicks EF. Why brain death is considered death and why there should be no confusion. Neurology 2014;83:1464-9. Crossref

15. Shewmon DA. Chronic “brain death” meta-analysis and conceptual consequences. Neurology 1998;51:1538-45. Crossref

16. López-Navidad A. Chronic “brain death”: meta-analysis and conceptual consequences. Neurology 1999;53:1369-70. Crossref

17. Truog RD, Miller FG. Changing the conversation about brain death. Am J Bioeth 2014;14:9-14. Crossref

18. Shah SK, Kasper K, Miller FG. A narrative review of the empirical evidence on public attitudes on brain death and vital organ transplantation: the need for better data to inform policy. J Med Ethics 2015;41:291-6. Crossref

19. Leung KK, Fung CO, Au CC, Chan DM, Leung GK. Knowledge and attitudes toward brain stem death among university undergraduates. Transplant Proc 2009;41:1469-72. Crossref

20. Truog RD. Brain death-too flawed to endure, too ingrained to abandon. J Law Med Ethics 2007;35:273-81. Crossref

21. Burns JP, Truog RD. Futility: a concept in evolution. Chest 2007;132:1987-93. Crossref