In 2017, Hong Kong had a low organ donation ratio of 6.0 deceased donors per million population, whereas the corresponding ratios were 46.9, 32.0, and 23.1 donors per million population in Spain, the United States, and the United Kingdom, respectively.1 Among all types of solid organs, there is the greatest shortage of donated kidneys in our locality. In 2018, there were 2318 patients on the waiting list for kidney transplantation.2 However, there were only 60 kidney donations from deceased donors and 16 from living donors in the same year.2 This marked mismatch has led to not only a long average waiting time for kidney transplantation of 51 months,3 but also the accompanying costs of prolonged dialysis, increased risk of dialysis-related complications, and adverse effects on patients’ quality of life.4 5 6

The majority of organ donations are from deceased donors. However, without knowing the wishes of deceased potential donors, it is often difficult to counsel their family members about organ donation. Therefore, it is important to engage the general public in prospective organ donor registration. Hong Kong has a population of approximately 7.39 million, but only 284 185 individuals had registered as organ donors via the Centralised Organ Donation Register (CODR) through June 2018, corresponding to a registration rate of 3.8%.7 In contrast, 52.6% of the respondents of the Behavioural Risk Factor Survey conducted in Hong Kong reported that they were willing to donate their organs after death.8 These results suggest that most people who are willing to donate organs after death have not yet registered as prospective organ donors. These individuals represent a group of passive-positive organ donors who would potentially become prospective organ donors if successfully engaged.9

We conducted a local survey to investigate the underlying reasons for the discrepancy between the number of individuals willing to donate organs and the number of registered donors. We postulate that appropriate use of social media may play a role in motivating people to register as prospective organ donors. Hence, we also investigated the use of smartphones and social media platforms by Hong Kong citizens. The results will be useful for planning our future directions and strategies for promoting organ donation.

A cross-sectional telephone survey of the general population of Hong Kong was conducted via the service provided by the Jockey Club School of Public Health and Primary Care, The Chinese University of Hong Kong. The survey was designed after consultation with doctors, nurses, living-related organ donors, organ recipients, and patient support groups. Demographic information including age, sex, marital status, education level, occupation, religion, smoking habits, drinking habits, exercise habits, and current health status was collected. Questions focusing on the respondents’ views about organ donation and their actions taken with regard to organ donor registration were asked. Questions regarding potential misconceptions about organ donation were also asked. The respondents’ habits of using smartphones and social media platforms were also evaluated. To minimise the sampling error, telephone numbers were first selected randomly from an updated telephone directory as seed numbers. Another set of three numbers was then generated by randomising the last two digits to recruit unlisted numbers. Duplicate numbers were then screened out, and the remaining numbers were mixed in random order to become the final sample.

Interviews were conducted by experienced interviewers between 10:00 and 22:00 on weekdays and other periods, including weekends and public holidays, should appointments with suitable subjects be arranged. The inclusion criteria for the study were Chinese Hong Kong residents aged ≥18 years. Upon successful contact with a target household, one qualified member of the household was selected among the family members using the last-birthday random selection method (ie, the respondent aged ≥18 years in the household who had his/her birthday most recently was selected) to participate in the telephone interview. We aimed for the survey to have 1000 respondents. All results were analysed and presented descriptively.

From 15 April 2019 to 8 May 2019, telephone numbers were sampled for the survey until 1000 valid responses from eligible individuals were received. Of 16 373 telephone numbers called, 14 514 were invalid for various reasons: 6556 were facsimile/invalid lines, 555 were non-residential lines, 1008 cut the line immediately, 6366 did not pick up the phone after three attempts, and 29 were non-Chinese persons. Among the remaining 1859 eligible individuals, 750 refused to participate in the survey, seven terminated the survey mid-way, and we failed to contact the remaining 102 after three attempted calls each. The overall response rate was 53.8% (1000/1859).

The majority of our respondents were female (68.3%) and within the age-group of 51 to 60 years (24%) or ≥61 years (43.6%). In total, 73.6% of the respondents did not have any religious beliefs. The vast majority of them were non-smokers (95.2%) and non-drinkers (93.1%). In all, 65.6% of the respondents exercised regularly, with 65.7% and 10.3% considering themselves “healthy” and “very healthy,” respectively (Table 1).

Table 2 shows the survey results on the respondents’ views about organ donation and their actions taken with respect to organ donor registration. Unexpectedly, a relatively large proportion of interviewees (30.6%) had never heard of CODR, and 89.1% of the respondents had never visited the CODR website. Only 31.3% of the respondents were willing to donate their organs after death, and 43.3% were indecisive, while 25.4% refused. When interviewees were asked if they would support a family member’s decision to become a prospective organ donor, 56.4% said they would be supportive, 7% would object, and 36.6% were uncertain. Looking further at the 313 respondents who were willing to become prospective organ donors, only 34.2% of them had registered on CODR, whereas 55.3% had expressed such wishes to their family members. Of those who had registered, 98.1% did so in the hope of rekindling others’ lives, 93.9% believed that their organs would become useless after death, and 44.7% were influenced by successful organ donation stories publicised by the media.

Of the 107 respondents who had registered to be prospective organ donors, 47.7% did so via organ donation promotional booths, 28% filled in the application forms and mailed them back to the Department of Health, and 24.3% registered online. Among the 206 respondents who were willing to donate organs after death but had not yet registered on CODR, 52.2% said they were not determined enough to take action, 47.8% said they were too busy, 37.8% admitted that they were too lazy to do so, 20.4% said they were always forgetful about registering, and 32.8% said they were not aware of the ways to register as a prospective organ donor.

A total of 687 respondents were indecisive or refused to become organ donors. In all, 30.7% hoped to keep their bodies intact after death, 7.1% refused to register because of religious beliefs, 7.1% were worried that organ donation might increase their suffering, and 6.8% worried that by agreeing with organ donation, they would receive suboptimal or inadequate medical care. In total, 22% were not keen to donate organs owing to objections from family (Table 2).

The questionnaire also studied potential misconceptions about organ donation (Table 3). Of the respondents, 6.6% believed that the process of organ harvesting would induce unnecessary pain to the deceased person, and 4.5% worried that organ harvesting would hinder funeral arrangements. A total of 60.6% thought that only perfectly healthy individuals could donate organs after death, 34% believed only young people could donate organs after death, and 7.6% believed that the organ recipients would always know the identity of the organ donor. Moreover, 1.7% were under the impression that there was an adequate supply of organs in Hong Kong.

Respondents were interviewed about their use of smartphones and social media platforms (Table 4). In all, 77% of the respondents often use their smartphones to access social media platforms. Among the various non-messenger types of social media platforms, Facebook, YouTube, and Instagram were the most commonly used. The majority of the respondents believed that Facebook and YouTube were effective at engaging the audience.

There has been a great demand for organ donations in Hong Kong, yet the number of organ transplantations conducted is very small. The organ donation rate in Hong Kong is much lower than that of many European countries, perhaps because of cultural, religious, governmental, legal, and regulatory differences, as well as differences in the level of intensive care unit support and organ donation criteria.10 11 12 13 Hong Kong currently follows the opt-in approach to organ donation, as opposed to the opt-out approach, which is the standard in countries like Singapore and Spain. Although the opt-out approach may increase the availability of suitable organs for donation, there is a reasonable concern about differing views between members of the general public and the potential ethical issues related to that approach. The degree of knowledge, awareness, and attitude towards organ donation is also important for an individual to take action to become a prospective organ donor.14 More efforts should be made in these areas to improve the organ donation rate in Hong Kong.

In 2015, a Behavioural Risk Factor Survey with a total of 4253 respondents was conducted in Hong Kong.8 Among the respondents, 52.6% reported that they were willing to donate their organs after death, 11.3% refused to donate their organs after death, and the rest remained undecided. However, until June 2018, the registration rate in Hong Kong was only 3.8%.7 This represents a huge area of potential improvement if we are able to engage these potential organ donors successfully.

Our survey showed that only 34.2% of the respondents who were willing to donate their organs after death actually completed registration at CODR. A large proportion of respondents said they were too busy, too lazy, too forgetful, or simply not determined enough to take action to register at CODR. Of the respondents, 32.8% were not aware of the ways to register as a prospective organ donor. The majority of the respondents had not heard about CODR, and only 5.9% had browsed the CODR website. We need better ways to reach out to these passive-positive donors and to provide convenient methods for immediate registration after engaging them successfully. Our survey showed that the majority of the respondents use smartphones to access social media platforms every day and that Facebook, YouTube, and Instagram are the major social media platforms being used in Hong Kong. These social media platforms should be used for any organ donation promotion activities in the future.

Our survey showed that 47.7% of registered organ donors completed their registration via organ donation promotional booths. Face-to-face settings such as promotional booths allow the best engagement and interaction with the audience, and this definitely yields better results, especially for older adults who may not be familiar with the use of internet or social media platforms. Booths provide opportunities for educators to clarify people’s misconceptions, resolve their queries, and provide live guidance regarding their registration. Our survey reflected the effectiveness of organ donation promotion booths established by the Hong Kong government in the past. It would be worth investing more resources to set up regular and frequent promotion booths in more diverse areas owing to their promising effects.

Two systematic reviews and meta-analyses have been conducted to identify effective community-based interventions to increase organ donor registration.15 16 Among all studies reviewed, four randomised controlled trials demonstrated the effectiveness of an intervention based on an increase in verified organ donor registrations.17 18 19 20 The first study investigated the role of group discussions about organ donation in a church setting together with a 32-minute video featuring organ donation, organ transplantation, and the personnel involved during the whole process.17 The second study investigated the effects of a 5-minute video using an iPod Classic or iPod Touch with noise-cancelling headphones.18 The video was designed to address a number of concerns related to organ donation. The participants were then interviewed and given written information about organ donation. The third study investigated the role of a brief motivational intervention by hair stylists that encouraged organ donation.19 Hair stylists received training on communication skills, motivational interviewing, and discussion of ways to integrate organ donation into their client interaction. Each client was given a package containing organ donor registration cards. The fourth study investigated the use of the IIFF model (Immediate opportunity to register, Information, Focused engagement, and Favourable activation) to increase the rate of organ donor registration in the setting of Secretary of State branch offices.20 Participants were gathered at the Town Hall, where organ donation was discussed, and there were registration cards at the end of the session.

The four successful studies have common features.17 18 19 20 First, all four studies successfully engaged the public with motivational interactions.17 18 19 20 Intervention participants were 1.23- to 7.02-times more likely than comparison participants to report positive registration status.17 18 19 20 Passive-positive organ donors already have beliefs and attitudes that favour organ donation; what they need is additional motivational interactions to convert their belief into action. Second, three studies adopted a face-to-face approach in their interventions, which yielded positive results.17 19 20 This is consistent with our findings, in that we have also seen the positive effects of face-to-face promotion (ie, booths) in Hong Kong. Third, two studies used video media to provide information about organ donation and organ transplantation.17 18 Mass media alone are unlikely to produce any substantial effects, but the combination of media with motivational interaction can have synergistic effects in engaging the general public. Media intervention must also be innovative enough to attract the general public for better engagement. Fourth, three studies provided immediate opportunities for organ donor registration.18 19 20 The opportunity for organ donor registration must be immediately present following successful engagement of an individual, and it must be rapid and convenient enough for the individual to complete the process. All four studies demonstrated the effectiveness of their promotion strategies. Although it might not be suitable to duplicate and apply those interventions directly in Hong Kong due to discrepancies in promotion setting and target audience, by learning from these successful examples, we can identify the essential components of a successful organ donation promotion project.

Our study has several limitations. First, we only randomly selected 1000 Chinese Hong Kong residents to complete the survey, and this cannot represent the views of all Hong Kong citizens. Second, as this was a telephone survey, the majority of our respondents are aged ≥51 years. The results may not be a good reflection of the younger generation. Third, although our survey enables us to understand more about the situation of passive-positive donors in Hong Kong and the appropriate channels for engaging these potential organ donors, the exact ways to achieve audience engagement cannot be ascertained based only on the results of our survey. We intend to conduct future studies to investigate the effectiveness of social media platforms for interventions such as short videos, online challenge campaigns, online question-and-answer forums, online polling, live interviews, and live talks.

There are many passive-positive organ donors in Hong Kong. Many of those surveyed were not aware of the ways to register as prospective organ donors. The majority also lacked the determination to register as organ donors. Engaging these individuals and providing immediate opportunities for registration is necessary. Promotional booths are most effective at providing this face-to-face, and social media platforms can provide this on an online setting.

Concept or design: JYC Teoh, NY Far, SKK Yuen, BSY Lau.
Acquisition of data: JYC Teoh, NY Far, SKK Yuen, BSY Lau.
Analysis of data: JYC Teoh, NY Far, SKK Yuen, BSY Lau.
Drafting of the manuscript: JYC Teoh, NY Far.
Critical revision of the manuscript for important intellectual content: CH Yee, SSM Hou, TSC Teoh, CF Ng.

As editors of the journal, JYC Teoh and CF Ng were not involved in the peer review process. Other authors have disclosed no conflicts of interest.

This survey was supported by the Health and Medical Research Fund, Health Care and Promotion Scheme (Ref: 02180248).

The study has been approved by Survey and Behavioural Research Ethics Committee (Ref SBRE-18-241). Survey respondents provided verbal consent to participate in the telephone interview.

1. International Registry of Organ Donation and Transplantation. Donation activity charts. Available from: http://www.irodat.org/?p=database. Accessed 7 Mar 2020.

2. Department of Health, Hong Kong SAR Government. Organ donation. Statistics (milestones of Hong Kong organ transplantation). Available from: https://www. organdonation.gov.hk/eng/statistics.html. Accessed 1 Oct 2019.

3. Hospital Authority, Hong Kong SAR Government. Smart Patient. Chronic renal failure. Available from: https:// www21.ha.org.hk/smartpatient/SPW/en-US/Disease- Information/Disease/?guid=368b30e4-cc1c-4185-b673- 7dfb3ea8f74b. Accessed 1 Oct 2019.

4. Karlberg I, Nyberg G. Cost-effectiveness studies of renal transplantation. Int J Technol Assess Health Care 1995;11:611-22. Crossref

5. Whiting JF, Kiberd B, Kalo Z, Keown P, Roels L, Kjerulf M. Cost-effectiveness of organ donation: evaluating investment into donor action and other donor initiatives. Am J Transplant 2004;4:569-73. Crossref

6. Bakewell AB, Higgins RM, Edmunds ME. Quality of life in peritoneal dialysis patients: decline over time and association with clinical outcomes. Kidney Int 2002;61:239-48. Crossref

7. Surveillance and Epidemiology Branch, Centre for Health Protection, Department of Health, Hong Kong SAR Government. Registrations recorded in the centralised organ donation register. Available from: https://www. organdonation.gov.hk/eng/home.html. Accessed 1 Jun 2018.

8. Surveillance and Epidemiology Branch, Centre for Health Protection, Department of Health, Hong Kong SAR Government. Behavioural risk factor survey (April 2015). Available from: https://www.chp.gov.hk/files/pdf/ brfs_2015apr_en.pdf. Accessed 1 Oct 2019.

9. Siegel JT, Alvaro EM, Crano WD, Gonzalez AV, Tang JC, Jones SP. Passive-positive organ donor registration behavior: a mixed method assessment of the IIFF Model. Psychol Health Med 2010;15:198-209. Crossref

10. Cheung CY, Pong ML, Au Yeung SF, Chau KF. Factors affecting the deceased organ donation rate in the Chinese community: an audit of hospital medical records in Hong Kong. Hong Kong Med J 2016;22:570-5. Crossref

11. Cheung TK, Cheng TC, Wong LY. Willingness for deceased organ donation under different legislative systems in Hong Kong: population-based cross-sectional survey. Hong Kong Med J 2018;24:119-27. Crossref

12. Fan RP, Chan HM. Opt-in or opt-out: that is not the question. Hong Kong Med J 2017;23:658-60. Crossref

13. Tafran K. In search of the best organ donation legislative system for Hong Kong: further research is needed. Hong Kong Med J 2018;24:318-9. Crossref

14. Chung CK, Ng CW, Li JY, et al. Attitudes, knowledge, and actions with regard to organ donation among Hong Kong medical students. Hong Kong Med J 2008;14:278-85.

15. Deedat S, Kenten C, Morgan M. What are effective approaches to increasing rates of organ donor registration among ethnic minority populations: a systematic review. BMJ Open 2013;3:e003453. Crossref

16. Li AT, Wong G, Irving M, et al. Community-based interventions and individuals’ willingness to be a deceased organ donor: systematic review and meta-analysis. Transplantation 2015;99:2634-43. Crossref

17. Andrews AM, Zhang N, Magee JC, Chapman R, Langford AT, Resnicow K. Increasing donor designation through black churches: results of a randomized trial. Prog Transplant 2012;22:161-7. Crossref

18. Thornton JD, Alejandro-Rodriguez M, León JB, et al. Effect of an iPod video intervention on consent to donate organs: a randomized trial. Ann Intern Med 2012;156:483-90. Crossref

19. Resnicow K, Andrews AM, Beach DK, et al. Randomized trial using hair stylists as lay health advisors to increase donation in African Americans. Ethn Dis 2010;20:276-81.

20. Harrison TR, Morgan SE, King AJ, Williams EA. Saving lives branch by branch: the effectiveness of driver licensing bureau campaigns to promote organ donor registry sign-ups to African Americans in Michigan. J Health Commun 2011;16:805-19. Crossref

Family doctors, serving as the first point of professional contact for patients, are inevitably first to identify probable cases of coronavirus disease 2019 (COVID-19) among the many patients presenting with respiratory symptoms each day.1 Family doctors in Hong Kong have experience in dealing with the severe acute respiratory syndrome (SARS) epidemic in 20032 3 and the H1N1 pandemic in 2009.4 5 However, their preparedness in handling another outbreak of a novel infectious disease has not been explored. Furthermore, Hong Kong has a dual-track healthcare system in which 70% of primary medical care, especially acute episodic care, is provided in the private sector where practice settings and resources vary and differ from those of public clinics.6 7 8 Family doctors play a crucial role in the community to offer first contact and coordinated care for patients, and their preparedness, perceptions, and attitudes towards COVID-19 are particularly important to inform future strategies for responding to epidemics. Hence, the Hong Kong College of Family Physicians (HKCFP) conducted an online survey among its members to evaluate preparedness and to identify clinic-related challenges of private and public family doctors who were providing primary care services during the early phase of the evolving COVID-19 outbreak in Hong Kong.

All family doctors who are HKCFP members were invited to complete an online survey. The structured questionnaire (online supplementary Appendix 1) comprised 20 questions. Twelve closed-ended questions assessed the effects of the COVID-19 outbreak (at the time of the survey, the World Health Organization had not announced it as a pandemic) on clinical services and the preparedness of the responding family doctors, such as changes in infection control practice. An open-ended question invited respondents to express their concerns towards the COVID-19 outbreak and suggest measures that would facilitate their clinical practice. The last seven questions collected demographics of the respondents. The survey questions were modified from a previous survey for primary care doctors in Hong Kong and Canada9 10 and pilot-tested by a panel of experienced academic family doctors and HKCFP Research and Executive Committee members. Invitation e-mails and short message service reminders were sent to target participants between 31 January 2020 and 3 February 2020.

Descriptive statistics were used to summarise the characteristics of the respondents. Respondents were stratified by their practice sector (ie, public vs private). The differences in the effects of the COVID-19 outbreak on the clinical practices, clinic service, and infection control practices between public and private family doctors were evaluated by Pearson Chi squared test. Thematic analysis was performed on the respondents’ free comments and suggestions. The responses were reviewed independently by two investigators and consolidated into themes. Inconsistencies were resolved by discussion between the two investigators to reach consensus on a common theme. The consolidated themes from the respondents’ suggestions and concerns were further stratified by respondents’ practice sector using descriptive statistics.

All significance tests were two-tailed and those with a P value of <0.05 were considered statistically significant. The statistical analysis was executed by Stata (Version 16.0; StataCorp LLC, College Station [TX], US).

Since the survey was initially conducted to examine the needs and preparedness of frontline family doctors who are members of the HKCFP during the early phase of COVID-19 outbreak in Hong Kong, ethics approval was obtained from the Survey and Behavioural Research Ethics Committee of The Chinese University of Hong Kong subsequently for data analysis and presentation.

The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist for cross-sectional studies was used in the drafting of this article.11

Of 1589 HKCFP members invited to complete the survey, 491 (31%) provided a complete and valid response (Table 1). Of the respondents, 393 (80%) had attained higher qualifications in Family Medicine. Among the respondents, 236 (48%) worked at public primary care clinics operated by the Hospital Authority or Department of Health, and 242 (49%) worked in the private sector, half of whom were solo practitioners. The ratio of public to private sector respondents was approximately 1:1.

The vast majority of the respondents (n=482, 98%) continued to provide clinic services although most of their clinic practices (n=428, 87%) had been affected by the COVID-19 outbreak (Fig 1, online supplementary Appendix 2). Significantly a higher proportion of private than public family doctors reported reduced clinic service demands during this outbreak (n=111 [45%] vs n=60 [24%], P<0.001). Half of the surveyed family doctors adjusted non-acute consultation services and/or reduced consultation time. As of 4 February 2020, over 140 patients with suspected severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection had been encountered by 70 (14%) surveyed family doctors; one public family doctor reported a patient who was subsequently confirmed to have SARS-CoV-2 infection. Among the surveyed family doctors, 310 (63%) perceived needs for more training on how to deal with the COVID-19 outbreak. At time of the survey, to assist clinical decision making for diagnosing COVID-19, guidelines from the Centre for Health Protection or the Hospital Authority were used by public family doctors more frequently than by private family doctors (n=143 [58%] vs n=98 [40%], P<0.001). Conversely, 195 (80%) of the surveyed private family doctors encountered problems in stocking personal protective equipment (PPE).

Nearly all respondents wore masks during consultations (n=490, 99%) and washed hands between or before patient encounter (n=486, 99%) [Fig 2 and Supplementary Appendix 2]. A greater proportion of public than private family doctors insisted patients wear masks during consultations (n=210 [85%] vs n=165 [67%], P<0.001) and routinely screened patients’ body temperatures (n=211 [86%] vs n=183 [75%], P=0.002). In contrast, a greater proportion of private than public family doctors cleaned work surfaces with antiseptics at least once a day (n=223 [91%] vs n=200 [81%], P=0.002) and installed air purifiers (n=71 [29%] vs n=35 [14%], P<0.001).

Of the respondents, 159 (32%) answered the open-ended question, among which 135 (85%) suggested measures to be instituted by government and/or local health authorities to facilitate frontline family doctors to respond to the COVID-19 outbreak (Table 2). A significant proportion of respondents (n=52, 39%) appealed to the government, health service providers and/or professional bodies for securing adequate supply of PPE, especially surgical masks, for frontline healthcare workers as well as the general public. There was a strong call (n=49, 36%) for more effective public health policy to contain the outbreak, such as border control and/or quarantine measures for returning residents and travellers to reduce imported cases. Two respondents (1.5%) had expectations for better coordination between the public and private sectors with respect to role delineation and resource allocation, for example, setting up a Primary Care Authority to maximise efficiency and effectiveness of scattered primary healthcare delivery locally. Some family doctors (n=9, 7%) advocated for the introduction of designated clinics and requested availability of rapid diagnostic tests. A few respondents (n=8, 6%) stressed the importance of public education on infection control practice and reporting accurate travel and contact history during consultation.

Nineteen respondents (4%) expressed personal concerns that were consolidated into six themes (Fig 3 and online supplementary Appendix 3). The major concern was the risk of SARS-CoV-2 infection, as a result of the lack of PPE, consultation with an asymptomatic patient with SARS-CoV-2 infection, or dishonest patients with unreliable history. Owing to the lack of rapid tests and/or PPE, two private family doctors (11%) worried that they would be unable to provide clinic services, resulting in public healthcare system overload. Three respondents (16%) raised concerns about the need to handle and/or clarify “fake news” (ie, misinformation).

The vast majority of surveyed family doctors were committed to discharging their duties in the early phase of the COVID-19 outbreak despite clinical uncertainties, psychological distress arising from infectious risk to self and family, and corresponding significant effects on clinical services. Only 2% of the surveyed private family doctors had closed their clinics, compared with 8% during the SARS epidemic in 2003.9 These figures were also much lower than reported absenteeism rates of healthcare workers during influenza pandemics.12 13 14 In the 2009 H1N1 pandemic in 2009, 59% of local primary care doctors reported higher demands in clinical services.4 In contrast, in the present study, 25% of public family doctors and 45% of private family doctors reported reductions in clinical service demand. Different from an influenza outbreak when primary care doctors are tasked with providing confirmatory diagnostic tests and antiviral treatment, rapid diagnostic tests were not readily accessible in the primary care setting at time of our survey, and treatments for COVID-19 were available only in hospital settings. Patients with highly probable SARS-CoV-2 infection were sent directly to hospital isolation wards for further management. Patients suspecting themselves to have SARS-CoV-2 infection presented in large numbers to emergency departments instead of primary care clinics. Local citizens were also strongly encouraged to practise social distancing, especially avoiding areas of high contact risk, including clinics.15 Patients with other non-urgent health needs might opt to delay their clinic visits. Nevertheless, family doctors encountered probable SARS-CoV-2 infection, especially patients with milder, non-specific respiratory symptoms and with less clear travel and/or contact history. Thus, family doctors needed to remain vigilant in identifying suspected cases in the community during this period, while providing continuing care to other patients with unrelated medical conditions, mental health support for patients affected by the outbreak and educating healthy patients.

Although the mode of transmission of COVID-19 was not clearly understood in the early phase of outbreak, almost all surveyed family doctors readily adopted standard droplet and contact precautions, including wearing facemasks during consultation and washing hands before and/or between seeing patients as recommended by the Centre for Health Protection of Hong Kong.16 Wearing facemask during consultation became a common practice among family doctors in Hong Kong since the global outbreak of SARS.4 9 17 Conversely, hand hygiene practices of family doctors were less consistent and varied between 45% before the H1N1 influenza epidemic4 to 70% during SARS in 2003.9 In the present survey, 99% of family doctors reported washing their hands before patient encounters during the current outbreak, which has been proven more effective than facemask wearing alone in limiting the transmission of respiratory infections.18 19 The practice of other recommended infection control measures differed between public and private family doctors, reflecting practical challenges in their implementation. A particular infection control challenge for local small-sized clinics would be the required isolation of patients with suspected SARS-CoV-2 infection, where an extra single isolation room with or without negative pressure, or even a designated isolation area >1 m from the rest of the waiting area, was often unavailable.16 To protect other patients from possible cross-infection in clinics, respondents adjusted non-nonacute patient appointments, shortened consultation times to avoid crowding of patients in the clinic, or divert patients to other clinics. Despite the variations, these infection control measures might be contributory to the zero-infection rate observed among primary care providers in Hong Kong at the time of the survey.

Many respondents considered public health policies and interventions in response to the COVID-19 outbreak to be important. There has been an escalation of infection control responses to the COVID-19 outbreak, especially wearing of facemasks, in the healthcare20 and community settings.21 Consequently, an acute shortage of facemasks was experienced by respondents, similar to the situation observed in the US.22 Echoing the viewpoints of Australian general practitioners towards influenza pandemic management, family doctors from Hong Kong also considered that government and health authorities should be responsible for ensuring steady supply of PPE to frontline healthcare workers and/or the public.23 A few surveyed family doctors commented that they would cease to provide clinical service if PPE became unavailable, owing to the high infection risk. Moreover, a large proportion of respondents advocated for border controls and quarantine measures to limit cross-border transmission.24 Subsequently, border controls and mandatory quarantine were implemented on people arriving from mainland China in early February 2020,25 and extended to travellers from most regions around the globe in March 2020.26 These measures may have contributed to the relatively slow rise in the number of confirmed COVID-19 cases in Hong Kong.

Some private family doctors requested the introduction of designated fever clinics for the public, so that high-risk patients presenting with fever and/or respiratory symptoms could be diverted to a designated location and managed appropriately. Such arrangements would be particularly important for protecting the many small private clinics which lack the capacity to properly isolate high-risk patients. Designated clinics were successfully implemented in Hong Kong during the 2009 H1N1 pandemic5 and in China and the US during the current COVID-19 outbreak.1 22 Unfortunately, local citizens opposed these clinics owing to a fear of COVID-19 transmission in the neighbourhood. Instead, designated doctors were assigned to attend high-risk or febrile patients in certain public primary care clinics. However, the arrangement was not clear to the public nor frontline private family doctors and symptomatic patients continued to seek care from private family doctors. Despite repeated calls for coordinated care or clear role delineation of family doctors between public and private sectors at times of outbreak, this has still not been achieved.5 9

This study had two key strengths. First, our survey was conducted in the early phase of the COVID-19 outbreak, thus the survey rapidly captured the early effects of an emerging outbreak on primary care services and reflected the clinic preparedness and needs of frontline family doctors in Hong Kong. Second, our study covered family doctors from both public and private sectors, allowing for comparison between the two sectors. Possible service gaps in the current dual-track primary healthcare system could be readily identified to inform policy makers.

A major limitation of this study was the low response rate, attributable to the relatively short survey period. Although our response rate (31%) was lower than previous similar surveys among family doctors in Hong Kong during SARS (75%) and H1N1 pandemic (42%), the crude response rate was higher (n=491, vs 137 and 126, respectively). However, our respondents included only approximately 10% of the doctors listed in the Primary Care Directory.27 Also, only HKCFP members and fellows were targeted in this survey. Hence, the sample might not be representative of all primary care physicians in Hong Kong. Lastly, as an observational study, reporting bias existed.

Family doctors from both public and private sectors in Hong Kong reported willingness and preparedness to provide primary, continuous, and whole-person care to the community from the early phase of the COVID-19 outbreak. Despite limitations in clinic physical settings and potential for PPE shortages, most family doctors adopted standard precautions and effectively protected themselves and the public from cross-infection. Nevertheless, there is an obvious need for health authorities to improve role delineation and coordination between private and public primary care services and to provide relevant support during an outbreak, so that family doctors can continue to play their various roles in the community under the current dual-track primary healthcare system.

Concept or design: SYS Wong, EYT Yu, WLH Leung.
Acquisition of data: SYS Wong, EYT Yu, WLH Leung.
Analysis or interpretation of data: All authors.
Drafting of the manuscript: All authors.
Critical revision of the manuscript for important intellectual content: All authors.

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 would like to thank The Hong Kong College of Family Physicians (HKCFP) Research and Executive Committee members who have contributed to the design of questionnaire and review of the draft, including Dr Angus Chan, Dr David Chao, Dr Catherine Chen, Dr Lap-kin Chiang, Dr Billy Chiu, Dr Cecilia Fan, Dr Ho-lim Lau, Dr Jun Liang, Dr Shuk-yun Leung, Dr Lorna Ventura Ng, Professor Martin Wong, and Dr William Wong; and Miss Erica So, Miss Crystal Yung, and Miss Angel Fung who provided administrative support for the study. We would also like to thank all the participating family doctors who responded promptly to this survey.

This research has not been presented in any academic conference or published previously. Part of the findings from the survey was disseminated through a local press release on 10 March 2020.

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

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 Survey and Behavioural Research Ethics Committee of The Chinese University of Hong Kong (Ref SBRE-19-578).

1. Li DK, Zhu S. Contributions and challenges of general practitioners in China fighting against the novel coronavirus crisis. Fam Med Community Health 2020;8:e000361. Crossref

2. Zhong NS, Zheng BJ, Li YM, et al. Epidemiology and cause of severe acute respiratory syndrome (SARS) in Guangdong, People’s Republic of China, in February, 2003. Lancet 2003;362:1353-8. Crossref

3. Tsang KW, Ho PL, Ooi GC, et al. A cluster of cases of severe acute respiratory syndrome in Hong Kong. N Engl J Med 2003;348:1977-85. Crossref

4. Wong SY, Kung K, Wong MC, et al. Primary care physicians’ response to pandemic influenza in Hong Kong: a mixed quantitative and qualitative study. Int J Infect Dis 2012;16:e687-91. Crossref

5. Lee A, Chuh AA. Facing the threat of influenza pandemic—roles of and implications to general practitioners. BMC Public Health 2010;10:661. Crossref

6. Lee A. Seamless health care for chronic diseases in a dual health care system: managed care and the role of family physicians. J Manag Med 1998;12:398-405. Crossref

7. Working Party on Primary Health Care. Health for all, the way ahead: Report of the Working Party on primary health care. Hong Kong: Government Printer; 1990.

8. Wun YT, Lee A, Chan KK. Morbidity pattern in private and public sectors of family medicine/general practice in a dual health care system. Hong Kong Practitioner 1998;20:3-15. Crossref

9. Wong WC, Lee A, Tsang KK, Wong SY. How did general practitioners protect themselves, their family, and staff during the SARS epidemic in Hong Kong? J Epidemiol Community Health 2004;58:180-5. Crossref

10. Wong SY, Wong W, Jaakkimainen L, Bondy S, Tsang KK, Lee A. Primary care physicians in Hong Kong and Canada—how did their practices differ during the SARS epidemic? Fam Pract 2005;22:361-6. Crossref

11. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP; STROBE Initiative. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet 2007;370:1453-7. Crossref

12. Balicer RD, Omer SB, Barnett DJ, Everly GS, Jr. Local public health workers’ perceptions toward responding to an influenza pandemic. BMC Public Health 2006;6:99. Crossref

13. Ehrenstein BP, Hanses F, Salzberger B. Influenza pandemic and professional duty: family or patients first? A survey of hospital employees. BMC Public Health 2006;6:311. Crossref

14. Damery S, Wilson S, Draper H, et al. Will the NHS continue to function in an influenza pandemic? A survey of healthcare workers in the West Midlands, UK. BMC Public Health 2009;9:142. Crossref

15. Fung CS, Yu EY, Guo VY, et al. Development of a Health Empowerment Programme to improve the health of working poor families: protocol for a prospective cohort study in Hong Kong. BMJ Open 2016;6:e010015. Crossref

16. Ashton LM, Hutchesson MJ, Rollo ME, Morgan PJ, Collins CE. Motivators and barriers to engaging in healthy eating and physical activity. Am J Mens Health 2017;11:330- 43. Crossref

17. Wong CK, Yip BH, Mercer S, et al. Effect of facemasks on empathy and relational continuity: a randomised controlled trial in primary care. BMC Fam Pract 2013;14:200. Crossref

18. Cowling BJ, Chan KH, Fang VJ, et al. Facemasks and hand hygiene to prevent influenza transmission in households: a cluster randomized trial. Ann Intern Med 2009;151:437- 46. Crossref

19. Aiello AE, Murray GF, Perez V, et al. Mask use, hand hygiene, and seasonal influenza-like illness among young adults: a randomized intervention trial. J Infect Dis 2010;201:491-8. Crossref

20. Cheng VC, Wong SC, Chen JH, et al. Escalating infection control response to the rapidly evolving epidemiology of the coronavirus disease 2019 (COVID-19) due to SARS-CoV-2 in Hong Kong. Infect Control Hosp Epidemiol 2020;41:493-8. Crossref

21. Leung CC, Lam TH, Cheng KK. Mass masking in the COVID-19 epidemic: people need guidance. Lancet 2020;395:945. Crossref

22. Kamerow D. Covid-19: Don’t forget the impact on US family physicians. BMJ 2020;368:m1260.Crossref

23. Shaw KA, Chilcott A, Hansen E, Winzenberg T. The GP’s response to pandemic influenza: a qualitative study. Fam Pract 2006;23:267-72. Crossref

24. Jefferson T, Foxlee R, Del Mar C, et al. Physical interventions to interrupt or reduce the spread of respiratory viruses: systematic review. BMJ 2008;336:77-80. Crossref

25. Viner R, Macfarlane A. Health promotion. BMJ 2005;330:527-9. Crossref

26. Magnussen CG, Koskinen J, Chen W, et al. Pediatric metabolic syndrome predicts adulthood metabolic syndrome, subclinical atherosclerosis, and type 2 diabetes mellitus but is no better than body mass index alone: the Bogalusa Heart Study and the Cardiovascular Risk in Young Finns Study. Circulation 2010;122:1604-11. Crossref

27. Hong Kong SAR Government. Primary Care Directory. Available from: https://apps.pcdirectory.gov.hk. Accessed 15 Mar 2020.

Management of rheumatic diseases such as rheumatoid arthritis, psoriatic arthritis, and spondyloarthropathy has been revolutionised since the emergence of biologic and targeted synthetic disease-modifying antirheumatic drugs (b/tsDMARDs). However, evidence from various trials and registries has shown that these agents are associated with a heightened risk of tuberculosis (TB) or reactivation of latent tuberculosis infection (LTBI).1 2 3 4 Notably, a meta-analysis showed that patients receiving anti-tumour necrosis factor alpha (anti-TNF-α) exhibited a >3-fold increase in the risk of TB5; analysis of a post-marketing registry in Japan suggested that anti-interleukin-6 therapy increased the risk of TB to a similar extent.6

Screening for LTBI before initiation of b/tsDMARD treatment is strongly recommended universally; its importance cannot be overemphasised in TB endemic areas, such as Hong Kong.7 8 9 However, there is no gold standard screening test. Regional guidelines recommend the use of either tuberculin skin test (TST) or interferon gamma release assay (IGRA) for LTBI screening in patients with rheumatic diseases.10 The TST measures the delayed hypersensitivity response to tuberculin purified protein derivatives; however, its specificity is limited by cross-reactivity due to previous bacillus Calmette-Guérin vaccination or exposure to non-TB mycobacterium. In contrast, IGRA measures interferon gamma release by TB-specific effector T cells; this result does not exhibit cross-reactivity related to bacillus Calmette-Guérin or non-TB mycobacterium. However, both tests are known to exhibit reduced reliability in immunosuppressed patients.11 12

Agreement between the TST and IGRA has varied among studies. A meta-analysis showed that pooled concordances were 72% between the QuantiFERON-TB assay and TST, whereas they were 75% between the T-SPOT.TB and TST.13 Our centre in Hong Kong previously investigated the agreement between the TST and IGRA in patients with rheumatic diseases; the results showed that agreement was only fair (Cohen’s kappa=0.39).14 Therefore, we hypothesised that the use of dual testing with both IGRA and TST might improve sensitivity for detection of LTBI, thus reducing the incidence of TB in patients with rheumatic diseases receiving b/tsDMARDs. Here, we conducted a retrospective cohort study to compare the efficacy of dual testing with both TST and IGRA versus single testing in terms of reducing the incidence of TB among patients with rheumatic diseases in Hong Kong receiving b/tsDMARDs.

We reviewed the case records of all patients with rheumatic diseases who began receiving b/tsDMARDs in a regional rheumatology unit in Hong Kong, during the period from 1 January 2007 to 31 December 2018. The following b/tsDMARDs were included: abatacept, adalimumab, certolizumab, etanercept, golimumab, infliximab, rituximab, secukinumab, tocilizumab, tofacitinib, and ustekinumab. The following rheumatic diseases were included: rheumatoid arthritis, psoriatic arthritis, spondyloarthropathy, Behçet’s disease, adult-onset Still’s disease, and dermatomyositis. The patients had been diagnosed in accordance with the 2010 American College of Rheumatology/European League Against Rheumatism classification criteria for rheumatoid arthritis,15 the Classification Criteria for Psoriatic Arthritis,16 the Assessment of SpondyloArthritis International Society classification criteria,17 the 1990 International Study Group criteria for diagnosis of Behçet’s disease,18 the Yamaguchi criteria for adult Still’s disease,19 and Bohan and Peter’s criteria.20 Patients were included if they had undergone LTBI screening and had been followed up for at least 6 months after the initiation of therapy. Patients were excluded if their LTBI screening had been performed elsewhere, if their results were incomplete, or if they were lost to follow-up.

Patients underwent TST assessment alone prior to the commencement of biologic treatment before August 2013. The IGRA was available after August 2013; patients then underwent either single testing by TST or IGRA, or dual testing by both TST and IGRA for LTBI screening, at the discretion of the treating rheumatologists. Patients were divided into single testing group (TST or IGRA alone) or dual testing group (both TST and IGRA), according to the LTBI screening test they underwent prior to use of b/tsDMARDs. The following clinical data were collected: age; sex; co-morbidities including diabetes mellitus, chronic kidney/liver/lung disease, and heart failure; indication for use of b/tsDMARDs; and concurrent use of traditional DMARDs. Chest radiographs were performed for all patients; the results were considered normal if no abnormal changes were reported by the radiologists, who were blinded to the study conditions. The bacillus Calmette-Guérin vaccination status was documented if available. Patients were regularly followed up after the initiation of b/tsDMARDs; any development of active TB was recorded. Tuberculosis was defined by the 10th revision of the International Classification of Diseases, Clinical Modification (code A15-A19). All cases were confirmed by positive cultures or diagnostic pathological findings on tissue biopsy.

Tuberculin skin tests were performed by rheumatologists in the out-patient clinic. Either 0.1 mL (two tuberculin units) of Purified Protein Derivative (PPD)-RT23 (AJ Vaccines, Demark) or 0.1 mL (five tuberculin units) of PPD-S (JHP Pharmaceuticals, US) were injected intradermally to the forearms.14 The maximal transverse diameter of the induration was measured at 48 hours after injection. The test was regarded as positive if it was ≥10 mm, as recommended in local guidelines.10 If the area of induration was between 5 mm and 9 mm in patients on high-dose immunosuppressants, the test result was considered positive at the discretion of the treating rheumatologists. For patients who underwent the IGRA test, the following data were recorded: type of assay used, test results (positive, negative, or indeterminate), and test values if available. Two types of assays were used during the study period: the QuantiFERON-TB Gold (Qiagen, Germany) and A.TB IGRA (Haikou VTI Biological Institute, China).

Patients received isoniazid 300 mg daily for 9 months if they tested positive for LTBI. For patients with contra-indications or intolerance to isoniazid, rifampicin was used (450 mg daily for 4 months).10 Adverse events and early termination of drugs were recorded. Drug-induced liver toxicity was defined as an increase in alanine aminotransferase 5 times greater than the upper limit of normal in asymptomatic individuals, or 3 times greater than the upper limit of normal in symptomatic individuals.

The primary outcome of this study was the incidence of TB during b/tsDMARD therapy in the single testing and dual testing groups. Secondary outcomes included factors associated with TB, agreement between the TST and IGRA results, and the safety of LTBI treatment.

Descriptive data are presented as frequencies, means with standard deviations, or medians with ranges, as appropriate. Comparisons between clinical variables were performed using the Chi squared test or Fisher’s exact test for categorical variables, independent-samples t test for normally distributed continuous variables, or Mann-Whitney U test for non-normally distributed continuous variables. Survival analysis with the log rank test was used to control for differences in follow-up duration. Independent variables associated with the development of TB were identified by Cox regression modelling. Results were considered statistically significant if P<0.05. Agreement between the TST and IGRA results was evaluated by Cohen’s weighted kappa statistic. Kappa values represented the following degrees of agreement: >0.6, substantial agreement; 0.41 to 0.60, moderate agreement; 0.21 to 0.40, fair agreement; and <0.21, slight agreement.

In total, 248 patients were screened for eligibility. Thirty one patients were excluded from this study: 26 had incomplete TST or IGRA results, and five were lost to follow-up. Ultimately, 217 patients were included in the study: 121 in the single testing group and 96 in the dual testing group. The mean patient age was 53.2 years (standard deviation, 14.0 years), and 65% of included patients were women. Most patients had rheumatoid arthritis (56.7%), spondyloarthropathy (28.6%), or psoriatic arthritis (11.5%). Nearly half of the included patients (48.8%) were receiving prednisolone. Baseline clinical data were generally similar between the two groups; however, a significantly longer mean follow-up duration was observed after initiation of b/tsDMARDs in the single testing group (P<0.001), as shown in Table 1. The use of various b/tsDMARDs is also summarised in Table 1. Significantly more patients in the single testing group had ever been exposed to infliximab (P<0.001), whereas tofacitinib had been used more frequently in the dual testing group(P=0.021).

The frequencies of positive TST and IGRA test results in the single testing and dual testing groups are shown in Table 2. Two patients in the dual testing group had an inconclusive IGRA result. Isoniazid (28.9%) was administered less often in the single testing group than in the dual testing group (P=0.007). For the primary outcome, 10 of 217 (4.6%) patients developed TB during the follow-up. Significantly more patients in the single testing group developed TB, compared with patients in the dual test group (P=0.045; Table 2). However, the difference was not statistically significant when the follow-up duration was considered, using Kaplan-Meier analysis and log rank test (P=0.059) [Fig].

Of the 10 patients who developed TB during b/tsDMARD treatment, three had extrapulmonary or disseminated TB (Table 3). The median time from the initiation of b/tsDMARDs to the development of TB was 12.5 months (range, 2-32 months); five cases were diagnosed within the first 12 months. The indications for b/tsDMARDs in these patients were rheumatoid arthritis (five patients), spondyloarthropathy (three patients), psoriatic arthritis (one patient), and Behçet’s disease (one patient). Age at diagnosis of TB ranged from 23 to 73 years; the median age was 61 years. Among the nine patients who developed TB in the single testing group, eight underwent TST assessment for LTBI screening, while one underwent IGRA assessment. Eight of the nine patients had negative screening test results; one patient had positive TST results and was then administered isoniazid. However, that patient had abnormal liver function test results during treatment; thus, isoniazid was switched to rifampicin. The patient then developed pulmonary TB with TB lymphadenopathy and TB psoas abscess after 32 months of treatment with infliximab. Among the seven patients who had negative TST results in the single testing group, four had borderline TST results (area of induration, 6-9 mm); three had spondyloarthropathy and were receiving sulfasalazine alone before LTBI screening, whereas the remaining patient had rheumatoid arthritis and was receiving a combination of methotrexate and leflunomide. Regarding exposure to different b/tsDMARDs among patients with TB, nine had ever received anti-TNF-α treatment. The numbers of courses of b/tsDMARDs ever received and patients who ultimately developed TB are shown in Table 4. Among these b/tsDMARDs, only infliximab was significantly associated with development of TB (P<0.001). Cox regression analysis revealed that the use of infliximab was an independent predictor of TB (hazard ratio, 4.17; P=0.032; Table 5).

Levels of agreement between the TST and IGRA were calculated in subgroups of patients who received both tests and had definitive results (Table 6). Of the 94 patients, 24 (25%) had a discordant result: 18 had positive TST but negative IGRA results, while six had positive IGRA but negative TST results. The Cohen’s kappa value was 0.413 (moderate agreement); it decreased to 0.378 (fair agreement) in patients receiving any dose of prednisolone at baseline, and decreased further to 0.346 in patients receiving daily prednisolone ≥10 mg at the time of screening. Among traditional DMARDs used at the time of screening, leflunomide exhibited the lowest level of concordance (Cohen’s kappa=0.172).

In terms of adverse events, no mortality or major morbidity were observed in relation to isoniazid treatment. Three patients (2.4%) in the single testing group and five patients (5.2%) in the dual testing group required early termination of isoniazid; this termination was most commonly related to abnormal liver function test results, which were reversible after discontinuation of isoniazid. Two cases of allergy to isoniazid were noted in the single testing group, while one case of suspected isoniazid-induced lupus was noted in the dual testing group. There were no statistically significant between-group differences in side-effect profiles and rates of early termination of isoniazid therapy.

Tuberculosis remains a major complication related to use of b/tsDMARDs. Vigilant LTBI screening is needed to reduce the risk of TB reactivation. The overall prevalence of LTBI was 36.4% in this study; the test positivity rates of TST and IGRA across both groups were 33.6% and 22.5%, respectively. To the best of our knowledge, no local data are available regarding the prevalence of LTBI, because medical practitioners in Hong Kong are not required to report cases of LTBI. The estimated global burden of LTBI is 23.0%, although the prevalence of LTBI is expected to be higher in Asia.21

A significantly lower incidence of TB was observed in the dual testing group, which indicates that the dual test approach may be useful for prevention of TB during b/tsDMARD therapy, especially in TB endemic areas such as Hong Kong. The majority of patients in the single testing group underwent TST assessment alone. The number of IGRA-only cases in the single testing group was small; however, after 10 months of b/tsDMARD treatment, TB developed in one patient who had undergone IGRA alone. This patient was receiving prednisolone, methotrexate, and hydroxychloroquine at the time of screening; these results suggest that screening with the IGRA alone might exhibit reduced sensitivity in immunosuppressed patients. Indeed, a dual LTBI screening strategy with the TST and IGRA is described in different national guidelines.9 22

Nine of 10 patients with TB had received anti-TNF-α treatment before the development of TB. Multivariate analysis revealed that the use of infliximab was significantly associated with the development of TB. This association between use of infliximab and development of TB is consistent with the findings in a previous observational study in the United States.23 24 Previous studies assessing the associations of TB with infliximab, adalimumab, and etanercept also found higher TB incidence rates when patients received infliximab or adalimumab, compared with etanercept.25 26 The specific mechanisms underlying these differences in TB risk associated with the use of different anti-TNF agents are not fully understood. However, the differences may be related to the methods by which these agents neutralise TNF-α. Infliximab is a chimeric monoclonal antibody against TNF-α, which targets both soluble and membrane-bound forms of TNF-α. Similar binding activities are also exhibited by adalimumab and golimumab, which are both monoclonal antibodies against TNF-α. In contrast, etanercept is a fusion protein of the TNF-α receptor and human immunoglobulin G1 antibody, which has a much lower cytotoxicity for membrane-bound TNF-α expressing cells.27 An alternative explanation is that infliximab and adalimumab cause greater reduction in TB-responsive CD4 cells and suppression of antigen-induced interferon gamma production, when compared with etanercept; these results were demonstrated in a previous in-vitro study.28

In our study, the median time from initiation of b/tsDMARD treatment to the development of TB was 12.5 months, which differed from the median of 14 weeks reported by Keane et al.24 However, the majority of cases included in the report by Keane et al24 were from countries with low TB incidence, whereas TB is endemic in Hong Kong. Importantly, in Hong Kong, development of TB could result from reactivation of LTBI or from new TB infection. Therefore, it is important for rheumatologists to remain vigilant regarding the development of TB in all patients receiving b/tsDMARDs, regardless of the duration of b/tsDMARD treatment. Patients should receive education regarding possible symptoms associated with TB, and should be encouraged to seek medical attention when such symptoms arise. Regular chest X-ray surveillance should also be performed.

The level of agreement between the IGRA and TST was moderate in this study; discordant results were more commonly observed in patients receiving prednisolone or leflunomide at the time of screening. An immunosuppressed state has been suggested to induce a depressed T cell response, which may affect the accuracy of the tests; notably, a systematic review showed that both IGRA and TST results were significantly influenced by immunosuppressive therapy.29 In Hong Kong, local guidelines recommend a cut-off value of 10 mm for a positive TST result, whereas some experts recommend lowering the cut-off value to 5 mm in patients receiving high-dose prednisolone and/or immunosuppressants. Among the seven patients with TB in the single testing group, four had borderline TST results (area of induration, 6-9 mm). Unfortunately, we did not re-examine the results later (ie, at 72 hours), which might have enabled us to identify patients with delayed tuberculin response. Thus, TST sensitivity might have been impaired. Furthermore, both tests exhibit other limitations. The TST requires an intradermal injection and induration assessment, both of which are highly operator-dependent. The IGRA can yield a confusing indeterminate result in a considerable number of patients. Therefore, although universal dual testing might be difficult due to limited resources, its use should be strongly considered in select groups of patients. In patients with borderline TST results, additional IGRA testing might be helpful. Similarly, TST assessment should be considered if IGRA results are indeterminate.

Our study had a few limitations. First, the median follow-up duration was significantly longer in the single testing group. Notably, the incidence of TB during b/tsDMARD therapy was not significantly different between groups when the follow-up duration was considered, indicating that follow-up duration was an important confounder. The IGRA was only introduced in our centre in August 2013; all patients who began b/tsDMARD therapy before August 2013 only underwent single testing with the TST, which led to a longer follow-up duration in the single testing group. However, the risk of TB reactivation was highest in the first 90 days after initiation of b/tsDMARD therapy23; in the current study, the median follow-up duration of 27.5 months in the dual testing group was presumably sufficient to assess patients during that period. Second, there was a disproportionately high rate of infliximab use in the single testing group, compared with the dual group, such that infliximab use constituted an important confounder. Because of the retrospective nature of this study, all patients underwent TST assessment alone before August 2013, and infliximab was one of the first biologic agents available; this sequence of events is likely to explain the higher usage of infliximab in the single testing group. Third, the number of IGRA-only patients in this study was relatively small; therefore, we could not draw robust conclusions regarding the reliability of LTBI screening by IGRA alone, compared with dual testing. However, among the six patients who underwent IGRA assessment alone, one developed TB after 10 months of sequential tocilizumab and tofacitinib treatment, despite an initial negative screening result. Further multicentre prospective studies are needed to compare the efficacy and safety of either TST or IGRA alone with dual testing, in terms of preventing the development of TB during b/tsDMARD treatment. Lastly, some patients on immunosuppressants underwent TST assessment and exhibited borderline area of induration; these test results might have been regarded as positive at the discretion of the attending rheumatologists. The lack of a clear definition might have constituted another source of bias.

Although the difference in follow-up duration may have been a confounding factor, the results of this study suggest that dual LTBI screening with the TST and IGRA might be a useful and safe strategy to reduce the incidence of TB in patients with rheumatic diseases receiving b/tsDMARDs in an endemic area. Use of infliximab was significantly associated with development of TB. The level of agreement between the TST and IGRA was moderate; it was lower in patients receiving prednisolone or leflunomide at time of LTBI screening. Dual LTBI screening should be strongly considered in patients with rheumatic diseases receiving steroid or leflunomide treatment at the time of LTBI screening, when either test result is equivocal, or when use of infliximab is anticipated.

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

Concept or design: I Tang, H So, R Yip.
Acquisition of data: I Tang, H So, L Luk, V Wong, S Pang, V Lao.
Analysis or interpretation of data: I Tang, H So, L Luk, V Wong, S Pang, V Lao.
Drafting of the manuscript: I Tang, H So.
Critical revision of the manuscript for important intellectual content: L Luk, V Wong, S Pang, V Lao, R Yip.

All authors have disclosed no conflicts of interest.

This research was presented in the Hong Kong Hospital Authority Convention 2019, East Asian Group of Rheumatology 2019 and Annual Meeting of American College of Rheumatology 2019.

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

This study conformed to the provisions of the Declaration of Helsinki and was approved by the Hospital Authority Research Ethics Committee (Kowloon Central/Kowloon East) (Ref KC/KE-19-0241/ER-3). The requirement for patient consent was waived by the ethics committee. The study contains no identifiable personal or medical information.

1. Gómez-Reino JJ, Carmona L, Valverde VR, Mola EM, Montero MD; BIOBADASER Group. Treatment of rheumatoid arthritis with tumor necrosis factor inhibitors may predispose to significant increase in tuberculosis risk: a multicenter active-surveillance report. Arthritis Rheum 2003;48:2122-7. Crossref

2. Winthrop KL, Baxter R, Liu L, et al. Mycobacterial diseases and antitumour necrosis factor therapy in USA. Ann Rheum Dis 2013;72:37-42. Crossref

3. Ridker PM, Everett BM, Thuren T, et al. Antiinflammatory therapy with canakinumab for atherosclerotic disease. N Engl J Med 2017;377:1119-31. Crossref

4. Winthrop KL, Mariette X, Silva JT, et al. ESCMID Study Group for Infections in Compromised Hosts (ESGICH) Consensus Document on the safety of targeted and biological therapies: an infectious diseases perspective (Soluble immune effector molecules [II]: agents targeting interleukins, immunoglobulins and complement factors). Clin Microbiol Infect 2018;24 Suppl 2:S21-40. Crossref

5. Kourbeti IS, Ziakas PD, Mylonakis E. Biologic therapies in rheumatoid arthritis and the risk of opportunistic infections: a meta-analysis. Clin Infect Dis 2014;58:1649-57. Crossref

6. Koike T, Harigai M, Inokuma S, et al. Effectiveness and safety of tocilizumab: postmarketing surveillance of 7901 patients with rheumatoid arthritis in Japan. J Rheumatol 2014;41:15-23. Crossref

7. Hasan T, Au E, Chen S, Tong A, Wong G. Screening and prevention for latent tuberculosis in immunosuppressed patients at risk for tuberculosis: a systematic review of clinical practice guidelines. BMJ Open 2018;8:e022445. Crossref

8. Chen DY, Su WJ, Shen GH, et al. 2012 Screening and management of tuberculosis infection in patients scheduled for tumor necrosis factor-alpha inhibitors: consensus recommendations from the Taiwan Rheumatology Association. Formos J Rheumatol 2012;26:8-14.

9. Mazurek GH, Jereb J, Vernon A, et al. Updated guidelines for using interferon gamma release assays to detect Mycobacterium tuberculosis infection—United States, 2010. MMWR Recomm Rep 2010;59:1-25.

10. Lau CS, Chia F, Dans L, et al. 2018 Update of the APLAR recommendations for treatment of rheumatoid arthritis. Int J Rheum Dis 2019;22:357-75. Crossref

11. Kleinert S, Tony HP, Krueger K, et al. Screening for latent tuberculosis infection: performance of tuberculin skin test and interferon-gamma release assays under real-life conditions. Ann Rheum Dis 2012;71:1791-5. Crossref

12. Bartalesi F, Vicidomini S, Goletti D, et al. QuantiFERON-TB Gold and the TST are both useful for latent tuberculosis infection screening in autoimmune diseases. Eur Respir J 2009;33:586-93. Crossref

13. Ruan Q, Zhang S, Ai J, Shao L, Zhang W. Screening of latent tuberculosis infection by interferon-γ release assays in rheumatic patients: a systemic review and meta-analysis. Clin Rheumatol 2016;35:417-25. Crossref

14. So H, Yuen CS, Yip RM. Comparison of a commercial interferon-gamma release assay and tuberculin skin test for the detection of latent tuberculosis infection in Hong Kong arthritis patients who are candidates for biologic agents. Hong Kong Med J 2017;23:246-50. Crossref

15. Aletaha D, Neogi T, Silman AJ, et al. 2010 Rheumatoid arthritis classification criteria: an American College of Rheumatology/European League Against Rheumatism collaborative initiative. Ann Rheum Dis 2010;69:1580-8. Crossref

16. Tillett W, Costa L, Jadon D, et al. The ClASsification for Psoriatic ARthritis (CASPAR) Criteria—a retrospective feasibility, sensitivity, and specificity study. J Rheumatol 2012;39:154-6. Crossref

17. Rudwaleit M, van der Heijde D, Landewé R, et al. The Assessment of SpondyloArthritis International Society classification criteria for peripheral spondyloarthritis and for spondyloarthritis in general. Ann Rheum Dis 2011;70:25-31. Crossref

18. International Study Group for Behçet’s Disease. Criteria for diagnosis of Behçet’s disease. Lancet 1990;335:1078-80.

19. Yamaguchi M, Ohta A, Tsunematsu T, et al. Preliminary criteria for classification of adult Still’s disease. J Rheumatol 1992;19:424-30.

20. Bohan A, Peter JB. Polymyositis and dermatomyositis (first of two parts). N Engl J Med 1975;292:344-7. Crossref

21. Houben RM, Dodd PJ. The global burden of latent tuberculosis infection: a re-estimation using mathematical modelling. PLoS Med 2016;13:e1002152. Crossref

22. Cantini F, Nannini C, Niccoli L, et al. Guidance for the management of patients with latent tuberculosis infection requiring biologic therapy in rheumatology and dermatology clinical practice. Autoimmun Rev 2015;14:503-9. Crossref

23. Wallis RS, Broder MS, Wong JY, Hanson ME, Beenhouwer DO. Granulomatous infectious diseases associated with tumor necrosis factor antagonists. Clin Infect Dis 2004;38:1261-5. Crossref

24. Keane J, Gershon S, Wise RP, et al. Tuberculosis associated with infliximab, a tumor necrosis factor alpha–neutralizing agent. N Engl J Med 2001;345:1098-104. Crossref

25. Navarra SV, Tang B, Lu L, et al. Risk of tuberculosis with anti-tumor necrosis factor-α therapy: substantially higher number of patients at risk in Asia. Int J Rheum Dis 2014;17:291-8. Crossref

26. Yamauchi PS, Bissonnette R, Teixeira HD, Valdecantos WC. Systematic review of efficacy of anti-tumor necrosis factor (TNF) therapy in patients with psoriasis previously treated with a different anti-TNF agent. J Am Acad Dermatol 2016;75:612-8. Crossref

27. Baddley JW, Cantini F, Goletti D, et al. ESCMID Study Group for Infections in Compromised Hosts (ESGICH) Consensus Document on the safety of targeted and biological therapies: an infectious diseases perspective (Soluble immune effector molecules [I]: anti-tumor necrosis factor-α agents). Clin Microbiol Infect 2018;24 Suppl 2:S10-20. Crossref

28. Saliu OY, Sofer C, Stein DS, Schwander SK, Wallis RS. Tumor-necrosis-factor blockers: differential effects on mycobacterial immunity. J Infect Dis 2006;194:486-92. Crossref

29. Shahidi N, Fu YT, Qian H, Bressler B. Performance of interferon-gamma release assays in patients with inflammatory bowel disease: a systematic review and meta-analysis. Inflamm Bowel Dis 2012;18:2034-42. Crossref

In 2016, prostate cancer was the third most common type of cancer in men in Hong Kong, with 1912 new cases diagnosed.1 Increased disease awareness and the use of prostate-specific antigen (PSA) screening have increased the frequency of diagnosis of early-stage prostate cancers.2

While active surveillance has become increasingly common for low-risk localised prostate cancer, robot-assisted laparoscopic prostatectomy (RALP) and external beam radiotherapy (RT) remain the mainstay treatments for localised prostate cancer. Despite the effectiveness of both treatments, treatment-related adverse effects are not uncommon. Prostatectomy can result in urinary incontinence and erectile dysfunction, which can greatly affect patients’ quality of life.3 4 The adverse effects of RT include irritative voiding symptoms, radiation cystitis, and proctitis.3 4 Therefore, comparing the functional outcomes of different treatments is an important task for clinicians and patients during treatment planning. However, information about the functional and patient-reported outcomes (PROs) of RALP and RT is lacking in Asia. Therefore, in this study, we aimed to evaluate the intermediate-term functional and PROs of prostatectomy and RT for patients diagnosed with localised prostate cancer.

We conducted a retrospective review of the treatment outcomes of patients with localised prostate cancer diagnosed from January 2010 to December 2011 and treated by either RALP or RT at a university hospital centre in Hong Kong. As the review took place in December 2016, the follow-up period for all patients was 5 to 7 years, allowing us to measure intermediate-term outcomes.5 This cohort of patients was chosen to ensure a relatively stable group of patients with at least 5 years’ follow-up but to avoid symptoms related to disease recurrence and progression. The study was approved by the institution’s ethics committee (the Joint Chinese University of Hong Kong–New Territories East Cluster Clinical Research Ethics Committee, Ref CREC 2016.373) and conducted according to the Declaration of Helsinki.

Patients’ demographic characteristics, pretreatment disease characteristics, treatment outcomes, and complications were retrieved from hospital records. The patients’ quality of life was assessed using a validated PRO measure, the Expanded Prostate Cancer Index Composite (EPIC) questionnaire.6 The EPIC questionnaire has been validated as a measure of disease-specific function. The full version of the EPIC questionnaire consists of 50 questions divided into four treatment complication-related domains: urinary symptoms, bowel symptoms, sexual symptoms, and hormonal treatment-related symptoms. Each domain is further subdivided into different functional components that evaluate the severity of adverse effects and level of bother (ie, distress) related to the symptoms and adverse effects. Most of the questions are scored on a Likert scale, with the majority using a 5-point scale. The calculation of final quality of life was based on the EPIC scoring guideline; the final score ranged from 0 to 100, with higher scores representing better quality of life. A validated Chinese version of this questionnaire was used.7 Copies of the questionnaire were mailed to the patients’ home addresses in early December 2016, and the patients were given 2 weeks to complete the questionnaire. Patients who participated in the PRO segment consented to their participation in the study.

All of the statistical analyses were performed using IBM SPSS Statistics (Windows version 25.0; IBM Corp., Armonk [NY], United States). Personnel involved in the collection and input of EPIC data were blinded to the patients’ clinical information. The results were deemed statistically significant at a P value of <0.05. The results for the RALP and RT groups were compared using the Chi squared test, Fisher’s exact test, or Student’s t test. The EPIC scores for the two treatment arms were compared using the Mann-Whitney U test. The details of individual symptoms and their impact on quality of life were also assessed. Missing values were excluded fromanalysis.

Ninety three men aged 58 to 84 years were recruited to take part in the study. Eighty five of these patients were still alive during the study period, but eight died due to non-prostate cancer–related causes (RALP n=1, RT n=7). Thirty (32.3%) patients had been treated by RALP, whereas 63 (67.7%) had received RT. Fifty two patients returned the questionnaires, giving a response rate of 61.2% (Table 1).

The mean ages of the patients who underwent RALP and RT were 65.8±6.07 and 69.81±5.48 years, respectively, and the RALP patients were significantly younger than the RT patients (P=0.002). More than half of the patients in each group were defined as Eastern Cooperative Oncology Group category 1.

The pretreatment serum PSA levels of the patients who underwent RALP (9.6±4.18 ng/mL) were also significantly lower than those of the patients who underwent RT (29.44±32.03 ng/mL, P<0.001). One third of the patients in the RT (33.3%) and half of the patients in the RALP (50.0%) group were diagnosed as T1c. Sixty percent of the patients who underwent RALP and 38.1% of the patients who underwent RT had Gleason scores <7; only 23.3% of the RALP patients and 36.5% of the RT patients had scores >7. The Gleason scores for the patients who underwent RALP were lower than those for the patients who underwent RT, although the difference was not statistically significant (P=0.185). There was a trend of more patients with D’Amico high-risk cancer in the RT group than the RALP group (Table 2).

All patients who underwent RALP achieved undetectable serum PSA levels after surgery, whereas 46.0% of the patients who underwent RT had undetectable PSA levels. During the follow-up period, 12 (19.0%) patients from the RT group and two (6.7%) patients from the RALP group developed biochemical recurrence, but this difference was statistically insignificant (P=0.213). The majority of the patients who underwent RT and had biochemical recurrence (75%) opted for androgen deprivation therapy for further treatment, whereas all RALP patients with biochemical recurrence (100%) chose salvage RT as an additional therapy. About 5% of the patients in the RT group developed metastatic disease during the follow-up period, while none in the RALP group did so (Table 3).

The clinical information indicates that the majority of the patients in both groups suffered urinary symptoms to some degree during the first 3 months after treatment. However, short-term bowel dysfunction was more prevalent in the patients who underwent RT (46%) than in the patients who underwent RALP (6.7%; P<0.001).

Regarding intermediate-term complications reported during physician consultation, 70% of the RALP group patients reported urinary incontinence, compared with only 3.2% of the RT group patients (P<0.001). In contrast, more RT group patients (87.3%) than RALP group patients (50%) experienced lower urinary tract symptoms other than urinary incontinence (P<0.001). Furthermore, 36.5% of the RT group patients complained of perirectal bleeding, which was not reported by the RALP group patients (P<0.001). Erectile dysfunction was more prevalent in the RALP group (85.2%) than in the RT group (23.2%), with 60.9% of the RALP group patients requesting treatment for this condition (P<0.001).

Six (9.5%) of the patients who underwent RT had unplanned hospitalisations related to their prostate cancer or its treatment during the follow-up period. Four were due to haematuria, one was due to rectal bleeding, and one was due to acute urine retention. No unplanned hospital admissions were observed in the RALP group (Table 3).

The EPIC questionnaire was used to assess the patients’ reported outcomes with regard to their choice of treatment and their urinary, bowel, and sexual function over the 4 weeks before the survey (Table 4). The overall response rate was 61.2% (52 out of 85 patients), and the response rates for the RT group (33 out of 56, 58.9%) and the RALP group (19 out of 29, 65.5%) were similar.

The RALP group patients had poorer median (interquartile range) EPIC urinary summary scores (81.5 [18.3]) than the RT group patients had (88.9 [17.9]; P=0.016). Significantly poorer urinary function and incontinence results were observed in the RALP group (75.9 [20.4] and 60.5 [31.8], respectively] than in the RT group (93.6 [16.2] and 91.8 [14.5], respectively) [both P<0.001]. However, the two groups reported similar EPIC scores for urinary bother and urinary irritation/obstruction (Table 4 and Fig 1).

The patients who underwent both RT and RALP reported good bowel function, with median EPIC bowel summary scores of 92.9 (15.6) and 92.0 (19.6), respectively. There was no statistically significant difference between the two groups in terms of either bowel function or bowel bother (Fig 2).

Poor median EPIC sexual summary and function scores were reported by both the RT (22.4 [23.1] and 2.8 [13.4], respectively) and the RALP (31.7 [32.2] and 10.6 [32.1], respectively) groups. Although low scores on the sexual bother subscale were also reported, they were less poor than those for the function subscale. However, no significant differences were observed between the RT and RALP groups on any of the sexual function scales (Fig 3).

This study was the first to investigate the PROs of treatment for localised prostate cancer in a Chinese cohort and the intermediate-term complications experienced by patients. As expected, the complication profiles and PROs differed between patients who underwent radical prostatectomy and those who received external beam RT as treatment for localised prostate cancer. Compared with the RT group, the RALP group had less bowel disturbance immediately after treatment and less rectal bleeding during follow-up. The RALP group reported more urinary incontinence during follow-up, whereas the RT group experienced a greater frequency of other urinary symptoms. In addition, more unplanned admissions (mainly due to haematuria and rectal bleeding) were observed among the RT patients. However, the PRO results showed that the two groups achieved similar scores in other domains, except that the patients who underwent RALP had significantly lower incontinence and function scores.

Prostate cancer is the second most commonly diagnosed cancer in men in the world and the fifth leading cause of death from cancer in men.8 Although prostate cancer has a lower overall incidence in Asia than in Western countries, its incidence in Asia is rising.9 The rapid increase in incidence has been partly related to the increased usage of serum PSA for early cancer detection, which has resulted in increased detection of earlier-stage cancers. As a result, more patients can receive treatment with curative intent. This may be one cause of the decreasing prostate cancer mortality rate in many countries.8 9

Currently, most Asian patients with localised prostate cancer choose active intervention as treatment; active surveillance is not commonly practised in the region.10 The two most commonly performed treatments for prostate cancer in Asia are RT and RALP. Therefore, our study focused on assessing the outcomes of these two treatment modalities. As these two treatments have similar oncological outcomes for localised cancer,11 the choice of treatment depends on other factors, such as the risk to patients associated with surgery and possible post-treatment complications. We also noticed that patients’ age and performance status may affect the choice of treatment, as in our cohort, older patients and those with more co-morbidities were more likely to choose RT than surgery. Potential treatment-related complications were also a major consideration during the decision making process.

Some cross-sectional12 and prospective3 13 studies have investigated the effects of different types of treatment on the PROs of patients with localised prostate cancer. In general, these studies have found that different treatments are associated with different patterns of complications. Compared with RT, RALP has been shown to have greater negative effects on sexual function and urinary incontinence, whereas RT is associated with more bowel symptoms, especially bloody stools. Interestingly, sexual function has been found to gradually decline even under active surveillance, which is related to the natural ageing process.3 Our results indicated similar patterns of complications, with more RALP group patients reporting urinary incontinence and more RT group patients reporting bloody stools during follow-up. Similar to the results of the ProtecT study,4 we found that by around 6 years (72 months) after treatment, the two treatment groups achieved similar scores for voiding symptoms, bowel bother, and bowel function. In terms of sexual function, although both treatment groups had low sexual summary and function scores, the RT group generally had lower scores than the RALP group, although the difference was statistically insignificant. This finding was inconsistent with the results of previous studies.4 However, the higher mean age and greater reported use of hormonal therapy of our RT cohort compared with our RALP cohort may have resulted in poorer sexual function. Nevertheless, the two groups’ sexual bother scores were similar, which suggests that the poorer sexual function reported by the RT group compared with the RALP group did not cause more bother to the patients.

There is little information available about the differences in PROs yielded by different modes of treatment for localised prostate cancer in Asian populations. Only one early report has been conducted with a Japanese population, in which urinary domain scores initially worsened after prostatectomy and brachytherapy, gradually improving later.14 Compared with other treatment modalities, external beam RT resulted in significantly lower bowel summary scores. Therefore, our study provides important information about the outcomes of treatment of localised prostate cancer, particularly the use of a relatively new treatment modality, RALP. More prospective multicentre studies would be helpful to provide more information to support patient counselling.

We not only investigated PROs, but also recorded symptoms/complications after treatment (ie, physician-reported outcomes) and unplanned admissions due to disease- or treatment-related complications. Although some studies have suggested that PROs are more accurate than physician-reported outcomes,15 we believed that combining the two would provide a more comprehensive picture of patients’ post-treatment course. Our findings revealed not only differences in the patterns of urinary and bowel symptoms experienced by the two groups of patients, but also a trend of more unplanned hospital admissions in the RT group than the RALP group. Most of these admissions were related to the adverse effects of RT, such as haematuria and perirectal bleeding. Along with PROs, this information is important for patient counselling.

The study had several limitations. The sample was relatively small and involved only local Chinese participants. Currently, data on PROs for the local population are lacking. We hope our data can help to provide some information for patients and physicians during their decision making process. Our results were quite similar to the reported PROs for localised prostate cancer in the literature. Further, Tyson et al16 suggested there is not much racial difference in PROs amongst patients with prostate cancer. Second, as data on treatment-related complications were collected retrospectively, the information about the incidence of complications may have been biased, particularly for erectile dysfunction. Further prospective studies should provide more information on this topic. Furthermore, the number of patients was unbalanced between the two groups, as more RT-treated patients were recruited. This may have reflected the clinicians’ and patients’ preference. In addition, baseline disease characteristics differed between the two groups, with more high-risk patients in the RT group than in the RALP group. Therefore, oncological outcomes were not analysed further in this study. Finally, as active surveillance has only become common clinical practice in recent years, the number of patients receiving this treatment a few years ago was small. Therefore, we did not include this treatment modality in our initial planning, and no information about this treatment was available for comparison with findings for the other treatments.

Robot-assisted laparoscopic prostatectomy and external beam RT are associated with different patterns of complications and PROs with respect to urinary, bowel, and sexual function. Our results provide valuable information for counselling patients regarding treatment choices and outcome expectations.

Concept or design: CF Ng.
Acquisition of data: KY Kong, CY Li, JKT Li, NY Li, BPK Ng.
Analysis or interpretation of data: KY Kong, CY Li, JKT Li, NY Li, BPK Ng, SCH Leung.
Drafting of the manuscript: CF Ng, SCH Leung.
Critical revision of the manuscript for important intellectual content: All authors.

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 editors of the journal, CF Ng and JYC Teoh were not involved in the peer review process. Other authors have disclosed no conflicts of interest.

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

The study was conducted in accordance with good clinical practice guidelines and the Declaration of Helsinki. Ethics approval was granted by the Joint Chinese University of Hong Kong–New Territories East Cluster Clinical Research Ethics Committee (Ref CREC 2016.373). Informed consent was obtained from the patients when they completed the questionnaires.

1. Hong Kong Cancer Registry, Hospital Authority, Hong Kong SAR Government. Available from: http://www3. ha.org.hk/cancereg/pdf/factsheet/2016/prostate_2016. pdf. Accessed 6 Sep 2019.

2. Wong HF, Yee CH, Teoh JY, et al. Time trend and characteristics of prostate cancer diagnosed in Hong Kong (China) in the past two decades. Asian J Androl 2019;21:104-6. Crossref

3. Wallis CJ, Mahar A, Cheung P, et al. New rates of interventions to manage complications of modern prostate cancer treatment in older men. Eur Urol 2016;69:933-41. Crossref

4. Donovan JL, Hamdy FC, Lane JA, et al. Patient-reported outcomes after monitoring, surgery, or radiotherapy for prostate cancer. N Engl J Med 2016;375:1425-37. Crossref

5. Prabhu V, Lee T, McClintock TR, Lepor H. Short-, intermediate-, and long-term quality of life outcomes following radical prostatectomy for clinically localized prostate cancer. Rev Urol 2013;15:161-77.

6. Wei JT, Dunn RL, Litwin MS, Sandler HM, Sanda MG. Development and validation of the expanded prostate cancer index composite (EPIC) for comprehensive assessment of health-related quality of life in men with prostate cancer. Urology 2000;56:899-905. Crossref

7. Lee TK, Poon DM, Ng AC, et al. Cultural adaptation and validation of the Chinese version of the expanded prostate cancer index composite. Asia Pac J Clin Oncol 2018;14 Suppl 1:10-5. Crossref

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

9. Wong MC, Goggins WB, Wang HH, et al. Global incidence and mortality for prostate cancer: analysis of temporal patterns and trends in 36 countries. Eur Urol 2016;70:862- 74. Crossref

10. Lojanapiwat B, Lee JY, Gang Z, et al. Report of the third Asian Prostate Cancer Study Meeting. Prostate Int 2019;7:60-7. Crossref

11. Hamdy FC, Donovan JL, Lane JA, et al. 10-Year outcomes after monitoring, surgery, or radiotherapy for localized prostate cancer. N Engl J Med 2016;375:1415-24. Crossref

12. Nicolaisen M, Müller S, Patel HR, Hanssen TA. Quality of life and satisfaction with information after radical prostatectomy, radical external beam radiotherapy and postoperative radiotherapy: a long-term follow-up study. J Clin Nurs 2014;23:3403-14. Crossref

13. Barocas DA, Alvarez J, Resnick MJ, et al. Association between radiation therapy, surgery, or observation for localized prostate cancer and patient-reported outcomes after 3 years. JAMA 2017;317:1126-40 Crossref

14. Kakehi Y, Takegami M, Suzukamo Y, et al. Health related quality of life in Japanese men with localized prostate cancer treated with current multiple modalities assessed by a newly developed Japanese version of the Expanded Prostate Cancer Index Composite. J Urol 2007;177:1856- 61. Crossref

15. Gordon BE, Chen RC. Patient-reported outcomes in cancer survivorship. Acta Oncol 2017;56:166-73. Crossref

16. Tyson MD, Alvarez J, Koyama T, et al. Racial variation in patient-reported outcomes following treatment for localized prostate cancer: results from the CEASAR Study. Eur Urol 2017;72:307-14. Crossref

Hypertension is highly prevalent and a key risk factor for cardiovascular disease.1 Less well recognised by patients and health professionals alike is that some of the pharmacotherapies used to treat hypertension may adversely impact other cardiovascular risk factors by causing weight gain. More specifically, the weight gain effects of beta-adrenergic antagonists (beta-blockers) have been highlighted by many studies of Caucasian patients.2 3 4 5 Many guidelines no longer list beta-blockers as first-line antihypertensives.6 7 8 However, the 2018 hypertension guidelines published by the European Society of Cardiology and the European Society of Hypertension,9 the joint statement published in 2012 by the European Society of Hypertension and the European Association for the Study of Obesity,10 and a position paper of The Obesity Society and the American Society of Hypertension11 still advocate the use of beta-blockers in patients with both hypertension and obesity because beta-blockade is more effective in lowering blood pressure (BP) in patients with obesity than in patients who are thinner.12 In Hong Kong, two recent large database studies found that beta-blockers are still the most commonly prescribed antihypertensives, although they are used relatively less in younger patients aged <55 years.13 14 Furthermore, only 4% of Hong Kong patients have their antihypertensive treatment upgraded by changing from beta-blockers to firstline agents.15

The generally increasing prevalence of obesity among older adults is an important factor in drug-induced weight gain.16 Studies have shown that obesity in older people is associated with functional impairment and co-morbidity, including hypertension, type 2 diabetes, coronary heart disease, heart failure, and dementia.17 18 19 These chronic conditions, for which being overweight is a risk factor, are also worryingly increasing in prevalence.20 21 22 This renders weight management in older persons an important health issue.

To date, there has been a lack of research to confirm the weight gain effects of beta-blockers in non-Caucasian populations, and this is particularly germane because so many older Chinese patients with hypertension are still receiving beta-blocker therapy.

The objective of this study was to explore the association between beta-blocker use, hypertension, and overweight/obesity in a cohort of older Chinese people. The specific objectives were to: (1) identify the proportion of patients prescribed beta-blockers; (2) compare the body mass index (BMI) of beta-blocker users with non-users; and (3) compare the effects of different beta-blockers on BMI.

A post-hoc analysis was undertaken using an existing dataset comprising the medication profiles of a cohort of community-dwelling older adults in Hong Kong. The data were originally collected (July to August in 2016) for a primary study seeking to explore the relationship between diet and the prevalence of atrial fibrillation.23 The study was approved by the Survey and Behavioural Research Ethics Committee of The Chinese University of Hong Kong and was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all study participants. Confidentiality agreement forms were signed by all data collectors. STROBE reporting guidelines were implemented in this manuscript.

The original study cohort comprised 1665 people attending one of 32 neighbourhood elderly recreational community centres in Hong Kong. The inclusion criteria included those who were Hong Kong Chinese residents, aged ≥65 years, living in the community, and able to speak and understand Cantonese. The exclusion criteria included those with terminal health conditions and/or significant cognitive impairment that would preclude participation because of communication barriers (eg, severe mental illness, dementia).

All primary study data were collected during a summer community outreach programme—a territory-wide medical outreach service in Hong Kong provided by volunteer students from the Faculty of Medicine, The Chinese University of Hong Kong. The data were originally collected via face-to-face interviews with the participants in Cantonese and recorded on paper-based questionnaires by trained volunteer students.

At each outreach visit, participants’ body weight, height, and BP were measured, BMI calculated, medication profile extracted, and demographic data recorded. Calculated BMIs were compared with the Asian BMI classification set by the World Health Organization, in which BMI ≥23 and ≥25 are considered as overweight and obese, respectively.24 A self-reported diagnosis of hypertension and diabetes was recorded and verified against the patient’s current medication profile. The self-reported diagnosis of hypertension was the only evidence to determine whether the patient had hypertension or not. As part of the logistics and service provided by our outreach, BP was also measured using an Omron HEM-7011^TM electronic blood pressure monitor (Omron Healthcare, Kyoto Japan), which has an “A/A” performance classification under British Hypertension Society criteria (ie, indicating that at least 80% and 95% of readings are within an absolute difference of 5 to 10 mm Hg from each other, respectively25). The readings were compared against 2017 American Heart Association guidelines,6 but such readings were not used to diagnose hypertension. To ensure accuracy, for each patient, the BP measurements were repeated after 10 minutes of rest when the first reading was found to be elevated (ie, BP >120/80 mm Hg). When a beta-blocker had been prescribed, the name, dose, frequency, and route of administration of the agent were additionally recorded. The consumption of other drugs known to alter body weight was also recorded and listed in Table 1.

Data handling (data entry, verification, and analysis) was computerised using SPSS (Windows version 23.0; IBM Corp, Armonk [NY], United States). Only the patients with complete drug profiles and all variables recorded were included. Any incomplete drug profile or missing variables (eg, BMI) were considered missing data and were not included in the analysis. Independent samples t tests were used to compare BMI values between people taking different beta-blockers. Binary logistic regression was used to verify the contributions of various drug classes and co-morbidities on a stratified binary BMI parameter (normal and underweight vs overweight and obesity). Chi squared tests were used to detect any significant differences in the proportion of patients with high BMI across different types of beta-blockers. A P value of <0.05 was considered statistically significant.

Among the 1665 individuals participating in the summer community outreach programme, data pertaining to 1053 were included in the analysis after screening against the inclusion and exclusion criteria. The participants’ demographics and characteristics are shown in Table 1. People taking beta-blockers were generally heavier (mean BMI 24.6±3.7 vs 23.6±3.6, P<0.001). From the perspective of BP measurement readings, around 80% of the participants (804/1053) had elevated BP measurements during the outreach visit; of these, one third (286/804) had no history of hypertension and were not on any medications likely to be used for treating hypertension. The remaining two thirds (518/804) self-reported having hypertension. From the perspective of self-reported hypertension, 81.9% of those with self-reported hypertension (513/626) had hypertensive readings, and only 6% (40/626) had normotensive readings; the remaining ones had borderline hypertensive readings (12%, 73/626).

Overall, 185 (18%) of the 1053 participants were using one beta-blocker. Among the range of beta-blockers used, atenolol and metoprolol were the most frequently prescribed (40% and 53%, respectively), followed by propranolol (5%) and bisoprolol (2%). Among the beta-blocker users, the majority (95%, 175/185) had a self-reported history of hypertension. In those participants with self-reported hypertension who were not using beta-blockers (n=451), the main antihypertensive agents prescribed were calcium channel blockers (71.3%), angiotensin-converting enzyme inhibitors (22.3%), angiotensin receptor blockers (12.1%), alpha-blockers (7.9%), methyldopa (4.2%), hydralazine (1.0%), and reserpine (0.2%).

Overall, 54.1% (570/1053) of the participants were overweight or obese. The summary of patients’ BMI with or without beta-blockers is summarised in Table 2. After adjusting for consumption of various weight-altering drug classes, binary logistic regression showed that beta-blockers were the only drug class that made a significant contribution to stratified BMI status (ie, classification as overweight or obese) [Table 3]. Among those prescribed beta-blockers (n=185), a significantly higher proportion was either overweight or obese compared with those who were not taking beta-blockers (n=868) [64.8% vs 51.8%, P=0.002; Table 2]. This difference is most evident when comparing the proportion of patients with obesity across beta-blocker users and non-users (45.9% vs 32.1%, respectively; P=0.002).

Participants deemed to have hypertension (based solely on self-reported diagnosis of hypertension plus verification against medication profile, but not on BP measurement during the outreach service) had a significantly higher BMI than those who were normotensive (mean BMIs: 24.3±3.6 vs 22.9±3.5, respectively; P<0.001). Although this difference may not be clinically significant, it triggered further sub-analysis on participants with hypertension diagnosis. A sub-analysis on the patients with self-reported diagnosis of hypertension was performed to confirm the association between the use of beta-blockers and BMI (Table 4). Among these participants with self-reported diagnosis of hypertension (n=626), a significantly higher proportion of patients with obesity (BMI >25) was observed in those using atenolol compared with those taking metoprolol (58.9% vs. 38.5%, P=0.031) or those who did not use any beta-blockers (58.9% vs 38.4%, P=0.007). Those using atenolol had a significantly higher BMI than those who did not use beta-blockers (mean BMIs: 25.3±3.5 vs 24.1±3.6, P=0.01).

Binary logistic regression analysis found that loop diuretics were associated with BMI reduction. However, concerning mechanism and indication, loop diuretics function by enhancing salt and water excretion and are clinically used to maintain euvolaemia or prevent volume expansion.26 In other words, it makes no significant contribution to alteration of dry body weight, which is used for BMI determination and obesity evaluation. Therefore, despite the above findings, loop diuretics were excluded from our further analysis.

Our study presents preliminary findings regarding the potential real-world impact of beta-blockers on weight, noting the high proportion of community-dwelling older adults using these agents. We found a significant association between obesity/overweight with the use of beta-blockers in older Chinese adults. The difference was largely driven by the strong association between obesity and atenolol (rather than other beta-blockers). This may have important ramifications on therapeutic choice if the association is causal.

The potential mechanisms by which beta-blockers may induce weight gain include reduction of total energy expenditure (by 5%-10%), which may involve (1) decreased resting energy expenditure; (2) increased feelings of tiredness, with subsequent reduction of non-exercise-associated thermogenesis; (3) inhibition of lipolysis; and (4) enhancement of insulin resistance.27

Many studies of Caucasian populations have reported the weight gain effects of beta-blockers, being associated with a mean weight gain of 1.2 kg (range, -0.4 kg to 3.5 kg),27 which could explain our findings of 2.6 kg higher mean body weight in people taking beta-blockers (mean body weight of people who did vs did not take beta-blockers: 57.8±9.7 kg vs 55.2±9.8 kg, respectively; P=0.002 respectively), but we do not have longitudinal data to make pre-versus post-drug commencement comparisons. Several studies have compared impact on weight between selected beta-blockers and alternative antihypertensive medications, ie, atenolol versus captopril,2 metoprolol versus thiazide diuretics,3 atenolol versus chlorthalidone,4 and atenolol versus nifedipine,5 with all reporting weight gain (or reduced weight loss) in the beta-blocker treatment group. A long-term follow-up study also reported sustained weight gain in a propranolol treatment group compared with a placebo group.28

Despite the findings of previous studies, to date, there have been no intraclass head-to-head studies in humans regarding beta-blocker-induced weight gain. The two beta-blockers of note, atenolol and metoprolol, both being β1-selective beta-blockers, seemingly have no plausible cause that could account for such differences between them. Our study suggests that atenolol and metoprolol (the two most commonly used beta-blockers in Hong Kong13 14) may have different effects on weight. We found a significantly higher BMI and a higher proportion of patients with obesity in those taking atenolol compared with those who did not, especially among those defined as being hypertensive. This difference in obesity was also significant for the comparison with metoprolol. Patients taking metoprolol did not show any significant difference from patients who were not on any beta-blockers.

In our study, the mean difference in BMI between those using atenolol and no beta-blockers was about 1.2. Converting the mean BMI difference in a 60-kg woman reveals a mean body weight difference of 3.1 kg. Whether such a degree of weight gain is clinically significant in light of age-related weight gain is worthwhile to discuss. Indeed, age-related weight gain is an important factor of concern. According to a study of older Chinese adults, the median weight change from 20 years old to baseline was 11.8 kg and 11.5 kg for men and women, respectively.29 Another 10-year follow-up study also reported that a modest weight gain (2.5-5 kg) was not associated with an increase in mortality.30 Therefore, a gain of 3.1 kg alone may not be clinically significant. Yet, such a modest weight gain can be additive to physiological age-related weight gain and contribute to obesity. While age-related weight gain may not be a modifiable factor, the selection of pharmacotherapy is definitely one. Particularly, the use of beta-blockers may not actually be the best therapeutic option for hypertension management. Switching to other first-line agents that have no weight gain effects would reduce the possibilities to become obese.

Given our findings, it is important to note the variable recommendations around the use of beta-blockers in hypertension management. Although a number of international guidelines advocate the use of beta-blockers in patients with both obesity and hypertension,9 10 11 the 2017 American Heart Association guideline criticised atenolol for its inferior efficacy in treatment of hypertension.6 Moreover, a meta-analysis of atenolol versus other antihypertensive treatments also reported higher overall and cardiovascular mortality and more frequent strokes with atenolol treatment.31 Given these reservations about atenolol and our findings of higher prevalence of obesity in those taking atenolol, it is uncertain whether atenolol should continue to be the most-used drug for hypertension. This is relevant to many parts of the world, including Hong Kong, where atenolol appears to be the first-line therapy for hypertension in older patients and may be associated with an adverse effect on weight and BMI.

To the best of our knowledge, this is the first study to evaluate the association between BMI and beta-blocker use in a Chinese population. Additionally, this is the first study highlighting the intraclass differences between beta-blockers in terms of possible weight gain effects (as illustrated by the proportion of obesity), specifically in the hypertensive Chinese population. Therefore, our findings have implications for local clinical practice given the high rate of use of beta-blockers, particularly atenolol, in the older adults in Hong Kong, despite guideline recommendations.

In considering this study’s findings, it is important to acknowledge some of its limitations. First, given the cross-sectional nature of the study, no temporal or causal relationships can be fully assessed or confirmed. We only found an association between beta-blocker usage and obesity in hypertensive patients. We did not assess weight gain, as the term weight gain has a temporal element that should be validated with duration of drug therapy and weight changes throughout a certain period. Second, there is an issue with ‘confounding by indication’: it is possible that some study patients were obese when their antihypertensive therapy was initiated and, consequently, their physicians elected to use beta-blockers, given the recommendation by the few international consensuses and guidelines.9-12 However, since this was a post-hoc analysis based on a cross-sectional study, it is difficult to figure out the temporal sequence—whether beta-blockers were initiated because the patients were obese, or the patients became obese after taking beta-blockers. Third, the results were subject to selection bias because (a) study participation was voluntary, likely representing those who were more physically and/or socially active, and (b) those who were home-bound or had limited access to outdoor environments were not available for inclusion. Third, although the selection criteria were not designed in favour of women, the greater participation of women in our study is common in community-based investigations in Hong Kong,32 33 34 35 36 37 probably reflecting their greater participation in community-based and health-related activities.32 33 Although the self-reported diagnoses of hypertension were verified against the medication profiles, it is possible that beta-blockers were prescribed for an alternative indication, such as ischaemic heart disease.

Our study has reported a high proportion of beta-blocker use among Hong Kong older adults with hypertension. Beta-blocker users, and more specifically atenolol users, have a significantly higher BMI, as well as a higher propensity towards obesity compared with non-users. Our results can remind clinicians of the possibility that beta-blockers, particularly atenolol, may worsen weight control in hypertensive patients with obesity or cause significant weight gain in those who are not yet obese. Our findings taken together with other guideline reservations cast doubt on whether beta-blockers, particularly atenolol, should be the major drug prescribed to older adults with hypertension.

All authors contributed to the concept of study, acquisition and analysis of data, wrote the article, and had critical revision for important intellectual content. All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

All authors have disclosed no conflicts of interest.

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

The study was approved by the Survey and Behavioural Research Ethics Committee of The Chinese University of Hong Kong (Ref 14610518) and was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all study participants. Confidentiality agreement forms were signed by all data collectors.

1. World Health Organization. A global brief on hypertension: silent killer, global public health crisis. Available from: http://www.who.int/cardiovascular_diseases/publications/global_brief_hypertension/en. Accessed 5 Mar 2019.

2. UK Prospective Diabetes Study Group. Efficacy of atenolol and captopril in reducing risk of macrovascular and microvascular complications in type 2 diabetes: UKPDS 39. BMJ 1998;317:713-20. Crossref

3. Wikstrand J, Warnold I, Olsson G, Tuomilehto J, Elmfeldt D, Berglund G. Primary prevention with metoprolol in patients with hypertension. Mortality results from the MAPHY study. JAMA 1988;259:1976-82. Crossref

4. Davis BR, Oberman A, Blaufox MD, et al. Effect of antihypertensive therapy on weight loss. The Trial of Antihypertensive Interventions and Management Research Group. Hypertension 1992;19:393-9. Crossref

5. Houston MC, Olafsson L, Burger MC. Effects of nifedipine GITS and atenolol monotherapy on serum lipids, blood pressure, heart rate, and weight in mild to moderate hypertension. Angiology 1991;42:681-90. Crossref

6. Whelton PK, Carey RM, Aronow WS, et al. 2017 ACC/ AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/ NMA/PCNA guideline for the prevention, detection, evaluation, and management of high blood pressure in adults: a report of the American College of Cardiology/ American Heart Association Task Force on Clinical Practice Guidelines. J Am Coll Cardiol 2018;71:e127-248. Crossref

7. National Institute of Health and Care Excellence. Hypertension in adults: diagnosis and management (Clinical Guideline 127). Available from: https://www.nice.org.uk/guidance/cg127. Accessed 5 Mar 2019.

8. National Heart Foundation of Australia. Guideline for diagnosis and management of hypertension in adults—2016. Melbourne: National Heart Foundation of Australia; 2016.Crossref

9. Williams B, Mancia G, Spiering W, et al. 2018 ESC/ESH Guidelines for the management of arterial hypertension. Eur Heart J 2018;39:3021-104. Crossref

10. Jordan J, Yumuk V, Schlaich M, et al. Joint statement of the European Association for the Study of Obesity and the European Society of Hypertension: obesity and difficult to treat arterial hypertension. J Hypertens 2012;30:1047-55. Crossref

11. Landsberg L, Aronne AJ, Beilin LJ, et al. Obesity-related hypertension: pathogenesis, cardiovascular risk, and treatment: a position paper of The Obesity Society and the American Society of Hypertension. J Clin Hypertens (Greenwich) 2013;15:14-33. Crossref

12. Dentali F, Sharma AM, Douketis JD. Management of hypertension in overweight and obese patients: a practical guide for clinicians. Curr Hypertens Rep 2005;7:330-6. Crossref

13. Wong MC, Tam WW, Cheung CS, et al. Antihypertensive prescriptions over a 10-year period in a large Chinese population. Am J Hypertens 2013;26:931-8. Crossref

14. Wong MC, Tam WW, Wang HH, et al. Predictors of the incidence of all-cause mortality and deaths due to diabetes and renal diseases among patients newly prescribed antihypertensive agents: a cohort study. Int J Cardiol 2013;168:4705-10. Crossref

15. Wong MC, Tam WW, Cheung CS, et al. Initial antihypertensive prescription and switching: a 5 year cohort study from 250,851 patients. PLoS One 2013;8:e53625. Crossref

16. Mathus-Vliegen EM. Obesity and the elderly. J Clin Gastroenterol 2012;46:533-44. Crossref

17. Han TS, Tajar A, Lean ME. Obesity and weight management in the elderly. Br Med Bull 2011;97:169-96. Crossref

18. Zamboni M, Mazzali G, Zoico E, et al. Health consequences of obesity in the elderly: a review of four unresolved questions. Int J Obes (Lond) 2005;29:1011-29. Crossref

19. Bui AL, Horwich TB, Fonarow GC. Epidemiology and risk profile of heart failure. Nat Rev Cardiol 2011;8:30-41. Crossref

20. Klonoff DC. The increasing incidence of diabetes in the 21st century. J Diabetes Sci Technol 2009;3:1-2. Crossref

21. Hajjar I, Kotchen JM, Kotchen TA. Hypertension: trends in prevalence, incidence, and control. Annu Rev Public Health 2006;27:465-90. Crossref

22. Savarese G, Lund LH. Global public health burden of heart failure. Card Fail Rev 2017;3:7-11. Crossref

23. Leung K, Fong W, Lau PS, et al. Impact of dining out and food intake pattern on atrial fibrillation prevalence in Hong Kong Chinese elders. Value Health 2018;21:S57. Crossref

24. WHO Expert Consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet 2004;363:157-63. Crossref

25. Coleman A, Freeman P, Steel S, Shennan A. Validation of the Omron 705IT (HEM-759-E) oscillometric blood pressure monitoring device according to the British Hypertension Society protocol. Blood Press Monit 2006;11:27-32. Crossref

26. Casu G, Merella P. Diuretic therapy in heart failure—current approaches. Eur Cardiol 2015;10:42-7. Crossref

27. Pischon T, Sharma AM. Use of beta-blockers in obesity hypertension: potential role of weight gain. Obes Rev 2001;2:275-80. Crossref

28. Rössner S, Taylor CL, Byington RP, Furberg CD. Long term propranolol treatment and changes in body weight after myocardial infarction. BMJ 1990;300:902-3. Crossref

29. Zhu J, Xiang YB, Cai H, et al. Associations of obesity and weight change with physical and mental impairments in elderly Chinese people. Maturitas 2018;108:77-83. Crossref

30. Park SY, Wilkens LR, Maskarinec G, Haiman CA, Kolonel LN, Marchand LL. Weight change in older adults and mortality: the Multiethnic Cohort Study. Int J Obes (Lond) 2018;42:205-12. Crossref

31. Carlberg B, Samuelsson O, Lindholm LH. Atenolol in hypertension: is it a wise choice? Lancet 2004;364:1684-9. Crossref

32. Leung GT, Leung KF, Lam LC. Classification of late-life leisure activities among elderly Chinese in Hong Kong. East Asian Arch Psychiatry 2011;21:123-7.

33. Woo J, Mak B, Yeung F. Age-friendly primary health care: an assessment of current service provision for older adults in Hong Kong. Health Serv Insights 2013;6:69-77. Crossref

34. Su EX, Lin YQ, Zhang SL, Leung GT, Lam LC, Chiu HF. Physical activity and cognitive function of community Chinese elderly in Hong Kong (HK) and Guangzhou (GZ). Int Psychogeriatr 2015;27:959-66. Crossref

35. Kwok T, Wong A, Chan G, et al. Effectiveness of cognitive training for Chinese elderly in Hong Kong. Clin Interv Aging 2013;8:213-9. Crossref

36. Shen C, Schooling CM, Chan WM, Lee SY, Leung GM, Lam TH. Self-reported diabetes and mortality in a prospective Chinese elderly cohort study in Hong Kong. Prev Med 2014;64:20-6. Crossref

37. Cheung MC, Ting W, Chan LY, et al. Leisure participation and health-related quality of life of community dwelling elders in Hong Kong. Asian J Gerontology Geriatr 2009;4:15-23.