Out-of-hospital cardiac arrest is a public health problem and leads to the highest proportion of deaths in many parts of the world.1 2 According to the American Heart Association (AHA), in the US and Canada, approximately 350 000 people per year suffer out-of-hospital cardiac arrests for which cardiopulmonary resuscitation (CPR) is attempted.1 3 4 In Hong Kong, although no such direct epidemiological information can be referred to, more than 1000 persons are believed to die suddenly and unexpectedly each year; many of which are presumed to be primarily due to cardiac arrests.5

For those who endure sudden cardiac arrests, early, high-quality CPR can greatly improve chances of survival.6 7 Nowadays, the importance of CPR is well recognised and emphasised. Accordingly, the AHA even recommended that CPR training and familiarisation with automated external defibrillators (AEDs) should be included in secondary school curricula.8 Thus, equipping the public with such skills becomes one of the essential strategies to increase the success of CPR for cardiac arrest victims.

In recent years, studies have been conducted to examine the knowledge and attitude of the public regarding CPR. In general, people had poor knowledge on this subject and the proportion of the public who had received the CPR training was low.9 10 11 12 Besides, many individuals did not want to perform cardiac compression with mouth-to-mouth ventilation, due to fear of acquiring transmitted diseases.13 These factors are likely to limit the numbers of bystander CPRs carried out and contribute to the low survival rates from out-of-hospital cardiac arrests. A local study showed that for out-of-hospital cardiac arrests, the frequency of bystander CPR was only about 15.7% and the survival rate to eventual discharge from hospital was as low as 1.3% in Hong Kong.10

To identify effective measures to promote CPR, the current situation should be evaluated. This study aimed to explore the Hong Kong public’s knowledge and attitudes about CPR. Its findings could inform the community regarding preferences to perform bystander CPR and more importantly it could indicate directions for future training.

This was a cross-sectional population-based survey. The study population comprised the Chinese Hong Kong residents aged 15 to 64 years, who speak Cantonese in domestic households. Anonymous telephone interviews using a structured questionnaire were conducted and launched in the Telephone Survey Research Laboratory of the Hong Kong Institute of Asia-Pacific Studies of The Chinese University of Hong Kong. By using the Computer Assisted Telephone Interviewing system, telephone numbers were randomly selected from up-to-date residential telephone directories that covered over 95% of Hong Kong households. The interviews were conducted between 6:15 pm and 10:15 pm, to avoid over-representing the non-working population. For households with more than one eligible member, the one whose birthday was closest to the interview date was invited to join the study. At least three attempts were made to contact individuals in any given household. Such attempts were made at different times of the day and/or different days of the week, to avoid being labelled a non-contact status (with an assigned number) so as to ensure that survey results were not biased due to high non-contact/non-response rates. Eligible respondents were briefed about the study and verbal consent was sought. The study was approved by the Survey and Behavioural Research Ethics Committee of The Chinese University of Hong Kong.

According to a previous study,11 12% of the population had received CPR training. Owing to continuing efforts and CPR promotion programmes/campaigns by different associations and organisations in recent years, it was expected that around 20% of the study population had probably received prior CPR training. Depending on the possible prevalence of subjects with prior CPR training (ranging from 18 to 22%), it was estimated that 883 to 1025 subjects would be sufficient to estimate knowledge and attitudes with a margin of error of ± 2.5% at 5% level of significance. The sample size calculation was performed using PASS 11 (NCSS, Kaysville [UT], US). Thus, we aimed to recruit over 1000 subjects for this study.

In this study we used a structured questionnaire, which took about 5 to 10 minutes to complete, and was developed in January 2010 (Appendix). It was based on the 2005 AHA Guidelines for CPR and Emergency Cardiovascular Care,14 Basic Life Support for health care providers,15 and a review of the relevant literature.11 12 It consisted of three sections. The first entailed questions on demographics, including age, gender, education level, occupation, family history of heart disease, and ischaemic heart disease risk factors. The second entailed questions about previous CPR training. The third entailed questions on attitudes and knowledge regarding CPR. To evaluate respondents’ relevant attitudes and knowledge, questions were included about: willingness to perform CPR (2 items), the basic knowledge related to a victim’s response (1 item), management of airway (2 items), breathing (2 items), circulation (2 items), and AED usage (1 item). The anticipated answers for the CPR knowledge questions (victim’s response, management of airway, breathing, and circulation) were consistent with information in the latest AHA guidelines (2005 version). Content validity was established by an expert panel including four doctors and six nurses who were either AHA Basic Life Support providers or instructors. The content validity index rating item’s relevance to the underlying construct was 0.96.

Data were categorised and presented in frequencies (percentages). Univariate comparisons on demographics and ischaemic heart disease risk factors among those with and without CPR training were conducted, using Pearson Chi squared or Fisher’s exact tests, as appropriate. Logistic regression analysis was used to identify demographics and ischaemic heart disease risk factors (Table 1) that were associated with CPR training. Variables with a P value of <0.25 in the univariate analysis were selected for use in the stepwise multivariate logistic regression analysis, to delineate factors independently associated with CPR training.16

Logistic regression models were also employed to compare subjects with and without CPR training with respect to various outcome variables (attitude and knowledge about CPR), after adjustment for demographics and coronary heart disease risk factors. A ‘two-block stepwise’ logistic regression modelling approach was used to make adjusted comparisons of the two groups. The grouping factor (CPR training: Yes/No) was first entered into logistic regression model and then the demographics and ischaemic heart disease risk factors (Table 1) were entered in another block with stepwise selection. In the final model, the adjusted odds ratio (OR) to compare those with and without CPR training (reference group) was derived, taking account of demographics and ischaemic heart disease risk factors. All statistical analyses were conducted using SPSS 19.0 (Windows version 19.0; SPSS Inc, Chicago [IL], US) with two-sided tests; a P value of <0.05 was considered statistically significant.

In this study, 2703 phone calls were not picked up after three attempts, and 5669 calls were picked up but 2735 calls were disconnected immediately after knowing the purpose of the calls. A total of 2188 eligible respondents were identified, 1175 refused to participate. Finally, 1013 interviews were conducted (response rate, 46%). The demographics and ischaemic heart disease risk factors of these respondents are shown in Table 1.

Among the 1013 respondents, only 214 (21%) reported that they had received CPR training; the majority (72%, n=154) of whom had had their latest training more than 2 years earlier. A large proportion (63%, n=134) of the trained respondents received their training via the Hong Kong St John Ambulance (49%, n=104) and the Hong Kong Red Cross (14%, n=30). Another 35 (16%) participants had their training via their companies or workplaces. Their main reasons for taking CPR training were ‘job requirement’ (48%, n=102) and ‘personal interest’ (42%, n=90). For those who did not take CPR training (n=799), most of them (74%, n=589) claimed that they would not consider participating in CPR training in the future. Reasons for not taking CPR training could be multiple, and included ‘no time’ (41%, n=241), ‘not necessary’ (26%, n=156), and ‘not interested’ (19%, n=110). In addition, 104 (18%) participants picked ‘unable to learn CPR because of their low education level or being too old’.

Demographic and ischaemic heart disease risk factors listed in Table 1 with a P value of <0.25 in the univariate analysis were selected as candidate variables for multivariate stepwise logistic regression.16 Among them, age, education level, full-time working status, occupation, having dyslipidaemia and hypertension were associated with having CPR training in the univariate analysis. However, only having a full-time job (OR=2.2; 95% confidence interval [CI], 1.6-3.1; P<0.001), middle level education—Form 4-7/technical institute (OR=2.3; 95% CI, 1.5-3.6; P<0.001), and a high level of education—college or higher (OR=2.7; 95% CI, 1.7-4.2; P<0.001), were significantly associated with having CPR training in the multivariate analysis. Notably, having a low education level—Form 3 or below—was not significantly associated with such training (Table 2).

As shown in Table 3, the ratio of respondents with and without training willing to attempt CPR on family members at home was 72% vs 45% (P<0.001) and on strangers in the street was 42% vs 15% (P<0.001). Logistic regression analysis revealed that after adjusting for potentially confounding demographic and ischaemic heart disease risk factors, those with CPR training were also more likely to attempt CPR at home (OR=3.3; 95% CI, 2.4-4.6; P<0.001) and in the street (OR=4.3; 95% CI, 3.1-6.1; P<0.001) in emergencies (Table 3).

Regarding knowledge on CPR, trained respondents were more likely to give correct responses to each of the eight knowledge questions (all P<0.001). After adjusting for potential confounding demographic and ischaemic heart disease risk factors, logistic regression showed that the trained group was significantly more likely to give five or more appropriate responses to the eight knowledge items when compared with those without such training (OR=19.8; 95% CI, 11.4-34.4; P<0.001; Table 3). Although the trained respondents achieved higher scores on CPR knowledge (median=3) than those who were untrained (median=1), the overall CPR knowledge level of the respondents was low (median=1). Among all the 1013 respondents, only four (0.4%) answered all the questions correctly (score=8), which also represented 1.9% of those who had received CPR training (Table 4).

The present study showed that 21% of the respondents had received CPR training, which was higher than in a previous local study reporting 12%.11 Our rate was comparable to data reported from elsewhere (27% in New Zealand and 28% in Ireland),17 18 but much lower than in reports from Australia (58%),19 Poland (75%),20 and Washington (79%).21 Therefore, though the trend for CPR training in Hong Kong seems to be increasing, it seems far from sufficient, and the majority had received their training more than 2 years earlier. Although it is commonly believed that performing CPR without 100% accuracy is better than doing nothing, whether our respondents could perform appropriate CPR in an emergency was questionable. In our cohort, skills appeared to have deteriorated with time. One study suggested that 6-monthly reinstruction was needed to maintain adequate CPR skills22; the 2-year intervals noted in this study were much longer than what has been suggested. Thus, after their first training, it is suggested that individuals should attend refresher courses. Moreover, the training institutions should pay more attention to remind the trainees on the need for such reinstruction and updates.

The main reasons of taking CPR training were “job requirement” and “personal interest”, which were similar to reasons given in a previous study from Ireland.18 Therefore, the workplace might be considered a preferred place to conduct CPR training in conjunction with government and non-government organisations; in Hong Kong, these include St John Ambulance, the Hong Kong Red Cross, and the Auxiliary Medical Services. In fact, promoting CPR training in workplace seems an important strategy, as 16% of trained respondents in this study had already received such training in their workplaces, and this was also in line with the results of a study by Jennings et al.18

In this study, most non-trained respondents would not consider receiving CPR training, giving the following reasons: “no time”, “not necessary”, or “not interested”. Lack of time for CPR training is a common reason reported in different studies.11 23 24 To address this problem, self-instruction, such as via video or internet training, may be considered. Studies have shown that video self-instruction training was as good as traditional classroom training,25 26 which is not only cost-effective but also flexible compared to formal classroom training. In addition, as recommended by the AHA,8 CPR training could be incorporated into general education in secondary schools. Several studies have investigated knowledge and attitude towards CPR training, its feasibility and the impact of CPR or life-supporting first-aid training in primary and secondary schools in various countries (Austria, Japan, and Norway) and reported a positive experience.27 28 29 Either as part of the regular curriculum, as mandatory courses, or as an elective extra-curricular activity, it could be beneficial to the students and the general public. By providing students with CPR training, the first part of the chain of survival in out-of-hospital cardiac arrest could be enhanced for future generations, and increase survival after sudden cardiac arrest. To successfully carry out such a health and education policy, the Hong Kong SAR Government can learn from other Asian countries like Japan and Singapore, which have already gained experience in CPR education for secondary schools.

Those with full-time jobs and with higher levels of education were more likely to attend CPR training, which corresponded with the results of previous studies.11 18 Not surprisingly 48% of respondents in the present study were required to attend their CPR training in connection with their jobs, while 18% believed that they were unable to learn as they were too old or their level of education was too low. Accordingly, this misunderstanding about CPR needs correcting, and certainly CPR training should be made available to those who are not employed. Community centres could be used as possible teaching venues to promote CPR, in conjunction with the Hong Kong SAR Government and other health care organisations (Hospital Authority, Hong Kong Red Cross, and St John Ambulance). These health-related organisations could play critical roles in publicising the importance of CPR, and provide accessible trainings for the public. Encouragingly, the Resuscitation Council of Hong Kong was established in 2012, and has the power to promote high standards of training and public awareness on resuscitation.

In this study, respondents with CPR training were more willing to perform it at home and in the street (under emergency situations), presumably as they had acquired enough knowledge and skills to generate confidence and courage. The powerful impact of CPR training on saving lives should never be underestimated. Although only 15% of the respondents without CPR training would like to save others’ lives, nearly half of them (45%) expressed willingness to perform CPR for their family members if needed. The intimate relationship among family members may be the motivation in such cases. According to the AHA, 80% of sudden cardiac arrests happen at home.7 Therefore, it makes sense to exploit intimate emotions to facilitate and publicise the CPR training, especially for those with vulnerable members in their family.

In this study, the overall level of CPR knowledge of the respondents was very low, with a median of one correct answer out of eight questions, which was in agreement with previous studies.11 20 Knowledge was particularly weak related to the compression-to-ventilation ratio and appropriate number of cardiac compressions per minute. This could be because 79% of the respondents had not received any CPR training, whereas 72% out of the 214 who had, recalled receiving it more than 2 years earlier and 51% had received it more than 5 years earlier. The AHA recommends its frequently revised CPR guidelines based on rigorous scientific evidence and the consensus opinions of experts. Using a compression-to-ventilation ratio of 30:2 during CPR for victims of all ages was a major update in 2005.30 In addition, the sequence of ‘A-B-C’ (Airway, Breathing, Chest compression) was changed to ‘C-A-B’ (Chest compression, Airway, Breathing) in the 2010 Guidelines.30 Therefore, knowledge about up-to-date guidelines is likely to be most rewarding.

This survey did not explore why people refused to perform CPR, which could be crucial for raising bystander CPR rates in Hong Kong. As indicated in one study from Japan, people had fear of contracting transmitted diseases through mouth-to-mouth ventilations.13 Legal liability could be another concern. Therefore, public education and laws to protect CPR providers appear necessary, for which Good Samaritan laws need to be enacted. Certainly, the reasons why Hong Kong citizens opt not to undertake CPR warrant future surveys.

Knowledge of CPR in the Hong Kong public is still poor. The percentage of citizens that have had CPR training is relatively low. Unwillingness to perform CPR is particularly common, especially among those who have not received any CPR training. Government and non-government organisations need to promote educational activities and explore diverse approaches to reinforce and refresh the content of training. Government needs to increase public awareness of CPR and enact laws to protect bystanders undertaking CPR. Incorporating CPR training into the secondary school and college curricula has also been suggested.

The authors declare that there is no conflict of interest.

The study was supported by the Nethersole School of Nursing, Cardiovascular and Acute Care Research Group Funding.

1. Travers AH, Rea TD, Bobrow BJ, et al. Part 4: CPR Overview, 2010 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation 2010;122(18 Suppl 3):S676-84.

2. Roger VL, Go AS, Lloyd-Jones DM, et al. Heart disease and stroke statistics—2012 update: A report from the American Heart Association. Circulation 2012;125:e2-e220. CrossRef

3. Nichol G, Thomas E, Callaway CW, et al. Regional variation in out-of-hospital cardiac arrest incidence and outcome. JAMA 2008;300:1423-31. CrossRef

4. Chan PS, Jain R, Nallmothu BK, Berg RA, Susson C. Rapid response teams: a systematic review and meta-analysis. Arch Intern Med 2010;170:18-26. CrossRef

5. Lee K. The Hong Kong College of Cardiology automated external defibrillator programme. The Hong Kong Medical Diary 2008;13:14-6.

6. Sasson C, Rogers MA, Dahl J, Kellermann AL. Predictors of survival from out-of-hospital cardiac arrest: a systematic review and meta-analysis. Circ Cardiovasc Qual Outcomes 2010;3:63-81. CrossRef

7. Hazinski MF, Nolan JP, Billi JE, et al. Part 1: executive summary: 2010 international consensus on cardiopulmonary resuscitation and emergency cardiovascular care science with treatment recommendations. Circulation 2010;122(16 Suppl 2):s250-75. CrossRef

8. Cave DM, Aufderheide TP, Beeson J, et al. Importance and implementation of training in cardiopulmonary resuscitation and automated external defibrillation in schools: a science advisory from the American Heart Association. Circulation 2011;123:691-706. CrossRef

9. Wong TW, Yeung KC. Out-of-hospital cardiac arrest: two and a half years experience of an accident and emergency department in Hong Kong. J Accid Emerg Med J 1995;12:34-9. CrossRef

10. Leung LP, Wong TW, Tong HK, Lo CB, Kan PG. Out-ofhospital cardiac arrest in Hong Kong. Prehosp Emerg Care 2001;5:308-11. CrossRef

11. Cheung BM, Ho C, Kou KO, et al. Knowledge of cardiopulmonary resuscitation among the public in Hong Kong: telephone questionnaire survey. Hong Kong Med J 2003;9:323-8.

12. Konstandions HD, Evangelos KI, Stamatis K, Thyresia S, Zacharenia AD. Community cardiopulmonary resuscitation training in Greece. Res Nurs Health 2008;31:165-71. CrossRef

13. Taniguchi T, Omi W, Inaba H. Attitudes toward the performance of bystander cardiopulmonary resuscitation in Japan. Resuscitation 2007;75:82-7. CrossRef

14. American Heart Association. 2005 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Part 3: overview of CPR. Circulation 2005;112:s12-8.

15. Hazinski MF, Gonzales L, O'Neill L, editors. BLS for healthcare providers: student manual. Dallas, TX: American Heart Association; 2006.

16. Hosmer DW, Lemeshow S. Applied logistic regression. 2nd ed. New York: John Wiley and Sons; 2000. CrossRef

17. Larsen P, Pearson J, Galletly D. Knowledge and attitudes towards cardiopulmonary resuscitation in community. N Z Med J 2004;117:U870.

18. Jennings S, Hara TO, Cavanagh B, Bennett K. A national survey of prevalence of cardiopulmonary resuscitation training and knowledge of the emergency number in Ireland. Resuscitation 2009;80:1039-42. CrossRef

19. Celenza T, Gennat HC, O'Brien D, Jacobs IG, Lynch DM, Jelinek GA. Community competence in cardiopulmonary resuscitation. Resuscitation 2002;55:157-65. CrossRef

20. Rasmus A, Czekajlo MS. A national survey of the Polish population's cardiopulmonary resuscitation knowledge. Eur J Emerg Med 2000;7:39-43. CrossRef

21. Sipsma K, Stubbs BA, Plorde M. Training rates and willingness to perform CPR in King Country, Washington: a community survey. Resuscitation 2011;82:564-7. CrossRef

22. Berden HJ, Willems FF, Hendrick JM, Pijls NH, Knape JT. How frequently should basic cardiopulmonary resuscitation training be repeated to maintain adequate skills? BMJ 1993;306:1576-7. CrossRef

23. Liu H, Clark AP. Cardiopulmonary resuscitation training for family members. Dimens Crit Care Nurs 2009;28:156-63. CrossRef

24. Blewer AL, Leary M, Decker CS, et al. Cardiopulmonary resuscitation training of family members before hospital discharge using video self-instruction: a feasibility trial. J Hosp Med 2011;6:428-32. CrossRef

25. Einspruch EL, Lynch B, Aufderheide TP, Nichol G, Becker L. Retention of CPR skills learned in a traditional AHA Heart-saver course versus 30-min video self-training: A controlled randomized study. Resuscitation 2007;74:476-86. CrossRef

26. Chung CH, Siu AY, Po LL, Lam CY, Wong PC. Comparing the effectiveness of video self-instruction versus traditional classroom instruction targeted at cardiopulmonary resuscitation skills for laypersons: a prospective randomised controlled trial. Hong Kong Med J 2010;16:165-70.

27. Uray T, Lunaer A, Ochsenhofer A, et al. Feasibility of lifesupporting first-aid (LSFA) training as a mandatory subject in primary schools. Resuscitation 2003;59:211-20. CrossRef

28. Hamasu S, Morimoto T, Kuramoto N, et al. Effects of BLS training on factors associated with attitude toward CPR in college students. Resuscitation 2009;80:359-64. CrossRef

29. Kanstad BK, Nilsen SA, Fredriken K. CPR knowledge and attitude to performing bystander CPR among secondary school students in Norway. Resuscitation 2011;82:1053-9. CrossRef

30. Field JM, Hazinski MF, Sayre MR, et al. Part 1: executive summary: 2010 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation 2010;112(18 Suppl 3): S640-56. CrossRef

Several recent surveys have documented low levels of knowledge about chronic kidney disease among patients.1 2 3 In a US survey of almost 400 patients at all stages of chronic kidney disease not on dialysis,1 more than half reported no or little knowledge about the symptoms of kidney disease, or medications that can be harmful to the kidney. Furthermore, awareness of chronic kidney disease in the community is low, and limited knowledge about this disorder in the general public poses an even more significant hurdle in disease prevention. A large representative sample of Chinese adults yielded a 10.8% prevalence of chronic kidney disease, whereas only 12.5% of them were aware they had this condition.4

Data on general public knowledge on chronic kidney disease are essential to understanding knowledge gaps and formulating education programmes. Without knowing knowledge gaps, public health education programmes cannot be planned in a strategic manner. To examine knowledge on kidney disease and identify areas of misconception in the general population, we conducted a cross-sectional telephone survey in Hong Kong. We anticipated that assessment of knowledge gaps would be important to improve the public education and has the potential of preventing chronic kidney disease.5

Between 4 and 7 March 2013, we carried out a telephone survey of adults in Hong Kong. Respondents were required to be adults aged 18 years or older, and to speak Cantonese or Mandarin. The sampling method entailed selecting telephone numbers randomly from the latest Hong Kong Residential Telephone Directories (both Chinese and English versions) as seed numbers. In order to include unpublished telephone numbers, the last two digits of the selected seed numbers were replaced by two new and random digits generated by computer. When telephone contact was established successfully with a target household, only a person aged 18 years or more was chosen for an interview. When there was more than one eligible subject in the household, only one was chosen for the survey (by convenience).

As a result, a total of 11 600 telephone calls were made, and 2659 families were successfully contacted. Ineligible contacts included invalid lines, non-residential lines, voice machines, facsimile numbers, and language problems. Of those successfully connected, 1383 cut the line before confirmation, 562 targeted persons refused the interview, 185 families had no eligible participant, and 13 were not interviewed because the target participants were not available. Finally, a total of 516 respondents were successfully interviewed, yielding a response rate of 47.3% (out of the 1091 eligible families contacted).

Trained interviewers from Hong Kong Institute of Asia-Pacific Studies administered the survey by telephone, and each interview lasted 10 minutes. Participants were asked close-ended questions on general knowledge about chronic kidney disease. The survey domains and instrument were developed to assess knowledge of the respondents on the general function of the kidneys, causes and symptoms of chronic kidney disease, and management and treatment of kidney disease. Some of the multiple-choice questions were refined and modified from a questionnaire previously tested in Singapore.2 Pre-testing of the questionnaire was carried out on members of the public through focus group discussions. The questionnaire was tested for face validity as well as content saturation. The finalised questionnaire was administered to patients at primary care public medical centres in persons with no known chronic kidney disease.2 We also collected demographic information (age, sex, education level, and self-reported health conditions) from the respondents.

Statistical analyses were performed using the Statistical Package for the Social Sciences (Windows version 16.0; SPSS Inc, Chicago [IL], US). Numerical data were expressed as mean ± standard deviation. Percentages were compared by means of Fisher’s exact test or Chi squared test. A two-tailed P value of <0.05 was regarded as statistically significant.

Table 1 lists the characteristics of the respondents, who had a median age of 50 years. Over 80% of them had received a secondary or higher level of education. Only 3.7% (n=19) of the respondents were aware of any personal history of chronic kidney disease, and 19.0% (n=98) admitted they had hypertension. A family history of hypertension and diabetes was reported in 42.8% and 29.1% of the respondents, respectively.

Table 2 lists the answers to questions on the function of the kidneys and general knowledge on kidney disease. Only 27.9% of the respondents knew that only one kidney is needed for a human being to lead a normal life, although most (84.7%) were aware of the kidney’s function. Most respondents (79.5%) listed high dietary sodium as a risk factor for chronic kidney disease, but hypertension and diabetes were selected less frequently. Less than half of the respondents knew that hypertension (43.8%) and diabetes (44.0%) can cause kidney disease, but 52.7% answered that frothy urine can be an early manifestation of kidney disease. However, only 17.8% correctly identified the asymptomatic nature of chronic kidney disease.

We further analysed factors that are associated with a lack of knowledge that hypertension can cause chronic kidney disease. There were no significant differences between groups in terms of age and gender. On the other hand, respondents with higher levels of education were more likely to self-report a personal or family history of hypertension and a personal history of diabetes mellitus (Table 3), and were more likely to know that hypertension is a cause of kidney disease. Similarly, a higher education level and a personal history of diabetes mellitus were associated with better knowledge about the causal relationship between diabetes and chronic kidney disease (Table 4).

Table 5 shows the perceived sequelae of chronic kidney disease. Overall, more than 80% of respondents said that kidney disease could be prevented, and could be controlled by medication. On the other hand, over 60% of them did not identify hypertension as a complication of chronic kidney disease. Furthermore, nearly half (43.6%) of the subjects did not know the importance of checking the blood pressure in patients with chronic kidney disease, and less than a quarter (22.5%) believed that chronic kidney disease can increase the risk of atherosclerosis.

The main findings from this survey define the key knowledge gaps concerning the kidney disease. In particular, the public in Hong Kong is unfamiliar with the relationship between hypertension and kidney disease. Almost four in five knew that high dietary sodium intake can be associated with kidney disease, but the risk of hypertension causing kidney disease was underestimated. Only 43.8% of respondents considered hypertension as a factor that increases the risk of kidney disease, and 43.6% did not perceive the need for patients with kidney disease to have blood pressure monitored. Another key issue was that close to 20% of the respondents self-reported a diagnosis of hypertension. Quantifying such knowledge deficit indicates that high blood pressure is relatively neglected and provides useful input for future public education.

Low public awareness of hypertension as a cause of kidney damage has been demonstrated in other national surveys. A cross-sectional survey of 1435 primary care patients without kidney disease in Singapore reported that only 51.2% knew that chronic kidney disease could be caused by diabetes, hypertension, and hereditary conditions.2 Overall, the public remains relatively unaware of the two leading causes of chronic kidney disease (hypertension and diabetes) in all developed and many developing countries. Similar to hypertension, the rising worldwide prevalence of diabetes and the lack of knowledge about its relationship with chronic kidney disease are of great concern.6 The AusDiab study involving a survey of 852 Australian subjects from the general population found an even lower level of understanding of hypertension; 25.7% of respondents reported poor diet as a cause of kidney disease but only 2.8% identified hypertension as risk factor.7 According to a cross-sectional survey of 2017 African Americans, 12.1% knew that having hypertension was a risk factor for kidney disease.8 A strikingly prevailing theme among all these surveys (including ours) was the tendency of the public to name aspects of lifestyle instead of medical conditions as risk factors for kidney disease. In other words, there is relatively higher awareness of dietary risk factors for kidney disease compared with high-risk medical condition, notably hypertension. This was affirmed in a community-based qualitative exploratory analysis on kidney disease knowledge among rural populations in the US.9 Analysis of the audiotape scripts identified a representative theme: lifestyle choices, such as drinking sodas and diet, were routinely brought up as a means to explain the occurrence of kidney disease.9

More accurate and prioritised knowledge of kidney disease risk factors will lead to better disease awareness and increase chances of opportunistic screening. It is of concern that the general public underestimates the importance of blood pressure control. The most recent Global Burden of Disease Study launched by the World Bank and the World Health Organization announced that high blood pressure has shifted from the fourth to the top risk factor in terms of the global disability-adjusted life-years.10 Inability to consider hypertension as the risk factor for kidney disease implies that many subjects perceive themselves at lower risk of kidney disease, do not get screened, and have less concern for certain health behaviours. In fact, insufficient knowledge can drive the problem of antihypertensive medication non-adherence, which has recently been confirmed to confer an increased risk of end-stage renal disease. Using the Canadian health insurance databases of 185 476 patients with hypertension, among those who were in possession of their prescribed medication, more than 80% of the time had a 33% lower risk of end-stage renal disease.11 Targeting health care professionals is probably not the utmost concern, because only 3.4% of primary care physicians failed to recognise hypertension as a risk factor for chronic kidney disease according to a cross-sectional representative survey of primary care providers in the community.12 On the other hand, targeting public education to prevent asymptomatic renal disease should be explored. We have previously confirmed a high frequency of abnormal blood pressure readings and subclinical urinalysis abnormalities (17.4%) in a screening programme of 1201 apparently healthy community-dwelling adults in Hong Kong.13 A successful public educational programme should therefore aim at better informing the asymptomatic nature of early chronic kidney disease, and address the risk factors such as hypertension. For this reason, the role of hypertension in kidney disease was chosen as the key message for World Kidney Day 2009.14

One important limitation of our survey was that individual-level data of subjects who declined the interview were missing. The fact that we could not compare the baseline characteristics of participants and those who declined raises the possibility of response bias. Response bias implies that the small percentage of subjects who responded could have differed systematically from the majority who did not answer telephone calls or cut the line before confirmation. Moreover, the sample of respondents in this residential telephone registry may not be generalised to other populations, such as those who mostly use mobile phones. Thus, requirement of a landline telephone in order to be sampled by the current random digit-dial telephone survey raised the possibility of non-coverage bias. Furthermore, our survey tool was not developed through experts in health literacy and psychometric analyses, and the questions were not field-tested and validated. We are aware of better constructed chronic kidney disease–specific knowledge survey tools in other populations with known kidney disease.15 Future research to assess kidney disease knowledge in the Chinese community should follow similar developments to improve the reliability and validity of questionnaires. In addition, the diagnosis of hypertension and chronic kidney disease in our telephone survey respondents was not validated, instead it was based entirely on self-reporting.

The general public in Hong Kong did not recognise that the kidney is both a cause and victim of hypertension. Public health education efforts that target knowledge of kidney disease risk factors may help reduce the burden of kidney disease.

No conflicts of interest were declared by authors.

1. Wright Nunes JA, Wallston KA, Eden SK, Shintani AK, Ikizier TA, Cavanaugh KL. Associations among perceived and objective disease knowledge and satisfaction with physician communication in patients with chronic kidney disease. Kidney Int 2011;80:1344-51. CrossRef

2. Chow WL, Joshi VD, Tin AS, et al. Limited knowledge of chronic kidney disease among primary care patients—a cross-sectional survey. BMC Nephrol 2012;13:54. CrossRef

3. Tan AU, Hoffman B, Rosas SE. Patient perception of risk factors associated with chronic kidney disease morbidity and mortality. Ethn Dis 2010;20:106-10.

4. Zhang L, Wang F, Wang L, et al. Prevalence of chronic kidney disease in China: a cross-sectional survey. Lancet 2012;379:815-22. CrossRef

5. Collins AJ, Gilbertson DT, Snyder JJ, Chen SC, Foley RN. Chronic kidney disease awareness, screening and prevention: rationale for the design of a public education program. Nephrology (Carlton) 2010;15(Suppl 2):37-42. CrossRef

6. Jha V, Garcia-Garcia G, Iseki K, et al. Chronic kidney disease: global dimension and perspectives. Lancet 2013;382:260-72. CrossRef

7. White SL, Polkinghorne KR, Cass A, Shaw J, Atkins RC, Chadban SJ. Limited knowledge of kidney disease in a survey of AusDiab study participants. Med J Aust 2008;188:204-8.

8. Waterman AD, Browne T, Waterman BM, Gladstone EH, Hostetter T. Attitudes and behaviors of African Americans regarding early detection of kidney disease. Am J Kidney Dis 2008;51:554-62. CrossRef

9. Jennette CE, Vupputuri S, Hogan SL, Shoham DA, Falk RJ, Harward DH. Community perspectives on kidney disease and health promotion from at-risk populations in rural North Carolina, USA. Rural Remote Health 2010;10:1388.

10. Murray CJ, Lopez AD. Measuring the global burden of disease. N Engl J Med 2013;369:448-57. CrossRef

11. Roy L, White-Guay B, Dorais M, Dragomir A, Lessard M, Perreault S. Adherence to antihypertensive agents improves risk reduction of end-stage renal disease. Kidney Int 2013;84:570-7. CrossRef

12. Lea JP, McClellan WM, Melcher C, Gladstone E, Hostetter T. CKD risk factors reported by primary care physicians: do guidelines make a difference? Am J Kidney Dis 2006;47:72-7. CrossRef

13. Li PK, Kwan BC, Leung CB, et al. Prevalence of silent kidney disease in Hong Kong: the screening for Hong Kong Asymptomatic Renal Population and Evaluation (SHARE) program. Kidney Int Suppl 2005;94:S36-40. CrossRef

14. Bakris GL, Ritz E. The message for World Kidney Day 2009: hypertension and kidney disease: a marriage that should be prevented. Nephrology (Carlton) 2009;14:49-51. CrossRef

15. Wright JA, Wallston KA, Elasy TA, Ikizler TA, Cavanaugh KL. Development and results of a kidney disease knowledge survey given to patients with CKD. Am J Kidney Dis 2011;57:387-95. CrossRef

The annual incidence of renal cell carcinoma (RCC) in Hong Kong has increased steadily over the past 10 years reaching a rate of 5.9 cases per 100 000 inhabitants.1 Surgical management remains the main treatment modality. With advances and ready availability of imaging, including screening by ultrasonography, more RCCs are diagnosed at an early stage (ie T1). The treatment modalities of these localised renal masses include radical nephrectomy or partial nephrectomy, in the form of an open or laparoscopic (with or without robotic-assisted) approach, as well as other form of ablative therapy. Several large, retrospective studies and the recently published European Organization for Research and Treatment of Cancer randomised trial2 have confirmed that the oncological outcomes of partial nephrectomy and radical nephrectomy are equivalent. The advantages of radical nephrectomy include better preservation of renal function and prevention of renal failure, lower cardiovascular morbidity, and better overall survival.3 Although nephron-sparing surgery has slightly higher complication rate compared with radical nephrectomy,4 most international guidelines recommend the former as the standard treatment for solitary renal tumours up to a diameter of 7 cm, whenever technically feasible.5 6 In the US population, utilisation of such techniques has recently been reported to be low, partly due to lack of technical advancements and publicity about possible adverse long-term consequences.7

Decisions on the choice of surgery mostly depend on the size and location of the tumour. Other external factors, such as the surgeon’s training, practice pattern, operating centre facilities, and hardware available, have a major impact on the choice of approaches and operation to be performed. In the presence of multiple treatment options, an objective way to describe the complexity of renal masses and to accurately assess the risks of postoperative complications is important for patient counselling and clinical decision-making. Scoring systems have therefore been developed and validated, and to date three are available for clinical use.8 9 10

Herein, we report our investigation into using the R.E.N.A.L. nephrometry score, as developed by Kutikov and Uzzo in 2009,8 and its relationship to the choice of treatment and postoperative complications.

Data about patients having renal tumours treated by total nephrectomy in Queen Mary Hospital during the period of January 2006 to December 2011 were retrieved retrospectively from a clinical database and analysed. Patients who had not had preoperative computed tomography and three-dimensional reconstruction (available in the Queen Mary Hospital radiological department) were excluded, so as to standardise the radiographic characteristic of the renal tumours under study. This involved allocating a R.E.N.A.L. nephrometry score to each renal tumour utilising computerised imaging systems (GE Advantage Workstations; General Electric Healthcare, US) by a blinded qualified radiologist. The R.E.N.A.L. score was described in 2009 and includes the assessment of tumour (R)adius (size at the maximal diameter), (E)xophytic/ endophytic properties, (N)earness of tumour to the collecting system or sinus, (A)nterior/posterior descriptor, and (L)ocation relative to polar lines. Standardised points (1-3 points per descriptor) were assigned onto each parameter, except the anterior or posterior component as originally described by Kutikov and Uzzo8 (Table 1). Radius was measured as the maximum diameter of the tumour in centimetres and points were allocated as 1 (≤4 cm), 2 (>4 but <7 cm), and 3 (≥7 cm). Exophytic/endophytic points assigned were 1 when 50% or more of the tumour was exophytic, 2 when less than 50% was exophytic, and 3 when it was entirely endophytic. For non-spherical or asymmetrically located tumours, the predominant feature on any axis (not just the axial or coronal axis) was considered with reference to the renal cortex. The N component was measured as the distance of the deepest portion of the tumour to the collecting system and points were allocated as 1 (≥7 mm), 2 (>4 but <7 mm), and 3 (invading, touching or within 4 mm). Anterior/posterior location of the tumour was designated as a non-numerical suffix that describes the location of the tumour with respect to the kidney midline plane as assessed on axial images. When the mass was located at the tip of the renal poles or lay on the coronal plane where a meaningful anterior or posterior designation was not possible, the suffix “x” was assigned. The location score was assigned as the position of the mass relative to polar lines. The polar line was assigned as the plane of the kidney above or below which the medial lip of parenchyma was interrupted by the renal sinus fat, vessels or the collecting system and best located in the coronal plane. Two polar lines were measured for each renal unit. The position of the renal tumour with respect to the polar lines was measured and a score allocated as described in Table 1. Nephrometry classes in terms of complexity were allocated as low (4-6), moderate (7-9), and high (10-12) based on the sum of scores allocated to each parameter. Patient demographics, including age, gender, preoperative renal function, and estimated glomerular filtration rate (eGFR) as calculated by Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equations were logged.11 In addition, the American Society of Anesthesiologists (ASA) class,12 chronic kidney disease stage, mode of surgery (radical vs partial), approaches (open vs minimally invasive surgery [MIS]), and ischaemic time were analysed with respect to their R.E.N.A.L. score and classes. The 90-day postoperative morbidity and mortality were retrieved according to the Clavien-Dindo system.13 Continuous variables were analysed with Student’s t test and categorical variables by the Chi squared and Fisher’s exact tests. Any P value of <0.05 was taken as statistically significant. All data were analysed with the Statistical Package for the Social Sciences (Windows version 18.0; SPSS Inc, Chicago [IL], US).

There were 74 patients included during this study period, of which 38 underwent partial nephrectomy (group 1) and 36 underwent radical nephrectomy (group 2). There were 41 males and 33 females. No statistical differences were found between the groups in terms of gender distribution, age, preoperative creatinine level, ASA class, or chronic kidney disease stage, although the mean eGFR was significantly lower in the radical nephrectomy group (65 vs 77 mL/min, P=0.039; Table 2). The final pathology of the majority of our patients was clear-cell RCC (n=52), and the remainder suffered from angiomyolipoma (n=10), oncocytoma (n=4), chromophobe RCC (n=3), and others (n=5). There were significant differences between the partial and radical nephrectomy groups in terms of their mean nephrometry score (6.9 vs 9.3, P<0.001). Individual parameters of the R.E.N.A.L. score in terms of radius (P<0.001), nearest to the collecting system (P<0.001), and locations relative to polar lines (P=0.017) were significantly different in the two groups, but there was no significant difference in terms of exophytic components or anterior/posterior location (Table 3).

Further analysis of the partial nephrectomy patients revealed that respective mean nephrometry scores of open versus MIS were 7.8 vs 6.0 (P=0.001), and in particular the nearest components were significantly different (P<0.001; Table 4). Such a difference was evident for the radical nephrectomy group. The overall 90-day morbidity in our study cohort was low, and included urinary leakage (n=1) and bleeding warranting embolisation (n=1) in the partial nephrectomy group, and intestinal obstruction (n=1) in the radical nephrectomy group. None of our patients received a postoperative transfusion. Mortality at 90 days in the radical nephrectomy group (n=1) was in a patient with metastatic RCC undergoing cytoreductive nephrectomy. There was no difference in postoperative 90-day morbidity and mortality between the two groups, even after stratification according to mean nephrometry score or with respect to different classes (Table 5). Ischaemic time was significantly higher for patients in higher nephrometry classes in the partial nephrectomy group (36 mins vs 51 mins vs 80 mins, P=0.008; Table 6); all three patients with high nephrometry scores underwent open surgery using cold ischaemia with ice sludge surface cooling, thus explaining the difference in ischaemic time.

The standard care of patients with a solid renal mass is excision. Partial nephrectomy has become the standard for T1a RCCs and more recent data support its use in larger tumours of up to 7 cm (ie T1b). Most internationally recognised guidelines support recourse to partial nephrectomy for T1a tumours whenever technically feasible,5 6 as data suggest comparable oncological outcomes with more favourable outcomes in terms of risk of renal failure warranting dialysis, cardiovascular morbidity, and even mortality. Approaches to the management of a solid renal mass include consideration of whether to remove the whole kidney or resect the tumour only and achieve a margin clear of pathology. Secondary consideration is given to the approach of the surgery, be it a traditional open one or MIS (purely laparoscopic or robotic-assisted laparoscopic). Although the latter is technically more demanding and has more postoperative complications (blood loss, recourse to transfusions, and urinary leakage), many high-volume centres show favourable results in experienced hand.14

Many factors contribute to the choice of surgery and mode of approach. They include hospital infrastructures and patient volume, experience and training history of the relevant surgeons, patient preference, and most importantly tumour characteristics. Traditionally, clinical decisions were based mostly on the first of these factors, resulting in heterogeneous clinical choices and operative results. Even when only tumour characteristics were taken into account, there was wide heterogeneity in definitions, such as centrality or hilar location, and makes direct comparison of results between studies difficult and impractical.

The concept of nephrometry was proposed as a tool to objectively assess the complexity of a solid renal mass. To date there are three studies of largely nephrometric systems. They are the R.E.N.A.L nephrometry score proposed by Kutikov and Uzzo in 2009,8 the preoperative aspects and dimensions used for an anatomical (PADUA) classification of renal tumours by Ficarra et al in 2009,9 and the C-index method proposed by Simmons et al in 2010.10 Most studies utilise the nephrometry scales in patients undergoing partial nephrectomy. The three methods made use of different parameters to assess the locations of the tumour in relation to various important structures of the kidney, and to predict the technical difficulty that might be encountered during nephron-sparing surgery of the target lesion. They have been reviewed as new tools that can guide surgical decision-making to improve academic reporting, risk assessment of complications, and prediction of functional outcomes.

The R.E.N.A.L. nephrometry score is one of the most studied scoring systems with numerous articles describing its use in clinical practice. The original description of the score was to set a standard reporting system, and its use suggested a relationship between renal mass anatomy, pathology, and prognosis.8 Assessments of inter-observer variability confirm their reproducibility and inter-observer agreement was robust across specialties and levels of training.15 16 17 18 Later studies showed that high R.E.N.A.L. scores were associated with higher major complication rates than those with intermediate or low scores.15 19 Moreover, multivariate analysis revealed that prolonged operating time and high-complexity nephrometry score category were independent predictors of major complications.19 Other reports demonstrated that the R.E.N.A.L. score correlated with both tumour grade (P<0.0001) and histology (P<0.0001), such that as tumour size increases there would be a greater probability of malignancy, including high-grade and clear-cell tumour on histology.20 21 Nomograms have been developed based on study results to preoperatively predict the likelihood of malignant and high-grade pathology of an enhancing renal mass,20 and such systems have been externally validated.22 Other studies have demonstrated the association of nephrometry scores with use of ischaemia in partial nephrectomy,15 warm ischaemia time,23 choice of surgery (partial vs radical nephrectomy),17 24 25 need of conversion to radical nephrectomy,23 changes in the percent functional volume preserved and perioperative functional decrease,26 long-term renal functional outcome following partial nephrectomy,27 and postoperative urinary leakage.28 In particular R.E.N.A.L. scores were higher in patients with partial nephrectomy who developed complications than in partial nephrectomy patients who did not (6.9 vs 6.0, P=0.02). No corresponding differences were found in patients having radical nephrectomy (P=0.99).29 Other studies investigating their applications on robotic partial nephrectomy have shown incongruent results. In one study, Mufarrij et al30 did not show the ability of this scoring system to predict perioperative outcomes in robotic-assisted partial nephrectomy. Others found significant correlations of the score with increased warm ischaemia time, blood loss, complications, and length of hospital stay31 32 in patients undergoing robotic and laparoscopic partial nephrectomy. Clinical application of such anatomical classification systems has gained popularity in selecting cases suitable for alternative treatment of small renal masses (such as by thermal ablation).33 Available data so far show more evidence to support the use of this scoring system to make treatment decision more objective for renal masses.34 35

The results of our study clearly demonstrate a positive correlation of R.E.N.A.L. scores with the choice of nephrectomy (partial vs radical), in terms of the total summed scores and individual parameters including radius (size), location nearest to the collecting system, and relationship to polar lines. These findings support the idea that clinical decisions based solely on the size of tumours are oversimplified and other anatomical factors should enter overall considerations. We did not find significant correlations for other individual parameters, such as exophytic components and anterior/posterior location. This was in contrast to a previous study which espoused the relevance of such components to the choice of ablative therapy (radiofrequency, cryoablation, or partial nephrectomy) as originally described by Kutikov and Uzzo.8 Another significant finding was the correlation between the score and the choice of approach in partial nephrectomy. It was shown that with an increase in mean nephrometry score or class, there was a trend towards choosing open rather than a MIS approach. This signifies that whenever partial nephrectomy is feasible, the open method is preferred for more complex tumours and that this practice can be based on an objective scoring system. However, this was not observed in our radical nephrectomy group, which echoed a previous study finding and like the original description aimed at partial nephrectomy (not radical nephrectomy). The significant correlation of R.E.N.A.L. class with ischaemic time may be useful to guide the choice of open approaches for partial nephrectomy in the presence of a renal tumour with a high score. This could facilitate the safe use of cold ischaemia so as to maximise preservation of renal function.

Our results were contrary to previous investigators reporting that the R.E.N.A.L. score was not associated with presence or severity of complications in both patient groups in terms of their mean score or class. This could be explained by the relatively low frequency of major complications in our study cohort (5.4%) and in the small sample size. With more prospective data available, we believe similar correlations of the score with the frequency of postoperative complications and perioperative outcomes would be revealed.

An inherent limitation of our study was that it was retrospective with respect to data collection and analysis. A second limitation was the exclusion of many patients due to unavailability of satisfactory quality images for the calculation of scores to make direct comparisons. A third limitation of the R.E.N.A.L. score per se was that the weight given to individual components contributed to the total score; numerical values were allocated arbitrarily and still await validation. Although ours is one of the few studies that demonstrate the association of this score and individual parameters on the choice of surgery rather than sole reliance on tumour size, we still have to define a single value in this scoring system below which we can confidently recommend partial nephrectomy. Moreover, other confounding factors such as the surgeon’s experience and learning curve data were not available for analysis, and may heavily influence clinical decisions.

Future directions of studies and clinical utilisation of such a scoring system will aim to define different weightings for individual components contributing to the total score. Other studies may aim at enhancing the reproducibility and predictability of such tools, so that direct comparison can be made with other centres. Are we doing better than eyeballing when managing a solid renal mass? Maybe we are, but the use of the nephrometry score will enhance communication, documentation, and education for the coming younger generation of urologists. Lately, Simmons et al36 have described the integration of the R.E.N.A.L. and C-index scoring systems as diameter-axial-polar nephrometry (DAP). Initial results demonstrate the DAP scoring system to be simpler, to decrease measurement error, to improve performance characteristic, to make interpretation easier, and to exhibit a clear association with volume loss and late function after partial nephrectomy. More mature data will allow us to choose the best tools for our patients.

The R.E.N.A.L. nephrometry score of a solid renal mass shows a significant association with our choice of surgery (partial vs radical) and our approach to surgery (open vs MIS), particularly in patients receiving partial nephrectomy. Its association with postoperative complications was not demonstrated in this study. The score provides a useful tool to define the character of a renal mass objectively, aid clinical decision-making, and enhance communication between professionals with respect to the management of solid renal masses.

1. Hospital Authority. Hong Kong Cancer Registry. Available from: http://www3.ha.org.hk/cancereg/statistics.html. Accessed Jun 2013.

2. Van Poppel H, Da Pozzo L, Albrecht W, et al. A prospective randomized EORTC intergroup phase 3 study comparing the oncologic outcome of elective nephron-sparing surgery and radical nephrectomy for low-stage renal cell carcinoma. Eur Urol 2011;59:543-52. Crossref

3. Go AS, Chertow GM, Fan D, McCulloch CE, Hsu CY. Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med 2004;351:1296-305. Crossref

4. Van Poppel H, Da Pozzo L, Albrecht W, et al. A prospective randomized EORTC intergroup phase 3 study comparing the complications of elective nephron-sparing surgery and radical nephrectomy for low-stage renal cell carcinoma. Eur Urol 2007;51:1606-15. Crossref

5. Ljungberg B, Cowan NC, Hanbury DC, et al. EAU guidelines on renal cell carcinoma: the 2010 update. Eur Urol 2010;58:398-406. Crossref

6. Campbell SC, Novick AC, Belldegrun A, et al. Guideline for management of the clinical T1 renal mass. J Urol 2009;182:1271-9. Crossref

7. Smaldone MC, Kutikov A, Egleston B, et al. Assessing performance trends in laparoscopic nephrectomy and nephron-sparing surgery for localized renal tumors. Urology 2012;80:286-92. Crossref

8. Kutikov A, Uzzo RG. The R.E.N.A.L. nephrometry score: a comprehensive standardized system for quantitating renal tumor size, location and depth. J Urol 2009;182:844-53. Crossref

9. Ficarra V, Novara G, Secco S, et al. Preoperative aspects and dimensions used for an anatomical (PADUA) classification of renal tumors in patients who are candidates for nephron-sparing surgery. Eur Urol 2009;56:786-93. Crossref

10. Simmons MN, Ching CB, Samplaski MK, Park CH, Gill IS. Kidney tumor location measurement using the C index method. J Urol 2010;183:1708-13. Crossref

11. Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med 2009;150:604-12. Crossref

12. American Society of Anesthesiologists. New classification of physical status. Anesthesiology 1963;24:111.

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

14. Gill IS, Kavoussi LR, Lane BR, et al. Comparison of 1,800 laparoscopic and open partial nephrectomies for single renal tumors. J Urol 2007;178:41-6. Crossref

15. Hew MN, Baseskioglu B, Barwari K, et al. Critical appraisal of the PADUA classification and assessment of the R.E.N.A.L. nephrometry score in patients undergoing partial nephrectomy. J Urol 2011;186:42-6. Crossref

16. Kolla SB, Spiess PE, Sexton WJ. Interobserver reliability of the RENAL nephrometry scoring system. Urology 2011;78:592-4. Crossref

17. Weight CJ, Atwell TD, Fazzio RT, et al. A multidisciplinary evaluation of inter-reviewer agreement of the nephrometry score and the prediction of long-term outcomes. J Urol 2011;186:1223-8. Crossref

18. Montag S, Waingankar N, Sadek MA, Rais-Bahrami S, Kavoussi LR, Vira MA. Reproducibility and fidelity of the R.E.N.A.L. nephrometry score. J Endourol 2011;25:1925-8. Crossref

19. Simhan J, Smaldone MC, Tsai KJ, et al. Objective measures of renal mass anatomic complexity predicts rates of major complications following partial nephrectomy. Eur Urol 2011;60:724-30. Crossref

20. Kutikov A, Smaldone MC, Egleston BL, et al. Anatomic features of enhancing renal masses predicts malignant and high-grade pathology: a preoperative nomogram using the RENAL nephrometry score. Eur Urol 2011;60:241-8. Crossref

21. Satasivam P, Sengupta S, Rajarubendra N, Chia PH, Munshey A, Bolton D. Renal lesions with low R.E.N.A.L nephrometry score are associated with more indolent renal cell carcinomas (RCCs) or benign histology: findings in an Australian cohort. BJU 2012;109 Suppl 3:44-7.

22. Wang HK, Zhu Y, Yao XD, et al. External validation of a nomogram using RENAL nephrometry score to predict high grade renal cell carcinoma. J Urol 2012;187:1555-60. Crossref

23. Long JA, Arnoux V, Fiard G, et al. External validation of the RENAL nephrometry score in renal tumours treated by partial nephrectomy. BJU Int 2013;111:233-9. Crossref

24. Satasivam P, Rajarubendra N, Chia PH, Munshey A, Sengupta S, Bolton D. Trends in the use of nephron-sparing surgery (NSS) at an Australian tertiary referral centre: an analysis of surgical decision-making using the R.E.N.A.L. nephrometry scoring system. BJU Int 2012;109:1341-4. Crossref

25. Broughton GJ, Clark PE, Barocas DA, et al. Tumour size, tumour complexity, and surgical approach are associated with nephrectomy type in small renal cortical tumours treated electively. BJU Int 2012;109:1607-13. Crossref

26. Simmons MN, Hillyer SP, Lee BH, Fergany AF, Kaouk J, Campbell SC. Nephrometry score is associated with volume loss and functional recovery after partial nephrectomy. J Urol 2012;188:39-44. Crossref

27. Cha EK, Ng CK, Jeun B, et al. Preoperative radiographic parameters predict long-term renal impairment following partial nephrectomy. World J Urol 2013;31:817-22. Crossref

28. Bruner B, Breau RH, Lohse CM, Leibovich BC, Blute ML. Renal nephrometry score is associated with urine leak after partial nephrectomy. BJU Int 2011;108:67-72. Crossref

29. Rosevear HM, Gellhaus PT, Lightfoot AJ, Kresowik TP, Joudi FN, Tracy CR. Utility of the RENAL nephrometry scoring system in the real world: predicting surgeon operative preference and complication risk. BJU Int 2012;109:700-5. Crossref

30. Mufarrij PW, Krane LS, Rajamahanty S, et al. Does nephrometry scoring of renal tumors predict outcomes in patients selected for robot-assisted partial nephrectomy? J Endourol 2011;25:1649-53. Crossref

31. White MA, Haber GP, Autorino R, et al. Outcomes of robotic partial nephrectomy for renal masses with nephrometry score of ≥7. Urology 2011;77:809-13. Crossref

32. Hayn MH, Schwaab T, Underwood W, Underood W, Kim HL. RENAL nephrometry score predicts surgical outcomes of laparoscopic partial nephrectomy. BJU Int 2011;108:876-81.

33. Reyes J, Canter D, Putnam S, et al. Thermal ablation of the small renal mass: case selection using the R.E.N.A.L-Nephrometry Score. Urol Oncol 2013;31:1292-7. Crossref

34. Canter D, Kutikov A, Manley B, et al. Utility of the R.E.N.A.L. nephrometry scoring system in objectifying treatment decision-making of the enhancing renal mass. Urology 2011;78:1089-94. Crossref

35. Tobert CM, Kahnoski RJ, Thompson DE, Anema JG, Kuntzman RS, Lane BR. RENAL nephrometry score predicts surgery type independent of individual surgeon’s use of nephron-sparing surgery. Urology 2012;80:157-61. Crossref

36. Simmons MN, Hillyer SP, Lee BH, Fergany AF, Kaouk J, Campbell SC. Diameter-axial-polar nephrometry: integration and optimization of R.E.N.A.L. and centrality index scoring systems. J Urol 2012;188:384-90. Crossref

Obstetric admissions to the intensive care unit (ICU) and maternal mortality continue to have a significant impact on maternal health care, despite the low rate of such admissions in developed countries.1 Unlike others, obstetric patients pose a major management challenge to ICU physicians and obstetricians due to altered physiology during pregnancy, consideration of fetal wellbeing, and the unique type of disorders to be dealt with.

Despite ongoing improvements in obstetric care, more patients were admitted to ICU in the reviewed period compared with decades earlier.2 Thus, the purpose of this study was to review and compare the characteristics of obstetric patients admitted to the ICU over the recent 20 years using historical controls, with respect to their epidemiology, medical background, antenatal and peripartum risks, durations of ICU stay, interventions in the ICU, and predictability of the Acute Physiology and Chronic Health Evaluation (APACHE II) score, as well as maternal and fetal outcomes.

This was a retrospective case series of obstetric patients admitted to the ICU of Kwong Wah Hospital, Hong Kong, over a 5-year period from 1 January 2006 to 31 December 2010. Our hospital provides joint care with seven other hospitals in the Kowloon West Cluster to residents of six districts, which account for about 1.9 million inhabitants. Our obstetric service is available for 24 hours each day for women in parts of the Kowloon West and Wong Tai Sin districts. We provide out-patient and in-patient services, including antenatal check-ups, prenatal diagnoses, elective and emergency operations and services that are supported by a blood bank and various laboratory test facilities available for patients in hospital and in the community. Moreover, 24-hour midwifery, and perinatal and anaesthetic services are available in our delivery suite. Our team consists of consultants, associate consultants, as well as senior and junior residents. Three staff (one specialist, two residents) are always available on site for emergency admissions. Annually, we manage 5000 to 6000 deliveries, which is one of the highest delivery rates for a Hong Kong hospital. Our ICU was established in 1968, currently has 14 beds, and admits 500 to 600 patients every year. The ICU team consists of a critical care physician, a resident anaesthetist, medical and surgical residents, and a nursing team with critical care–registered nurse specialists.

This study was approved by the Ethics Committee of the Kowloon West Cluster, Hospital Authority. No patient consent was required as the study only involved review of medical records.

Obstetric patients from 24 weeks of gestation onwards to 6 weeks postpartum admitted to the ICU were reviewed. They were identified via the computerised database system adopted by the ICU. All corresponding medical records were reviewed in detail. Supplementary information was retrieved from the Clinical Management System, Electronic Patient Record, and Obstetrics Clinical Information System.

Data retrieved for analysis included patient demographics (age, ethnicity, smoking and drinking status, parity, order of pregnancy, and body mass index [BMI] at booking visit), antenatal booking status, number of antenatal visits, medical history, perinatal risks, gestation at and mode of delivery, indications for caesarean section, interventions involved at and after delivery, indications and admission status to the ICU, and maternal and fetal outcomes. Patient mortality was predicted by recourse to the APACHE II score.

Indications for ICU admission were divided into obstetric and non-obstetric causes. Obstetric causes were those unique to pregnancy or liable to occur within 6 weeks of delivery. Non-obstetric causes were those not specifically related in pregnancy.

Interventions provided by ICU physicians were classified into non-invasive and invasive. Those deemed non-invasive included insertion of arterial or central lines, blood product transfusion, use of continuous positive airway pressure (CPAP) ventilation, and use of inotropes. Invasive procedures included invasive mechanical ventilation, cardiopulmonary resuscitation (CPR), defibrillation, and haemodialysis.

Immediate and long-term complications of the mothers and neonates were assessed up to 6 to 8 weeks post-delivery. Maternal and perinatal mortalities were also calculated.

Data were entered manually into Excel and analysed using the Statistical Package for the Social Sciences (SPSS version 17, Chicago [IL], US). The data were compared with those from the results of a historical review in the same hospital (1989-1995, by Tang et al2) and a review in another regional hospital (1998-2007, by Leung et al3). Chi squared or Fisher’s exact tests were used to compare proportions and Student’s t test to compare continuous variables.

In all, 67 relevant patients were admitted to the ICU and reviewed during the period of 1 January 2006 to 31 December 2010, which amounted to 0.23% of the total hospital maternities and 2.34% of all ICU admissions (Table 1). Their demographic features are shown in Table 2. The mean age of women at delivery was 34 (standard deviation [SD], 5; range, 20-42) years. Thirty (45%) of the patients were of advanced maternal age (ie age at confinement of ≥35 years). The majority of them were Chinese (n=65, 97%), one was Filipino and one an Indonesian. Nine (13%) patients were visitors from mainland China. Seven (10%) were smokers, two (3%) drank alcohol regularly and two (3%) had a history of substance abuse. In all, 38 (57%) were nulliparous and six (9%) carried twin pregnancies. Most of the patients (n=61, 91%) were booked in our unit; 19 (28%) had three antenatal check-ups or less.

Among these 67 patients, 39 (58%) enjoyed good past health, and six (9%) had a BMI of more than 25 kg/m2 at their booking visit. The most common co-existing diseases were gynaecological (n=11, 16%) and haematological (n=9, 13%). Their antenatal and peripartum risks are summarised in Table 3.

The mean gestational age at delivery was 37 (SD, 6; range, 27-41) weeks. Most of them were delivered by emergency caesarean section (n=34, 51%), including one transferred to us after delivery in the private sector. One patient remained undelivered and died antenatally. Placenta praevia and pregnancy-induced hypertensive disorders were the main indications for elective and emergency caesarean sections, respectively. Other indications are listed in Table 4.

The reasons for ICU admission and procedures undertaken therein are listed in Table 5. Most were admitted to the ICU postpartum (n=65, 97%) and for obstetric problems (n=58, 87%), of which postpartum haemorrhage (PPH) was the leading cause (n=39, 58%) followed by pre-eclamptic toxaemia (PET) or eclampsia (n=17, 25%). The mean duration of ICU stay was 1.8 (SD, 1.2; range, 0.5-10) days; four (6%) of the patients stayed for more than 3 days.

In all, 39 patients were admitted to the ICU due to a PPH, the mean estimated blood loss was 4852 mL. Major causes of PPH were related to placenta praevia (n=16), uterine atony (n=12), and perineal trauma (n=5). Blood products given included packed cells (mean, 12 units), platelet concentrate (mean, 5 units), fresh frozen plasma (mean, 7 units), and cryoprecipitate (mean, 1 unit). Three patients received recombinant factor VIIa (NovoSeven; Novo Nordisk A/S, Bagsværd, Denmark). Procedures to control PPH included compression sutures (n=10), uterine artery embolisation (n=9), insertion of a Sengstaken-Blakemore tube (n=6), and uterine artery ligation (n=2). There were 13 patients who underwent hysterectomy despite multiple other interventions and use of multiple uterotonics, and eight patients with PPH were complicated with disseminated intravascular coagulation, one had a ventricular tachycardia, and one had a urinary tract injury. One of the patients with a PPH and anaphylactic shock suffered a cavernous sinus thrombosis and a cranial nerve VI palsy, for which she received therapeutic doses of low-molecular-weight heparin. Another patient was admitted 2 weeks after delivery due to delirium secondary to sepsis.

There were 17 patients admitted to the ICU for PET or eclampsia (9 of whom had eclampsia) and were all stabilised in the ICU. Two patients had HELLP (haemolysis, elevated liver enzymes, low platelets) syndrome and two had hypertensive encephalopathy diagnosed on the basis of computed tomography. Other complications included acute pulmonary oedema (n=1), deranged renal function (n=3), deranged liver function (n=1), and aspiration pneumonitis (n=1). Another eight had persistent hypertension 6 weeks postpartum and were referred to physicians.

Regarding the 10 patients (15% of the cohort) admitted to the ICU for non-obstetric reasons, two had an epileptic seizure, three had cardiovascular problems (cardiomyopathy, heart failure, and pulmonary hypertension), and one each had renal disease, ethanol toxicity, acute pulmonary oedema, myasthenia gravis, and anaphylactic shock.

Invasive procedures performed in the ICU were CPR (n=2, 3%) and mechanical ventilation (n=7, 10%). Non-invasive procedures were blood product transfusions (n=36, 54%), central line insertion (n=18, 27%), arterial line insertion (n=24, 36%), use of inotropes (n=2, 3%), and CPAP ventilation (n=1, 2%).

The mean APACHE II score was 17 (range, 4-37) and the mean predicted mortality rate was 28% (range, 4-85%). The actual mortality rate in this series was 3% (Table 6). The maternal mortality ratio (MMR; actual/predicted mortality) was 0.11.

In our study period, there were two maternal deaths in the 28 976 maternities or 7 per 100 000 births, both in ICU patients. One was a patient who enjoyed good health but suspected to have pulmonary hypertension at 27 weeks of gestation, who rapidly deteriorated and died 1 day after admission. Her diagnosis was confirmed at postmortem examination. The other maternal death ensued in the postpartum period due to multi-organ failure and brain death, secondary to eclampsia and intraventricular haemorrhage.

Regarding these ICU admissions, three (5%) of the fetuses endured intrauterine death (IUD) and one (2%) whose neonate died (due to necrotising enterocolitis). The IUDs were associated with abruptio placentae, pulmonary hypertension, and severe pre-eclampsia with early intrauterine growth restriction.

The health care system of Hong Kong aims to protect/ improve maternal and child health, by means of antenatal, intrapartum, and postnatal services that are readily available at very low costs. Whilst the MMR fluctuated between 1.0 and 11.2 per 100 000 live births over the past 31 years,4 5 the above-mentioned services have contributed to the decreasing and now very low maternal mortality rates.

Despite advances in obstetric care, the admission rate to the ICU had doubled compared with a decade ago (from 0.12% to 0.23%).2 Whereas such ICU utilisation rates for obstetric cases were also higher compared with Tang et al’s data2 (2.34% vs 0.61%), nevertheless they were low compared to reports from overseas.6 7 The rates were also higher than those reported by Leung et al (admission, 0.13%; utilisation, 0.65%).3 One of the reasons for the rise in ICU admission rates was changes in patient allocation in our hospital, and over the Hong Kong Special Administrative Region. The number of beds in our ICU was reduced from 18 to 14 after the severe acute respiratory syndrome epidemic in 2003. The number of surgical admissions was also much lower than a decade earlier. Moreover, the number of trauma cases dropped significantly, since two other nearby tertiary hospitals became trauma centres. Changing attitudes of obstetricians and anaesthetists also contributed to the increase in ICU admission rate. Given the fact that our patients were most commonly delivered by elective caesarean section for placenta praevia, a proper preoperative management plan with a multidisciplinary approach involving anaesthetist, intensive care physician, and obstetricians should have been available before the operation, which included booking of the ICU bed. With the increasing trend of placenta praevia, it was expected that more and more patients would be admitted to the ICU electively for monitoring rather than any future active intervention. The shorter duration of ICU stays, compared with those detailed earlier by Tang et al,2 is probably consistent with this trend towards elective admissions.

The mean age of our patients at delivery was higher than that in the patient series described by Tang et al2 and Leung et al.3 Indeed, patients of advanced maternal age were more likely to be admitted to the ICU when compared with our background population, though this was not shown for such ICU admissions reported by Selo-Ojeme et al.8 Increasing maternal age implies that our patients were more likely to have co-existing diseases complicating pregnancy, as reflected by our data, even though the medical problems in question were generally mild and stable.

According to Tang et al’s2 and Leung et al’s3 reports about non-booked cases (NBCs), patients from mainland China used to be admitted via the emergency department very late when they went into advanced labour. As a result, potential or present obstetric complications were known to us only when they were admitted. With the commencement of the policy to allow these mothers to register and deliver in Hong Kong (since 2007), the number of NBCs decreased significantly, as did their number of ICU admissions.

In the literature there are conflicting data when parity is considered one of the risk factors for ICU admission. In our study, nulliparity was not related to ICU admission, which was also what Pollock et al noted.6

During our data analysis, twin pregnancy was more likely in our ICU patients compared with the background population. However, such data cannot be retrieved from Tang et al’s or Leung et al’s reports.2 3 Twin pregnancy is known to confer a higher risk of gestational diabetes, hypertension, premature delivery, operative deliveries, and postpartum complications (including PPH).9 10 Our findings also supported the need of a specialised twin pregnancy clinic to look after this high-risk group.

Placenta praevia was the most frequent risk factor identified in our patient series, being much more common than in Tang et al’s study.2 Increasing popularity of evaluation by ultrasound has raised the detection rate of placenta praevia early in the antenatal period. All our patients with placenta praevia were delivered electively with proper preoperative arrangements. These entailed booking of ICU facilities, standby uterine artery embolisation, preparation of recombinant factor VIIa and Sengstaken-Blakemore tubes, and involvement of obstetric consultants to make decisions. One consequence was a significantly higher number of elective caesarean sections for placenta praevia compared with decades ago, though the overall section rate remained relatively stable.11 This also correlated with placenta praevia being the commonest causes of PPH in our ICU patients.

As in Tang et al’s2 and Leung and et al’s3 studies, in our series admissions due to obstetric problems remained the main cause of obstetric ICU admissions. Postpartum haemorrhage was consistently the most common indication for ICU admission, which was also noted in Tang et al’s series.2 Although the mean estimated blood loss of our patients was apparently higher than that reported by Tang et al,2 and abdominal delivery is known to increase the risk of hysterectomy following PPH,12 the number of hysterectomies performed was significantly lower than before. The increasing use of compression sutures and uterine artery embolisation together with strategies to retain the placenta in cases of placenta accreta might account for the decreasing recourse to hysterectomy compared with 20 years ago.

The current series had more patients with pre-eclampsia or eclampsia admitted to the ICU than those reported by Tang et al,2 although the difference was not statistically significant. As suggested by the National Institute for Health and Clinical Excellence guideline,12 our management protocol was updated to incorporate the more liberal use of antihypertensives and magnesium sulphate. Our use of the modified early obstetric warning scoring system allowed early detection of potential complications to prevent poor obstetric outcomes. Intensive care is indicated in patients with severe hypertension, or moderate hypertension with symptoms of impending eclampsia or any suggestion of organ dysfunction. These innovations lead to the rising trend of ICU admissions to monitor for pre-eclampsia.

Active involvement of anaesthetists plays a role in the changing pattern of obstetric ICU admissions. There was a drastic reduction of admissions for anaesthesia-related causes compared with those reported by Tang et al.2 Only one of our patients was admitted due to anaphylactic shock, which can be explained by the significant improvements in anaesthetic care and mechanical ventilation in our hospital. Invasive and non-invasive procedures (eg intubation and insertion of arterial and central venous lines) undertaken in the ICU were significantly fewer than decades ago, as most of them had been performed before admission to ICU by anaesthetists.

In our series, the mean duration of ICU stay was 1.8 days, which was shorter than 4.1 days reported in Tang et al’s study.2 The change in attitude and approach to management of both obstetricians and anaesthetists made ICU admission a more elective occurrence than before. As a result, patients admitted to the ICU tended to be more stable and fewer invasive interventions were warranted. These observations highlight the need for obstetric high-dependency units to cater for patients requiring more intensive care, but not to the extent of ICU support.13

When compared with the findings reported by Leung et al,3 over the decades there was no significant increase in perinatal mortality, nor was there an increased rate of fetal loss when compared with our background population. Nevertheless, maternal mortality had not decreased. In our series, there were two maternal deaths that amounted to a mortality rate of 7 per 100 000 maternities. In Tang et al’s series2 the maternal mortality was 5 per 100 000, and in the UK it was reported to be 14 and 11 per 100 000 in 2003-2005 and 2006-2008, respectively.14 However, these differences between series were not statistically significant.

One limitation of our study was that data collection from the computerised system might have omitted pregnant women admitted to the ICU from other specialties with diagnoses that were not obstetrically related. A second limitation was that the causes of maternal ICU admission may not relate directly to the causes of maternal mortality. For example, thromboembolism, one of the leading causes of maternal death in the UK, was not a major cause of ICU admission. In the UK, only 30% of such maternal deaths were in patients admitted to the ICU,15 and on this issue there is no global consensus on the optimal indications for ICU admission. A third limitation was that the frequency of obstetric ICU admissions is also affected by the standard of obstetric care and the threshold admission criteria determined by obstetricians, anaesthetists, and intensive care physicians. As a result, ICU admissions may not truly reflect the standard of obstetric care.16 A composite performance indicator of obstetric care by combining the frequency of ICU admission, numbers of emergency admissions and/or proportions of emergency/elective admissions, and proportions having prolonged stays (eg >3 days) could be a more useful measure of the standard of obstetric practice in the future.

The APACHE II scoring system has been used as a quantitative predictor of mortality, mainly in medical and surgical patients admitted to the ICU using several physiological measures. Its potential has also been evaluated when applied to obstetric patients. According to different overseas reviews, it may overestimate risks in pregnant patients, as in them normal physiology can differ, often in subtle ways, and can undergo abrupt changes in various emergency conditions.17 18 Thus, to date, there is still no proper screening system for obstetric emergencies. A specific scoring system for obstetric patients should be developed and warrants a large-scale international prospective study for this purpose.

The majority of patients discharged from our ICU enjoyed satisfactory recoveries in the puerperal period. In all, eight patients with hypertensive disorder in pregnancy had persistent hypertension for which they were referred for medical assessment. In our series, long-term outcome was not determined. Leung et al3 found that women admitted to the ICU had lower mean scores for quality of life than normal Hong Kong females of similar age, but these authors commented that the relationship of low scores to the obstetric illnesses was unclear and might be resolved by long-term patient follow-up.

In some respects, Leung et al’s study3 also provided us with geographical controls. However, they included patients with gestational ages of <24 weeks, which complicated any comparison of risk factors. Moreover, being intensive care physicians they emphasised quality of life after discharge from the ICU. In contrast, we obstetricians looked for indicators to prevent/reduce maternal ICU admissions, as advocated in modern obstetrical care guidelines.

Our findings illustrate the various changes in ICU admission practice of obstetric cases in the last 20 years, and are comparable to those in other developed countries. Elective caesarean section for placenta praevia and PPH were the major reasons for ICU admission. More conservative management of placenta praevia and PPH appeared to reduce resorting to hysterectomy. These ongoing changes in practice may make emergency obstetric admissions to the ICU less likely in the future. Maternal mortality in our unit has remained low over the years, and can hardly be reduced any further. It is therefore more important to refine and improve obstetric practice to reduce maternal morbidity. The ‘near-miss’ in terms of obstetric ICU admission rates, together with measures targeting the duration of ICU stay, and a potential obstetric morbidity scoring system will no doubt better reflect our clinical performance standards in the future.

1. Public Health Agency of Canada. Make every mother and child count: report on maternal and child health in Canada. 2005. Available from: http://publications.gc.ca/collections/ Collection/H124-13-2005E.pdf. Accessed Jun 2010.

2. Tang LC, Kwok AC, Wong AY, Lee YY, Sun KO, So AP. Critical care in obstetrical patients: an eight-year review. Chin Med J (Engl) 1997;110:936-41.

3. Leung NY, Lau AC, Chan KK, Yan WW. Clinical characteristics and outcomes of obstetric patients admitted to the intensive care unit: a 10-year retrospective review. Hong Kong Med J 2010;16:18-25.

4. Department of Health, Hong Kong. Chapter 1: Health of community. Annual Report 2009/2010.

5. Department of Health, Hong Kong. Number of maternal deaths by major disease group, 2001-2009. Available from: www.healthyhk.gov.hk/phisweb/reports/2009_01_01_ MOT00101.xls. Accessed Aug 2010.

6. Pollock W, Rose L, Dennis CL. Pregnant and postpartum admissions to the intensive care unit: a systematic review. Intensive Care Med 2010;36:1465-74. Crossref

7. Togal T, Yucel N, Gedik E, Gulhas N, Toprak HI, Ersoy MO. Obstetric admissions to the intensive care unit in a tertiary referral hospital. J Crit Care 2010;25:628-33. Crossref

8. Selo-Ojeme DO, Omosaiye M, Parijat Battacharjee P, Kadir RA. Risk factors for obstetric admissions to the intensive care unit in a tertiary hospital: a case-control study. Arch Gynecol Obstet 2005;272:207-10. Crossref

9. Liu AL, Yung WK, Yeung HN, et al. Factors influencing the mode of delivery and associated pregnancy outcomes for twins: a retrospective cohort study in a public hospital. Hong Kong Med J 2012;18:99-107.

10. Yung WK, Liu AL, Lai SF, et al. A specialised twin pregnancy clinic in a public hospital. Hong Kong J Gynaecol Obstet Midwifery 2012;12:21-32.

11. Department of Obstetrics and Gynaecology, Kwong Wah Hospital. Annual Report 2006-2010.

12. National Institute for Health and Clinical Excellence. Caesarean section. NICE Clinical Guidelines, No 132. 2010.

13. Mirghani HM, Hamed M, Ezimokhai M, Weerasinghe DS. Pregnancy-related admissions to the intensive care unit. Int J Obstet Anesth 2004;13:82-5.  Crossref

14. Cantwell R, Clutton-Brock T, Cooper G, et al. Saving mothers’ lives: reviewing maternal deaths to make motherhood safer: 2006-2008. The eighth report on confidential enquiries into maternal deaths in the United Kingdom. BJOG 2011;118(Suppl 1):1S-203S. Crossref

15. Lewis G, Clutton-Brock T, Cooper G, et al. Saving mothers’ lives: reviewing maternal deaths to make motherhood safer–2003-2005. The seventh report on confidential enquiries into maternal deaths in the United Kingdom. London: CEMACH; 2007.

16. Baskett T, Sternadel J. Maternal intensive care and near-miss mortality in obstetrics. Br J Obstet Gynaecol 1998;105:981-4.  Crossref

17. Vasquez DN, Estenssoro E, Canales HS, et al. Clinical characteristics and outcomes of obstetric patients requiring ICU admission. CHEST 2007;131:718-24.  Crossref

18. Gilbert TT, Smulian JC, Martin AA, et al. Obstetric admissions to the intensive care unit: outcomes and severity of illness. Obstet Gynecol 2003;102:897-903.  Crossref

One of the major mechanisms of ventilator-associated pneumonia (VAP) is microaspiration of bacteria-colonised oropharyngeal secretions that collect above the inflated cuff of the endotracheal tube (ETT). In Hong Kong, for several decades, the cuff of the most commonly used ETT has been made of polyvinyl chloride (PVC) and has a globular shape. This type of cuff protects against microaspiration poorly, due to microchannels formed from infolding of redundant cuff material after inflation.1 Novel designs of the ETT cuff attempt to overcome this problem by modifying the material from the thicker (50- to 80-micron) PVC to the ultrathin (10-micron) polyurethane (PU),2 and the cuff shape from globular to tapered or cylindrical. In addition, a subglottic secretion drainage port for aspiration of secretions collected above the cuff is available in some ETTs. Although these novel ETTs have been available for many years,3 4 they are not widely used locally. Reasons include inadequate cost-effectiveness data, even though most studies reported favourable efficacy in reducing VAP, though not necessarily mortality.1

Many previous benchtop studies have shown the benefits of these novel designs, but mostly under a limited number of conditions, or under a static condition without the dynamic effect of different levels of positive pressures from mechanical ventilation or negative pressures associated with spontaneous breathing and/or suction.5 6 Moreover, they mostly reported the amount of leakage at a particular time point without showing continuous trends.5 The aim of the present study was to compare these novel ETTs under a comprehensive set of simulated clinical conditions, and to find the bestperforming tube in which downward leakage of fluid across the cuff was minimal.

The experimental setup is shown in Figure 1. The three types of ETTs under test were the Portex Endotracheal Tube (Smiths Medical International Ltd, UK) with a globular PVC cuff; the Mallinckrodt TaperGuard Endotracheal Tube (Covidien, US) with a tapered PVC cuff; and the KimVent Microcuff Endotracheal Tube (Kimberly-Clark Health Care, US) with a cylindrical PU cuff (Fig 2). A transparent, hollow silicone cylinder of length 20 cm and an internal diameter of 2 cm was used as the model trachea. An internal diameter of 2 cm was chosen because from autopsy studies, the mean diameters of male and female tracheas were 2.2 cm and 1.8 cm, respectively.7 A flexible and extensible tube was added to the proximal end of the model trachea to prevent fluid from splashing out where significant upward leakage results from high positive ventilatory pressure.5 The ETT under study was connected to a SERVO-i Adult ventilator (Maquet GmbH & Co. KG, Germany). The model trachea was inclined at 35 degrees to the horizontal to simulate the semi-recumbent position for VAP prevention. Cuff pressure (Pcuff) was maintained by an automated maintenance setup as devised and modified from a previous study.8 In short, compressed air in the range of 2 to 3 L/min was used to inflate the cuff and the pressure was altered with a leakage port along the circuit, to maintain the desired Pcuff within a range of ± 1 cm H₂O at end-expiration. The PVC cuff was monitored continuously using a calibrated electronic pressure transducer (Model HCLA0050EU; Sensortechnics GmbH, Germany), with signals digitally transformed by an analogue-to-digital converter (NI USB-6212; National Instruments, US), so as to display on a computer using the LabVIEW 2010 software (National Instruments Corporation, US). The proximal end of a Y-piece was connected to the distal end of the model trachea. To the Y-piece’s vertical end, a water trap made of a 20-mL syringe was connected. For the study of positive pressure mechanical ventilation, it was connected to a lung simulator (SMS Lung Simulator; SMS Technologies, UK) at its distal end. For the study of spontaneous breathing, it was connected to a Huszczuk-Whipp-Wasserman Gas Exchange System Validator (MedGraphics, US)9 which generated breaths to mimic different metabolic rates; and for the study of total disconnection, it was opened to atmospheric pressure.

The scenarios simulated are shown in Table 1. Clear water (20 mL) was added above the ETT cuff. The whole process was recorded by a video recorder and leakage was measured as observed in the syringe for 20 minutes. Each scenario was studied under different Pcuffs of 10, 20, and 30 cm H₂O. For each scenario at each Pcuff, two tubes of the same ETT type were tested, and each tube was studied repeatedly for 4 times, therefore making a total of 8 measurements for each ETT type per scenario and Pcuff.

The same scenarios were then repeated under sustained tracheal suction by placing a suction catheter (12-Fr closed suction catheter) inside the ETT near the Murphy eye,10 and a suction pressure of 200 cm H₂O was applied continuously for 3 minutes.

The primary measurement was the downward leakage across the cuff, defined as the amount of fluid collected in the syringe (the fluid trap) every minute during the observation period.

Between- and within-group analysis of variance (ANOVA) was used for analysis of the amount of downward leakage during the whole observation period, with the aim of comparing the difference in leakage between the three types of ETTs for each scenario. Each scenario at each Pcuff was analysed separately, in which between-group data were different types of ETTs and within-group data were the cumulative amount (in mL) of leakage over each observation period. Statistical analysis was performed using IBM SPSS software version 20 (SPSS, Inc, IBM, US).

Results of the five scenarios without suction (1NS to 5NS) are shown in Figure 3. The P values represent analyses by ANOVA of the cumulative leakage over the whole 20-minute observation period. The cumulative leakages at the 20-minute time point are shown in Table 2. In summary, in scenario 1NS (positive end-expiratory pressure [PEEP] of 0 cm H₂O), the Microcuff outperformed the other two ETTs at all Pcuffs. At a Pcuff of 30 cm H₂O, only the Portex showed leakage, which was early and massive. At a Pcuff of 20 cm H₂O, only Microcuff did not leak. At a Pcuff of 10 cm H₂O, all three ETTs leaked, but the rate was fastest in Portex and lowest in Microcuff. In scenario 2NS (PEEP of 5 cm H₂O), at Pcuffs of 30 and 20 cm H₂O, all three ETTs showed no leakage, while at a Pcuff of 10 cm H₂O, only Portex showed minimal leakage. In scenario 3NS (PEEP of 10 cm H₂O), none leaked. In scenario 4NS (disconnection), significant leakage occurred only in Portex at Pcuffs of 30 and 20 cm H₂O. At a Pcuff of 10 cm H₂O, all leaked, but the rate was lowest with Microcuff. In scenario 5NS (spontaneous breathing), addition of spontaneous breathing led to leakage in all ETTs at Pcuffs of 10 and 20 cm H₂O, but the rate remained the lowest in Microcuff. At a Pcuff of 30 cm H₂O, only Microcuff showed minimal leakage.

Results of the five scenarios with suction (1S to 5S) are shown in Figure 4. The P values pertained to analyses by ANOVA of the whole 3-minute observation period. The cumulative leakages at the 3-minute time point are shown in Table 2. In summary, in scenario 1S (PEEP of 0 cm H₂O), at Pcuffs of 30 and 20 cm H₂O, only Microcuff was protective, while the other two leaked almost instantly. At a Pcuff of 10 cm H₂O, all three ETTs leaked, but Microcuff leaked very slowly. In scenario 2S (PEEP of 5 cm H₂O), at a Pcuff of 30 cm H₂O, leakage occurred instantly with Portex, in contrast to zero leakage in the corresponding scenario without suction. At a Pcuff of 20 cm H₂O, only Microcuff did not leak. At a Pcuff of 10 cm H₂O, all three ETTs leaked, but Microcuff leaked very slowly. In scenario 3S (PEEP of 10 cm H₂O), at a Pcuff of 30 cm H₂O, none leaked. At a Pcuff of 20 cm H₂O, minimal leakage occurred with Portex. In scenario 4S (disconnection), at a Pcuff of 30 cm H₂O, significant leakage was found in Portex. At a Pcuff of 20 cm H₂O, only Microcuff was protective whilst TaperGuard leaked slowly. At a Pcuff of 10 cm H₂O, Microcuff was still protective with minimal leakage at 3 minutes. In scenario 5S (spontaneous breathing), at a Pcuff of 30 cm H₂O, significant leakage was found in Portex. At a Pcuff of 20 cm H₂O, Portex leaked significantly, TaperGuard leaked very slowly, while Microcuff was protective. At a Pcuff of 10 cm H₂O, all three leaked.

The present benchtop study showed that under the various simulated scenarios studied (positive pressure ventilation, disconnection, spontaneous breathing, with or without the application of suction), the Microcuff ETT consistently outperformed the others with the least downward leak and the lowest sealing pressure, whereas the TaperGuard ETT was in second place. The Portex ETT performed the worst, with significant leakage whenever there was a loss of positive airway pressure even at the recommended Pcuffs of 20 to 30 cm H₂O.

Among the limitations of the present study, many in-vivo factors were not or could not be fully simulated but might have affected the leakage rate. In-vivo leakage could be greater when there is a change in tracheal dimensions during inspiration and/or imperfect conformation of the circular cuff to the trachea (which could have different anteroposterior compared with transverse dimensions), and when there is a sudden change in airway pressure or the cuff position inside the trachea as the patient coughs or moves. Moreover, the seal between the cuff and the moist tracheal mucosa might actually be better, especially with the use of lubricant at the time of intubation.4 11 12 The upward sweeping movement of the ciliated mucosa might also decrease aspiration. Furthermore, subglottic collections are more viscous than water, and their volume is not likely to be as high as 20 mL. This volume was chosen in the current experiment to better discriminate ETT performance and minimise measurement errors. Dynamic patient factors may also affect leakage. For example, leakage in the first scenario (zero PEEP) might be less if significant auto-PEEP develops in severe airflow obstruction, and the size of the cuff relative to the trachea might increase or decrease leakage. Nor did we test the effect of subglottic suction, which is extremely effective in removing fluid collected above the cuff. However, the effectiveness of subglottic suction might differ in vivo, depending on the viscosity of secretions and apposition of the posterior mucosal wall obstructing the suction lumen and interfering with such efforts.13 According to guidelines on endotracheal suction for adults,14 the lowest suction pressure that can effectively clear secretions should be used, which should not exceed 150 cm H₂O, and for not more than 15 seconds. We deliberately used sustained suction at 200 cm H₂O for 3 minutes to better discriminate the performance of each ETT. Notably though, in the present experiment it was observed that leakage occurred within the first second of suction. Although it can be argued that a 20-minute observation period may be too short, we found that prolonging this period led to excessive evaporation of the water above the cuff. We did not use coloured water because we measured water collected in the syringe, simulating the volume of fluid actually aspirated into the lower airway, and not the minute amounts that might just leak and stay around the cuff. Fluid more viscous than water was not used because firstly, human secretions can never be fully simulated, and secondly, should there be any leakage, fluid of a lower viscosity (like water) was considered more liable to leak. Furthermore, the need for thorough cleansing of viscous fluid by dismantling the connections of the apparatus after each set of experiments was another consideration. Such cleansing was regarded as not feasible, because each connection needed to be secured with glue and tapes to withstand the high positive airway pressure and this would take an unrealistic amount of time to do so repeatedly.

The present experiment clearly discriminated the performance characteristics of the three ETTs. Infolding of the excess and thick PVC material in the Portex ETT formed micro-channels through which leakage occurred. By modifying the cuff to a tapered shape, leakage in the TaperGuard ETT was minimised, as there was a point along the length of the cuff where it fitted the trachea perfectly without the infolding of excess cuff material. In the Microcuff ETT, infolding of excess cuff material was still observed. However, micro-channels were not formed because of the much thinner PU material. Its cylindrical shape also provided a larger contact area between the cuff and tracheal wall, thus providing a better seal than the other two types of ETTs.

One of the clinical implications of our findings was that PEEP guards against downward leakage across cuffs, irrespective of the ETT type and Pcuff. In a previous study, it was shown that only the application of a 35 cm H₂O PEEP prior to cuff deflation at extubation was protective against aspiration, but not open or closed suction (that supposedly removes aspirated water).15 This protective effect of a higher PEEP was independent of the peak inspiratory pressure (PIP) as leakage still occurred in scenario 1NS (with a PIP of 15 cm H₂O); such finding was in line with another study showing that only higher PEEP values and not higher inspiratory pressures were protective.16 However, the protective effect of PEEP, especially with low pressures of around 5 cm of H₂O, was counteracted by the application of suction, as shown by the appearance of leakage when suction was applied to the Portex and TaperGuard ETTs as in scenario 2S. Another study showed that when suction at 200 or 300 mbar was applied via ETTs with a PVC cuff, leakage could only be reduced by transiently increasing Pcuff to 50 cm H₂O, and not at all by increasing the PEEP from 5 to 10 cm H₂O or the PIP from 15 to 25 cm H₂O.17 This same study also showed that the PU cuff almost eliminated leakage under all suction pressures, and all PEEP or PIP values.17 Therefore, to prevent leakage during suction, a Pcuff of 50 cm H₂O may be necessary unless a PU cuff is used, while the suction duration and force should be reduced to a minimum, and routine suction should always be avoided. Given that suctioning results in the loss of PEEP and recruitment manoeuvres are recommended thereafter,14 it may be worth studying whether the application of high PEEP during suction can prevent both derecruitment and leakage across the cuff.

Similar to the effect of applying no PEEP during mechanical ventilation, disconnection from such ventilation results in significant leakage across the cuff, and should be avoided as far as possible. Harnessing the portability of an intensive care unit ventilator to avoid circuit disconnections should therefore be considered when patients are transported. When disconnection is necessary, its duration should be kept to a minimum with the maintenance of PEEP. For example, a PEEP valve from a bag-valve device may minimise leakage across the cuff. Spontaneous breathing during disconnection, which creates a negative intrathoracic pressure, further exacerbates downward leakage (scenario 5NS). Another study has also found that leakage increased with increasing inspiratory effort.16 Therefore, during disconnection for spontaneous breathing trials, microaspiration is to be expected if the conventional globular PVC cuff is used. Even with the use of novel ETTs, a high patient inspiratory effort during a trial of spontaneous breathing is conducive to microaspiration and should be anticipated. Extrapolating these results, a high patient inspiratory effort while on mechanical ventilation may prove to be another scenario at high risk of microaspiration.

Under situations with a high risk of microaspiration, namely zero or low PEEP, circuit disconnection, in the presence of high patient inspiratory effort and application of a suction force, the type of ETT used will make a difference to the rate of downward leakage. The Microcuff ETT was shown to offer the best protection in these situations. The TaperGuard ETT was protective if a higher Pcuff could be maintained, and a Pcuff maintenance device (keeping it between 20 and 30 cm of H₂O) may be helpful. Notably, the most commonly used Portex ETT provided the least protection against microaspiration, and leakage occurred in these situations despite maintaining a Pcuff at the recommended 20 to 30 cm H₂O.

Based on the findings of the present study, further clinical trials on VAP prevention using novel ETTs less prone to microaspiration are needed. These should control for confounding factors including PEEP, airway suction, use of automated Pcuff maintenance devices, and airway disconnections. At the time of writing, the cost of a Microcuff ETT was around 6 times that of a Portex ETT (HK$68 vs HK$11), and a TaperGuard ETT was around double (HK$20). Although cost-effectiveness analysis is worthwhile as a basis for wider promotion of the novel ETTs, taking into account the small absolute cost difference, there may be a case for just switching ETTs to those with a lesser tendency to leak until evidence to the contrary appears.

The present benchtop study showed that a higher PEEP, avoidance of unnecessary circuit disconnections and suctioning, and maintenance of adequate Pcuff are important in minimising microaspiration. The Microcuff ETT was shown to be superior to TaperGuard and Portex ETTs in preventing leakage across the cuff. As microaspiration is one of the major mechanisms of VAP, more widespread use of ETTs with a PU cuff, combined with other prevention measures (eg bedhead elevation, oral hygiene) may help to reduce the frequency of associated pneumonias.

No conflicts of interest were declared by the authors.

1. Fernandez JF, Levine SM, Restrepo MI. Technologic advances in endotracheal tubes for prevention of ventilator-associated pneumonia. Chest 2012;142:231-8. Crossref

2. Bulpa P, Bouhon S, Schryvers F, et al. Microaspirations during mechanical ventilation: polyurethane versus polyvinyl cuffed endotracheal tubes. Crit Care 2010;14(Suppl 1):228S. Crossref

3. Dullenkopf A, Gerber A, Weiss M. Fluid leakage past tracheal tube cuffs: evaluation of the new Microcuff endotracheal tube. Intensive Care Med 2003;29:1849-53. Crossref

4. Dullenkopf A, Schmitz A, Frei M, Gerber AC, Weiss M. Air leakage around endotracheal tube cuffs. Eur J Anaesthesiol 2004;21:448-53. Crossref

5. Pitts R, Fisher D, Sulemanji D, Kratohvil J, Jiang Y, Kacmarek R. Variables affecting leakage past endotracheal tube cuffs: a bench study. Intensive Care Med 2010;36:2066-73. Crossref

6. Zanella A, Scaravilli V, Isgrò S, et al. Fluid leakage across tracheal tube cuff, effect of different cuff material, shape, and positive expiratory pressure: a bench-top study. Intensive Care Med 2011;37:343-7. Crossref

7. Mehta S, Myat HM. The cross-sectional shape and circumference of the human trachea. Ann R Coll Surg Engl 1984;66:356-8.

8. Farré R, Rotger M, Ferre M, Torres A, Navajas D. Automatic regulation of the cuff pressure in endotracheally-intubated patients. Eur Respir J 2002;20:1010-3. Crossref

9. Huszczuk A, Whipp BJ, Wasserman K. A respiratory gas exchange simulator for routine calibrator in metabolic studies. Eur Respir J 1990;3:465-8..

10. Dave MH, Frotzler A, Madjdpour C, Koepfer N, Weiss M. Massive aspiration past the tracheal tube cuff caused by closed tracheal suction system. J Intensive Care Med 2011 Feb 7. Epub ahead of print. Crossref

11. Dave MH, Koepfer N, Madjdpour C, Frotzler A, Weiss M. Tracheal fluid leakage in benchtop trials: comparison of static versus dynamic ventilation model with and without lubrication. J Anesth 2010;24:247-52. Crossref

12. Blunt MC, Young PJ, Patil A, Haddock A. Gel lubrication of the tracheal tube cuff reduces pulmonary aspiration. Anesthesiology 2001;95:377-81. Crossref

13. Dragoumanis CK, Vretzakis GI, Papaioannou VE, Didilis VN, Vogiatzaki TD, Pneumatikos IA. Investigating the failure to aspirate subglottic secretions with the Evac endotracheal tube. Anesth Analg 2007;105:1083-5. Crossref

14. American Association for Respiratory Care. AARC Clinical Practice Guidelines. Endotracheal suctioning of mechanically ventilated patients with artificial airways 2010. Respir Care 2010;55:758-64.

15. Hodd J, Doyle A, Carter J, Albarran J, Young P. Increasing positive end expiratory pressure at extubation reduces subglottic secretion aspiration in a bench-top model. Nurs Crit Care 2010;15:257-61. Crossref

16. Ouanes I, Lyazidi A, Danin PE, et al. Mechanical influences on fluid leakage past the tracheal tube cuff in a benchtop model. Intensive Care Med 2011;37:695-700. Crossref

17. Dave MH, Frotzler A, Weiss M. Closed tracheal suction and fluid aspiration past the tracheal tube. Impact of tube cuff and airway pressure. Minerva Anestesiol 2011;77:166-71.