Colonoscopy is an efficient, invasive, and commonly used diagnostic tool with promising therapeutic capacity. Common colonoscopy-related complications include prolonged pain and distension, and rarely draw medical attention or lead to hospitalisation. Severe complications, including bleeding and perforation, are potentially life-threatening and require urgent management. Although death is uncommon, occurring in no more than 3 per 10 000 procedures, the incidence of post-polypectomy bleeding and perforation ranges from 1.6 to 14.8 and 0.2 to 1.0 per 1000 procedures, respectively.1 2 3 4 5 6 It is difficult to accurately benchmark direct colonoscopy-related complications due to the different outcome measure definitions used in studies. For example, some studies include immediate complications only, while others extend the complication period to 7 or 30 days, and some studies include an extensive list of complications while others include only bleeding and perforation.1 2 5 7 8 9 10 Furthermore, the efficacy and safety of the procedure vary across clinical settings and the targeted populations. With no available local data, consensus for complication incidence remains inconclusive.

Intravenous sedation is routinely used during colonoscopy to minimise the discomfort and pain associated with the procedure. Endoscopists are equipped to give sedatives and to monitor their side-effects, but anaesthetists are often invited to provide monitored anaesthetic care (MAC) when the patient is considered to be at high risk for complications, for instance, older patients and those with multiple co-morbidities are particularly vulnerable to complications. Systemic complications vary from prolonged drowsiness to fatal events such as cardiovascular or cerebrovascular events. Cardiovascular events following sedation, such as hypotension and myocardial infarction, during colonoscopy have been reported to range from 0.1 to 59.1 per 1000 procedures. Cerebrovascular events such as stroke range from 0.1 to 1.3 per 1000 procedures.3 6 10 The outcome variables are highly heterogeneous, for example, Nelson et al’s study3 included myocardial infarction, vasovagal event, and arrhythmia in the cardiovascular incidents, while Ma et al’s study10 recorded hypotension only.10 Some endoscopists opted to study complications related to the use of carbon dioxide (CO₂) insufflation and absence of sedative use.11 12 13 This study aimed to record all complications systematically and to determine the relevant risk factors.

This prospective study collected data for all colonoscopies done from 1 June 2011 to 31 May 2012 at the Endoscopy Centre of the Hong Kong Sanatorium & Hospital (HKSH), which is a private hospital in Hong Kong. Prior to colonoscopy, patients were invited to give their written consent for their participation in the study, including for the 30-day follow-up telephone interview. The Hospital Management Committee involving the Research Ethics Committee of the HKSH approved the study.

The complications were recorded by nurses on a standard form during and immediately after the procedure. The standard audit forms for immediate and delayed colonoscopy complications were designed by a research doctor, with the most common complications based on literature review.

The immediate complications audit form included patients’ demographics, use of sedation/analgesic/antispasmodic, use of MAC, gross indications for colonoscopy (therapeutic or diagnostic), type of therapeutic procedures performed such as polypectomy, the reason for incomplete colonoscopy (caecum intubation failure), quality of bowel preparation (adequate – good/adequate or not – fair/poor, which was rated by the endoscopist), and the use of CO₂ insufflation. Complication data were divided into systemic and colonoscopy-related complications. For systemic complications, we captured data for nausea/vomiting, hypotension (systolic blood pressure <100 mm Hg), bradycardia/tachycardia (heart rate <50 to >100 beats/min), vasovagal fainting, and other cardiovascular or cerebrovascular events. For colonoscopy-related complications, data for perforation, persistent pain/discomfort, abdominal distension, and haemorrhage were gathered.

Delayed complications were defined as the above events happening from the day after the initial colonoscopy to the 30th day that required readmission or admission to other hospitals. For those readmissions, we automatically inspected the records for the reasons and interventions if the readmissions were complication-related. Otherwise, trained student nurses or a research doctor telephoned all consented participants to interview for the 30-day complications using the delayed complication audit form. A participant was declared lost to follow-up after three telephone attempts. The student nurses were trained by senior nurses and the research doctor with standard instructions.

Data analysis was performed by the Statistical Package for the Social Sciences (Windows version 14.0; SPSS Inc, Chicago [IL], US). Descriptive statistics (mean, percentage, incidence, and/or 95% confidence interval [CI]) were used to display the characteristics of the sample. For those complications with zero events, only 95% CIs were given.14 Backward logistic regression analyses were performed to draw prediction models for immediate complications, including colonoscopy-related complications or systemic complications, and overall complications, including immediate and delayed complications, from the sample; entering variables were chosen from age, sex (male or female), use of MAC (yes or no), sufficiency of bowel preparation (adequate or not), and completion of colonoscopy (yes or no) according to the results of the univariate analyses by Pearson Chi squared tests of all potential independent variables, with a significance level set at 10%; variable inclusion in the iteration was set at P<0.1 for backward logistic regression analyses. Returning coefficients of the variables were interpreted as adjusted odds ratio (OR_adjusted) with 95% CI provided. All significance levels were set at two-sided α=0.05.

A total of 6196 colonoscopies (3143 women; mean age 53.7 years; standard deviation, 12.7 years) were done during the study period. Most patients were aged between 45 and 64 years, were in-patients, had undergone diagnostic colonoscopy, and received intravenous sedation (60.5%, 70.5%, 53.0%, and 99.4% respectively; Table 1). The data for immediate complications were complete, while the 30-day follow-up was completed for 3657 procedures (803 were lost to follow-up and 1736 refused; compliance rate 59.0%; Table 2).

Of the 6196 colonoscopies, 2912 were therapeutic with 99.7% dedicated to polypectomy (Table 1). There were 73 (1.2%) cases of incomplete colonoscopy, 18 (24.7%) of which were due to inadequate preparation. Other reasons for incomplete colonoscopy were tumour obstruction (15 of 73; 20.5%) and intention of sigmoidoscopy or stent insertion (26 of 73; 35.6%). A total of 149 patients were readmitted within 30 days after the procedure, of which six (4.0%) were related to complications. The other reasons were cancer, gastro-intestinal disease, or cardiac events (Table 2).

Regarding the choice of sedation, midazolam and pethidine were the most used at 81.5% and 82.0%, respectively, while 15.9% of patients underwent MAC. Immediate complications reported included hypotension, vasovagal fainting, and nausea/vomiting, or a combination (6.1, 0.6, 0.3, and 0.5 per 1000 procedures, respectively; Table 3). There were no severe cardiovascular events such as heartbeat irregularity or myocardial infarction or cerebrovascular events such as stroke. Furthermore, none of the patients reported delayed systemic complications in the 30-day follow-up.

Modelling to study the potential risk factors showed that being female, use of MAC, and inadequate bowel preparation were the significant independent predictors for systemic immediate complications (OR_adjusted=2.0, 2.6, and 3.7, respectively; all were P<0.05; Table 4).

Immediate colonoscopy-related complications were recorded for 48 patients (7.7 per 1000 procedures), including extensive pain/discomfort, abdominal distension, and perforation (Table 3). The only perforation was at the sigmoid-rectal junction and was due to adhesion, which might be related to previous abdominal surgery (total hysterectomy and bilateral salpingo-oophorectomy was done 10 years previously). In the 30-day follow-up, six patients reported complication-related readmissions, five of whom were for bleeding after discharge from hospital and one was for inflammation; all were caused by polypectomy.

Modelling was done to formulate a predictive algorithm for significant independent risk factors and outcome events (Table 4). Inadequate bowel preparation and incomplete colonoscopy were the significant predictors for immediate colonoscopy-related complications (OR_adjusted=3.5 and 6.2, respectively; both were P<0.05) and for all immediate complications (OR_adjusted=3.5 and 4.5, respectively; both were P<0.05). In the model for all immediate complications, being female and use of MAC were also predictors (OR_adjusted=1.6 and 1.8, respectively; all were P<0.05).

A predictive model was also done for overall complications, including all immediate and delayed complications among those who completed follow-up at 30 days. The effect of previous significant predictors was transient in that they did not predict the overall complications occurring in the 30-day post-colonoscopy period. Monitored anaesthetic care was the only predictor during the 30-day period (OR_adjusted=2.0; P=0.019).

Inadequate bowel preparation and incomplete colonoscopy were identified as risk factors for colonoscopy-related complications. Other complications were mostly hypotension and abdominal distension. No myocardial infarction, transient ischaemic attack, or death relating to colonoscopy was reported.

Colonoscopy is rarely a complication-free procedure, but a good understanding of the possible complications can help to minimise them. While experienced endoscopists, diagnostic procedures, and young patients are protective factors for colonoscopy complications, trainee endoscopists, therapeutic procedures, advanced age, female sex, obesity, co-morbidity, anticoagulant use, and previous abdominal surgery are risk factors for complications.3 5 6 10 15 16 17 The relatively low incidence of complications recorded in this study could be attributed to the fact that more than half of the procedures were diagnostic, thus reducing the potential for polypectomy-associated complications.

Other events—such as abdominal pain and distension, hypotension, vasovagal fainting, and nausea/vomiting—accounted for 98.9% of all immediate complications; these events were mostly reported by women (60.6%; OR_adjusted=1.6; Table 4). This result is consistent with the literature.17 This effect could possibly be explained by different perceptions of somato-sensation, which could be traced back to the socio-emotional cultivation and cultural expectation of the different sexes.18

Despite the sex effect, use of MAC, inadequate bowel preparation, and incomplete colonoscopy were all related to immediate complications. The use of MAC, in particular, requires more interpretation because it has not been found to be a risk factor in other studies and its use ought not be a cause of complications. However, MAC is designed for patients who are vulnerable to the complications of sedation and the procedure, especially those with co-morbidities and who are at an advanced age. In this study, co-morbidity was not reviewed, but patients who underwent MAC were significantly older than those who did not (mean age, 56.1 vs 53.3 years; P<0.001, t test) which may be partly contributory. However, age was not significantly associated with any of the complications after adjustment for other factors. Age might have a greater impact if co-morbidity was also considered and this may be an area for further study.

Inadequate bowel preparation, in which faeces obscure the inner lining of the colon, impedes the vision and heightens the risk for complications. Likewise, incomplete colonoscopy due to inadequate bowel preparation or unbearable discomfort inevitably increases the risk for complications. These results are consistent with other studies.2 3 10

A large proportion (59%) of the study population completed the study. Per protocol univariate analyses showed that use of MAC was the sole significant factor related to overall complications (Table 4).

Sedation-free colonoscopy is a feasible alternative that could reduce the risk of systemic complications; use of CO₂ insufflation might increase its tolerability while maintaining visibility. In this study, a sedation-free procedure was performed in only 36 patients, with no complications; CO₂ insufflation was done for 647 (10.9%) patients instead of gas (room air), with only minor complications encountered (two patients reported abdominal distension and seven reported hypotension/vasovagal fainting). Insufflation with CO₂ could be adopted more widely because it is non-explosive, absorbable, and does not affect the mucosal blood flow, thus minimising discomfort and the risk of colonic ischaemia. According to Bretthauer et al,19 CO₂ insufflation leads to quicker recovery, and less pain and complications. These advantages are supported by other studies including local research.11 19 20 These factors might inspire greater use of CO₂ insufflation and minimise use of sedation for selected patients.12 13 21

Inadequate bowel preparation not only hampers completion of the procedure, but also increases the risk for complications. As evidenced from the prediction models, inadequate bowel preparation significantly increases the occurrence of complications. In addition to implementing the current standardised bowel preparation protocol, enforcement of patient education, compliance, and early admission for monitored bowel preparation might help to further suppress the rate for inadequate bowel preparation.

Many of the patients were lost to contact by telephone. In this study, we assumed that the pattern of missing patients was random and was not affected by whether or not the patients had complications. However, self-selection bias might exist. If patients without complications were more likely to be lost to contact, the rate of delayed complications would be overestimated. However, the rate of delayed complications would be underestimated if the patients had been admitted to other hospitals or had died. Endoscopist experience, and patient characteristics of co-morbidities and severe symptoms before the procedure were not recorded for analysis in this study. These are potential risk factors for complications. In view of the importance of monitoring the complications of colonoscopy, further study might include the missing variables of co-morbidity, endoscopist experience, symptoms at presentation and oxygen level, and explore factors such as the influence of body mass index and medical history, sedation dose, indication for colonoscopy, use of laxatives and compliance, and post-colonoscopy diagnosis/pathology that might help to minimise the variance on prediction of complications.

We would like to thank Dr Sheri Lim, former Administrative Officer (Medical) who proposed and developed the audit. We are grateful to Dr KM Lai, Head of Department of Anaesthesiology for his expert advice on the anaesthetic terminology; Dr Raymond Yung and Dr KN Lai, Assistant Medical Superintendents for their valuable suggestions and encouragement; Ms Grace Wong, Medical Records Manager for coordinating with different departments; and Ms Sara Fung, former Research Nurse for coordinating the project. Also, thanks to all doctors who contributed their knowledge and data, for this study would not have been accomplished without them.

1. Lorenzo-Zúñiga V, Moreno de Vega V, Doménech E, Mañosa M, Planas R, Boix J. Endoscopist experience as a risk factor for colonoscopic complications. Colorectal Dis 2010;12(10 Online):e273-7.

2. Gastrointestinal endoscopy, version 1. Australasian Clinical Indicator Report 2003-2010. Australian Council on Healthcare Standards; 2010.

3. Nelson DB, McQuaid KR, Bond JH, Lieberman DA, Weiss DG, Johnston TK. Procedural success and complications of large-scale screening colonoscopy. Gastrointest Endosc 2002;55:307-14. CrossRef

4. Rathgaber SW, Wick TM. Colonoscopy completion and complication rates in a community gastroenterology practice. Gastrointest Endosc 2006;64:556-62. CrossRef

5. Rabeneck L, Paszat LF, Hilsden RJ, et al. Bleeding and perforation after outpatient colonoscopy and their risk factors in usual clinical practice. Gastroenterology 2008;135:1899-1906, 1906.e1.

6. Viiala CH, Zimmerman M, Cullen DJ, Hoffman NE. Complication rates of colonoscopy in an Australian teaching hospital environment. Intern Med J 2003;33:355-9. CrossRef

7. Ko CW, Riffle S, Michaels L, et al. Serious complications within 30 days of screening and surveillance colonoscopy are uncommon. Clin Gastroenterol Hepatol 2010;8:166-73. CrossRef

8. Singh H, Penfold RB, DeCoster C, et al. Colonoscopy and its complications across a Canadian regional health authority. Gastrointest Endosc 2009;69:665-71. CrossRef

9. Teruki O, Keiichi O. The safety of endoscopic day surgery for colorectal polyps. Dig Endosc 2008;20:92-5. CrossRef

10. Ma WT, Mahadeva S, Kunanayagam S, Poi PJ, Goh KL. Colonoscopy in elderly Asians: a prospective evaluation in routine clinical practice. J Dig Dis 2007;8:77-81. CrossRef

11. Wong JC, Yau KK, Cheung HY, Wong DC, Chung CC, Li MK. Towards painless colonoscopy: a randomized controlled trial on carbon dioxide-insufflating colonoscopy. ANZ J Surg 2008;78:871-4. CrossRef

12. Bayupurnama P, Nurdjanah S. The success rate of unsedated colonoscopy examination in adult. Internet Journal of Gastroenterology 2010;9:2.

13. Ylinen ER, Vehviläinen-Julkunen K, Pietilä AM, Hannila ML, Heikkinen M. Medication-free colonoscopy—factors related to pain and its assessment. J Adv Nurs 2009;65:2597-607. CrossRef

14. Ho AK. When the numerator is zero: another lesson on risk. Am Biol Teach 2009;71:531-3. CrossRef

15. Miller A, McGill D, Bassett ML. Anticoagulant therapy, anti-platelet agents and gastrointestinal endoscopy. J Gastroenterol Hepatol 1999;14:109-13. CrossRef

16. Dobbins C, DeFontgalland D, Duthie G, Wattchow DA. The relationship of obesity to the complications of diverticular disease. Colorectal Dis 2006;8:37-40. CrossRef

17. Ko CW, Riffle S, Shapiro JA, et al. Incidence of minor complications and time lost from normal activities after screening or surveillance colonoscopy. Gastrointest Endosc 2007;65:648-56. CrossRef

18. Kroenke K, Spitzer RL. Gender differences in the reporting of physical and somatoform symptoms. Psychosom Med 1998;60:150-5. CrossRef

19. Bretthauer M, Lynge AB, Thiis-Evensen E, Hoff G, Fausa O, Aabakken L. Carbon dioxide insufflation in colonoscopy: safe and effective in sedated patients. Endoscopy 2005;37:706-9. CrossRef

20. Welchman S, Cochrane S, Minto G, Lewis S. Systematic review: the use of nitrous oxide gas for lower gastrointestinal endoscopy. Aliment Pharmacol Ther 2010;32:324-33. CrossRef

21. Takahashi Y, Tanaka H, Kinjo M, Sakumoto K. Sedation-free colonoscopy. Dis Colon Rectum 2005;48:855-9. CrossRef

Urinary incontinence (UI), defined as “the complaint of any involuntary leakage of urine”,1 is a major clinical problem, and a significant cause of disability and dependency in the aged population. It is a condition with heterogeneous pathology and commonly classified as stress urinary incontinence (SUI), urge urinary incontinence (UUI), and mixed urinary incontinence (MUI) depending on the symptom behaviour. While the prevalence of UI in older women, globally, is estimated to range from 15% to 30%,2 the reported prevalence rate of UI in Hong Kong ranges from 20% to 52%.3 It has been acknowledged that UI is associated with profound adverse impact on the quality of life (QoL) of the sufferers.3 4 The impact of UI is so substantial that community-dwelling elderly with UI reported inferior physical and mental health, worse self-perceived health status, greater disability, and more depressive symptoms.5 In addition, the extent of the impact was shown to be associated with the severity of UI. Therefore, it is important to investigate a safe and effective treatment strategy in this population, especially in a community setting.

Conservative management has been recommended as the first-line management for UI. It is acknowledged that a variety of conservative management strategies which require patient’s active participation shows promising results for patients with UI. These include pelvic floor muscle training (PFMT),6 7 vaginal cones,8 bladder training (BT),9 and even combination of PFMT and general lumbopelvic mobilisation exercises.10 A recent Cochrane review7 suggested that PFMT, the ‘knack’ manoeuvre (a voluntary counterbracing type of contraction during physical stress), and BT are effective strategies in the management of UI in general. In particular, a combination of PFMT and BT was shown to have superior outcome than BT alone for the management of UUI and MUI.9 However, to date, there is insufficient conclusive evidence on the best approach for PFMT.11 In addition, the applicability and effectiveness of PFMT and BT for treating UI have not been properly evaluated in the elderly Chinese population, especially in randomised controlled studies. The aim of this study was to evaluate the effectiveness of a Urinary Continence Physiotherapy Programme (UCPP), which is a comprehensive programme involving education and exercise (PFMT and BT) components for managing SUI, UUI, and MUI in older Chinese women in a community setting.

A total of 60 subjects were recruited for screening by convenience sampling from six Elderly Health Centres (EHCs), Department of Health, Hong Kong. Inclusion criteria were Chinese females aged 65 years or older who had a clinical diagnosis of SUI, UUI, or MUI (with reference to the definition from International Continence Society1) of a mild-to-moderate severity (based on the scoring system by Lagro-Janssen et al12) which is made by the EHC medical officers in-charge. Exclusion criteria were active urinary tract infection, patients on diuretic medication, presence of bladder pathology or dysfunction due to genitourinary fistula, tumour, pelvic irradiation, neurological or other chronic conditions (eg diabetes mellitus, Parkinson’s disease), previous anti-incontinence surgery, significant cognitive impairment assessed by the Cantonese version of Mini-Mental State Examination Score (CMMSE13 with cutoffs of: ≤18 for illiterate subjects, ≤20 for those who had had 1 to 2 years of schooling, ≤22 for those who had had more than 2 years of schooling out of a maximum score of 30), obesity (body mass index [BMI] of >30 kg/m²), and use of concomitant treatments during the trial.

Randomisation was performed prior to the study by an off-site investigator using a computerised randomisation programme with allocation concealment by sequentially numbered, opaque, and sealed envelopes. After taking consent, grouping of the individual participants was revealed to the principal investigator by phone. Overall, 55 eligible participants were assigned to the intervention (n=27) or control (n=28) groups. The trial period lasted for 12 weeks. The study was approved by the Institutional Medical Research Ethics Committee and was conducted in accordance with the Declaration of Helsinki.14

One physiotherapist was responsible for delivering assessment and treatment to all subjects during the trial period of 12 weeks, and she was not blinded to the intervention.

The intervention group received a 30-minute individual training session at a pre-decided time of the day, once weekly for the first 4 weeks; and then once bi-weekly for the remaining 8 weeks. A total of eight treatment sessions were given to each recipient. There were three major components in the UCPP: education (anatomy of the pelvic floor muscle [PFM] and urinary tract, urinary continence mechanism, and bladder care), PFMT with the aid of vaginal palpation, and BT. Pelvic floor muscle training included Kegel exercise programme and neuromuscular re-education (the ‘knack’).15 Bladder training involved strategies to increase the time interval between voids by a combination of progressive void schedules, urge suppression, distraction, self-monitoring, and reinforcement.

Four stages of Kegel’s PFMT programme were adopted in this study, including (1) muscle awareness, (2) strengthening, (3) endurance, and (4) habit building and muscle utilisation.16 The exercise regimen was designed to progressively strengthen both type I and type II muscle fibres of the pelvic floor. In the first 2 weeks (muscle awareness phase), one set of 10 (week 1) and 15 (week 2) slow submaximal contractions for 5 seconds each and five fast maximal contractions with a 10-second relaxation between contractions was performed in lying down position. In the strengthening phase (weeks 3 to 4), the muscle-strengthening element was reinforced by gradually increasing the number of submaximal contractions to 25 per session with an increment of five repetitions per week in gravity-dependent position including sitting and standing. The number of fast maximal contractions (5) remained the same as in the awareness phase. In the endurance phase (weeks 5 to 8), the training was more focused on improving the performance of slow and sustained contractions of PFMs by increasing the contraction time to 10 seconds with submaximal contraction while keeping the exercise position and number of both submaximal and maximal contractions as in week 4. In the habit-building and muscle utilisation phase (weeks 9 to 12), the learnt neuromuscular re-education technique (the ‘knack’) and urge suppression strategies were reinforced. In this period, one set of 30 slow submaximal contractions for 10 seconds each and 10 fast maximal contractions with a 10-second relaxation between contractions was practised. Participants were asked to perform three sets of the above-mentioned exercise at specific periods as part of the home programme.

The control group was given advice and received an educational pamphlet with information about management of UI at baseline. Participants were given an appointment for a follow-up visit in 12 weeks.

A number of indicators were employed to assess different aspects of outcomes. First, the number of UI episodes in the previous 7 days (UI7) was examined using a weekly bladder diary log sheet, which was also the primary outcome measure of this study. Information for UI7 was collected at baseline and then weekly until the end of the programme (12 weeks). Second, a validated condition-specific QoL assessment tool—Incontinence Impact Questionnaire Short Form (IIQ-7) Chinese (Taiwan) version17—was used to study the impact of UI on QoL and its change with the intervention after minor modifications were made to align with the local culture. Content validation was made by a panel of doctors who reviewed the instrument and determined if the questions satisfied the content domain. Seven items were included in the questionnaire to examine if the subjects were suffering from urine leakage under those specific situations. The questionnaire was administered by the same physiotherapist and the subjects were asked to choose the most appropriate response to those 7 items on a 4-point ordinal scale, with 0 meaning “not at all affected”, 1 “slightly affected”, 2 “moderately affected”, and 3 “greatly affected”. The maximum score of 21 indicated a great impact of UI on QoL. Third, subjective perception of improvement was assessed with a 10-cm visual analogue scale (VAS) rated from 0 to 10, with 0 suggesting “no improvement” and 10 “complete relief” at the end of the intervention period. Fourth, another VAS was used to assess subjects’ satisfaction to treatment (on a 0 to 10 rating), with 0 being “totally dissatisfied” and 10 “totally satisfied”. The IIQ-7 was collected at baseline and at the end of the programme. Compliance with treatment in the intervention group was evaluated from two perspectives: attendance and compliance with home exercises. Attendance was reviewed by calculating the proportion of sessions that were attended by an individual. Compliance with home exercises was also reviewed by calculating the reported frequency of exercises being executed. Any drawbacks and adverse effects during the intervention period were also monitored.

The sample size calculation based on the power estimation and results of a similar study,18 with a power of 0.8 and α = 0.05 and an attrition rate of 30%, revealed that a recruitment sample of 60 participants (30 participants in each group) was required. Statistical analysis was performed with the Statistical Package for the Social Sciences (Windows version 13.0; SPSS Inc, Chicago [IL], US). Any missing data were treated with “the last observation carried forward” approach. Independent t tests (for parametric data) or Mann-Whitney test (for nonparametric or non-normally distributed data) and Chi squared tests (for nominal/ordinal data) were used to compare the demographic data and outcome variables between the intervention and control groups at baseline. Pairwise comparisons with the P value adjusted using Bonferroni correction was used to examine for the differences in UI7 between week 1 and the subsequent time points during the intervention period (eg between week 1 and 2, week 1 and 3, etc). The results were presented as mean (standard deviation [SD]).

The demographics and baseline measurements for the participants are shown in Table 1. Of the 60 participants recruited for screening, three did not turn up for assessment, one declined to participate in the study, and one was excluded due to impaired mental status (Fig 1). The majority of the participants were in their 70s with a mean (± SD) age of 74.3 ± 4.6 years. There was no significant difference between the groups in terms of age, BMI, parity, education level, mental status (CMMSE), and the characteristics of UI at baseline.

There was a significant interaction between time and groups in UI7 (F_(1,53) = 33.14; P<0.001). A significant reduction in UI7 was noted only in the intervention group. There was significant difference between the control and intervention groups (t = –5.3; P<0.001) at 12 weeks with a mean difference of -6.4 (95% confidence interval [CI], -8.9 to -3.9). The mean numbers of UI7 for intervention and control groups were 1.0 ± 1.9 (95% CI, 0.3-1.7) and 7.4 ± 6.2 (95% CI, 5.0-9.8), respectively, at 12 weeks (Table 2). When comparing the percentage reduction ([pre-treatment frequency – post-treatment frequency] / [pre-treatment frequency] x 100%) in UI7, the intervention group demonstrated a mean of more than 90% reduction versus 7.2% in the control group. A similar trend of improvement was shown in subjects with SUI, UUI, and MUI; however, statistical analysis was not performed due to insufficient power (Fig 2). Post-hoc pairwise comparisons (a total of 11) suggested a significant improvement from week 1 to week 5 and onwards (P<0.001) in the intervention group.

A significant interaction between time and groups was noted (F_(1,53) = 54.56; P<0.001). As IIQ-7 was non-normally distributed, Mann-Whitney test was used and revealed a significant reduction of IIQ-7 (ie improvement in QoL) only in the intervention group (P<0.001), with a mean difference of -3.9 (95% CI, -5.1 to -2.8) at 12 weeks (Table 2).

The results of the subjective perception of improvement and level of satisfaction with treatment after 12 weeks are shown in Table 2. The majority of the participants in the intervention group were satisfied with the interventions and perceived a subjective improvement. The mean attendance and exercise compliance rates in the intervention group were 97.7% ± 5.0% and 99.4% ± 1.9%, respectively. The attrition rate was zero in both groups during the whole study period. No adverse effect or discomfort was reported during the intervention period.

In line with results from previous studies, this study further confirms that PFMT is an effective and safe treatment for women suffering from various types of UI. However, we observed a mean of >90% reduction in UI episodes which is noticeably higher than that reported in previous similar studies (32% to 73% reduction).18 19 Although recommendation of PFMT for women with UI is strongly supported by previous research findings,7 the best approach of PFMT programme continues to remain unclear.11 One possible explanation for the superior outcome in this study might be the combination of a few effective features in this programme, including programme duration of 12 weeks with gradual exercise progression, combination of PFMT and BT, manual vaginal palpation, and one-on-one supervised training session. First, the design of this UCPP incorporated some important concepts of exercise therapy. The training period of this study was 12 weeks, which could optimise the effect of the neural adaptation (recruitment of efficient motor units and frequency of excitation) and muscle hypertrophy according to the recommendations made by the American College of Sports Medicine.20 In addition, a recent systematic review7 also revealed that implementation of PFMT programme for at least 3 months (ie around 12 weeks) is more likely to result in greater treatment effect versus that lasting for <12 weeks. This programme also adopted the concepts raised by Kegel16 in which the progression of the exercise regimen is designed according to different stages, namely muscle awareness, strengthening, endurance, habit building, and muscle utilisation. Previous studies suggested a negative correlation between increased PFM strength and UI symptoms.21 22 The improvement of UI7 at the end of the UCPP might be a direct result of the muscle training programme, although PFM strength was not measured in this study. Theoretically, skeletal muscle strengthening should be facilitated by using additional resistance20 and, therefore, it is questionable whether muscle strength could be increased by UCPP using only maximal voluntary contractions. However, a previous study has indicated muscle strength improvement with daily practice of voluntary PFM contraction without resistance,20 and absence of extra improvement in UI patient groups with intravaginal resistance as compared with the group without resistance.23

Second, a combination of PFMT and BT was used in this UCPP. Although PFMT has been recommended as first-line management, even for women with UUI and MUI,7 there is some evidence suggesting superior outcome with combined PFMT and BT in this population.9 The result of this study confirmed these suggestions as a similar pattern of improvement was observed across the three groups (UUI, SUI, and MUI), although statistical analysis of the difference between subgroups was not performed due to the small sample size.

Third, vaginal palpation was used to facilitate and ensure correct PFM contraction. It was reported that approximately 30% of women are unable to perform isolated pelvic floor contractions with only written or verbal instructions.21 We consider the extra proprioceptive cue and specific verbal feedback are an integral part of the PFMT, and consider these to play a crucial role, especially in the initial (muscle awareness) stage. Ensuring feedback may also increase exercise adherence and compliance, apart from improving the treatment outcomes.

Fourth, the exercise sessions were conducted on a one-on-one basis for 30 minutes each. It has been reported that the amount of contact with health care professionals is positively correlated with reported cure and improvement (eg perception of change and incontinence-specific QoL) in patients with UI.11 It is argued that women receiving more attention may overestimate their improvement to please the treatment provider (ie experimenter effect),11 and it is strongly suggested to include more ‘objective’ data such as leakage episode outcomes in all PFMT trials. The result of this study revealed improvement in UI7 as well as other subjective measures (IIQ-7, perception of improvement, and treatment satisfaction), which could be considered as additional evidence base.11

The subjects’ compliance with the treatment programme was excellent, as reflected by the high compliance rate with exercise regimens, high attendance rate, and zero dropouts; the dropout rates reported in previous studies were relatively high, ranging from 12% to 41%.6 18 24 A possible explanation for such good compliance might be the significant improvement in the early stage of the protocol, which in turn increased the participants’ motivation for and confidence in adhering with the PFMT programme. Regular meetings (weekly or bi-weekly) with the same physiotherapist, who offered continuity of care, could be another possible explanation for the favourable compliance.

The main limitations in this study were: (1) potential selection bias due to use of convenience sampling, (2) absence of assessor blinding, (3) possibility of over-reporting, and (4) the use of the modified IIQ-7 Chinese (Taiwan) version. It has been acknowledged that convenient sampling might not be representative of the whole population suffering from UI. On the other hand, our subject group might have similar care-seeking behaviour as the client group in clinical practice. Although statistically insignificant, the data showed a small difference in some aspects of the demographic characteristics. In general, the control group tended to be slightly older (75.4 years vs 73.0 years), more likely to be illiterate, and have milder severity, and shorter duration of UI versus the intervention group. As these slight differences in the demographics might induce confounding, their effects warrant further investigation. Nevertheless, interpretation of the results of this study deserves some caution. In this study, five participants recruited for screening did not join the programme due to various reasons (one denied, three failed to turn up, and one due to impaired mental status). Although the specific reason for the absence of three participants was not investigated, the possibility of self-selection bias should be considered. In addition, all involved parties (the assessor, treatment provider, and the participants) were not blinded to the intervention, as opposed to the ideal experimental setup. However, it is widely acknowledged that given the nature of the treatment programme, it is difficult and often impossible to blind the treatment provider and participants during treatment.11 Due to resource limitation, it was not possible to include an independent, blinded assessor for outcome assessment. We were well aware of the possible ‘experimenter effect’, and therefore used UI7 as our primary outcome measure which is considered a more objective measure to minimise the possible effect of over-reporting of subjective improvement,11 although its ‘objectivity’ remains controversial. In addition, a significant proportion of participants (approximately 44%) required assistance for completing the outcome questionnaires due to illiteracy. This could possibly lead to over-reporting of improvement. Furthermore, the possibility of over-reporting of compliance by participants using self-reported weekly exercise diary should not be overlooked. There is, however, no better measure available to monitor the performance of this type of exercise accurately. A recent randomised controlled trial25 reported that severity of SUI symptoms at baseline and extent of PFM strength improvement, rather than exercise adherence, were correlated with symptom reduction for women with SUI. The result suggested a complex interaction between subject’s health condition, exercise compliance and treatment effectiveness, which warrant further investigation. Therefore, we intended not to account the improvement observed in our intervention group to the high self-reported compliance rate. Finally, a modified Chinese (Taiwan) version of IIQ-7 was used in this study. We are aware of the fact that this version has not been properly validated. However, we do not believe this affects our results as the modification was minor and the version was highly comparable with the Hong Kong version which was validated subsequent to the current study.

This study examined the immediate effectiveness of the verbally instructed UCPP just after cessation of supervised training. No follow-up data were collected. It is recommended that the long-term effectiveness of UCPP be explored, especially in the light of fairly extensive literature which reported poor long-term adherence and relapse at 3 to 5 years following pelvic floor rehabilitation programme.26

This study demonstrated that a structured 12-week programme of PFMT with gradual exercise progression, BT with urgency suppression, and enhanced education is likely to improve episodes of urinary leakage and QoL in older Chinese women with various kinds of UI in a community setting.

1. Abrams P, Cardozo L, Fall M, et al. The standardisation of terminology of lower urinary tract function: report from the Standardisation Sub-committee of the International Continence Society. Neurourol Urodyn 2002;21:167-78. CrossRef

2. Klausner AP, Vapnek JM. Urinary incontinence in the geriatric population. Mt Sinai J Med 2003;70:54-61.

3. Pang MW, Leung HY, Chan LW, Yip SK. The impact of urinary incontinence on quality of life among women in Hong Kong. Hong Kong Med J 2005;11:158-63.

4. Cheung RY, Chan S, Yiu AK, Lee LL, Chung TK. Quality of life in women with urinary incontinence is impaired and comparable to women with chronic diseases. Hong Kong Med J 2012;18:214-20.

5. Aguilar-Navarro S, Navarrete-Reyes AP, Grados-Chavarria BH, Garcia-Lara JM, Amieva H, Avila-Funes JA. The severity of urinary incontinence decreases health-related quality of life among community-dwelling elderly. J Gerontol A Biol Sci Med Sci 2012;67:1266-71. CrossRef

6. Fan HL, Chan SS, Law TS, Cheung RY, Chung TK. Pelvic floor muscle training improves quality of life of women with urinary incontinence: a prospective study. Aust N Z J Obstet Gynaecol 2013;53:298-304. CrossRef

7. Dumoulin C, Hay-Smith J. Pelvic floor muscle training versus no treatment, or inactive control treatments, for urinary incontinence in women. Cochrane Database Syst Rev 2010;(1):CD005654.

8. Pereira VS, de Melo MV, Correia GN, Driusso P. Long-term effects of pelvic floor muscle training with vaginal cone in post-menopausal women with urinary incontinence: a randomized controlled trial. Neurourol Urodyn 2013;32:48-52. CrossRef

9. Wallace SA, Roe B, Williams K, Palmer M. Bladder training for urinary incontinence in adults. Cochrane Database Syst Rev 2004;(1):CD001308.

10. Kim H, Yoshida H, Suzuki T. The effects of multidimensional exercise treatment on community-dwelling elderly Japanese women with stress, urge, and mixed urinary incontinence: a randomized controlled trial. Int J Nurs Stud 2011;48:1165-72. CrossRef

11. Hay-Smith EJ, Herderschee R, Dumoulin C, Herbison GP. Comparisons of approaches to pelvic floor muscle training for urinary incontinence in women. Cochrane Database Syst Rev 2011;(12):CD009508.

12. Lagro-Janssen TL, Debruyne FM, Smits AJ, van Weel C. Controlled trial of pelvic floor exercises in the treatment of urinary stress incontinence in general practice. Br J Gen Pract 1991;41:445-9.

13. Chiu HF, Lee HC, Chung WS, Kwong PK. Reliability and validity of the Cantonese version of Mini-Mental State Examination: a preliminary study. J Hong Kong Coll Psych 1994;4:25-8.

14. World Medical Association. World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA 2013;310:2191-4. CrossRef

15. Miller JM, Sampselle C, Ashton-Miller J, Hong GR, DeLancey JO. Clarification and confirmation of the Knack maneuver: the effect of volitional pelvic floor muscle contraction to preempt expected stress incontinence. Int Urogynecol J Pelvic Floor Dysfunct 2008;19:773-82. CrossRef

16. Kegel AH. Physiologic therapy for urinary stress incontinence. J Am Med Assoc 1951;146:915-7. CrossRef

17. Tsai CH. The effectiveness of a pelvic floor muscle rehabilitation program in managing urinary tract incontinence among Taiwanese middle age and older women [dissertation]. Pittsburgh: University of Pittsburgh; 2001.

18. Castro RA, Arruda RM, Zanetti MR, Santos PD, Sartori MG, Girao MJ. Single-blind, randomized, controlled trial of pelvic floor muscle training, electrical stimulation, vaginal cones, and no active treatment in the management of stress urinary incontinence. Clinics (Sao Paulo) 2008;63:465-72. CrossRef

19. Zahariou A, Karamouti M, Georgantzis D, Papaioannou P. Are there any UPP changes in women with stress urinary incontinence after pelvic floor muscle exercises? Urol Int 2008;80:270-4. CrossRef

20. Garber CE, Blissmer B, Deschenes MR, et al. American College of Sports Medicine position stand. Quantity and quality of exercise for developing and maintaining cardiorespiratory, musculoskeletal, and neuromotor fitness in apparently healthy adults: guidance for prescribing exercise. Med Sci Sports Exerc 2011;43:1334-59. CrossRef

21. Bø K. Pelvic floor muscle strength and response to pelvic floor muscle training for stress urinary incontinence. Neurourol Urodyn 2003;22:654-8. CrossRef

22. Dannecker C, Wolf V, Raab R, Hepp H, Anthuber C. EMG-biofeedback assisted pelvic floor muscle training is an effective therapy of stress urinary or mixed incontinence: a 7-year experience with 390 patients. Arch Gynecol Obstet 2005;273:93-7. CrossRef

23. Herbison GP, Dean N. Weighted vaginal cones for urinary incontinence. Cochrane Database Syst Rev 2013;(7):CD002114.

24. Hay-Smith EJ, Bø Berghmans LC, Hendriks HJ, de Bie RA, van Waalwijk van Doorn ES. Pelvic floor muscle training for urinary incontinence in women. Cochrane Database Syst Rev 2001;(1):CD001407.

25. Hung HC, Chih SY, Lin HH, Tsauo JY. Exercise adherence to pelvic floor muscle strengthening is not a significant predictor of symptom reduction for women with urinary incontinence. Arch Phys Med Rehabil 2012;93:1795-800. CrossRef

26. Bø K, Hilde G. Does it work in the long term?—A systematic review on pelvic floor muscle training for female stress urinary incontinence. Neurourol Urodyn 2013;32:215-23. CrossRef

Cancer patients receiving chemotherapy sufficient to cause myelosuppression and adverse effects on the integrity of gastro-intestinal mucosa are at high risk of invasive infections. Patients with profound, prolonged neutropenia are at particularly high risk of serious infections. Prolonged neutropenia is most likely to occur in patients undergoing induction chemotherapy for acute leukaemia. More than 80% of those with haematological malignancies will develop fever during more than one chemotherapy cycle.1 Since neutropenic patients are unable to mount a strong inflammatory response to infections, fever may be the only sign. Infection in neutropenic patients can progress rapidly, leading to serious complications and even death with a mortality rate ranging from 2% to 21%.2 3 It is critical to recognise neutropenic fever patients early and initiate empirical, broad-spectrum antibiotics. Major international guidelines advocate early administration of empirical antibiotics within 1 hour of emergency department (ED) presentation, sometimes even without cytological proof of neutropenia.4 5 6 7 However, management of febrile neutropenic patients varies across different EDs, and even among different physicians. A recent audit performed in the EDs of the United Kingdom showed that only 26% of the audited patients received intravenous antibiotics within the target time of 1 hour.8 Another study in French EDs showed that management of febrile neutropenia was inadequate and severity was under-evaluated in the critically ill.9 In order to improve and standardise the care of post-chemotherapy sepsis in haematological malignancy patients, the Accident and Emergency Department (A&E) and Department of Medicine of Queen Elizabeth Hospital (QEH) initiated a treatment protocol in 2011. This is the first hospital in Hong Kong to implement such a treatment protocol. It included febrile patients with haematological malignancy who had received chemotherapy within 1 month of ED visit. These patients were identified at triage station and provided with a fast-track consultation. The ED physician would verify the history and perform a thorough physical examination and targeted investigations. Empirical antibiotics were administered after taking appropriate culture samples aiming at a door-to-antibiotic (DTA) time of less than 1 hour (Fig).

Local publication on post-chemotherapy patients mainly focused on solid tumour patients, in-patient management and their outcomes.10 There is a paucity of literature concerning the initial ED management of haematological malignancy patients. The objective of this study was to examine the protocol compliance rate among ED physicians, the DTA time before and after implementation of the protocol, and the impact of the protocol on A&E and hospital services. It also serves to provide invaluable epidemiological data regarding the haematological malignancy patients in Hong Kong.

This is a before-and-after study of the impact of a protocol on the management of post-chemotherapy sepsis in haematological malignancy patients. A 2-year retrospective chart review was conducted. The first chart review was performed from June 2010 to May 2011. These patients were admitted through ED to the haematological ward prior to implementation of the protocol and served as historical referents. Data were retrieved from the admission book of the haematology ward. A diagnosis of post-chemotherapy fever or neutropenic fever was shortlisted. Cases that were admitted through ED were analysed. The second year started from June 2011. The intervention group included patients recruited in the protocol. There were two patients who fulfilled the inclusion criteria but were excluded from the study since they refused any investigation or treatment in ED despite explanation. The charts were reviewed by two emergency physicians and two senior nurses. Any discrepancy was resolved by discussion among investigators. The protocol was implemented on a 24-hour basis. According to the protocol, fever was defined by a single measurement of oral temperature of >38.3°C either at the triage station or self-reported at home. Neutropenia was defined as absolute neutrophil count (ANC) of <1 x 10^-9 /L. Sepsis was defined by Bone criteria11 (ie >2 out of 4 of the following: leukocyte count <4 or >12 x 10^-9 /L, respiratory rate >20/min, oral temperature >38°C or <35°C, pulse >90 beats/min). Door-to-antibiotic time was charted in the medical record. Lengths of stay in the A&E and hospital were retrieved from the Clinical Data Analysis and Reporting System. The primary outcome was mean DTA time. Secondary outcomes included compliance of the ED physician with the protocol, mean ED length of stay, mean hospital length of stay, and the adverse outcome rate. Adverse outcomes included occurrence of a serious medical complication or death during index admission; these criteria are commonly cited in oncology literature.12 Adverse outcome was charted from patients’ medical record during the index admission.

Chi squared tests were performed when comparing categorical parameters between the protocol and referent groups. Student’s t tests were performed for parametric variables. All statistical analyses were performed using the Statistical Package for the Social Sciences (Windows version 17; SPSS Inc, Chicago [IL], US). A P value of less than 0.05 was regarded as statistically significant.

The study was conducted in the A&E of QEH, Hong Kong, a tertiary referral centre for haematological malignancy patients. The A&E of QEH is an urban ED with a daily attendance of 500 and is one of the busiest EDs in Hong Kong. This study was approved by the chief of service of the department.

A total of 69 patients were recruited; 19 patients were referents while 50 belonged to the protocol group. Baseline demographic data are shown in Table 1. Overall, 49% of the patients were male. Their mean age was 56 years (range, 20-81 years). Leukaemia was the most commonly encountered haematological malignancy, accounting for 51% of cases (n=35/69). Among these, acute myeloid leukaemia was the most prevalent subtype. Lymphoma was the second most common haematological malignancy, making up 42% (n=29/69) of the cases. The mean duration of the last chemotherapy dose to ED visit was 12 days in both groups of patients. At least one co-morbidity was present in 47% of patients in the referent group and in 52% of patients in the protocol group (P=0.18).

During the index ED visit, the mean door-to-consultation time was 15 and 12 minutes in referent group and protocol group, respectively (P=0.40). Overall, 88% (n=16/19 in referent and n=45/50 in protocol group) of the patients fulfilled the sepsis criteria; 64% (n=44/69) had ANC of <1 x 10^-9 /L, although the result was not known at the time of consultation. All protocol group patients received antibiotics after blood cultures were taken during their ED stay compared to none in the control group. Tazobactam-piperacillin (Tazocin; Pfizer, Taiwan) was the most commonly prescribed antibiotic in ED. The mean DTA time in the protocol group was 47 minutes compared to 300 minutes in referent group (P<0.05). Overall, 86% (n=43/50) of protocol group patients could achieve the target DTA time of less than 1 hour (P<0.05). The shortest time required for antibiotic administration in the referent group was 70 minutes. The mean length of stay in ED was 105 minutes in the referent group versus 76 minutes in the protocol group (P=0.46). The major outcomes are shown in Table 2.

The duration of fever, which was defined as oral temperature of >38°C for 24 hours, was 4 days in the referent group and 3 days in the protocol group (P=0.09). One patient from the referent group suffered from septic shock and required intensive care unit (ICU) admission; no patient from this group died. Six patients in the protocol group had adverse outcomes; three had septic shock requiring inotropic support, one of them required ICU admission, while three patients died during index admission. Adverse event rate was 5% in the referent group versus 14% in the protocol group (P=0.45). Overall, 25% (n=17/69) of patients had bacteraemia. Escherichia coli was recovered in five samples of which two were extended-spectrum beta-lactamase (ESBL)–producing bacteria. Streptococcus mitis was the second most common pathogen and was found in four samples. Overall, 43% (n=30/69) of the patients had microbiologically documented infection. The mean length of hospital stay was 15 days in the referent group compared with 11 days in the protocol group (P=0.15).

Chemotherapy-induced sepsis is a medical emergency that requires urgent assessment and treatment with antibiotics. Our study shows that 88% of post-chemotherapy febrile patients fulfilled the sepsis criteria. Overall, 25% of patients had bacteraemia, a rate similar to that reported in the literature.4 Hence, prompt identification and early administration of broad-spectrum empirical antibiotics is the cornerstone of management. In a retrospective study of 2731 patients with septic shock (only 7% of whom were neutropenic), each hour delay in initiating effective antimicrobials decreased survival by around 8%.13 Another cohort study showed that the in-hospital mortality among adult patients with severe sepsis or septic shock decreased from 33% to 20% when time from triage to appropriate antimicrobial therapy was ≤1 hour compared with >1 hour.14 Our protocol suggested Tazocin as the first-line antibiotic, in accordance with the 2010 Infectious Diseases Society of America guideline.4 However, the rising trend of ESBL E coli infection may raise concern of antibiotic resistance. A larger-scale cohort study should be carried out to update the local microbiology prevalence and amend the empirical antibiotic recommendations accordingly.

Implementation of the protocol in our department could significantly reduce the mean DTA time from 300 minutes to 47 minutes (P<0.05). Furthermore, 86% of patients could achieve the target DTA time of <1 hour. The result was satisfactory when compared with similar studies conducted in Europe and North America where reported median DTA ranged from 154 minutes to 3.9 hours.15 16 17 Audits from the UK report that only 18% to 26% of patients receive initial antibiotic within the target DTA of 1 hour.8 According to the authors, the most common reasons for failure to comply with this time frame included failure to administer the initial dose of the empirical antibacterial regimen until the patient has been transferred from ED to the inpatient ward, prolonged time between arrival and clinical assessment, lack of awareness of the natural history of neutropenic fever syndrome and its evolution to severe sepsis and shock, failure of the ED to stock appropriate antibacterial medications, and non-availability of neutropenic fever protocols in the ED for quick reference.8 The last two points were further supported by studies. A chart review of 201 febrile neutropenic patients in Canada showed that the electronic clinical practice guideline could decrease the DTA time by 1 hour (3.9 hours vs 4.9 hours).16 Another retrospective observational study of timeliness of antibiotic administration in severely septic patients presenting to a US community ED showed that storing key antibiotics could decrease the mean DTA time by 70 minutes (167 minutes vs 97 minutes).18 The percentage of severely septic patients receiving antibiotics within 3 hours of arrival to the ED increased from 65% pre-intervention to 93% post-intervention.18 Before the implementation of this protocol, multiple briefing sessions were held with nurses and physicians to increase awareness about prompt treatment of post-chemotherapy fever. Antibiotics were stocked in the ED and were readily available. The protocol could be easily downloaded from the department website. Regular collaboration existed between the nursing manager and the pharmacist to replenish the antibiotic stock. Thus, successful implementation of the protocol involved a joint effort by different parties.

There was a trend towards reducing the duration of fever and length of hospital stay in the intervention group. Although this does not imply causation, especially in view of the small sample size, the correlation makes one ponder whether a delay in antibiotic delivery indeed increases the length of hospital stay. Similar correlation was demonstrated in a UK review.15 However, we could not demonstrate an impact on mortality and adverse outcome. The reason may partly be related to the small sample size, heterogeneous nature of haematological malignancies, and overall low incidence of mortality (4%) in our study as compared to 49.8% in-hospital mortality rate reported in an 11-year review.19 Our result shows that the length of ED stay was similar between control and intervention groups, thus, demonstrating that this protocol did not add further burden to the overcrowding ED.

This study has two limitations that need to be discussed. First, the study used a retrospective chart review design and there were inherent challenges with missing information and poor documentation. Second, prior to implementation of the protocol, the febrile haematological malignant patients were often instructed to either attend the day ward or A&E; this partly explains the relatively small case number in the control group. In addition, we relied on the diagnosis coding for case identification in the control group; some cases might have been missed as a result of error in coding. Even if they attended ED, the lack of awareness and reluctance of physicians to prescribe antibiotics led to a significant delay in administration of the first dose of antibiotic. Although the number of control cases was small, mean DTA time of 300 minutes echoed the same in a similar study performed overseas.16 Efforts were made to use accepted chart review methods to assess outcomes that were automatically recorded in the electronic ED information systems, and to examine the nurse records of antibiotic administration.

Implementation of a treatment protocol in post-chemotherapy febrile haematological malignancy patients can significantly shorten the mean DTA time to <1 hour, which is now the standard of care worldwide. The key to effective implementation lies in orchestration of efforts between administrators, physicians, nurses, and pharmacists. We can prove that the protocol is feasible even in a busy urban ED.

We would like to thank Ms Kelly Choy for statistical analysis.

No conflicts of interests were declared by authors.

1. Klastersky J. Management of fever in neutropenic patients with different risks of complications. Clin Infect Dis 2004;39 Suppl 1:S32-7. CrossRef

2. Smith TJ, Khatcheressian J, Lyman GH, et al. 2006 Update of recommendations for the use of white cell growth factors: an evidence-based clinical practice guideline. J Clin Oncol 2006;24:3187-205. CrossRef

3. Herbst C, Naumann F, Kruse EB, et al. Prophylactic antibiotics or G-CSF for the prevention of infections and improvement of survival in cancer patients undergoing chemotherapy. Cochrane Database Syst Rev 2009;(1):CD007107.

4. Freifeld AG, Bow EJ, Sepkowitz KA, et al. Clinical practice guideline for the use of antimicrobial agents in neutropenic patients with cancer: 2010 update by the Infectious Diseases Society of America. Clin Infect Dis 2011;52:e56-93. CrossRef

5. Dellinger RP, Levy MM, Rhodes A, et al. Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock: 2012. Crit Care Med 2013;41:580-637. CrossRef

6. Bate J, Gibson F, Johnson E, et al. Neutropenic sepsis: prevention and management of neutropenic sepsis in cancer patients (NICE Clinical Guideline CG151). Arch Dis Child Educ Pract Ed 2013;98:73-5. CrossRef

7. Flowers CR, Seidenfeld J, Bow EJ, et al. Antimicrobial prophylaxis and outpatient management of fever and neutropenia in adults treated for malignancy: American Society of Clinical Oncology clinical practice guideline. J Clin Oncol 2013;31:794-810. CrossRef

8. Clarke RT, Warnick J, Stretton K, Littlewood TJ. Improving the immediate management of neutropenic sepsis in the UK: lessons from a national audit. Br J Haematol 2011;153:773-9. CrossRef

9. Andr&eacute S, Taboulet P, Elie C, et al. Febrile neutropenia in French emergency departments: results of a prospective multicentre survey. Crit Care 2010;14:R68. CrossRef

10. Hui EP, Leung LK, Poon TC, et al. Prediction of outcome in cancer patients with febrile neutropenia: a prospective validation of the Multinational Association for Supportive Care in Cancer risk index in a Chinese population and comparison with the Talcott model and artificial neural network. Support Care Cancer 2011;19:1625-35. CrossRef

11. Bone RC, Balk RA, Cerra FB, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest 1992;101:1644-55. CrossRef

12. Klastersky J, Paesmans M, Rubenstein EB, et al. The Multinational Association for Supportive Care in Cancer Risk Index: a multinational scoring system for identifying low-risk febrile neutropenic cancer patients. J Clin Oncol 2000;18:3038-51.

13. Kumar A, Roberts D, Wood KE, et al. Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock. Crit Care Med 2006;34:1589-96. CrossRef

14. Gaieski DF, Mikkelsen ME, Band RA, et al. Impact of time to antibiotics on survival in patients with severe sepsis or septic shock in whom early goal-directed therapy was initiated in the emergency department. Crit Care Med 2010;38:1045-53. CrossRef

15. Sammut SJ, Mazhar D. Management of febrile neutropenia in an acute oncology service. QJM 2012;105:327-36. CrossRef

16. Lim C, Bawden J, Wing A, et al. Febrile neutropenia in EDs: the role of an electronic clinical practice guideline. Am J Emerg Med 2012;30:5-11, 11.e1-5.

17. Nirenberg A, Mulhearn L, Lin S, Larson E. Emergency department waiting times for patients with cancer with febrile neutropenia: a pilot study. Oncol Nurs Forum 2004;31:711-5. CrossRef

18. Hitti EA, Lewin JJ 3rd, Lopez J, et al. Improving door-to-antibiotic time in severely septic emergency department patients. J Emerg Med 2012;42:462-9. CrossRef

19. Legrand M, Max A, Peigne V, et al. Survival in neutropenic patients with severe sepsis or septic shock. Crit Care Med 2012;40:43-9. CrossRef