Sepsis is a life-threatening condition that contributes to nearly 850 000 emergency department (ED) visits annually in the US.1 According to the practice guidelines published by the Surviving Sepsis Campaign, a care bundle of sepsis treatment—including fluid resuscitation, antimicrobial therapy, and source control—is recommended as life-saving treatment for patients with sepsis.2 Computed tomography (CT) scanning is a popular method for identifying the focus of infection and guiding the implementation of an appropriate antimicrobial strategy in emergency medical care settings.3 The utilisation of CT scans in the ED has increased considerably, such that more than 70 million CT scans are performed in the US annually.4 Approximately one in seven patients undergoes a CT scan during evaluation in the ED.5

Although the use of iodinated contrast media is an important method for improving the diagnostic accuracy of CT examination,6 there are concerns regarding the potential for precipitating renal dysfunction, especially in patients who already have impaired renal function.7 8 The third leading cause of acute kidney injury (AKI) in hospitalised patients is reported to be contrast-associated (CA)-AKI9; CA-AKI is associated with increased risks of major adverse events, including myocardial infarction, renal failure, and mortality.7 10 Nevertheless, it remains controversial whether an association exists between intravenous administration of contrast media during CT scans and the development of CA-AKI.11 12 13 This controversy exists largely because the introduction of refined iso- or low-osmolar contrast agents has reduced the risk of AKI14 and because the majority of previous studies on CA-AKI were performed in patients who underwent coronary angiography,15 16 17 which utilises different dosages and routes of contrast administration relative to those of conventional contrast-enhanced CT scans.18 Previous studies on CA-AKI in an emergency setting have been inconclusive.6 7 19 20 21 22 23 24 Although those studies investigated the benefits and risks of contrast administration in many clinical settings, including acute stroke, pulmonary embolism, and trauma, very few of them evaluated the impact of contrast administration on patients with sepsis. Notably, sepsis remains a leading cause of mortality in critically ill patients2 and CT imaging studies play important roles in both identifying the source of infection and facilitating infection control in patients with sepsis. Therefore, the aim of the current study was to investigate whether intravenous contrast administration in patients with sepsis is associated with an increased risk of AKI and increased incidences of other adverse clinical outcomes.

This retrospective cohort study was conducted at a tertiary referral medical centre with approximately 50 000 ED visits per year. The study population included all adult (age ≥18 years) patients who visited the ED and underwent CT scans (including brain, chest, abdomen or extremities) and serial serum creatinine measurements during their initial ED visits and any follow-ups within 48 to 72 hours from 1 January 2012 to 31 December 2016. Patients with sepsis were identified by principal diagnosis and serum lactate measurement, in accordance with Sepsis-3 guidelines.25 Patients who received haemodialysis, underwent contrast-enhanced CT scan within 3 months, or experienced a cardiac arrest event before ED arrival were excluded from the analysis. This study protocol followed the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines.

Demographic characteristics of the enrolled patients (ie, age and sex) and clinical information (eg, co-morbidities, chronic medications, laboratory results, acute illness, types and dosage of contrast agent, and initial and final diagnoses) were obtained from written medical charts and electronic medical records. Co-morbidities were coded based on International Classification of Diseases, Ninth Edition, Clinical Modification diagnostic codes reported in medical records. In accordance with World Health Organization criteria, anaemia was defined as baseline haematocrit values below 39% and below 36% for men and women, respectively.26 Chronic kidney disease was defined as a baseline estimated glomerular filtration rate (eGFR) <60 mL/min/1.73 m², calculated using the Modification of Diet in Renal Disease equation.27 Baseline renal function was calculated according to each patient’s serum creatinine level at 24 hours before the CT scan. The presence of shock was identified by the need for vasopressors to maintain haemodynamic stability despite adequate fluid administration during ED stay.

We divided the eligible patients for this study into two groups: contrast-enhanced CT and non-enhanced CT; primary and secondary outcomes were recorded and compared between groups. The primary outcome was the incidence of AKI, which was defined as an absolute increase of 0.5 mg/dL or >50% increase in baseline serum creatinine concentration within 48 to 72 hours after CT scan.28 The secondary outcomes included the incidences of emergent dialysis (defined as initiation of dialysis during the hospital stay) and short-term mortality (defined as death within 30 days after CT scan), as well as the difference in length of hospital stay for survivors.

The estimation of sample size was performed with PASS 11 software in accordance with the results of previous studies regarding AKI incidence in patients with sepsis29 and odds ratio (OR) of CA-AKI.30 With a 30% incidence of AKI in patients with sepsis and an OR of 2.7 for CA-AKI, we determined that 109 patients were needed to detect a significant association with probability (power) of 0.8 and Type 1 error of 0.05.

Data are presented as means±standard deviations or medians with 25th to 75th percentiles (ie, interquartile range) for continuous variables, and as numbers (%) for categorical variables. Two-sample t tests and Chi squared tests were used to compare continuous and categorical variables, respectively. A two-tailed P value of <0.05 was considered statistically significant. Propensity score matching was performed to reduce potential selection bias and other confounding factors. We calculated the propensity score for each patient by modelling the probability of receiving contrast medium. Variables in the model were composed of factors that influence outcomes related to renal function or influence the selection of contrast medium. We used total 21 variables including age, sex, co-morbidities (ie, diabetes mellitus, hypertension, liver cirrhosis, coronary artery disease, left heart failure, chronic kidney disease, anaemia, chronic obstructive pulmonary disease, dyslipidaemia, and malignancy), nephrotoxic medications (ie, statins, non-steroidal anti-inflammatory drugs, angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers, nephrotoxic antibiotics such as aminoglycosides and vancomycin), laboratory data (ie, initial serum creatinine, eGFR, and serum lactate), measures of illness severity (ie, initial presence of septic shock and need for intensive care unit [ICU] admission) to calculate the propensity scores for all patients. A multivariable logistic regression analysis model using nearest-neighbour matching, calliper 0.1, was generated to predict the probability of receiving contrast medium. We used the resulting propensity scores to match the contrast-enhanced CT group members with non-enhanced CT group members at a ratio of 1:1. Patients without a corresponding match were excluded. All statistical analyses were performed using SPSS (Windows version 22.0; IBM Corp, Armonk [NY], US).

During the study period, 200 427 adult patients visited the ED; of these, 712 met the criteria for inclusion in this study. After further exclusion of patients with elevated serum lactate levels from shock with non-septic aetiology, the remaining 587 patients (enhanced CT group: 105; non-enhanced CT group: 482) were analysed (Fig). In the contrast-enhanced CT group, 23 patients received intravenous iopromide (Ultravist 370; Bayer Parma AG, Berlin, Germany) and 82 patients received intravenous iohexol (Omnipaque; Bayer Parma AG, Berlin, Germany). Only one patient received a contrast volume >100 mL (120 mL).

Prior to propensity score matching, patients in the contrast-enhanced CT group were significantly younger; moreover, they had lower prevalences of hypertension, chronic kidney disease, and chronic obstructive pulmonary disease, compared to patients in the non-enhanced CT group. Patients in the enhanced CT group also had significantly lower initial and follow-up serum creatinine levels, and had higher initial serum lactate levels than those in the non-enhanced CT group. There were no significant differences in the incidences of shock and ICU admission between the two groups (Table 1).

By using propensity score with 1:1 matching, 101 patients with sepsis in the contrast-enhanced CT group were successfully paired with an equal number of patients in the non-enhanced CT group. After matching, there were no statistically significant differences between the two groups in any covariates (Table 2).

Before propensity score matching, the risks of AKI, emergent dialysis, and short-term mortality were not significantly greater in the contrast-enhanced CT group than in the non-enhanced CT group. Five of 44 patients with sepsis in the non-enhanced CT group who received emergency haemodialysis subsequently required chronic dialysis; however, no patients required chronic dialysis in the contrast-enhanced CT group. Furthermore, there was no significant difference in the length of hospital stay between the two groups (Table 3). The same results were observed after propensity score matching: there were no notable differences in the risks of AKI, emergent dialysis, or short-term mortality; the median length of hospital stay was also similar between the matched contrast-enhanced and non-enhanced CT groups (Table 4).

In this ED-based single-centre retrospective study, we performed a subgroup analysis to investigate the possible adverse clinical impacts of contrast agent administration in patients with sepsis. By using propensity score matching, we demonstrated that intravenous administration of contrast media in patients with sepsis was not associated with increased risks of AKI or other adverse outcomes, following contrast-enhanced CT scans to identify foci of infection. During revision of this manuscript, Hinson et al31 reported a retrospective cohort study; they concluded that contrast medium administration was not associated with increased incidence of AKI in patients with sepsis, consistent with our findings. Compared with the study by Hinson et al, the patients in our study had more severe sepsis (ie, higher incidences of shock and ICU admission); moreover, our findings revealed that administration of reasonable volumes of contrast medium did not increase the risks of emergency dialysis or short-term mortality. Thus, clinicians can use contrast-enhanced CT scans in a reasonable manner in septic patients, in order to identify occult infection foci earlier and facilitate prompt medical management.

Our patients had surprisingly high prevalences of hypertension, diabetes mellitus, and chronic kidney disease, which could have been related to their older age, as reported in a prior study.32 Before propensity score matching, patients in the contrast-enhanced CT group were significantly younger and had fewer co-morbidities, including hypertension, chronic kidney disease, and chronic obstructive pulmonary disease. Moreover, patients in the contrast-enhanced CT group had lower initial serum creatinine levels and higher eGFRs, as observed in other studies.6 32 This could be related to the common clinical practice of using contrast-enhanced CT for younger patients with few co-morbidities and relatively good renal function, based on considerations of the potential nephrotoxicities of the contrast agents7; a few patients with poor renal function (24 of 482 patients with sepsis in the non-enhanced CT group) may also have avoided contrast agents following an explanation of the potential for nephrotoxicity. Clinicians may have hesitated to administer contrast media to patients with respiratory disease because of the risk of immediate hypersensitivity reaction; however, asthma and chronic obstructive pulmonary disease have not been established as consistent risk factors for contrast media-related adverse drug reactions.33 The lack of significant differences in risks of AKI, emergent dialysis, and short-term mortality between the non-enhanced and enhanced CT groups before propensity score matching in our study may have been influenced by the above-mentioned tendency for clinicians to perform contrast-enhanced CT in presumably healthier patients. However, it is difficult to evaluate the causal relationship between administration of contrast agents and risk of AKI in patients with sepsis by comparing two patient groups with many different demographic and characteristics; therefore, we used propensity score matching to minimise the impacts of potential confounders.

Although the mean initial serum lactate level in the contrast-enhanced CT group was slightly but significantly higher than that in the non-enhanced CT group, this difference was not correlated with the incidences of acute illness (eg, shock), ICU admission, and short-term mortality between the two groups. In addition, the levels of renal function, reflected by serum creatinine levels and eGFRs within 48 to 72 hours after CT scans, improved relative to initial measurements in both groups. These seemingly paradoxical findings suggested that sepsis, rather than the administration of contrast media, was the determinant of clinical outcomes in the present study.

Previous studies in emergency medical settings have shown wide variation in the incidence of post-contrast AKI (3.2%-12%); this may be partially explained by the variety of diseases encountered in the ED, as well as differences in the definitions of AKI adopted in each study.6 7 19 20 21 22 23 24 31 Nearly half of the patients (49%) in the present study experienced septic shock; thus, the increased incidence of post-contrast AKI in our patients (12.4%), compared with that observed in prior studies, may be attributed to the impaired physical status of our patients. This may also explain the considerably higher rates of emergent dialysis and short-term mortality, as well as the increased median length of hospital stay for survivors among our patients, compared to those parameters measured in other studies that did not focus on patients with sepsis.21 34

Thus far, the pathophysiology of CA-AKI remains poorly characterised. Based on the results of some animal studies, proposed mechanisms include acute tubular necrosis caused by medullary hypoxia from vasoconstriction, as well as direct cytotoxic effects of the contrast agent on renal tubular cells.35 36 Compared with AKI caused by other aetiologies, CA-AKI involves relatively rapid recovery of renal function; this is potentially because of the reduced extent of tubular necrosis, which leads to minor and transient functional impairment of tubular epithelial cells.37 Nevertheless, sepsis is the leading cause of AKI in critically ill patients and is associated with a higher mortality rate among patients in the ICU, compared with patients who have AKI caused by other aetiologies.38 Therefore, hesitation to perform contrast-enhanced CT scans for patients with sepsis, in order to identify occult infection foci, could result in delayed diagnoses of life-threatening conditions that carry considerable risks of morbidity and mortality, even in patients with serum creatinine up to 4.0 mg/dL.6

A number of studies performed in the past several years have been designed to maintain a balance between the benefits and adverse effects of contrast-enhanced CT scans in many clinical settings.6 7 12 20 21 22 The vast majority of those studies showed no significant association between the use of contrast agents and an increased risk of AKI. Consistent with the prior findings, contrast-enhanced CT scans of our patients with sepsis were not associated with increased risks of AKI and other adverse clinical outcomes. Among all aetiologies of AKI in patients requiring emergent medical attention, such as sepsis, dehydration, and nephrotoxic medication use,18 the contribution of CA-AKI is regarded as considerably less important37; notably, our findings support this view. Furthermore, it has been consistently shown that the performance of a contrast-enhanced CT scan is justified in patients for whom the examination is indicated, provided that other risk factors of AKI are well controlled.39

There are several limitations in our study, largely in relation to its single-centre and retrospective design. First, the non-enhanced CT group consisted of older patients with a higher prevalence of hypertension and worse renal function; this suggested a selection bias. Although we routinely checked serum lactate for patients with suspected sepsis in the ED, there were a few patients diagnosed with sepsis who did not have lactate measurement data; this may also have resulted in selection bias. Second, although propensity score matching was used to minimise the impacts of potential confounders, unmeasured confounding variables remained, leading to potentially biased results. Therefore, further large-scale cohort or well-controlled prospective randomised studies are warranted. Finally, the definition of AKI used in this study (elevation of serum creatinine concentration by 0.5 mg/dL or by 50% increase relative to baseline within 48 to 72 hours after contrast administration) may not accurately reflect the clinical condition because the relationship between increases in serum creatinine level and deterioration of renal function is reportedly non-linear.40

Our study demonstrated that the intravenous administration of contrast media during CT scans was not associated with increased risks of AKI, emergent dialysis, or short-term mortality for patients with sepsis in the ED; moreover, the use of contrast-enhanced CT was not associated with prolonged length of hospital stay in these patients. The lack of a significant correlation between the administration of contrast agents and the risk of AKI in patients with sepsis conflicts with the tendency to withhold contrast-enhanced CT for the diagnostic assessment and management of sepsis in the emergency setting. Further studies are necessary to confirm these findings and provide further guidance for clinical practice.

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

Concept or design: YC Hsu.
Acquisition of data: HY Su, CW Hsu, CY Liang.
Analysis or interpretation of data: TB Chen.
Drafting of the article: YC Hsu.
Critical revision for important intellectual content: CK Sun, CW Hsu.

Dr Chi-feng Hsieh is acknowledged for providing technical support in sample size calculation.

All authors have disclosed no conflicts of interest.

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

The study was approved by the Institutional Review Board of E-Da hospital (EMRP-106-037) and the requirement for informed patient consent was waived because of the retrospective observational nature of the study.

1. Wang HE, Jones AR, Donnelly JP. Revised national estimates of emergency department visits for sepsis in the United States. Crit Care Med 2017;45:1443-9. Crossref

2. Rhodes A, Evans LE, Alhazzani W, et al. Surviving Sepsis Campaign: international guidelines for management of sepsis and septic shock: 2016. Intensive Care Med 2017;43:304-77. Crossref

3. Jimenez MF, Marshall JC; International Sepsis Forum. Source control in the management of sepsis. Intensive Care Med 2001;27 Suppl 1:S49-62. Crossref

4. Berrington de González A, Mahesh M, Kim KP, et al. Projected cancer risks from computed tomographic scans performed in the United States in 2007. Arch Intern Med 2009;169:2071-7. Crossref

5. Kocher KE, Meurer WJ, Fazel R, Scott PA, Krumholz HM, Nallamothu BK. National trends in use of computed tomography in the emergency department. Ann Emerg Med 2011;58:452-62.e3. Crossref

6. Hinson JS, Ehmann MR, Fine DM, et al. Risk of acute kidney injury after intravenous contrast media administration. Ann Emerg Med 2017;69:577-86.e4. Crossref

7. Mitchell AM, Kline JA, Jones AE, Tumlin JA. Major adverse events one year after acute kidney injury after contrast-enhanced computed tomography. Ann Emerg Med 2015;66:267-74.e4. Crossref

8. McCullough PA, Adam A, Becker CR, et al. Epidemiology and prognostic implications of contrast-induced nephropathy. Am J Cardiol 2006;98:5K-13K. Crossref

9. Nash K, Hafeez A, Hou S. Hospital-acquired renal insufficiency. Am J Kidney Dis 2002;39:930-6. Crossref

10. Coca SG, Peixoto AJ, Garg AX, Krumholz HM, Parikh CR. The prognostic importance of a small acute decrement in kidney function in hospitalized patients: a systematic review and meta-analysis. Am J Kidney Dis 2007;50:712-20. Crossref

11. Katzberg RW, Newhouse JH. Intravenous contrast medium–induced nephrotoxicity: is the medical risk really as great as we have come to believe? Radiology 2010;256:21-8. Crossref

12. McDonald JS, McDonald RJ, Comin J, et al. Frequency of acute kidney injury following intravenous contrast medium administration: a systematic review and meta-analysis. Radiology 2013;267:119-28. Crossref

13. Kashani K, Levin A, Schetz M. Contrast-associated acute kidney injury is a myth: We are not sure. Intensive Care Med 2018;44:110-4. Crossref

14. Lameire N, Kellum JA; KDIGO AKI Guideline Work Group. Contrast-induced acute kidney injury and renal support for acute kidney injury: a KDIGO summary (Part 2). Crit Care 2013;17:205. Crossref

15. Mehran R, Aymong ED, Nikolsky E, et al. A simple risk score for prediction of contrast-induced nephropathy after percutaneous coronary intervention: development and initial validation. J Am Coll Cardiol 2004;44:1393-9. Crossref

16. Mehran R, Nikolsky E. Contrast-induced nephropathy: definition, epidemiology, and patients at risk. Kidney Int Suppl 2006;(100):S11-5. Crossref

17. Rihal CS, Textor SC, Grill DE, et al. Incidence and prognostic importance of acute renal failure after percutaneous coronary intervention. Circulation 2002;105:2259-64. Crossref

18. Aycock RD, Westafer LM, Boxen JL, Majlesi N, Schoenfeld EM, Bannuru RR. Acute kidney injury after computed tomography: a meta-analysis. Ann Emerg Med 2018;71:44- 53.e4. Crossref

19. Mitchell AM, Kline JA. Contrast nephropathy following computed tomography angiography of the chest for pulmonary embolism in the emergency department. J Thromb Haemost 2007;5:50-4. Crossref

20. Sonhaye L, Kolou B, Tchaou M, et al. Intravenous contrast medium administration for computed tomography scan in emergency: a possible cause of contrast-induced nephropathy. Radiol Res Pract 2015;2015:805786. Crossref

21. Heller M, Krieger P, Finefrock D, Nguyen T, Akhtar S. Contrast CT scans in the emergency department do not increase risk of adverse renal outcomes. West J Emerg Med 2016;17:404-8. Crossref

22. Ehrlich ME, Turner HL, Currie LJ, Wintermark M, Worrall BB, Southerland AM. Safety of computed tomographic angiography in the evaluation of patients with acute stroke: a single-center experience. Stroke 2016;47:2045-50. Crossref

23. Lima FO, Lev MH, Levy RA, et al. Functional contrast-enhanced CT for evaluation of acute ischemic stroke does not increase the risk of contrast-induced nephropathy. AJNR Am J Neuroradiol 2010;31:817-21. Crossref

24. Tremblay LN, Tien H, Hamilton P, et al. Risk and benefit of intravenous contrast in trauma patients with an elevated serum creatinine. J Trauma 2005;59:1162-6. Crossref

25. Singer M, Deutschman CS, Seymour CW, et al. The third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA 2016;315:801-10. Crossref

26. World Health Organ Tech Rep Ser. Nutritional anaemias: report of a WHO scientific group. World Health Organ Tech Rep Ser 1968;405:5-37.

27. National Kidney Foundation. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Am J Kidney Dis 2002;39:S1-266.

28. Mehta RL, Kellum JA, Shah SV, et al. Acute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury. Crit Care 2007;11:R31. Crossref

29. Bagshaw SM, George C, Bellomo R; ANZICS Database Management Committee. Early acute kidney injury and sepsis: a multicentre evaluation. Crit Care 2008;12:R47. Crossref

30. Song W, Zhang T, Pu J, Shen L, He B. Incidence and risk of developing contrast-induced acute kidney injury following intravascular contrast administration in elderly patients. Clin Interv Aging 2014;9:85-93. Crossref

31. Hinson JS, Al Jalbout N, Ehmann MR, Klein EY. Acute kidney injury following contrast media administration in the septic patient: A retrospective propensity-matched analysis. J Crit Care 2019;51:111-6. Crossref

32. McDonald, JS, McDonald RJ, Williamson EE, Kallmes DF, Kashani K. Post-contrast acute kidney injury in intensive care unit patients: a propensity score-adjusted study. Intensive Care Med 2017;43:774-84. Crossref

33. Bettmann MA, Heeren T, Greenfield A, Goudey C. Adverse events with radiographic contrast agents: results of the SCVIR Contrast Agent Registry. Radiology 1997;203:611-20. Crossref

34. McDonald RJ, McDonald JS, Carter RE, et al. Intravenous contrast material exposure is not an independent risk factor for dialysis or mortality. Radiology 2014;273:714-25. Crossref

35. Persson PB, Hansell P, Liss P. Pathophysiology of contrast medium-induced nephropathy. Kidney Int 2005;68:14-22. Crossref

36. Heyman SN, Rosenberger C, Rosen S. Regional alterations in renal haemodynamics and oxygenation: a role in contrast medium-induced nephropathy. Nephrol Dial Transplant 2005;20 Suppl 1:i6-11. Crossref

37. Molitoris BA, Dahl R, Geerdes A. Cytoskeleton disruption and apical redistribution of proximal tubule Na(+)-K(+)-ATPase during ischemia. Am J Physiol 1992;263:F488-95. Crossref

38. Bellomo R, Kellum JA, Ronco C, et al. Acute kidney injury in sepsis. Intensive Care Med 2017;43:816-28. Crossref

39. Petek BJ, Bravo PE, Kim F, et al. Incidence and risk factors for postcontrast acute kidney injury in survivors of sudden cardiac arrest. Ann Emerg Med 2016;67:469-76.e1. Crossref

40. Ostermann M, Joannidis M. Acute kidney injury 2016: diagnosis and diagnostic workup. Crit Care 2016;20:299. Crossref

Sepsis is a life-threatening condition that contributes to nearly 850 000 emergency department (ED) visits annually in the US.1 According to the practice guidelines published by the Surviving Sepsis Campaign, a care bundle of sepsis treatment—including fluid resuscitation, antimicrobial therapy, and source control—is recommended as life-saving treatment for patients with sepsis.2 Computed tomography (CT) scanning is a popular method for identifying the focus of infection and guiding the implementation of an appropriate antimicrobial strategy in emergency medical care settings.3 The utilisation of CT scans in the ED has increased considerably, such that more than 70 million CT scans are performed in the US annually.4 Approximately one in seven patients undergoes a CT scan during evaluation in the ED.5

Although the use of iodinated contrast media is an important method for improving the diagnostic accuracy of CT examination,6 there are concerns regarding the potential for precipitating renal dysfunction, especially in patients who already have impaired renal function.7 8 The third leading cause of acute kidney injury (AKI) in hospitalised patients is reported to be contrast-associated (CA)-AKI9; CA-AKI is associated with increased risks of major adverse events, including myocardial infarction, renal failure, and mortality.7 10 Nevertheless, it remains controversial whether an association exists between intravenous administration of contrast media during CT scans and the development of CA-AKI.11 12 13 This controversy exists largely because the introduction of refined iso- or low-osmolar contrast agents has reduced the risk of AKI14 and because the majority of previous studies on CA-AKI were performed in patients who underwent coronary angiography,15 16 17 which utilises different dosages and routes of contrast administration relative to those of conventional contrast-enhanced CT scans.18 Previous studies on CA-AKI in an emergency setting have been inconclusive.6 7 19 20 21 22 23 24 Although those studies investigated the benefits and risks of contrast administration in many clinical settings, including acute stroke, pulmonary embolism, and trauma, very few of them evaluated the impact of contrast administration on patients with sepsis. Notably, sepsis remains a leading cause of mortality in critically ill patients2 and CT imaging studies play important roles in both identifying the source of infection and facilitating infection control in patients with sepsis; therefore, the current study aimed to investigate whether intravenous contrast administration in patients with sepsis was associated with an increased risk of AKI and increased incidences of other adverse clinical outcomes.

This retrospective cohort study was conducted at a tertiary referral medical centre with approximately 50 000 ED visits per year. The study population included all adult (age ≥18 years) patients who visited the ED and underwent CT scans (including brain, chest, abdomen or extremities) and serial serum creatinine measurements during their initial ED visits and any follow-ups within 48 to 72 hours from 1 January 2012 to 31 December 2016. Patients with sepsis were identified by principal diagnosis and serum lactate measurement, in accordance with Sepsis-3 guidelines.25 Patients who received haemodialysis, underwent contrast-enhanced CT scan within 3 months, or experienced a cardiac arrest event before ED arrival were excluded from the analysis. This study protocol followed the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines.

Demographic characteristics of the enrolled patients (ie, age and sex) and clinical information (eg, co-morbidities, chronic medications, laboratory results, acute illness, types and dosage of contrast agent, and initial and final diagnoses) were obtained from written medical charts and electronic medical records. Co-morbidities were coded based on International Classification of Diseases, Ninth Edition, Clinical Modification diagnostic codes reported in medical records. In accordance with World Health Organization criteria, anaemia was defined as baseline haematocrit values below 39% and below 36% for men and women, respectively.26 Chronic kidney disease was defined as a baseline estimated glomerular filtration rate (eGFR) <60 mL/min/1.73 m², calculated using the Modification of Diet in Renal Disease equation.27 Baseline renal function was calculated according to each patient’s serum creatinine level at 24 hours before the CT scan. The presence of shock was identified by the need for vasopressors to maintain hemodynamic stability despite adequate fluid administration during ED stay.

We divided the eligible patients for this study into two groups: contrast-enhanced CT and non-enhanced CT; primary and secondary outcomes were recorded and compared between groups. The primary outcome was the incidence of AKI, which was defined as an absolute increase of 0.5 mg/dL or >50% increase in baseline serum creatinine concentration within 48 to 72 hours after CT scan.28 The secondary outcomes included the incidences of emergent dialysis (defined as initiation of dialysis during the hospital stay) and short-term mortality (defined as death within 30 days after CT scan), as well as the difference in length of hospital stay for survivors.

The estimation of sample size was performed with PASS 11 software in accordance with the results of previous studies regarding AKI incidence in patients with sepsis29 and odds ratio (OR) of CA-AKI.30 With a 30% incidence of AKI in patients with sepsis and an OR of 2.7 for CA-AKI, we determined that 109 patients were needed to detect a significant association with probability (power) of 0.8 and Type 1 error of 0.05.

Data are presented as means±standard deviations or medians with 25th to 75th percentiles (ie, interquartile range) for continuous variables, and as numbers (%) for categorical variables. Two-sample t tests and Chi squared tests were used to compare continuous and categorical variables, respectively. A two-tailed P value of <0.05 was considered statistically significant. Propensity score matching was performed to reduce potential selection bias and other confounding factors. We calculated the propensity score for each patient by modelling the probability of receiving contrast medium. Variables in the model were composed of factors that influence outcomes related to renal function or influence the selection of contrast medium. We used total 21 variables including age, sex, co-morbidities (ie, diabetes mellitus, hypertension, liver cirrhosis, coronary artery disease, left heart failure, chronic kidney disease, anaemia, chronic obstructive pulmonary disease, dyslipidaemia, and malignancy), nephrotoxic medications (ie, statins, non-steroidal anti-inflammatory drugs, angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers, nephrotoxic antibiotics such as aminoglycosides and vancomycin), laboratory data (ie, initial serum creatinine, eGFR, and serum lactate), measures of illness severity (ie, initial presence of septic shock and need for intensive care unit [ICU] admission) to calculate the propensity scores for all patients. A multivariable logistic regression analysis model using nearest-neighbour matching, calliper 0.1, was generated to predict the probability of receiving contrast medium. We used the resulting propensity scores to match the contrast-enhanced CT group members with non-enhanced CT group members at a ratio of 1:1. Patients without a corresponding match were excluded. All statistical analyses were performed using SPSS (Windows version 22.0; IBM Corp, Armonk [NY], US).

During the study period, 200 427 adult patients visited the ED; of these, 712 met the criteria for inclusion in this study. After further exclusion of patients with elevated serum lactate levels from shock with non-septic aetiology, the remaining 587 patients (enhanced CT group: 105; non-enhanced CT group: 482) were analysed (Fig). In the contrast-enhanced CT group, 23 patients received intravenous iopromide (Ultravist 370; Bayer Parma AG, Berlin, Germany) and 82 patients received intravenous iohexol (Omnipaque; Bayer Parma AG, Berlin, Germany). Only one patient received a contrast volume >100 mL (120 mL).

Prior to propensity score matching, patients in the contrast-enhanced CT group were significantly younger; moreover, they had lower prevalences of hypertension, chronic kidney disease, and chronic obstructive pulmonary disease, compared to patients in the non-enhanced CT group. Patients in the enhanced CT group also had significantly lower initial and follow-up serum creatinine levels, and had higher initial serum lactate levels than those in the non-enhanced CT group. There were no significant differences in the incidences of shock and ICU admission between the two groups (Table 1).

By using propensity score with 1:1 matching, 101 patients with sepsis in the contrast-enhanced CT group were successfully paired with an equal number of patients in the non-enhanced CT group. After matching, there were no statistically significant differences between the two groups in any covariates (Table 2).

Before propensity score matching, the risks of AKI, emergent dialysis, and short-term mortality were not significantly greater in the contrast-enhanced CT group than in the non-enhanced CT group. Five of 44 patients with sepsis in the non-enhanced CT group who received emergency haemodialysis subsequently required chronic dialysis; however, no patients required chronic dialysis in the contrast-enhanced CT group. Furthermore, there was no significant difference in the length of hospital stay between the two groups (Table 3). The same results were observed after propensity score matching: there were no notable differences in the risks of AKI, emergent dialysis, or short-term mortality; the median length of hospital stay was also similar between the matched contrast-enhanced and non-enhanced CT groups (Table 4).

In this ED-based single-centre retrospective study, we performed a subgroup analysis to investigate the possible adverse clinical impacts of contrast agent administration in patients with sepsis. By using propensity score matching, we demonstrated that intravenous administration of contrast media in patients with sepsis was not associated with increased risks of AKI or other adverse outcomes, following contrast-enhanced CT scans to identify foci of infection. During revision of this manuscript, Hinson et al31 reported a retrospective cohort study; they concluded that contrast medium administration was not associated with increased incidence of AKI in patients with sepsis, consistent with our findings. Compared with the study by Hinson et al, the patients in our study had more severe sepsis (ie, higher incidences of shock and ICU admission); moreover, our findings revealed that administration of reasonable volumes of contrast medium did not increase the risks of emergency dialysis or short-term mortality. Thus, clinicians can use contrast-enhanced CT scans in a reasonable manner in septic patients, in order to identify occult infection foci earlier and facilitate prompt medical management.

Our patients had surprisingly high prevalences of hypertension, diabetes mellitus, and chronic kidney disease, which could have been related to their older age, as reported in a prior study.32 Before propensity score matching, patients in the contrast-enhanced CT group were significantly younger and had fewer co-morbidities, including hypertension, chronic kidney disease, and chronic obstructive pulmonary disease. Moreover, patients in the contrast-enhanced CT group had lower initial serum creatinine levels and higher eGFRs, as observed in other studies.6 32 This could be related to the common clinical practice of using contrast-enhanced CT for younger patients with few co-morbidities and relatively good renal function, based on considerations of the potential nephrotoxicities of the contrast agents7; a few patients with poor renal function (24 of 482 patients with sepsis in the non-enhanced CT group) may also have avoided contrast agents following an explanation of the potential for nephrotoxicity. Clinicians may have hesitated to administer contrast media to patients with respiratory disease because of the risk of immediate hypersensitivity reaction; however, asthma and chronic obstructive pulmonary disease have not been established as consistent risk factors for contrast media-related adverse drug reactions.33 The lack of significant differences in risks of AKI, emergent dialysis, and short-term mortality between the non-enhanced and enhanced CT groups before propensity score matching in our study may have been influenced by the above-mentioned tendency for clinicians to perform contrast-enhanced CT in presumably healthier patients. However, it is difficult to evaluate the causal relationship between administration of contrast agents and risk of AKI in patients with sepsis by comparing two patient groups with many different demographic and characteristics; therefore, we used propensity score matching to minimise the impacts of potential confounders.

Although the mean initial serum lactate level in the contrast-enhanced CT group was slightly but significantly higher than that in the non-enhanced CT group, this difference was not correlated with the incidences of acute illness (eg, shock), ICU admission, and short-term mortality between the two groups. In addition, the levels of renal function, reflected by serum creatinine levels and eGFRs within 48 to 72 hours after CT scans, improved relative to initial measurements in both groups. These seemingly paradoxical findings suggested that sepsis, rather than the administration of contrast media, was the determinant of clinical outcomes in the present study.

Previous studies in emergency medical settings have shown wide variation in the incidence of post-contrast AKI (3.2%-12%); this may be partially explained by the variety of diseases encountered in the ED, as well as differences in the definitions of AKI adopted in each study.6 7 19 20 21 22 23 24 31 Nearly half of the patients (49%) in the present study experienced septic shock; thus, the increased incidence of post-contrast AKI in our patients (12.4%), compared with that observed in prior studies, may be attributed to the impaired physical status of our patients. This may also explain the considerably higher rates of emergent dialysis and short-term mortality, as well as the increased median length of hospital stay for survivors among our patients, compared to those parameters measured in other studies that did not focus on patients with sepsis.21 34

Thus far, the pathophysiology of CA-AKI remains poorly characterised. Based on the results of some animal studies, proposed mechanisms include acute tubular necrosis caused by medullary hypoxia from vasoconstriction, as well as direct cytotoxic effects of the contrast agent on renal tubular cells.35 36 Compared with AKI caused by other aetiologies, CA-AKI involves relatively rapid recovery of renal function; this is potentially because of the reduced extent of tubular necrosis, which leads to minor and transient functional impairment of tubular epithelial cells.37 Nevertheless, sepsis is the leading cause of AKI in critically ill patients and is associated with a higher mortality rate among patients in the ICU, compared with patients who have AKI caused by other aetiologies.38 Therefore, hesitation to perform contrast-enhanced CT scans for patients with sepsis, in order to identify occult infection foci, could result in delayed diagnoses of life-threatening conditions that carry considerable risks of morbidity and mortality, even in patients with serum creatinine up to 4.0 mg/dL.6

A number of studies performed in the past several years have been designed to maintain a balance between the benefits and adverse effects of contrast-enhanced CT scans in many clinical settings.6 7 12 20 21 22 The vast majority of those studies showed no significant association between the use of contrast agents and an increased risk of AKI. Consistent with the prior findings, contrast-enhanced CT scans of our patients with sepsis were not associated with increased risks of AKI and other adverse clinical outcomes. Among all aetiologies of AKI in patients requiring emergent medical attention, such as sepsis, dehydration, and nephrotoxic medication use,18 the contribution of CA-AKI is regarded as considerably less important37; notably, our findings support this view. Furthermore, it has been consistently shown that the performance of a contrast-enhanced CT scan is justified in patients for whom the examination is indicated, provided that other risk factors of AKI are well controlled.39

There were several limitations in our study, largely in relation to its single-centre and retrospective design. First, the non-enhanced CT group consisted of older patients with a higher prevalence of hypertension and worse renal function; this suggested a selection bias. Although we routinely checked serum lactate for patients with suspected sepsis in the ED, there were a few patients diagnosed with sepsis who did not have lactate measurement data; this may also have resulted in selection bias. Second, although propensity score matching was used to minimise the impacts of potential confounders, unmeasured confounding variables remained, leading to potentially biased results. Therefore, further large-scale cohort or well-controlled prospective randomised studies are warranted. Finally, the definition of AKI used in this study (elevation of serum creatinine concentration by 0.5 mg/dL or by 50% increase relative to baseline within 48 to 72 hours after contrast administration) may not accurately reflect the clinical condition because the relationship between increases in serum creatinine level and deterioration of renal function is reportedly non-linear.40

Our study demonstrated that the intravenous administration of contrast media during CT scans was not associated with increased risks of AKI, emergent dialysis, or short-term mortality for patients with sepsis in the ED; moreover, the use of contrast-enhanced CT was not associated with prolonged length of hospital stay in these patients. The lack of a significant correlation between the administration of contrast agents and the risk of AKI in patients with sepsis conflicts with the tendency to withhold contrast-enhanced CT for the diagnostic assessment and management of sepsis in the emergency setting. Further studies are necessary to confirm these findings and provide further guidance for clinical practice.

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

Concept or design: YC Hsu.
Acquisition of data: HY Su, CW Hsu, CY Liang.
Analysis or interpretation of data: TB Chen.
Drafting of the article: YC Hsu.
Critical revision for important intellectual content: CK Sun, CW Hsu.

Dr Chi-feng Hsieh is acknowledged for providing technical support in sample size calculation.

All authors have disclosed no conflicts of interest.

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

The study was approved by the Institutional Review Board of E-Da hospital (EMRP-106-037) and the requirement for informed patient consent was waived because of the retrospective observational nature of the study.

1. Wang HE, Jones AR, Donnelly JP. Revised national estimates of emergency department visits for sepsis in the United States. Crit Care Med 2017;45:1443-9. Crossref

2. Rhodes A, Evans LE, Alhazzani W, et al. Surviving Sepsis Campaign: international guidelines for management of sepsis and septic shock: 2016. Intensive Care Med 2017;43:304-77. Crossref

3. Jimenez MF, Marshall JC; International Sepsis Forum. Source control in the management of sepsis. Intensive Care Med 2001;27 Suppl 1:S49-62. Crossref

4. Berrington de González A, Mahesh M, Kim KP, et al. Projected cancer risks from computed tomographic scans performed in the United States in 2007. Arch Intern Med 2009;169:2071-7. Crossref

5. Kocher KE, Meurer WJ, Fazel R, Scott PA, Krumholz HM, Nallamothu BK. National trends in use of computed tomography in the emergency department. Ann Emerg Med 2011;58:452-62.e3. Crossref

6. Hinson JS, Ehmann MR, Fine DM, et al. Risk of acute kidney injury after intravenous contrast media administration. Ann Emerg Med 2017;69:577-86.e4. Crossref

7. Mitchell AM, Kline JA, Jones AE, Tumlin JA. Major adverse events one year after acute kidney injury after contrast-enhanced computed tomography. Ann Emerg Med 2015;66:267-74.e4. Crossref

8. McCullough PA, Adam A, Becker CR, et al. Epidemiology and prognostic implications of contrast-induced nephropathy. Am J Cardiol 2006;98:5K-13K. Crossref

9. Nash K, Hafeez A, Hou S. Hospital-acquired renal insufficiency. Am J Kidney Dis 2002;39:930-6. Crossref

10. Coca SG, Peixoto AJ, Garg AX, Krumholz HM, Parikh CR. The prognostic importance of a small acute decrement in kidney function in hospitalized patients: a systematic review and meta-analysis. Am J Kidney Dis 2007;50:712-20. Crossref

11. Katzberg RW, Newhouse JH. Intravenous contrast medium–induced nephrotoxicity: is the medical risk really as great as we have come to believe? Radiology 2010;256:21-8. Crossref

12. McDonald JS, McDonald RJ, Comin J, et al. Frequency of acute kidney injury following intravenous contrast medium administration: a systematic review and meta-analysis. Radiology 2013;267:119-28. Crossref

13. Kashani K, Levin A, Schetz M. Contrast-associated acute kidney injury is a myth: We are not sure. Intensive Care Med 2018;44:110-4. Crossref

14. Lameire N, Kellum JA; KDIGO AKI Guideline Work Group. Contrast-induced acute kidney injury and renal support for acute kidney injury: a KDIGO summary (Part 2). Crit Care 2013;17:205. Crossref

15. Mehran R, Aymong ED, Nikolsky E, et al. A simple risk score for prediction of contrast-induced nephropathy after percutaneous coronary intervention: development and initial validation. J Am Coll Cardiol 2004;44:1393-9. Crossref

16. Mehran R, Nikolsky E. Contrast-induced nephropathy: definition, epidemiology, and patients at risk. Kidney Int Suppl 2006;(100):S11-5. Crossref

17. Rihal CS, Textor SC, Grill DE, et al. Incidence and prognostic importance of acute renal failure after percutaneous coronary intervention. Circulation 2002;105:2259-64. Crossref

18. Aycock RD, Westafer LM, Boxen JL, Majlesi N, Schoenfeld EM, Bannuru RR. Acute kidney injury after computed tomography: a meta-analysis. Ann Emerg Med 2018;71:44- 53.e4. Crossref

19. Mitchell AM, Kline JA. Contrast nephropathy following computed tomography angiography of the chest for pulmonary embolism in the emergency department. J Thromb Haemost 2007;5:50-4. Crossref

20. Sonhaye L, Kolou B, Tchaou M, et al. Intravenous contrast medium administration for computed tomography scan in emergency: a possible cause of contrast-induced nephropathy. Radiol Res Pract 2015;2015:805786. Crossref

21. Heller M, Krieger P, Finefrock D, Nguyen T, Akhtar S. Contrast CT scans in the emergency department do not increase risk of adverse renal outcomes. West J Emerg Med 2016;17:404-8. Crossref

22. Ehrlich ME, Turner HL, Currie LJ, Wintermark M, Worrall BB, Southerland AM. Safety of computed tomographic angiography in the evaluation of patients with acute stroke: a single-center experience. Stroke 2016;47:2045-50. Crossref

23. Lima FO, Lev MH, Levy RA, et al. Functional contrast-enhanced CT for evaluation of acute ischemic stroke does not increase the risk of contrast-induced nephropathy. AJNR Am J Neuroradiol 2010;31:817-21. Crossref

24. Tremblay LN, Tien H, Hamilton P, et al. Risk and benefit of intravenous contrast in trauma patients with an elevated serum creatinine. J Trauma 2005;59:1162-6. Crossref

25. Singer M, Deutschman CS, Seymour CW, et al. The third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA 2016;315:801-10. Crossref

26. World Health Organ Tech Rep Ser. Nutritional anaemias: report of a WHO scientific group. World Health Organ Tech Rep Ser 1968;405:5-37.

27. National Kidney Foundation. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Am J Kidney Dis 2002;39:S1-266.

28. Mehta RL, Kellum JA, Shah SV, et al. Acute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury. Crit Care 2007;11:R31. Crossref

29. Bagshaw SM, George C, Bellomo R; ANZICS Database Management Committee. Early acute kidney injury and sepsis: a multicentre evaluation. Crit Care 2008;12:R47. Crossref

30. Song W, Zhang T, Pu J, Shen L, He B. Incidence and risk of developing contrast-induced acute kidney injury following intravascular contrast administration in elderly patients. Clin Interv Aging 2014;9:85-93. Crossref

31. Hinson JS, Al Jalbout N, Ehmann MR, Klein EY. Acute kidney injury following contrast media administration in the septic patient: A retrospective propensity-matched analysis. J Crit Care 2019;51:111-6. Crossref

32. McDonald, JS, McDonald RJ, Williamson EE, Kallmes DF, Kashani K. Post-contrast acute kidney injury in intensive care unit patients: a propensity score-adjusted study. Intensive Care Med 2017;43:774-84. Crossref

33. Bettmann MA, Heeren T, Greenfield A, Goudey C. Adverse events with radiographic contrast agents: results of the SCVIR Contrast Agent Registry. Radiology 1997;203:611-20. Crossref

34. McDonald RJ, McDonald JS, Carter RE, et al. Intravenous contrast material exposure is not an independent risk factor for dialysis or mortality. Radiology 2014;273:714-25. Crossref

35. Persson PB, Hansell P, Liss P. Pathophysiology of contrast medium-induced nephropathy. Kidney Int 2005;68:14-22. Crossref

36. Heyman SN, Rosenberger C, Rosen S. Regional alterations in renal haemodynamics and oxygenation: a role in contrast medium-induced nephropathy. Nephrol Dial Transplant 2005;20 Suppl 1:i6-11. Crossref

37. Molitoris BA, Dahl R, Geerdes A. Cytoskeleton disruption and apical redistribution of proximal tubule Na(+)-K(+)-ATPase during ischemia. Am J Physiol 1992;263:F488-95. Crossref

38. Bellomo R, Kellum JA, Ronco C, et al. Acute kidney injury in sepsis. Intensive Care Med 2017;43:816-28. Crossref

39. Petek BJ, Bravo PE, Kim F, et al. Incidence and risk factors for postcontrast acute kidney injury in survivors of sudden cardiac arrest. Ann Emerg Med 2016;67:469-76.e1. Crossref

40. Ostermann M, Joannidis M. Acute kidney injury 2016: diagnosis and diagnostic workup. Crit Care 2016;20:299. Crossref

Both the range of available drugs and the scope of drug markets are expanding and diversifying. Abuse of substances such as amphetamine-type stimulants, cannabis, and cocaine are major global health concerns. According to World Health Organization statistics, the number of cannabis users increased from 183 million in 2015 to 192 million in 2016 worldwide, whereas 34 million people abuse amphetamines and prescription stimulants.1

While the spectrum of substance abuse can be wide, many forms of illicit drug use inevitably induce toxicity and detrimental effects on the urinary tract. Smoking cannabis was found to have a significant association with bladder cancer in a hospital-based case-control study,2 attributed to the common carcinogens present in cannabis and tobacco smoke.3 Acute renal infarction has been observed in patients who used cocaine.4 Ketamine has received particular attention in the past few years for its impacts on both the upper and lower urinary tract.5 It has been one of the most commonly abused substances by teenagers since 2005 in Asian cities such as Hong Kong.6

In recent years, methamphetamine (MA) abuse has also become a serious and growing problem in Asia.7 The proportion of people abusing MA increased from 28.8% to 75.1% over a span of 5 years in China.8 Japan has seen its third epidemic of MA abuse since 1995.9 South Korea has also had an increase in psychotropic drug abuse, predominantly MA, from 7919 people in 2014 to 11 396 in 2016.10 In Hong Kong, 25.9% of people who abused drugs had exposure to amphetamine-type psychotropic substances in 2017.11 While the psychological and neurological effects of MA have been widely discussed, the urological aspects of the drug’s side-effects have not yet been well documented in the literature. We investigated the symptomatology and voiding function in a cohort of patients who abused the two most common psychotropic substances in our locality, namely MA and ketamine, and compared their urinary tract toxicity profiles.

In the period of 23 months from 1 October 2016, all consecutive new cases of patients who attended our centre for MA- or ketamine-related urological disorders were seen in a dedicated clinic and were recruited into a prospective cohort. Ethics committee approval was granted for the study (CREC Ref CRE-2011.454). Written informed consent was given by all participants before entering the study.

Basic demographic data were recorded before clinic attendance, including age, sex, employment status, drinking habits, and smoking history. Habits of substance abuse were characterised. Serum creatinine levels, urine microscopy and culture, and uroflowmetry were measured. Initial symptomatology enquiry included the presence and characteristics of frequency, urgency, suprapubic pain, haematuria, hesitancy, intermittency, and incomplete emptying. Functional bladder capacity was calculated by adding the voided volume to post-void urine residuals during the uroflowmetry assessment. Urological symptoms were assessed with the International Prostate Symptom Score (IPSS) or the Overactive Bladder Symptom Score (OABSS).12 The International Index of Erectile Function (IIEF) was used to assess sexual function in male respondents who were sexually active in the preceding 4 weeks. Another component of symptom assessment was the Pelvic Pain and Urgency/Frequency (PUF) patient symptom scale. The Chinese version of the PUF symptom scale is a validated assessment tool for cystitis.13 For cognitive dysfunction, we employed the Montreal Cognitive Assessment (MoCA) as an assessment tool. Chu et al14 proved the validity and reliability of the Cantonese Chinese MoCA as a brief screening tool for cognitive impairment.

Polysubstance abuse patients were excluded. Data were analysed by comparison between two groups of patients, namely those with ketamine abuse only and those with MA abuse only. Descriptive statistics were used to characterise the clinical characteristics of the study cohort. Chi squared tests were used for categorical data, and Mann-Whitney U tests were used for continuous data. A P value of <0.05 was considered to indicate statistical significance. The SPSS (Windows version 24.0; IBM Corp, Armonk [NY], United States) was used for all calculations.

From October 2016 to August 2018, 66 new patients attended our clinic for urological problems secondary to substance abuse. After excluding patients with substance abuse other than ketamine and MA, 38 patients were included for analysis (Table 1). Both genders contributed 19 patients. There was a statistically significant difference in mean age between the two groups of patients with MA and ketamine abuse (27.2 ± 7.2 years and 31.6 ± 4.8 years, respectively, P=0.011). Most patients were not active substance abusers upon presentation to the clinic. While all patients had a history of substance abuse, only two (5.3%) patients were consuming alcohol on a daily basis.

Urinary frequency was the single most common urological symptom in our patient cohort. Regardless of whether the patient was consuming ketamine alone, MA alone, or a combination of ketamine and MA, urinary frequency was found in 71.1% of the patients, with no statistically significant differences between these groups (Table 2). Other symptoms that shared similar distributions between both groups were urgency, suprapubic pain, intermittent stream, and sensation of incomplete emptying. There was a statistically significant difference in the prevalence of dysuria between the two groups (ketamine 43.5%, MA 6.7%, P=0.026). A trend was observed in the difference in prevalence of hesitancy (ketamine 4.3%, MA 26.7%, P=0.069).

To summarise the results of questionnaires that assess urinary storage symptoms and voiding symptoms as a whole, neither OABSS nor IPSS revealed a statistically significant difference between the two patient groups (Table 3). The mean IPSS score of the ketamine only group was 20.9 ± 8.1, falling into the severe symptom category, whereas that of the MA only group was 16.1 ± 8.9, falling into the moderate symptom group. No significant differences in maximal voiding velocity, voided volume, post-void residual, or bladder capacity were observed between the two groups. Similarly, no significant difference was observed in sexual function between the male patients of these three groups, as assessed by IIEF.

Pelvic pain assessment with the PUF symptom scale revealed higher scores in the ketamine group, especially in the Bother score domain (ketamine 6.7 ± 2.9, MA 4.7 ± 2.5, P=0.036). Cognitive assessment using MoCA revealed that both groups had impairment, but there was no significant difference between the MA group and the ketamine group (ketamine 24.8 ± 2.5, MA 23.6 ± 2.9, P=0.298). Serum creatinine did not differ significantly between the groups (ketamine 88.48 ± 55.44, MA 66.83 ± 16.92, P=0.138).

Substance abuse is a significant public health problem, with approximately 5.2% of the world population aged between 15 and 64 years having used illicit drugs at least once in the previous year.1 Southeast and East Asia have been a global hub for MA production and trafficking over the past decades, and its abuse is common in areas of South Korea, China, Taiwan, Japan, the Golden Triangle, and Iran.10 Psychotropic substance abuse is the most common form of drug abuse in Hong Kong, and since 2015, MA has taken over ketamine’s spot as the leading drug of abuse among all psychotropic substances.11 As MA has become the new trendy drug of abuse, and most drug abusers are young and had their first drug exposure at an early age, this can account for our finding that patients who used MA had a lower mean age than patients who used ketamine in the cohort (Table 1).

Methamphetamine belongs to the class of amphetamines that also includes other drugs such as MDMA (3,4-methylenedioxy-N-methylamphetamine). The stimulant, euphoric, anorectic, empathogenic, entactogenic, and hallucinogenic properties of MA drive its popularity for abuse. Kolbrich et al15 demonstrated the fast, widespread, and long-lasting distribution of MA in the human brain, paralleling the long-lasting behavioural and neurological effects of the drug. Our data on cognitive impairment in MA users echoed this finding, demonstrating impaired function in this group by MoCA assessment.

Much of the focus on MA in the literature has been placed on its neurological and behavioural aspects. Unlike ketamine, whose effects on the urinary tract and treatment modalities have been more commonly discussed,16 similar research endeavours have not been undertaken in the area of MA, even though it is a more widely abused drug. Thus, our study was an effort to investigate the clinical presentation of MA abuse on the urinary tract and compare it with ketamine abuse, another common drug of illicit use. As illustrated by our findings in the cohort, patients who used MA reported at least moderate severity of urinary symptoms by IPSS assessment. Because we studied a group of young patients with mean age 27.2 years, we conclude that the urological impact of MA abuse cannot be neglected.

In the current study, storage symptoms (particularly urinary frequency) had similar prevalence between patients who used MA and ketamine (Table 2). On assessment of storage symptoms by OABSS, patients in both groups attained similar scores to patients with overactive bladder syndrome.17 In the case of ketamine abuse, storage symptoms can be attributed to denuded mucosa and infiltration of inflammatory cells into the lamina propria of the bladder, eventually leading to chronic inflammation and fibrosis.5 It has been postulated that the storage symptoms from MA abuse can be the result of a dysfunctional dopamine pathway in detrusor control. The β-phenylethylamine core structure of MA allows it to cross the blood-brain barrier easily and to resist brain biotransformation. Furthermore, its structural similarity with monoamine neurotransmitters allows amphetamines to act as competitive substrates at dopamine’s membrane transporters. It also promotes dopamine release from storage vesicles. All these effects increase cytoplasmic dopamine concentrations and enhance reverse transport.18 However, long-term exposure to amphetamines may result in dopamine neuron terminal damage or loss. A post-mortem study of people who used MA chronically showed a mean 50% to 60% reduction in dopamine levels throughout the striatum.19 Another study on dopamine dysregulation reported a 25% to 30% decrease in the maximal extent of dopamine-induced stimulation of adenylyl cyclase activity in the striatum.20 These results suggested that dopamine signalling in the striatum of people who use MA chronically was impaired both presynaptically and postsynaptically. An animal study showed that a selective D1 antagonist decreased bladder capacity in rats.21 Taking Parkinson’s disease, which is the result of dopaminergic neuron degeneration with the basal ganglia failing to suppress micturition,22 as a reference, the pathological dopaminergic pathway could be one of the aetiologies behind MA-related urinary symptoms.

In our cohort, 26.7% of patients who used MA reported hesitancy during voiding, and 46.7% reported a sensation of incomplete emptying (Table 2). Case reports in the literature have drawn an association between amphetamines and urinary retention.23 24 These findings underlined the possibility that MA-related urological pathology might have multiple facets rather than purely concerning the dopaminergic axis and storage symptoms. The possible aetiology may be the increased release of norepinephrine from MA abuse, which may cause increased α-adrenergic stimulation of the bladder neck.25 In a study of medical treatment for urinary symptoms in people who used MA, Koo et al26 reported that α-blockers resulted in a 33% treatment success rate in terms of IPSS reduction for patients with predominant voiding symptoms. This observation could be employed as a reference for treatment options.

Patients with ketamine abuse more commonly experienced dysuria and pelvic pain. In our previous report of 319 patients with ketamine abuse, the mean PUF score was 22.2.16 In contrast, symptoms of dysuria and pelvic pain were not commonly observed in patients with MA abuse. This could be because the effects of MA on the urinary tract were more on neurology rather than local tissue destruction, resulting in less local nociceptor stimulation. Currently, there are no reports on histological assessment of urinary tract tissues in people who use MA. This would be useful to facilitate a more comprehensive investigation on the impact on MA on the urinary tract.

The relatively small sample size of our cohort limited our statistical analysis. As MA abuse has only gained popularity in recent years in our locality, and the urological sequelae of such abuse might take years before it becomes prominent and severe, the current study could act as an initial assessment highlighting the early observation of urological presentation. One of the potential limitations of our study is that the majority of the patients presenting to our clinic were not active substance abusers. At the time point of assessment, the use of a heterogeneous group of active and former abusers may introduce bias into our evaluation. However, our cohort included mostly patients with long abuse duration. Studies have demonstrated the persistent effects of drug abuse even after a period of abstinence, namely dysfunctional dopamine metabolism in patients who had used MA27 and urinary tract damage in patients who had used ketamine.5 Thus, the clinical picture captured by our study may still reflect the impact of drug abuse on the urinary tract.

In conclusion, MA is a common drug of choice for abuse in Asia. It causes urinary tract dysfunction, predominantly in the form of storage symptoms. Compared with ketamine, MA abuse was not commonly associated with dysuria or pelvic pain. Apart from the behavioural impacts of MA abuse, its urinary tract implications should not be neglected.

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

Concept or design: CH Yee, CF Ng, YH Tam.
Acquisition of data: YL Hong, PT Lai.
Analysis or interpretation of data: CH Yee, YL Hong.
Drafting of the article: CH Yee, YL Hong.
Critical revision for important intellectual content: CH Yee, CF Ng, YH Tam.

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

This research project was funded by the Beat Drugs Fund, The Government of the Hong Kong Special Administrative Region.

Ethics committee approval was granted for the study (CREC Ref CRE-2011.454). Written informed consent was given by all participants before entering the study.

1. United Nations Office on Drugs and Crime. World drug report. 2018. Available from: https://www.unodc.org/wdr2018/. Accessed 10 May 2019.

2. Chacko JA, Heiner JG, Siu W, Macy M, Terris MK. Association between marijuana use and transitional cell carcinoma. Urology 2006;67:100-4. Crossref

3. Bowles DW, O’Bryant CL, Camidge DR, Jimeno A. The intersection between cannabis and cancer in the United States. Crit Rev Oncol Hematol 2012;83:1-10. Crossref

4. Madhrira MM, Mohan S, Markowitz GS, Pogue VA, Cheng JT. Acute bilateral renal infarction secondary to cocaine-induced vasospasm. Kidney Int 2009;76:576-80. Crossref

5. Yee C, Ma WK, Ng CF, Chu SK. Ketamine-associated uropathy: from presentation to management. Curr Bladder Dysfunct Rep 2016;11:266-71. Crossref

6. Yiu-Cheung C. Acute and chronic toxicity pattern in ketamine abusers in Hong Kong. J Med Toxicol 2012;8:267-70. Crossref

7. Ren Q, Ma M, Hashimoto K. Current status of substance abuse in East Asia and therapeutic prospects. East Asian Arch Psychiatry 2016;26:45-51.

8. Sun HQ, Bao YP, Zhou SJ, Meng SQ, Lu L. The new pattern of drug abuse in China. Curr Opin Psychiatry 2014;27:251-5. Crossref

9. Wada K, Funada M, Shimane T. Current status of substance abuse and HIV infection in Japan. J Food Drug Anal 2013;21:S33-6. Crossref

10. Kwon NJ, Han E. A commentary on the effects of methamphetamine and the status of methamphetamine abuse among youths in South Korea, Japan, and China. Forensic Sci Int 2018;286:81-5. Crossref

11. Narcotics Division, Security Bureau, Hong Kong SAR Government. Central Registry of Drug Abuse Sixty-Seventh Report. Available from: https://www.nd.gov.hk/en/crda_report.htm. Accessed 10 May 2019.

12. Homma Y, Yoshida M, Seki N, et al. Symptom assessment tool for overactive bladder syndrome–overactive bladder symptom score. Urology 2006;68:318-23. Crossref

13. Ng CM, Ma WK, To KC, Yiu MK. The Chinese version of the Pelvic Pain and Urgency/Frequency Symptom Scale: a useful assessment tool for street-ketamine abusers with lower urinary tract symptoms. Hong Kong Med J 2012;18:123-30.

14. Chu LW, Ng KH, Law AC, Lee AM, Kwan F. Validity of the Cantonese Chinese Montreal Cognitive Assessment in Southern Chinese. Geriatr Gerontol Int 2015;15:96-103. Crossref

15. Kolbrich EA, Goodwin RS, Gorelick DA, Hayes RJ, Stein EA, Huestis MA. Plasma pharmacokinetics of 3,4-methylenedioxymethamphetamine after controlled oral administration to young adults. Ther Drug Monit 2008;30:320-32. Crossref

16. Yee CH, Lai PT, Lee WM, Tam YH, Ng CF. Clinical outcome of a prospective case series of patients with ketamine cystitis who underwent standardized treatment protocol. Urology 2015;86:236-43. Crossref

17. Hung MJ, Chou CL, Yen TW, et al. Development and validation of the Chinese Overactive Bladder Symptom Score for assessing overactive bladder syndrome in a RESORT study. J Formos Med Assoc 2013;112:276-82. Crossref

18. Carvalho M, Carmo H, Costa VM, et al. Toxicity of amphetamines: an update. Arch Toxicol 2012;86:1167-231. Crossref

19. Wilson JM, Kalasinsky KS, Levey AI, et al. Striatal dopamine nerve terminal markers in human, chronic methamphetamine users. Nat Med 1996;2:699-703. Crossref

20. Tong J, Ross BM, Schmunk GA, et al. Decreased striatal dopamine D1 receptor-stimulated adenylyl cyclase activity in human methamphetamine users. Am J Psychiatry 2003;160:896-903. Crossref

21. Seki S, Igawa Y, Kaidoh K, Ishizuka O, Nishizawa O, Andersson KE. Role of dopamine D1 and D2 receptors in the micturition reflex in conscious rats. Neurourol Urodyn 2001;20:105-13. Crossref

22. McDonald C, Winge K, Burn DJ. Lower urinary tract symptoms in Parkinson’s disease: prevalence, aetiology and management. Parkinsonism Relat Disord 2017;35:8-16. Crossref

23. Worsey J, Goble NM, Stott M, Smith PJ. Bladder outflow obstruction secondary to intravenous amphetamine abuse. Br J Urol 1989;64:320-1. Crossref

24. Delgado JH, Caruso MJ, Waksman JC, Honigman B, Stillman D. Acute, transient urinary retention from combined ecstasy and methamphetamine use. J Emerg Med 2004;26:173-5. Crossref

25. Skeldon SC, Goldenberg SL. Urological complications of illicit drug use. Nat Rev Urol 2014;11:169-77. Crossref

26. Koo KC, Lee DH, Kim JH, et al. Prevalence and management of lower urinary tract symptoms in methamphetamine abusers: an under-recognized clinical identity. J Urol 2014;191:722-6. Crossref

27. Ares-Santos S, Granado N, Moratalla R. The role of dopamine receptors in the neurotoxicity of methamphetamine. J Intern Med 2013;273:437-53. Crossref

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

All authors have disclosed no conflicts of interest.

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

All authors have disclosed no conflicts of interest.

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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