Radical radiotherapy is a standard treatment option for patients with early-stage and locally advanced non-metastatic prostate cancer. Advances in radiotherapy in the past 20 years include the use of androgen deprivation therapy for patients with this type of cancer, as well as the application of more precise radiotherapy techniques.1 2 Intensity-modulated radiation therapy (IMRT) has emerged as the standard radiotherapy technique.3 Its benefits have been explored in terms of the effects of dose escalation or hypofractionation on survival outcomes.4 5 For patients undergoing this type of treatment, toxicities are the primary concern. Long-term side-effects (ie, complications occurring ≥3 months after radiotherapy) have a major impact on the quality of life for affected patients; this is particularly important for patients with genitourinary or rectal toxicities. Late rectal toxicities, including per-rectal bleeding, faecal incontinence, and proctitis, have been reported to occur at rates of 5% to 21%.1 34 5 6 7 8 9

Associations have been reported between late rectal toxicities and various clinical and dosimetric parameters; however, most data were collected using the conventional three-dimensional conformal technique.8 1011 12 In addition, there have been limited reports of such associations among patients in Hong Kong. In particular, Poon et al8 reported that 8% of patients exhibited grade ≥2 late rectal toxicities following IMRT in a retrospective cohort study. Although several clinical parameters were assessed, most failed to show statistically significant associations, with the exception of the presence of acute rectal toxicities.8 To the best of our knowledge, pharmacological parameters following IMRT for prostate cancer have not yet been studied in local populations. Some previous reports showed a significant association between anticoagulant use and late rectal toxicities, whereas an association between antiplatelet use and androgen deprivation was inconsistent among studies.12 13 14

Multiple strategies have been used for the treatment of late rectal toxicities. The use of hyperbaric oxygen has shown promising results in some retrospective studies, but it has not been available in Hong Kong until recently.12 15 16 Treatments with sucralfate, prednisolone enaema, short-chain fatty acids, and antifibrinolytics have been evaluated in small trials.17 18 19 Thus far, no standard approach has been established, and there are no published data regarding local management practices.

Late rectal toxicities may represent clinically significant complications because of their non-negligible incidences. Insights regarding any factors predictive of their occurrence could aid in improved treatment planning and early identification of toxicity. This study was performed to assess the incidence of late rectal toxicities and to identify factors predictive for late proctitis in patients treated with prostate-specific IMRT in Hong Kong.

This retrospective longitudinal observational study included patients with prostate cancer who received IMRT with radical intent in a tertiary referral institution in Hong Kong from January 2007 to December 2011. Patients were excluded if they were followed up for fewer than 12 months from the start of radiotherapy, if they did not complete the course of radiotherapy, if they were not at risk of proctitis (eg, those with post-abdominoperineal resection), or if they did not have a retrievable radiotherapy plan due to technical difficulties. The cut-off date for data collection was 31 December 2018.

Patients underwent treatment with a comfortably full bladder and an empty rectum, with laxatives administered 1 day prior to simulation computed tomography. Patients were asked to empty the bladder prior to attending the radiotherapy suite, and then drink a comfortable volume of water. A pelvic thermoplastic mould was used for immobilisation. Intravenous contrast was administered prior to computed tomography. Re-simulation was performed automatically if bladder volume was below 150 cc, if prominent rectal gas was present, or upon request by the attending oncologist. Contouring was performed by designated oncologists with confirmation by at least one specialist. Tumour and whole prostate were contoured as a single volume; the clinical target volume (CTV) was the volume of the tumour, whole prostate, and base of the seminal vesicle (defined as 1 cm of the central seminal vesicle proximal to the base of the prostate). Whole seminal vesicle was included in the CTV if seminal vesicle involvement was observed. Planning target volume (PTV) was determined by expanding the CTV by a radial margin of 1.5 cm, except posteriorly where a smaller margin was used (0.7 cm). Pelvic lymph node irradiation was not performed. Patients received 70 Gy in 35 daily fractions over 7 weeks at 100% of the isodose level. Rectal volume was contoured in accordance with the Radiation Therapy Oncology Group Consensus Contouring Guidelines for normal male pelvic tissue. Dose constraints for organs at risk followed our departmental protocol: for the rectum, we classified the plan as fulfilling the first, second, or third criteria. First criteria were satisfied if V40 (% of organ volume receiving 40 Gy) <35% or V65 (% of organ volume receiving 65 Gy) <17%; second criteria were satisfied if V53 (% of organ volume receiving 53 Gy) <45% or V68 (% of organ volume receiving 68 Gy) <20%; and third criteria were satisfied if V60 (% of organ volume receiving 60 Gy) <50%, V65 (% of organ volume receiving 65 Gy) <35%, or V70 (% of organ volume receiving 70 Gy) <25%. Hormonal treatment was administered based on the risk stratification used in the United Kingdom National Institute for Health and Care Excellence guidelines. Patients were followed up at 3–6-month intervals until the patient died or defaulted, and data were censored at the last recorded follow-up. Dose distributions, doses administered to organs at risk, and dose volume histograms were evaluated by the Eclipse and Planning System (Varian Medical Systems; Palo Alto [CA], United States).

For each patient, basic demographic data were documented, including age; Eastern Cooperative Oncology Group performance score; smoking habit; pretreatment albumin level; co-morbidities such as hypertension, diabetes, lipid disorder, history of cerebrovascular disease, ischaemic heart disease, and/or chronic renal impairment; medical history of abdominal surgery; drug history including antihypertensives, oral glycaemic agents, antiplatelets, anticoagulants, lipid-lowering agents, and antipurine agents; androgen deprivation therapies, including medical or surgical castration; and use of immunosuppressants. Tumour characteristics were also recorded, including pretreatment prostate-specific antigen level, clinical T-staging determined by clinical and radiological findings (based on AJCC 7th edition20), and Gleason score.

Acute and late rectal toxicities, including proctitis, incontinence, and per-rectal bleeding, were recorded and classified in accordance with Common Terminology Criteria for Adverse Events version 4.21 Late rectal toxicities were defined as those that occurred at least 3 months after the completion of radiotherapy. Late proctitis was defined as either the presence of rectal symptoms listed in Common Terminology Criteria for Adverse Events version 4, or colonoscopy findings of proctitis (eg, telangiectasia, ulcers, or inflammation). If a patient presented with per-rectal bleeding, colonoscopy findings were referenced whenever present to differentiate proctitis or other causes of bleeding, such as diverticulosis or haemorrhoids. Per-rectal bleeding only was recorded if no endoscopic proctitis features were present; otherwise, both per-rectal bleeding and proctitis were recorded. Additional parameters recorded included time of onset of late rectal toxicities, as well as treatment modalities used.

Dosimetric parameters (eg, V40, V50, V60, V70, Dmax [maximum dose], mean dose to rectum, and contoured rectal volume) were evaluated with the radiotherapy planning system. The use of static beam or volumetric arc technique was recorded, as was the compliance with rectal dose constraints.

Incidences of grade ≥1 late rectal toxicities with 95% confidence interval (CI) were calculated at 1, 2, and 5 years after treatment. The Kaplan-Meier curve method was used to illustrate the time to onset of late rectal toxicities. The Chi squared test, Fisher’s exact test, independent t test, or Mann-Whitney U test were used to compare baseline patient characteristics, pharmacological and dosimetric parameters between patients in grades 0 and ≥1 late toxicities, as well as in patients with late proctitis. The association of each parameter with late proctitis was examined using a multivariable binary logistic regression model with a backward stepwise selection method, including variables with P<0.1 in univariable regression analyses. The presence of multicollinearity was determined by using variance inflation factors. Statistical analyses were performed using SPSS (Windows version 22.0; IBM Corp, Armonk [NY], United States). The threshold of statistical significance was set at P<0.05. The STROBE checklist was followed to ensure standardised reporting.

From January 2007 to December 2011, a total of 238 patients with prostatic cancer received radical radiotherapy in our institution. As shown in the Figure, 232 patients were included in the analysis. The mean age of patients was 72.3 ± 4.8 years at time of radiotherapy (Table 1). The mean follow-up period was 7.3 ± 2.1 years, and there were 157 (67.7%) surviving patients at the cut-off date for data collection. Forty-two (18.1%) patients had been diagnosed with biochemical recurrence during the study period, based on the Phoenix definition.22 In total, 229 patients received a PTV dose of ≤70 Gy. Owing to genuine bowel invasion, or as a component of individualised dose escalation, four patients received a PTV dose of 66 to 76 Gy, of which three were >70 Gy. Colonoscopy was performed in 103 (44.4%) patients during follow-up. Among patients with per-rectal bleeding, 93 (88.6%) had undergone colonoscopy.

Occurrences of acute and late rectal toxicities throughout the study period are shown in Table 2. The rates of all-grade acute and late rectal toxicities were 36.2% and 46.5%, respectively; the rates of grade ≥2 late rectal toxicities and proctitis were 28.4% and 25.0%, respectively. Nineteen (8.2%) patients had grade 3 per-rectal bleeding, with 15 (78.9%) requiring blood transfusion and eight (42.1%) requiring endoscopic coagulation. The cumulative incidences of rectal toxicities at 1, 2, and 5 years after treatment are shown in Table 3. The median times of onset of late proctitis, late faecal incontinence, and late per-rectal bleeding were 15, 21.8, and 18.4 months, respectively.

Patients’ detailed demographic, pharmacological, and dosimetric parameters are listed in Table 1. Factors including history of haemorrhoid, PTV dose, and V70 were significantly different between patients with and without late rectal toxicities. In addition, age was the sole demographic factor significantly associated with late proctitis. There was no significant association between antiplatelet use and late rectal toxicities (P=0.066). No associations were found between late proctitis and other demographic or pharmacological characteristics (eg, PTV dose and history of haemorrhoid) in this study.

Univariable and stepwise multivariable analyses were performed to identify factors predictive of

late proctitis (Table 4). In univariable analysis, the presence of acute rectal toxicities, antiplatelet use, age at radiotherapy, Dmax, and dose/volume histogram parameters (ie V50, V60, V70, and rectal constraints) were identified as potential risk factors. In the regression model with all potential risk factors included, multicollinearity was detected among the dose/volume histogram parameters (variance inflation factors of 7.21, 8.69, 3.05, and 4.97 for V50, V60, V70, and rectal constraints, respectively). Compared to V50 and V60, V70 (ie, the high-dose region) showed a stronger association with late proctitis in univariable analysis. Multicollinearity was resolved by exclusion of V50 and V60 from the multivariable regression model. The final multivariable regression model revealed increased odds of late proctitis in patients with older age, higher V70, and the presence of acute rectal toxicities. Antiplatelet use tended to show higher odds, but this finding was not statistically significant (odds ratio=1.98, 95% CI=0.95-4.14). Dmax and satisfaction of the 3rd criteria alone were associated with late proctitis in univariable analysis, but the associations were not significant in multivariable analysis.

Common treatment modalities among patients with grade ≥2 late proctitis were also recorded. Topical agents such as Ultraproct® (commercial preparation of fluocortolone pivalate, fluocortolone hexanoate, and cinchocaine hydrochloride), bismuth ointment, or an antifibrinolytic agent (eg, tranexamic acid) are commonly used as first-line treatment.23 More than half (53.4%) of the patients had been administered an antifibrinolytic agent, while 77.6% and 19% of the patients were prescribed Ultraproct® and bismuth, respectively. Prednisolone enaema was also administered in 22 (37.9%) patients; the median duration of enaema use was 3.5 months (interquartile range, 1-7.25 months). Subjective improvement was reported by eight (36.4%) patients who received enaema treatment.

Radiation proctitis and other long-term rectal toxicities are clinically significant complications of radiotherapy to the prostate, due to their detrimental effects on patients’ quality of life, as well as the expected long duration of post-treatment survival. In our cohort, the incidences of late proctitis (30.2%) and overall rectal toxicities (46.5%) were slightly higher than those in previous reports (5%-21%).1 3 4 5 6 7 8 9 Comparison of baseline characteristics showed that more patients had ≥T3 disease in our cohort, although we found no statistically significant association between T-staging and a higher incidence of proctitis; similarly, no association between these parameters were reported in other studies.8 12 Other variables with possible interactions were similar between our study and prior studies; these included age, dosimetric parameters (eg, V70, which was 14% in our study and 10% to 23% in previous studies), and the use of antiplatelets.8 11 12

There are two possible explanations for the higher incidences of late proctitis and overall rectal toxicities. First, our study involved frequent utilisation of colonoscopy for any rectal symptoms, which may lead to a higher rate of recognition; notably, the rate of utilisation was not reported in previous studies. Second, our study had a relatively long follow-up period. Previous studies described the incidence of toxicity throughout the study period. The mean follow-up period in our study was 7.3 years, whereas that of most previous studies was 38.9 to 66 months; in one notable exception, the follow-up period was 8.4 years (the incidence was 21% in that study).3 The longer study period may also have contributed to a higher number of late rectal toxicities.

Previous reports suggested that a variety of parameters are associated with late proctitis; knowledge of these parameters could help clinicians to predict the risk of proctitis in each patient. In our study, age, and dosimetric parameters including V50, V60, and V70 were associated with late proctitis; history of haemorrhoid and V70 were associated with overall late rectal toxicities. These findings are consistent with the results of previous studies.10 11 12 13 14 24 Some factors identified in prior studies, including diabetes, previous abdominal surgery, and the use of antiandrogen or anticoagulant medication,11 13 25 failed to demonstrate any associations in the present study. Of note, <10% of the patients in our study had a history of abdominal surgery or inflammatory bowel disease; this could have influenced our ability to identify a statistically significant association. Recall bias, incomplete documentation of coexisting medical conditions and pharmacological histories, and the relatively small sample size in our cohort may have influenced our conclusions regarding factors associated with overall late rectal toxicities and/or late proctitis.

Several dosimetric parameters and dose/volume histogram data (including V50, V60, and V70) were also associated with late proctitis, as in previous studies.8 Our in-house rectal constraints did not demonstrate significant associations with the occurrence of proctitis (P=0.092). Notably, in the present study, the PTV dose was associated with overall late toxicities, but not with late proctitis specifically. Most patients received 70 Gy in this study; therefore, the effects of PTV dose on complications were difficult to establish.

Regression analysis was used to predict the odds of late proctitis among patients in our study. As shown in Table 4, higher V70, older age, and the presence of acute rectal toxicities were found to increase the odds of late proctitis. Poon et al8 also reported similar findings concerning acute rectal toxicities; however, they did not find associations with V70 or age. The increased incidence of late proctitis in our study may have enhanced our ability to identify significantly associated factors. Nevertheless, both our present study and the study of Poon et al8 demonstrated that patients with acute rectal toxicities during radiotherapy had higher incidences of late proctitis than patients without acute rectal toxicities. Similar results were reported by Fellin et al.11 Taken together, the present and prior results indicate that the presence of acute toxicities is predictive for late proctitis. Clinicians should be vigilant and perform prompt investigations when patients with acute toxicities report any rectal symptoms during subsequent follow-up.

Theoretically, dosimetric parameters are expected to be associated with late proctitis. In our study, the dosimetric parameters exhibited modest associations with late proctitis. Notably, we did not find a significant association between our in-house rectal constraints and the occurrence of late proctitis. Fellin et al11 demonstrated similar associations between late proctitis and V70, as well as other dosimetric parameters, in their cohort. This suggests that the presence of confounding factors may reduce the strength of associations with late proctitis. A notable factor is the inter-fractional variation of rectal and bladder filling; specifically, Miralbell et al26 found that rectal filling was significantly associated with late rectal toxicities. Imaging-guided radiotherapy with inter-fractional bowel and bladder control has been suggested in accordance with the nomogram designed by Delobel et al9; this type of therapy could reduce the risks of acute and late rectal toxicities. In our study, there was no strict inter-fractional bowel or imaging control for bladder and rectal volumes during the course of IMRT. Although we found no statistically significant difference in the mean rectal volume during simulation computed tomography between patients with and without late proctitis, we could not retrieve the inter-fractional variation in rectal volumes for analysis in this study; this factor was also excluded from analysis in the study by Fellin et al.11 Although identical instructions were provided to patients during simulation and treatment, inter-fractional variations may have been statistically significant. To further confirm whether dosimetric parameters are predictive of late proctitis, a prospective study is needed in which strict interfractional rectal and bladder control are performed, in combination with improved treatment verification strategies (eg, the use of cone beam computed tomography).

There were a few weaknesses in this study. First, this was a retrospective study in which incomplete reporting may have occurred and data might have been missing. Second, the small sample size and the low prevalences of some clinical factors and events may have affected the statistical power to determine associations between rates of complications and potential predictive factors (eg, use of anticoagulants and presence of inflammatory bowel disease). Third, confounding factors might have been present as mentioned earlier in the Discussion, and could not be controlled because of the retrospective nature of this study. However, this study did identify factors that clinicians could use to predict the occurrence of late proctitis. The significant association of V70 with late proctitis should be applied to radiotherapy planning, in that high doses to the rectal volume should be limited where possible.

In summary, late rectal toxicities were frequent among patients in this study in Hong Kong. The occurrence of late proctitis was associated with age, V50, V60, and V70; the occurrence overall late rectal toxicities was associated with a history of haemorrhoid, PTV dose, and V70. Multivariable regression analysis suggested that age, V70, and the presence of acute rectal toxicities could predict the occurrence of late proctitis. Clinicians should closely monitor patients for the occurrence of late proctitis if they exhibit these high-risk factors.

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: BYH Ng, ACK Cheng.
Acquisition of data: BYH Ng.
Analysis or interpretation of data: BYH Ng, ELM Yu, TTS Lau.
Drafting of the article: BYH Ng, ELM Yu, TTS Lau, KS Law.
Critical revision for important intellectual content: BYH Ng, ELM Yu, KS Law, ACK Cheng.

All authors have disclosed no conflict of interest.

The initial abstract was presented at the “ESTRO meets Asia” Conference 2019, Singapore, 6-8 December 2019.

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

This study was approved by the Kowloon West Cluster research ethics committee (Ref KW/EX-19-020(131-08)) and the requirement for patient consent was waived by the committee.

1. Bolla M, Collette L, Blank L, et al. Long-term results with immediate androgen suppression and external irradiation in patients with locally advanced prostate cancer (an EORTC study): a phase III randomised trial. Lancet 2002;360:103-6. Crossref

2. Pilepich MV, Winter K, Lawton C, et al. Androgen suppression adjuvant to definitive radiotherapy in prostate carcinoma—long-term results of phase III RTOG 85-31. Int J Radiat Oncol Biol Phys 2005;61:1285-90. Crossref

3. Michalski J, Moughan J, Purdy J, et al. Effect of standard vs dose-escalated radiation therapy for patients with intermediate-risk prostate cancer: The NRG Oncology RTOG 0126 randomized clinical trial. JAMA Oncol 2018;4:e180039. Crossref

4. Dearnaley D, Syndikus I, Mossop H, et al. Conventional versus hypofractionated high-dose intensity-modulated radiotherapy for prostate cancer: 5-year outcomes of the randomized, non inferiority, phase 3 CHHiP trial. Lancet Oncol 2016;17:1047-60. Crossref

5. Aluwini S, Pos F, Schimmel E, et al. Hypofractionated versus conventionally fractionated radiotherapy for patients with prostate cancer (HYPRO): acute toxicity results from a randomised non-inferiority phase 3 trial. Lancet Oncol 2015;16:274-83. Crossref

6. Zelefsky MJ, Levin EJ, Hunt M, et al. Incidence of late rectal and urinary toxicities after three-dimensional conformal radiotherapy and intensity-modulated radiotherapy for localized prostate cancer. Int J Radiat Oncol Biol Phys 2008;70:1124-9. Crossref

7. Heemsbergen WD, Peeters ST, Koper PC, Hoogeman MS, Lebesque JV. Acute and late gastrointestinal toxicity after radiotherapy in prostate cancer patients: consequential late damage. Int J Radiat Oncol Biol Phys 2006;66:3-10. Crossref

8. Poon DM, Chan SL, Leung CM, et al. Efficacy and toxicity of intensity-modulated radiation therapy for prostate cancer in Chinese patient. Hong Kong Med J 2013;19:407-15. Crossref

9. Delobel J, Gnep K, Ospina JD, et al. Nomogram to predict rectal toxicity following prostate cancer radiotherapy. PLoS One 2017;12:e0179845. Crossref

10. Feng M, Hanlon AL, Pisansky T, et al. Predictive factors for late genitourinary and gastrointestinal toxicity in patients with prostate cancer treated with adjuvant or salvage radiotherapy. Int J Radiat Oncol Biol Phys 2007;68:1417-23. Crossref

11. Fellin G, Fiorino C, Rancati T, et al. Clinical and dosimetric predictors of late rectal toxicity after conformal radiation for localized prostate cancer: results of a large multicenter observational study. Radiother Oncol 2009;93:197-202. Crossref

12. Fuentes-Raspall R, Inoriza J, Rosello-Serrano A, Auñón- Sanz C, Garcia-Martin P, Oliu-Isern G. Late rectal and bladder toxicity following radiation therapy for prostate cancer: predictive factors and treatment results. Rep Pract Oncol Radiother 2013;18:298-303. Crossref

13. Takeda K, Ogawa Y, Ariga H, et al. Clinical correlations between treatment with anticoagulants/antiaggregants and late toxicity after radiotherapy for prostate cancer. Anticancer Res 2009;29:1831-4.

14. Sanguineti G, Agostinelli S, Foppiano S, et al. Adjuvant androgen deprivation impacts late rectal toxicity after conformal radiotherapy of prostate carcinoma. Br J Cancer 2002;86:1743-7. Crossref

15. Fuentes-Raspall R, Inoriza J, Martí-Utzet MJ, Auñón-Sanz C, Garcia-Martin P, Oliu-Isern G. Hyperbaric oxygen therapy for late rectal and bladder toxicity after radiation in prostate cancer patients. A symptom control and qualityof- life study. Clin Oncol (R Coll Radiol) 2012;24:e126. Crossref

16. Jones K, Evans A, Bristow R, Levin W. Treatment of radiation proctitis with hyperbaric oxygen. Radiother Oncol 2006;78:91-4. Crossref

17. Kochhar R, Sriram PV, Sharma SC, Goel RC, Patel F. Natural history of late radiation proctosigmoiditis treated with topical sucralfate suspension. Dig Dis Sci 1999;44:973-8. Crossref

18. Talley NA, Chen F, King D, Jones M, Talley NJ. Shortchain fatty acids in the treatment of radiation proctitis: a randomized double-blind, placebo-controlled, cross-over pilot trial. Dis Colon Rectum 1997;40:1046-50. Crossref

19. Denton A, Andreyev H, Forbes A, Maher EJ. Systematic review for non-surgical interventions for the management of late radiation proctitis. Br J Cancer 2002;87:134-43. Crossref

20. American Joint Committee on Cancer, Prostate Cancer Staging, 7th edition. Available from: https://cancerstaging.org/references-tools/quickreferences/Documents/ProstateSmall.pdf. Accessed 1 Apr 2019.

21. US Department of Health and Human Services, National Institutes of Health, National Cancer Institute, US Government. Common Terminology Criteria for Adverse Events (CTCAE) version 4.03. 2010. Available from: https://ctep.cancer.gov/protocolDevelopment/electronic_applications/ctc.htm#ctc_40. Accessed 1 Apr 2019.

22. Abramowitz M, Li T, Buyyounouski MK, et al. The Phoenix definition of biochemical failure predicts for overall survival in patients with prostate cancer. Cancer 2008;112:55-60. Crossref

23. Bansai N, Soni A, Kaur P, Chauhan AK. Exploring the management of radiation proctitis in current clinical practice. J Clin Diagn Res 2016;10:XE01-XE06. Crossref

24. Skwarchuk MW, Jackson A, Zelefsky MJ, et al. Late rectal toxicity after conformal radiotherapy of prostate cancer (I): multivariate analysis and dose-response. Int J Radiat Oncol Biol Phys 2000;47:103-13. Crossref

25. Herold DM, Hanlon AL, Hanks GE. Diabetes mellitus: a predictor for late radiation morbidity. Int J Radiat Oncol Biol Phys 1999;43:475-9. Crossref

26. Miralbell R, Taussky D, Rinaldi O, et al. Influence of rectal volume changes during radiotherapy for prostate cancer: a predictive model for mild-to-moderate late rectal toxicity. Int J Radiat Oncol Biol Phys 2003;57:1280-4. Crossref

With the ageing of the Hong Kong population, clinicians must monitor increasing numbers of patients with neurodegenerative disease. Progressive supranuclear palsy (PSP) is a common form of atypical parkinsonism,1 with a prevalence of up to 18 cases per 100 000 people.1 Pathologically, PSP is characterised by the presence of neurofibrillary tangles, neuropil threads, or both in the basal ganglia and brainstem.1 In patients with classical Richardson’s syndrome, the disease is characterised by early postural instability, falls, vertical gaze palsy, parkinsonism with poor response to levodopa, pseudobulbar palsy, and frontal release signs.2 Increasingly, patients with PSP have been reported to exhibit the following manifestations: parkinsonism, progressive gait freezing, corticobasal syndrome, apraxia of speech, frontal presentation, or cerebellar ataxia.2 Despite advancements in understanding the disease, there remains no approved treatment for PSP.

In addition to the need for accurate diagnosis of PSP, the natural clinical course of the disease is a concern for caregivers.3 4 Important problems for patients with PSP include increased risks of falls, dysphagia, aspiration pneumonia, pressure injuries, caregiver stress leading to institutionalisation, and long-term mortality.1 Given the lack of definitive treatment, it remains prudent for clinicians to educate caregivers regarding the natural course of PSP, which will ensure that caregivers are better prepared to care for their relatives; it will also allow implementation of different methods to avoid long-term complications, and may facilitate discussions of advanced care planning (ACP) at earlier stages of disease.3 In particular, clinicians need to identify the appropriate timing to discuss ACP. Previous studies have shown that the mean age at onset of PSP is 61 to 67.2 years, and that the disease affects both sexes equally; moreover, the median survival ranges from 5.3 to 10.2 years.5 Factors predictive of mortality included age at onset, early clinical milestones (eg, falls, vertical gaze palsy, neck or limb stiffness, dysphagia, and incontinence), cognitive impairment, language impairment, autonomic dysfunction, male sex, and certain subtypes of PSP, such as classical Richardson’s syndrome and pneumonia.6 7 8 9 10 11 12 13 14 15 16

To the best of our knowledge, there have been no studies of the natural clinical course of disease in Chinese patients with PSP. This information is important for clinical treatment and the design of future intervention studies (eg, for the purposes of sample size estimation). In the present study, we hypothesised that bulbar symptoms and pneumonia could predict mortality in patients with PSP. The aims of this study were (1) to calculate the prevalences of motor symptoms, cognitive symptoms, bulbar symptoms, other systemic symptoms, and long-term outcomes (eg, falls, tube feeding, pressure injuries, and institutionalisation) during the clinical course of PSP, and (2) to identify factors predictive of mortality among patients with PSP.

This retrospective study protocol was approved by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (HKU/HA HKW IRB; Approval No. UW 17-483) and New Territories West Cluster Clinical and Research Ethics Committee (NTWC CREC; Approval No. NTWC/CREC/17127); the requirement for informed consent was waived by the review board. This study comprised a retrospective review of the clinical records of all patients who presented to the geriatrics clinics of Queen Mary Hospital and Tuen Mun Hospital between 1 January 2008 and 30 December 2017. All patients had at least 1 year of clinical follow-up and fulfilled the latest Movement Disorder Society Criteria for clinical diagnosis of PSP.2 In addition, magnetic resonance imaging scans showed radiological evidence of midbrain atrophy (Hummingbird sign or Morning Glory sign) in all patients, according to radiological reports prepared by licensed radiologists.1 Twenty-five patients with clinically probable PSP and radiological evidence of midbrain atrophy were considered for inclusion in this study. Four patients were excluded because their clinical history and radiology findings were not suggestive of PSP. Finally, 21 patients were included: 19 had Richardson syndrome variant, one had PSP-corticobasal syndrome, and one had PSP with language impairment. The patient who had PSP with language impairment was previously described.17 Seven and 14 patients were recruited from Queen Mary Hospital and Tuen Mun Hospital, respectively.

Clinical information was retrospectively retrieved from clinical records, including age at onset, age at presentation, age at death, duration of symptoms, level of education, sex, presenting scores on Cantonese version of Mini-Mental State Examination (C-MMSE),18 clinical symptoms, and history of levodopa or dopamine agonists intake or response. Clinical symptoms were clustered into the following categories: motor symptoms (including limb or neck stiffness, slowness of movement, balance impairment, gait impairment, falls, tremor, and vertical gaze palsy), bulbar symptoms (including dysarthria, dysphagia, and drooling), cognitive symptoms (including memory impairment, apathy, apraxia, dysexecutive syndrome, behavioural disinhibition, repetitive motor behaviour, hyperorality, and visual hallucination), and others (including faecal or urinary incontinence, constipation, insomnia, depression, and caregiver stress).6 The dates of development of the above symptom clusters were retrospectively determined.

‘Age at presentation’ was defined as the age at which the patient first presented to the geriatrics clinic. ‘Duration of symptoms’ was defined as the time between the first appearance of clinical symptoms of neurodegenerative disease and the first presentation. ‘Age at onset’ was defined as the difference between the ‘age at presentation’ and the ‘duration of symptoms’. ‘Disease duration’ was defined as the difference between ‘age at onset’ and ‘age at death’ or the last date of follow-up. ‘Time to diagnosis’ was defined as the time between the date of disease onset and the date of diagnosis with PSP. The times to development of the above categories of symptoms were calculated in relation to both disease onset and presentation.

Long-term outcomes were recorded, including falls, dysphagia, pneumonia, pressure sore development, and mortality. ‘Time to event’ was defined by the difference between the onset of clinical symptoms and first appearance of these long-term events. The time to each event from the time of first presentation was also calculated.

These were defined as events that resulted in the patient’s body or body part inadvertently coming to rest on the ground or other surface lower than the body. The dates and numbers of falls were recorded. Geriatric day hospital training was recorded, including the pre-/post-training elderly mobility scale.19 Parkinsonism medications were often titrated in the geriatric day hospital.

A diagnosis of pneumonia was made based on the following criteria: clinical signs and symptoms, white cell count of ≥10×10⁹/L or proportion of neutrophils of ≥80%, fever (body temperature of ≥37.6℃), and new infiltrates or consolidations on chest radiography (X-ray or computed tomography).20 The dates and total numbers of pneumonia diagnoses were recorded.

Any documentation of dysphagia by a speech therapist was recorded; alternatively (if available), reports of video fluoroscopic swallowing studies were obtained. Penetration was defined as the entry of barium material into the airway without passage below the vocal cords; aspiration was defined as the passage of barium material below the level of the vocal cords.21 The dates of diagnosis of dysphagia and tube feeding were recorded.

The locations of pressure injuries and their stages, according to National Pressure Ulcer Advisory Panel guidelines, were recorded.22 The dates of discovery of pressure injuries were recorded.

This was defined as institutionalisation of the patient, regardless of the level of care (eg, personal care facility or health care facility). The dates of institutionalisation were recorded where possible.

The date and cause of death were recorded.

For descriptive statistics, continuous variables with normal distributions were expressed as means ± standard deviations; variables that did not exhibit a normal distribution were expressed as medians (interquartile ranges). Symptom prevalences (cumulative incidences) were estimated using Kaplan-Meier method, with the first presentation defined as time zero. Patients who did not exhibit a particular symptom by the most recent assessment were censored at that point. Median times to clinical events were used as cut-offs to define ‘early’ or ‘late’ development of those events (binary classification). Time zero was consistently defined as the first clinical presentation or onset of disease, while the event time was defined as the number of months from first presentation or disease onset to occurrence of the event. Cox proportional hazards modelling was used to identify factors predictive of mortality based on the above binary classification or clinical events as time-dependent covariates. Pearson correlation coefficients were used to study the correlations between age at presentation and time to mortality (from the date of presentation). A two-tailed P value of <0.05 was considered statistically significant. All statistical analyses were performed using SPSS for Windows (version 24; IBM Corp, Armonk [NY], United States).

Twenty-one patients were included in this analysis, with a total of 1671.4 months of follow-up from onset (1428.4 months from presentation). The baseline demographics are summarised in Table 1. The mean age of the patients at presentation was 67.6 ± 9.4 years, and most patients were men (76.2%). Fifty-seven percent of the patients received another diagnosis at the time of presentation: nine patients were diagnosed with Parkinson’s disease, one patient was diagnosed with Lewy body dementia, one patient was diagnosed with cervical myelopathy, and one patient was diagnosed with myasthenia gravis. None of the patients showed improvement when treated with levodopa or a dopamine agonist. Six patients (28.6%) exhibited dementia at the time of presentation. Seventeen patients (81%) had been referred to the geriatric day hospital for rehabilitation after a median of 20 months from presentation; they showed improvement in elderly mobility scale score (pre-geriatric day hospital elderly mobility scale score vs post-geriatric day hospital elderly mobility scale score: 14 ± 4.6 vs 16 ± 4.3, respectively, P=0.02). Fifteen patients (71.4%) died during follow-up with mean survival of 6.1 ± 2.6 years from onset (5.2 ± 3.2 years from presentation); 12 of these 15 patients (80%) died of pneumonia, while two (13.3%) died of sudden cardiac arrest.

Among the categories of potential symptoms, motor symptoms were most prevalent during initial presentation (specific symptoms affected up to 33.3% of patients) [Fig 1]. Motor manifestations were among the earliest clinical features observed in patients with PSP (online supplementary Appendix). The most frequent motor symptoms at the time of presentation were limb stiffness (33.3%) and gait impairment (28.6%). Vertical gaze palsy was present in 19% of patients at the time of presentation, but eventually affected all patients (100%). The prevalence of gait impairment increased rapidly, such that ≥80% of the patients were affected within 3 years after presentation. All motor features showed increased in prevalence over time, with final prevalences ranging from 47.6% to 100%. Regarding bulbar symptoms, dysarthria was the most frequent presenting symptom (9.5%). Both dysarthria and dysphagia reached 100% prevalence over time (Fig 1).

Regarding cognitive symptoms, memory impairment and apathy were the two most frequent presenting symptoms, with prevalences of 33.3% and 23.8%, respectively (Fig 2). The respective prevalences of memory impairment and apathy increased to >65% and >45% over time. The prevalence of dysexecutive syndrome reached 28.6% during the clinical course of disease. Other cognitive symptoms relevant to patients with PSP showed lower prevalence, including apraxia, behavioural disinhibition, repetitive motor behaviour, hyperorality, and visual hallucination; the highest prevalence for any of these symptoms was 9.5% throughout the clinical course of disease. The degree of caregiver stress also increased with progression of disease, such that it reached approximately 40% within 5 years after initial presentation.

Regarding systemic symptoms, faecal and urinary incontinence showed the highest prevalences (both reached approximately 40%), particularly in the later stages of disease (Fig 2). Regarding long-term outcomes, the prevalences of aspiration pneumonia and dysphagia requiring Ryle’s tube insertion both reached 100% over time (Fig 2).

Using median time to clinical events as cut-off (binary classification) [online supplementary Appendix] and with analysis in a Cox proportional hazards model, our results showed that earlier development of vertical gaze palsy (hazard ratio [HR]=4.4, 95% confidence interval [CI]=1.4-13.9, P=0.01) and earlier development of pneumonia (HR=10.9, 95% CI=2.2-53.3, P=0.003) were predictive of mortality from disease onset (Table 2). Multivariate analysis showed that earlier development of vertical gaze palsy (HR=3.8, 95% CI=1.1-13.0, P=0.04) and earlier development of pneumonia (HR=10.4, 95% CI=1.9-55.7, P=0.006) were predictive of mortality from disease onset (Table 2). Earlier development of vertical gaze palsy (HR=3.3, 95% CI=1.1-10.5, P=0.01), earlier development of dysphagia (HR=3.7, 95% CI=1.0-12.9, P=0.04), and earlier development of pneumonia (HR=10.6, 95% CI=2.2-52.0, P=0.004) were predictive of mortality from disease presentation (Table 2). Multivariate analysis showed that only earlier development of pneumonia (HR=9.9, 95% CI=1.9-51.8, P=0.007) was predictive of mortality from presentation (Table 2). The number of episodes of pneumonia was also predictive of mortality in patients with PSP, indicating that pneumonia is a major cause of mortality (Table 2). There was a significant negative correlation between the age at presentation and time to mortality from presentation (Pearson correlation=-0.54, P=0.04).

Using clinical events as time-dependent covariates in Cox modelling for prediction of mortality, we found that apathy, dysphagia, Ryle’s tube feeding, pneumonia, and pressure injuries were predictive of mortality from both disease onset and presentation (Table 3). Caregiver stress was only predictive of mortality from presentation (Table 3).

An accurate diagnosis of PSP is important for management of the disease in affected patients. However, only 43% of the patients in this study received a correct diagnosis at the time of initial presentation. This is potentially because vertical gaze palsy was not present initially and only developed during clinical follow-up (median time to develop, 19.6 months; online supplementary Appendix). In addition, 43% of patients were initially misdiagnosed with Parkinson’s disease; this group of patients may have had PSP with parkinsonism.2 Clinicians should regularly assess patients with parkinsonism for the presence of any vertical gaze palsy or poor response to levodopa, in order to correctly identify patients with PSP. Our reported mean time to diagnosis of 3 years was similar to the duration reported in previous studies (mean, 3.1-4 years).1 2 5 6 7 8 9 10 11 12 13 14 15 16 17 23 24 25 26 27 28

With regard to clinical features, our cohort of patients exhibited early development of motor symptoms, which was consistent with previous studies.1 2 5 6 7 8 9 10 11 12 13 14 15 16 17 23 24 25 26 27 28 Our cohort of patients showed evidence of improved mobility, as reflected by changes in elderly mobility scale score after training in the geriatric day hospital, with a median time to referral of 20 months from the initial clinical presentation. Patients with PSP should be referred to a physiotherapist and an occupational therapist for fall assessment, as well as guidance regarding the potential need for walking aids. Relatives should be educated to ensure close monitoring of environmental risks for falls. The patients in our cohort showed relatively early development of memory problems at a median duration of 9 months, which contrasted with the mean duration of 12 months observed in another study.6 There may have been bias in the current cohort, which recruited patients with PSP from a geriatrics clinic whereas the patients in the previous study were recruited from a neurology clinic.6

A previous meta-analysis showed mixed results with regard to whether the age at onset of PSP was predictive of mortality.5 Pooled results from six studies showed no prognostic effect of yearly increases in age at disease onset in either univariate or multivariate analyses.5 However, pooled results from nine studies using median age at onset as cut-off showed a pooled HR of 1.75 (95% CI=1.32-2.32) in multivariate analysis.5 Other predictors of mortality found in the present study, including vertical gaze palsy, dysphagia, pneumonia, or pressure injuries, were previously reported in other studies.6 7 8 9 10 11 12 13 14 15 16 It remains unknown whether resolution of dysphagia in patients with PSP can prevent pneumonia and reduce mortality. It is important for clinicians to refer patients with PSP involving dysphagia to a speech therapist for advice regarding food texture and appropriate swallowing posture (ie, chin-tuck) [Table 4].29 During the clinical course of disease, approximately 40% of primary caregivers complained of caregiver stress, which was also determined to be a significant predictor of mortality. Patient aggression and depression have been reported as sources of stress.4 When these positive predictors of mortality appear, clinicians should consider discussing ACP with patients or caregivers (Table 4).

The finding of apathy as a time-dependent covariate for prediction of mortality is notable. Recently, apathy was found to predict survival in a cohort of 124 patients with syndromes associated with frontotemporal lobar degeneration (including 35 patients with PSP, mean age of 72.2 ± 8.5 years).30 The development of apathy in patients with PSP was related to brainstem, midbrain, and frontal atrophy.30 It remains unknown whether apathy accelerates the decline to death or indirectly signifies the degree of brainstem degeneration, including the development of dysphagia, which is also related to greater mortality. Future clinical trials may consider the use of therapeutic measures to address apathy, in order to assess their impacts on the survival of patients with PSP.

Because there is currently no disease-modifying treatment for patients with PSP, ACP is an important component of clinical care, for which patients and their caregivers can reach a consensus during the clinical course of disease.4 In addition, symptomatic care plays an important role. Symptoms relevant to patients with PSP include dystonia, drooling, gaze palsy (also known as reduced blinking), constipation, and apathy.4 Drooling could be managed by the administration of sublingual atropine drops (Table 4).4 Reduced blinking could be managed by frequent application of lubricating eyedrops. Gaze palsy could be managed by the use of prisms or audiobooks (Table 4).4 Apathy could be minimised by addressing sensory deficits, such as through the use of eyeglasses or hearing aids (Table 4).4 Up to 75% of patients with PSP could be discharged home after stabilisation of symptoms.4

There are multiple limitations in the current study. First, it was a retrospective study involving reviews of clinical charts, and may be biased due to inconsistent documentation of symptoms. Second, there was a limited number of patients included, none of whom had autopsy and pathological confirmation of their diagnosis; however, we had radiological evidence of PSP. Third, our descriptive statistical results should be regarded as preliminary findings; only limited variables could be included in our Cox proportional hazards modelling for survival analysis. Notably, no specific scales were used to assess the severity of parkinsonism or other symptoms, including response to levodopa; most of our evaluations were subjective. Sequential C-MMSE scores were not recorded; thus, we were unable to examine the development of dementia over time. Fourth, we only included patients attending the geriatrics clinic; therefore, our cohort may not be fully representative of patients with PSP who present to most neurology clinics. Finally, the limited numbers of patients precluded stratified analyses based on subtypes of PSP.

In conclusion, important clinical milestones, including the development of dysphagia, vertical gaze palsy, significant caregiver stress, pressure injuries, and pneumonia, may be used to guide the ideal timing for discussions of ACP for patients with PSP, in order to facilitate long-term care.

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: YF Shea, ACK Shum.
Acquisition of data: YF Shea, ACK Shum.
Analysis or interpretation of data: All authors.
Drafting of the article: YF Shea, ACK Shum.
Critical revision for important intellectual content: All authors.

The authors have no conflicts of interest to disclose.

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

This retrospective study protocol was approved by the Institutional Review Board of the University of Hong Kong/ Hospital Authority Hong Kong West Cluster (HKU/HA HKW IRB; Ref UW 17-483) and New Territories West Cluster Clinical and Research Ethics Committee (NTWC CREC; Ref NTWC/CREC/17127). The requirement for informed consent was waived by the review board.

1. Boxer AL, Yu JT, Golbe LI, Litvan I, Lang AE, Höglinger GU. Advances in progressive supranuclear palsy: new diagnostic criteria, biomarkers, and therapeutic approaches. Lancet Neurol 2017;16:552-63. Crossref

2. Höglinger GU, Respondek G, Stamelou M, et al. Clinical diagnosis of progressive supranuclear palsy: The movement disorder society criteria. Mov Disord 2017;32:853-64. Crossref

3. Luk JK, Chan FH. End-of-life care for advanced dementia patients in residential care home—a Hong Kong perspective. Ann Palliat Med 2018;7:359-64. Crossref

4. Wiblin L, Lee M, Burn D. Palliative care and its emerging role in multiple system atrophy and progressive supranuclear palsy. Parkinsonism Relat Disord 2017;34:7-14. Crossref

5. Glasmacher SA, Leigh PN, Saha RA. Predictors of survival in progressive supranuclear palsy and multiple system atrophy: a systematic review and meta-analysis. J Neurol Neurosurg Psychiatry 2017;88:402-11. Crossref

6. Arena JE, Weigand SD, Whitwell JL, et al. Progressive supranuclear palsy: progression and survival. J Neurol 2016;263:380-9. Crossref

7. Golbe LI, Davis PH, Schoenberg BS, Duvoisin RC. Prevalence and natural history of progressive supranuclear palsy. Neurology 1988;38:1031-4. Crossref

8. Nath U, Ben-Shlomo Y, Thomson RG, Lees AJ, Burn DJ. Clinical features and natural history of progressive supranuclear palsy: a clinical cohort study. Neurology 2003;60:910-6. Crossref

9. Oliveira MC, Ling H, Lees AJ, Holton JL, De Pablo-Fernandez E, Warner TT. Association of autonomic symptoms with disease progression and survival in progressive supranuclear palsy. J Neurol Neurosurg Psychiatry 2019;90:555-61. Crossref

10. Cosseddu M, Benussi A, Gazzina S, et al. Natural history and predictors of survival in progressive supranuclear palsy. J Neurol Sci 2017;382:105-7. Crossref

11. Chiu WZ, Kaat LD, Seelaar H, et al. Survival in progressive supranuclear palsy and frontotemporal dementia. J Neurol Neurosurg Psychiatry 2010;81:441-5. Crossref

12. Dell’Aquila C, Zoccolella S, Cardinali V, et al. Predictors of survival in a series of clinically diagnosed progressive supranuclear palsy patients. Parkinsonism Relat Disord 2013;19:980-5. Crossref

13. Santacruz P, Uttl B, Litvan I, Grafman J. Progressive supranuclear palsy: a survey of the disease course. Neurology 1998;50:1637-47. Crossref

Litvan I, Mangone CA, McKee A, et al. Natural history of progressive supranuclear palsy (Steele-Richardson-Olszewski syndrome) and clinical predictors of survival: a clinicopathological study. J Neurol Neurosurg Psychiatry 1996;60:615-20. Crossref

15. O’Sullivan SS, Massey LA, Williams DR, et al. Clinical outcomes of progressive supranuclear palsy and multiple system atrophy. Brain 2008;131:1362-72. Crossref

16. Papapetropoulos S, Gonzalez J, Mash DC. Natural history of progressive supranuclear palsy: a clinicopathologic study from a population of brain donors. Eur Neurol 2005;54:1-9. Crossref

17. Shea YF, Ha J, Chu LW. Progressive supranuclear palsy presenting initially as semantic dementia. J Am Geriatr Soc 2014;62:2459-60. Crossref

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

19. Prosser L, Canby A. Further validation of the elderly mobility scale for measurement of mobility of hospitalized elderly people. Clin Rehabil 1997;11:338-43. Crossref

20. Tomita S, Oeda T, Umemura A, et al. Impact of aspiration pneumonia on the clinical course of progressive supranuclear palsy: a retrospective cohort study. PLoS One 2015;10:e0135823. Crossref

21. Rosenbek JC, Robbins JA, Roecker EB, Coyle JL, Wood JL. A penetration-aspiration scale. Dysphagia 1996;11:93-8. Crossref

22. Edsberg LE, Black JM, Goldberg M, McNichol L, Moore L, Sieggreen M. Revised national pressure ulcer advisory panel pressure injury staging system: revised pressure injury staging system. J Wound Ostomy Continence Nurs 2016;43:585-97. Crossref

23. Testa D, Monza D, Ferrarini M, Soliveri P, Girotti F, Filippini G. Comparison of natural histories of progressive supranuclear palsy and multiple system atrophy. Neurol Sci 2001;22:247-51. Crossref

24. Diroma C, Dell’Aquila C, Fraddosio A, et al. Natural history and clinical features of progressive supranuclear palsy: a clinical study. Neurol Sci 2003;24:176-7. Crossref

25. Goetz CG, Leurgans S, Lang AE, Litvan I. Progression of gait, speech and swallowing deficits in progressive supranuclear palsy. Neurology 2003;60:917-22. Crossref

26. Jecmenica-Lukic M, Petrovic IN, Pekmezovic T, Kostic VS. Clinical outcomes of two main variants of progressive supranuclear palsy and multiple system atrophy: a prospective natural history study. J Neurol 2014;261:1575-83. Crossref

27. Litvan I, Kong M. Rate of decline in progressive supranuclear palsy. Mov Disord 2014;29:463-8. Crossref

28. Ou R, Liu H, Hou Y, et al. Executive dysfunction, behavioral changes and quality of life in Chinese patients with progressive supranuclear palsy. J Neurol Sci 2017;380:182-6. Crossref

29. Luk JK, Chan DK. Preventing aspiration pneumonia in older people: do we have the ‘know-how’? Hong Kong Med J 2014;20:421-7. Crossref

30. Lansdall CJ, Coyle-Gilchrist IT, Vázquez Rodríguez P, et al. Prognostic importance of apathy in syndromes associated with frontotemporal lobar degeneration. Neurology 2019;92:e1547-57. 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 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