EJINME-02737; No of Pages 5 European Journal of Internal Medicine xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
European Journal of Internal Medicine journal homepage: www.elsevier.com/locate/ejim
Original Article
Timing of acute kidney injury — does it matter? A single-centre experience from the United Kingdom Ching Ling Pang ⁎, Dimitrios Chanouzas, Jyoti Baharani Heart of England NHS Foundation Trust, United Kingdom
a r t i c l e
i n f o
Article history: Received 20 January 2014 Received in revised form 23 May 2014 Accepted 4 June 2014 Available online xxxx Keywords: Acute kidney injury Nephrology Acute medicine
a b s t r a c t Background: Acute kidney injury (AKI) requiring renal replacement therapy (RRT) is associated with high mortality and long-term dependence on RRT. However, there is limited information about the difference in outcome between patients who develop AKI in the community (c-AKI), and those who develop AKI in hospital (h-AKI). Aim: Identify differences in short- and long-term outcomes between patients admitted with c-AKI and h-AKI who require intermittent haemodialysis, and to identify factors that predict poor outcome. Design & methods: Single-centre, retrospective analysis of 306 patients with AKI who received intermittent haemodialysis between 2009 and 2011. Follow-up: six months. Primary endpoints: patient and renal survival. Secondary endpoints: time on dialysis, length of hospital stay, and admission to the intensive care unit (ICU). Results: Survival for patients in the h-AKI group was significantly lower, at 42.9% (compared to 72%). They had a significantly longer length of stay. However, at 6-month follow-up, the survival benefit of the c-AKI group was no longer significant. Patients with h-AKI were more likely to be dialysis independent at discharge and six months although this result did not reach statistical significance. Independent predictors of survival to discharge within the entire group included: renal/post-renal causes of AKI, younger age, pre-existing diabetes, and c-AKI. The only independent predictor for RRT dependence at discharge and six months was pre-existing chronic kidney disease. Conclusions: h-AKI is associated with high mortality and longer hospital stays during the acute admission. However, h-AKI patients who survive are more likely to be independent of RRT at discharge and follow-up. © 2014 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved.
1. Introduction Acute kidney injury (AKI) is a common clinical syndrome affecting patients in the community and in hospital. The incidence of AKI varies according to the definition used but has been reported to be up to 1811 per million population (pmp) per year [1–3]. The multiple aetiologies and risk factors [4] involved mean that it represents a spectrum of renal dysfunction ranging from small rises in serum creatinine [5], to the need for renal replacement therapy (RRT). Acute kidney injury in hospital inpatients is associated with significantly higher mortality rates both in the intensive care unit (ICU) [6] and amongst hospital inpatients [7,8]. It is also associated with longer hospital stays, and increased costs [9]. The associated mortality from AKI depends on the degree of kidney injury, but has been estimated as high as 26.3% for those in RIFLE class F [10]. This figure rises to over 50% for patients requiring RRT, and up to 70% for those requiring treatment in intensive care [7,11,12]. There is a distinction to be made between AKI that develops in the community (community acquired AKI; c-AKI) and AKI that sets in during an inpatient episode (hospital acquired AKI; h-AKI). It is well established that h-AKI is associated with significant morbidity and mortality that ⁎ Corresponding author. Tel.: +44 7783921024. E-mail address: [email protected] (C.L. Pang).
cannot be ascribed to comorbid conditions [13]. Despite this, identification of patients who develop AKI post-admission is poor in many centres, with a recent National Confidential Enquiry into Patient Outcome and Death (NCEPOD) study [14] finding that diagnosis was delayed or missed in 43% of admissions. Furthermore, for 20% of these patients, the cause was both predictable and preventable. NCEPOD concluded that the majority of cases had been inadequately assessed on admission for the severity of their illness and pre-existing risk factors, and that referral to nephrology services was delayed or did not happen in 20% of cases. The use of scoring systems to identify patients with AKI is well validated. Two such scoring systems are: the Acute Dialysis Quality Initiative's (ADQI) RIFLE [15] scoring system, and the Acute Kidney Injury Network [16] (AKIN) scoring system. Direct comparison [17] of these two scoring systems shows that they identify slightly different groups of patients. AKIN scoring identified 9% more Stage 1 patients than RIFLE scoring; and conversely, RIFLE scoring identified 26.9% more patients with AKI than AKIN. Nonetheless, both are independently associated with increased mortality rates [6,10]. Only a handful of studies have examined differences in outcome between c-AKI and h-AKI. Most of these studies were carried out in developing countries with the majority concluding that h-AKI is associated with higher mortality [18–21], although two studies from Brazil and Saudi Arabia found that c-AKI was linked with reduced survival
http://dx.doi.org/10.1016/j.ejim.2014.06.005 0953-6205/© 2014 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved.
Please cite this article as: Pang CL, et al, Timing of acute kidney injury — does it matter? A single-centre experience from the United Kingdom, Eur J Intern Med (2014), http://dx.doi.org/10.1016/j.ejim.2014.06.005
2
C.L. Pang et al. / European Journal of Internal Medicine xxx (2014) xxx–xxx
compared to h-AKI [5,22]. A recent North American study showed that patients with c-AKI were more likely to have a shorter length of hospital stay, fewer complications and better overall survival compared to h-AKI [23]. To assess the impact of the timing of acute kidney injury on patient outcome following the NCEPOD report, this study aims to identify and quantify differences in outcomes between patients admitted with c-AKI, and those who develop h-AKI during their hospital stay.
Multivariate analysis: Variables with p value ≤ 0.1 on univariate analysis were included in the multivariate models. Logistic regression was performed for binary outcome variables. Variables were entered in a single step. Odds ratios with confidence intervals and p values are reported. Linear regression was employed for log transformed continuous outcome variables. Variables were entered in a single step. B values with confidence intervals and p values are reported.
3. Results 2. Study cohort and methods 3.1. Patient demographics 2.1. Study cohort and patient selection This was a single centre, retrospective observational study. Our hospital operates across three sites, all of which have telephone access to a duty nephrologist. Two of the sites have an intensive care unit (ICU) with the capacity to provide continuous haemofiltration (CVVH); but only the main site has the facilities for intermittent haemodialysis (IHD) with a dedicated inpatient & outpatient nephrology presence. Using the renal patient database (PROTON), we identified a total of 306 consecutive patients that were admitted via one of our three hospital sites and received intermittent haemodialysis (IHD) at some point during their admission in the context of AKI between 2009 and 2011. We included patients with AKI on a background of chronic kidney disease (CKD). Patients were split into two groups: those who already had AKI on admission to hospital (c-AKI), and those in whom AKI developed 48 hours after admission (h-AKI). We identified AKI and differentiated between c-AKI and h-AKI based on the admission creatinine, available creatinine values prior to admission and serial creatinine measurements post-admission according to the AKIN staging system. We did not define AKI based on urine output as this data was not available for c-AKI patients. Data was collected from the renal unit's records, as well as the electronic patient record database (iCare), both of which are continuously updated and include all investigations and blood results performed at all 3 hospital sites, as well as by most GP practices within the trust catchment area. Data collected comprised basic patient demographics, including: age, gender, ethnicity, modified Charlson comorbidity index, the presence of pre-existing diabetes mellitus, and CKD. The presence of previous CKD was coded based on an eGFR value of b 60 ml/min/ 1.73 m2 within available blood results prior to hospital admission. Renal function was assumed to be normal if there were no prior blood tests available, and the patient gave no history of pre-existing kidney disease on admission. The cause of AKI was coded based on clinical information and biopsy results where available. Patients that suffered a pre-renal insult that might have led to clinical sequelae of acute tubular necrosis (ATN) were coded for their initial renal insult in terms of their AKI cause, i.e. pre-renal. Two primary outcomes were measured: patient survival and renal survival (dialysis independence) at discharge and at 6 months. Secondary outcomes evaluated were: length of hospital stay (LOS), the number of days spent on dialysis, and the rate of admission to a step-up unit (either high-dependency unit (HDU) or ICU).
The characteristics of the study population are shown in Table 1. The mean age of the cohort was 68.2 (+/−13.1) years. 67% of patients were male. The modified Charlson index was used as a measure of comorbidity. The majority of patients (62.7%) had a moderate risk score, with an estimated 10-year mortality of 53% or less [24]. Two hundred and fifty (81.7%) patients were admitted with AKI (c-AKI), whilst fifty-six (18.3%) patients developed AKI at least 48 hours after admission (h-AKI) (Table 1). Patients with c-AKI were most likely to present in Stage 1 AKI. By definition, patients in the h-AKI group had normal renal function, or renal function in keeping with their known baseline on admission. Pre-renal causes accounted for the majority of all AKI (70.6%). Renal and post-renal causes accounted for 21.5% and 7.9% of all AKI episodes respectively. The majority of pre-renal causes were multifactorial, and included conditions associated with effective reduced renal perfusion, such as sepsis associated with hypotension and hypovolaemia (Table 2). Most patients were admitted for an acute medical or surgical reason; however, some were admitted for an elective surgical procedure. Patients with h-AKI were significantly more likely to have sepsis (p = 0.012), and a pre-renal cause of AKI (p = 0.018). They were also more likely to present at the satellite sites (p = 0.021) where no on-site renal services are routinely available. There was no difference in the proportion of patients who had pre-existing diabetes mellitus, or CKD. The Charlson index score was also similar between both groups, indicating a similar co-morbid burden.
3.2. Survival and renal recovery outcomes On univariate analysis, patients in the h-AKI group had significantly higher mortality rates (p b 0.001; Table 3), with just 42.9% surviving to discharge. The initial survival advantage of the c-AKI was short-lived, as no difference in survival was detected amongst the survivors of the two groups at six months post-hospital discharge (p = 0.772). Hospital acquired AKI patients that survived had slightly better renal survival as judged by RRT independence at hospital discharge compared to c-AKI patients (Table 3), although this result did not reach statistical significance (p = 0.072). Nevertheless, h-AKI patients that survived to 6 months post-hospital discharge were still more likely to be dialysisindependent compared to c-AKI patients (Table 3), although again, this result did not reach statistical significance (p = 0.056). There was no difference in the mean eGFR amongst the RRT-independent patients between the two groups either at discharge or at 6 months.
2.2. Statistics
3.3. Secondary outcome analysis
All statistical analysis was carried out using SPSS. A p-value threshold of b0.05 was deemed statistically significant. Univariate analysis: Parametric tests (T-test) were used for transformed continuous variables (length of stay and eGFR at discharge). The chi-square test was used for categorical variables. Non-parametric tests (Mann–Whitney U and Spearman's correlation) were used for continuous variables that were not normally distributed.
Patients in the h-AKI group were significantly more likely to have an extended inpatient stay compared to c-AKI (p b 0.001; Table 3). On average, their length of stay was twenty-five days longer when compared to the c-AKI group. There were no significant differences in the length of time spent on haemodialysis; or on the proportion of patients that required admission to an ICU/HDU setting amongst the two groups (Table 3).
Please cite this article as: Pang CL, et al, Timing of acute kidney injury — does it matter? A single-centre experience from the United Kingdom, Eur J Intern Med (2014), http://dx.doi.org/10.1016/j.ejim.2014.06.005
C.L. Pang et al. / European Journal of Internal Medicine xxx (2014) xxx–xxx
3
Table 1 Characteristics of the study population. c-AKI (n = 250)
h-AKI (n = 56)
p-Value
Age (years +/− SD) Female gender (%) Charlson index score +/− SD Ethnicity (% Caucasian) Presence of CKD (%) Presence of DM (%) Presence of sepsis (%)
67.9 +/− 13.3 87 (34.8) 5.05 +/− 2.3 210 (80.6) 92 (36.8) 102 (40.8) 55 (22.0)
70.0 +/− 11.7 14 (25.0) 4.96 +/− 2.4 42 (75.5) 15 (26.8) 20 (35.7) 21 (38.2)
0.455 0.159 0.867 0.395 0.155 0.482 0.012
Cause of AKI Pre-renal (%) Renal (%) Post-renal (%)
169 (67.5) 60 (24.2) 21 (8.3)
47 (83.6) 6 (10.9) 3 (5.5)
0.018 0.031 0.587a
AKIN on admission Stage 1 (%) Stage 2 (%) Stage 3 (%) Presenting hospital (% that presented at main site) Presenting creatinine (mean)
102 (40.8) 41 (16.4) 107 (42.8) 182 (72.8) 541.2
N/A 32 (57.1) 136.7
0.021 N/A
Bold values indicate significant p-value of pb0.05. a Fisher's exact test.
4. Predictors of primary and secondary outcomes for the whole cohort
4.2. Renal recovery and RRT dependence Underlying CKD was an independent predictor of RRT dependence at both discharge and at 6 months (Table 4). The presence of sepsis predicted greater RRT independence at 6 months.
4.1. Patient survival Using a multivariate logistic regression model, the following variables were independent predictors of survival to discharge: a renal or post-renal cause of AKI as opposed to pre-renal cause, younger age, pre-existing diabetes, c-AKI, and the absence of sepsis (Table 4). The only predictor of survival to 6 months was the severity of the comorbid burden, with those patients who had lower Charlson indices being more likely to survive (Table 4).
4.3. Length of stay and length of RRT Using a multivariate linear regression model, h-AKI, and advancing age were independent predictors of an increased length of stay. Amongst patients that recovered renal independence, the presence of underlying CKD, as well as either a renal or a post-renal cause of AKI was associated with a longer length of RRT (Table 5). 5. Discussion
Table 2 Causes of acute kidney injury. Cause
Number of patients (%)
Pre-renal Sepsis associated with hypotension/hypovolaemia Dehydration/fluid losses Post-operative hypovolaemia/hypotension Diarrhoea and vomiting Prolonged hypotension due to cardiac dysfunction/cardiac arrest Prolonged hypotension due to other causes (e.g. ruptured AAA) Hypovolaemia secondary to blood loss Reduced renal perfusion associated with heart failure
216 (70.6) 74 (24.2) 49 (16) 28 (9.2) 20 (6.5) 20 (6.5) 12 (3.9) 8 (2.6) 5 (1.6)
Renal Glomerulonephritis Nephrotoxins (including contrast induced nephropathy) Cast nephropathy secondary to multiple myeloma Rhabdomyolysis Tubulointerstitial nephritis Biopsy-proven acute tubular necrosis with no clear precipitant Haemolytic uraemic syndrome/microangiopathic haemolytic anaemia
66 (21.5) 20 (6.5) 18 (5.9) 9 (2.9) 7 (2.3) 5 (1.6) 5 (1.6)
Post-renal Ureteric obstruction (e.g. secondary to malignancy) Bladder outflow obstruction
24 (7.9) 13 (4.2) 11 (3.6)
2 (0.7)
Inpatient mortality amongst patients that require RRT for AKI remains very high. Our findings are in keeping with data from other studies, where mortality associated with AKI requiring RRT has been reported to be as high as 70% [10]. Multivariate analysis of the entire cohort revealed that h-AKI, sepsis, a pre-renal cause of AKI and advancing age were all independent predictors of inpatient mortality. AKI has been shown to develop in over half of patients presenting with sepsis or septic shock, and mortality is directly correlated with the severity of the AKI [25]. Not having diabetes mellitus also predicted higher inpatient mortality. Patients with diabetes were slightly younger, less likely to have had sepsis and more likely to have had a pre-renal cause of AKI with a shorter duration of RRT (data not shown), possibly suggesting a somewhat less complicated course compared to patients that were not diabetic in this group of patients. We have previously observed a similar apparent survival benefit associated with the presence of diabetes in a cohort of patients with AKI requiring ICU admission and haemofiltration [8]. At 6 months, only a lower co-morbid burden as calculated by Charlson index was an independent predictor of survival. Focusing on differences between h-AKI and c-AKI, h-AKI was associated with a much higher mortality compared to c-AKI. In addition, h-AKI was a strong predictor of inpatient mortality even after controlling for other confounding variables. Only a handful of studies have compared the outcome of h-AKI to that of c-AKI. Furthermore, to the best of our knowledge our study is the first to examine the outcome of AKI requiring RRT in these two groups. Our findings are in agreement with a recent North American study that found c-AKI to be associated with a survival advantage [22]. They are also in agreement with two
Please cite this article as: Pang CL, et al, Timing of acute kidney injury — does it matter? A single-centre experience from the United Kingdom, Eur J Intern Med (2014), http://dx.doi.org/10.1016/j.ejim.2014.06.005
4
C.L. Pang et al. / European Journal of Internal Medicine xxx (2014) xxx–xxx
Table 3 Primary and secondary outcomes at discharge and at 6 months. c-AKI
h-AKI
p-Value
Primary outcomes Survival at hospital discharge (% survived) Dialysis dependence at hospital discharge (% dialysis dependent) eGFR at discharge +/− SD Survival at 6 months (% survived) Dialysis dependence at 6 months (% dialysis dependent) eGFR at 6 months +/− SD
72.0 35.0 33.1 +/− 24.6 82.8 32.2 44.7 +/− 22.9
42.9 16.7 38.2 +/− 25.8 87.5 12.5 52.3 +/− 22.5
b0.001 0.072⁎ 0.266 0.772a 0.056⁎, a 0.325
Secondary outcomes Length of RRT (days +/− SD) Length of stay (days +/− SD) ICU admission (%)
12.4 +/− 30.3 23.9 +/− 19.7 5.6
6.1 +/− 6.9 48.6 +/− 62.6 1.8
0.589 b0.001 0.321a
Bold values indicate significant p-value of pb0.05. ⁎ Difference that approaches statistical significance. a Fisher's exact test.
descriptive studies from India, which describe mortality amongst h-AKI and c-AKI patients, primarily on medical wards, to be 37.2% [26] and 17.39% [27] respectively. Interestingly, two studies from Brazil and Saudi Arabia found that c-AKI was linked with worse survival compared to h-AKI [5,21]. This may reflect specific population characteristics in these patient groups, as well as different access to secondary healthcare, which could potentially delay the presentation of c-AKI in those countries. Survivors in the h-AKI group were less likely to be dialysisdependent, both at discharge and at 6 months, although this difference did not reach statistical significance. The inclusion of an additional 30 patients overall with the same characteristics as observed in our analysis would render these results statistically significant (data not shown). We hypothesise that the renal survival benefit within the h-AKI group was due to a higher proportion of patients with sepsis and a pre-renal cause of AKI and therefore more of a chance to recover renal function amongst survivors. On the contrary, due to the high proportion of sepsis within these patients, they were more likely to have been more systemically unwell, and to have had higher mortality rates. It is also possible that failure to recover renal function might have contributed to higher inpatient mortality in the h-AKI group. Out of the thirty-two patients (57%) in the h-AKI group that died during their hospital admission, twenty-three (71.9%) died within 7 days of receiving the last dialysis session possibly suggesting that a clinical decision might have been taken to withdraw renal support in these patients although we did not have access to patient notes to confirm this. On the other hand, renal causes such as glomerulonephritides were associated with c-AKI. Therefore, c-AKI was linked with worse renal recovery. The frequency of CKD and diabetes was similar in both groups, and therefore unlikely to account for the worse renal recovery of the c-AKI group. This was confirmed in a post hoc analysis described
below. However, pre-existing CKD led to more overall RRT dependence at discharge and at 6 months. This is not a new finding as it has been noted in several studies including a recent observational study of a CKD disease cohort in Canada [28], where patients with CKD were not only more likely to develop AKI, but also more likely to require dialysis after discharge. Importantly, patients with h-AKI spent significantly longer amounts of time in hospital. This is probably a reflection of the severity of their illness, and may reflect the need for ongoing supportive therapy. With the average inpatient stay extended by almost a month; developing h-AKI has implications on patient experience. Finally, as AKI is estimated to account for up to 10% of all ICU bed days [29], AKI has significant logistic and financial implications for hospitals. The development of h-AKI was also associated with presentation in a hospital with no permanent nephrologist, and may reflect delayed referral to and involvement of the nephrology team. A recent pilot study showed improved outcomes for h-AKI with early nephrologist involvement [30]. Furthermore, delayed nephrology referral is linked to higher mortality and dialysis dependence at discharge in critically ill AKI patients [31]. We also carried out a post-hoc analysis after removing all patients with pre-existing CKD in order to determine differences between h-AKI and c-AKI in ‘pure’ AKI. This did not affect the direction or indeed significance of associations compared to our analysis of the whole group (data not shown). Interestingly, the only discrepancy was that by removing the CKD population, h-AKI patients were much more likely to be RRT independent compared to c-AKI patients at 6 months (p = 0.025). The remainder of the associations remained unchanged. One significant limitation of our study is the relatively low number of patients within the h-AKI group compared to the c-AKI group.
Table 4 Independent predictors of survival & RRT dependence. Odds ratio of survival (95% confidence interval)
p-Value
Survival to discharge Renal or post-renal cause of AKI Younger age Presence of diabetes c-AKI No sepsis
3.745 (1.745–8.000) 1.038 (1.012–1.066) 2.398 (1.300–4.425) 2.531 (1.269–5.048) 2.399 (1.246–4.618)
0.001 0.004 0.005 0.008 0.009
Survival to 6 months Lower Charlson index score
1.211 (1.027–1.429)
0.023
RRT dependence at discharge Presence of CKD RRT dependence at 6 months Presence of CKD No sepsis
Odds ratio of RRT dependence (95% confidence interval)
p-Value
3.247 (1.692–6.211)
b0.001
3.359 (1.423–7.929) 9.328 (1.167–74.546)
0.006 0.035
Please cite this article as: Pang CL, et al, Timing of acute kidney injury — does it matter? A single-centre experience from the United Kingdom, Eur J Intern Med (2014), http://dx.doi.org/10.1016/j.ejim.2014.06.005
C.L. Pang et al. / European Journal of Internal Medicine xxx (2014) xxx–xxx Table 5 Independent predictors of length of stay and length of RRT. B value (95% confidence interval)
p-Value
Increased length of stay h-AKI Advancing age
0.581 (0.271–0.891) 0.009 (0.002–0.017)
b0.001 0.013
Increased length of RRT Presence of CKD Renal or post-renal cause of AKI
0.511 (0.148–0.875) 0.509 (0.132–0.886)
0.006 0.008
Multivariate analysis was therefore only carried out within the entire group of 306 patients. Furthermore, as the data was collected retrospectively the quality of data could have been potentially affected. In order to avoid this we used prospectively collected data within the renal and hospital databases wherever possible. 6. Learning points • We have examined short and long-term outcomes in 306 consecutive patients with AKI that required RRT in the form of intermittent haemodialysis. • Patients with h-AKI had significantly higher mortality rates, compared with patients who had c-AKI. They were also more likely to have an extended hospital stay. However, survivors from the h-AKI group were less likely to be dependent on RRT at discharge, and at six months. • Independent predictors of mortality included sepsis, advancing age, a pre-renal cause of AKI, and h-AKI. • Early recognition and management by physicians, with timely input from nephrologists, are likely to be crucial in improving outcomes for patients with h-AKI. 7. Conclusions Acute kidney injury requiring RRT continues to be associated with high mortality, as well as a need for ongoing RRT, reaffirming the notion that AKI does indeed represent a ‘global health alert’ [32]. As we have shown, timing of AKI does matter, and h-AKI requiring RRT is associated with higher mortality, as well as an increased length of hospital stay. There is an ongoing need to further increase awareness of the multiple problems associated with AKI. Early recognition and management by clinicians is crucial, as the majority of AKI secondary to pre-renal causes is often avoidable, but linked with very high mortality. The early involvement of nephrologists is key, as are efforts to improve our diagnosis and prediction of incipient and established AKI with novel biomarkers [33]. Finally, there is a need for ongoing education to improve awareness and implement local strategies aimed at improving early diagnosis of AKI, prompt nephrology referral and appropriate management. References [1] Hegarty J, Middleton R, Krebs M, Hussain H, Cheung C, Ledson T, et al. Severe acute renal failure. Place of care, incidence, and outcomes. QJM 2005;98:661–6. [2] Stevens PE, Tamimi NA, Al Hasani MK, Mikhail AI, Kearney E, Lapworth R, et al. Nonspecialist management of acute renal failure. QJM 2001;94:533–40. [3] Ali T, Khan I, Simpson W, Prescott G, Townend J, Smith W, et al. Incidence and outcomes in acute kidney injury: a comprehensive population-based study. J Am Soc Nephrol 2007;18:1292–8. [4] Liangos O, Wald R, O'Bell JW, Price L, Pereira BJ, Jaber BL. Epidemiology and outcomes of acute renal failure in hospitalized patients: a national survey. Clin J Am Soc Nephrol 2006;1:43–51 [Epub 2005 Oct 26].
5
[5] Praught ML, Shlipak MG. Are small changes in serum creatinine an important risk factor? Curr Opin Nephrol Hypertens 2005;14:265–70. [6] Daher EF, Silva Junior GB, Santos SQ, R Bezerra CC, Diniz EJ, Lima RS, et al. Differences in community, hospital and intensive care unit-acquired acute kidney injury: observational study in a nephrology service of a developing country. Clin Nephrol 2012;78:449–55. http://dx.doi.org/10.5414/CN107167. [7] Meier P, Bonfils RM, Vogt B, Burnand B, Burnier M. Referral patterns and outcomes in noncritically ill patients with hospital-acquired acute kidney injury. Clin J Am Soc Nephrol 2011;6:2215–25. http://dx.doi.org/10.2215/CJN.01880211. Epub 2011 Aug 4. [8] Ng KP, Chanouzas D, Fallouh B, Baharani J. Short and long-term outcome of patients with severe acute kidney injury requiring renal replacement therapy. QJM 2012;105:33–9. [9] Chertow GM, Burdick E, Honour M, Bonventre JV, Bates DW. Acute kidney injury, mortality, length of stay, and costs in hospitalized patients. J Am Soc Nephrol 2005;16:3365–70. [10] Hoste EA, Clermont G, Kersten A, Venkataraman R, Angus DC, De Bacquer D, et al. RIFLE criteria for acute kidney injury are associated with hospital mortality in critically ill patients: a cohort analysis. Crit Care 2006;10:R73. [11] Liaño F, Junco E, Pascual J, Madero R, Verde E. The spectrum of acute renal failure in the intensive care unit compared with that seen in other settings. The Madrid Acute Renal Failure Study Group. Kidney Int 1998;53:S16–24. [12] Karsou SA, Jaber BL, Pereira BK. Impact of intermittent hemodialysis variables on clinical outcomes in acute renal failure. Am J Kidney Dis 2000;35:980–91. [13] Nolan CR, Anderson RJ. Hospital-acquired acute renal failure. JASN 1998;9:710–8. [14] National Confidential Enquiry into Patient Outcome, Death. Acute kidney injury: adding insult to injury; 2009. [15] Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P. Acute renal failure — definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care 2004;8:R204–12. [16] Mehta RL, Kellum JA, Shah SV, Molitoris BA, Ronco C, Warnock DG, et al. Acute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury. Crit Care 2007;11:R31. [17] Joannidis M, Metnitz B, Bauer P, Schusterschitz N, Moreno R, Druml W, et al. Acute kidney injury in critically ill patients classified by AKIN versus RIFLE using the SAPS 3 database. Intensive Care Med 2009;35:1692–702. [18] Wang Y, Cui Z, Fan M. Hospital-acquired and community-acquired acute renal failure in hospitalized Chinese: a ten-year review. Ren Fail 2007;29:163–8. [19] Kohli HS, Bhat A, Jairam A, Aravindan AN, Sud K, Jha V, et al. Predictors of mortality in acute renal failure in a developing country: a prospective study. Ren Fail 2007;29:463–9. [20] Dharan KS, John GT, Antonisamy B, Kirubakaran MG, Jacob CK. Prediction of mortality in acute renal failure in the tropics. Ren Fail 2005;27:289–96. [21] Sesso R, Roque A, Vicioso B, Stella S. Prognosis of ARF in hospitalized elderly patients. Am J Kidney Dis 2004;44:410–9. [22] Al-Homrany M. Epidemiology of acute renal failure in hospitalized patients: experience from southern Saudi Arabia. East Mediterr Health J Sep-Nov 2003;9:1061–7. [23] Schissler MM, Zaidi S, Kumar H, Deo D, Brier ME, McLeish KR. Characteristics and outcomes in community-acquired versus hospital-acquired acute kidney injury. Nephrology (Carlton) 2013;18:183–7. http://dx.doi.org/10.1111/nep.12036. [24] Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–83. [25] Suh SH, Kim CS, Choi JS, Bae EH, Ma SK, Kim SW. Acute kidney injury in patients with sepsis and septic shock: risk factors and clinical outcomes. Yonsei Med J 2013;54:965–72. [26] Singh TB, Rathore SS, Choudhury TA, Shukla VK, Singh DK, Prakash J. Hospitalacquired acute kidney injury in medical, surgical, and intensive care unit: a comparative study. Indian J Nephrol 2013;23:24–9. [27] Prakash J, Singh TB, Ghosh B, Malhotra V, Rathore SS, Vohra R, et al. Changing epidemiology of community-acquired acute kidney injury in developing countries: analysis of 2405 cases in 26 years from eastern India. Clin Kidney J 2013;6:150–5. [28] Lafrance JP, Djurdjev O, Levin A. Incidence and outcomes of acute kidney injury in a referred chronic kidney disease cohort. Nephrol Dial Transplant 2010;25:2203–9. [29] Intensive-Care National Audit Research Centre. www.icnarc.org; 2005. [30] Balasubramanian G, Al-Aly Z, Moiz A, Rauchman M, Zhang Z, Gopalakrishnan R, et al. Early nephrologist involvement in hospital acquired acute kidney injury: a pilot study. Am J Kidney Dis 2011;57:228–34. [31] Costa e Silva VT, Liano F, Muriel A, Diez R, De Castro I, Yu L. Nephrology referral and outcomes in critically ill acute kidney injury patients. PLoS One 2013;8:e70482. [32] Li PK, Burdmann EA, Mehta RL. World Kidney Day Steering Committee 2013: acute kidney injury: global health alert. 2013;83:372–6. [33] Obermuller N, Geiger H, Weipert C, Urbschart A. Current developments in early diagnosis of acute kidney injury. Int Urol Nephrol 2014;46(1):1–7.
Please cite this article as: Pang CL, et al, Timing of acute kidney injury — does it matter? A single-centre experience from the United Kingdom, Eur J Intern Med (2014), http://dx.doi.org/10.1016/j.ejim.2014.06.005
Contents lists available at ScienceDirect
European Journal of Internal Medicine journal homepage: www.elsevier.com/locate/ejim
Original Article
Timing of acute kidney injury — does it matter? A single-centre experience from the United Kingdom Ching Ling Pang ⁎, Dimitrios Chanouzas, Jyoti Baharani Heart of England NHS Foundation Trust, United Kingdom
a r t i c l e
i n f o
Article history: Received 20 January 2014 Received in revised form 23 May 2014 Accepted 4 June 2014 Available online xxxx Keywords: Acute kidney injury Nephrology Acute medicine
a b s t r a c t Background: Acute kidney injury (AKI) requiring renal replacement therapy (RRT) is associated with high mortality and long-term dependence on RRT. However, there is limited information about the difference in outcome between patients who develop AKI in the community (c-AKI), and those who develop AKI in hospital (h-AKI). Aim: Identify differences in short- and long-term outcomes between patients admitted with c-AKI and h-AKI who require intermittent haemodialysis, and to identify factors that predict poor outcome. Design & methods: Single-centre, retrospective analysis of 306 patients with AKI who received intermittent haemodialysis between 2009 and 2011. Follow-up: six months. Primary endpoints: patient and renal survival. Secondary endpoints: time on dialysis, length of hospital stay, and admission to the intensive care unit (ICU). Results: Survival for patients in the h-AKI group was significantly lower, at 42.9% (compared to 72%). They had a significantly longer length of stay. However, at 6-month follow-up, the survival benefit of the c-AKI group was no longer significant. Patients with h-AKI were more likely to be dialysis independent at discharge and six months although this result did not reach statistical significance. Independent predictors of survival to discharge within the entire group included: renal/post-renal causes of AKI, younger age, pre-existing diabetes, and c-AKI. The only independent predictor for RRT dependence at discharge and six months was pre-existing chronic kidney disease. Conclusions: h-AKI is associated with high mortality and longer hospital stays during the acute admission. However, h-AKI patients who survive are more likely to be independent of RRT at discharge and follow-up. © 2014 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved.
1. Introduction Acute kidney injury (AKI) is a common clinical syndrome affecting patients in the community and in hospital. The incidence of AKI varies according to the definition used but has been reported to be up to 1811 per million population (pmp) per year [1–3]. The multiple aetiologies and risk factors [4] involved mean that it represents a spectrum of renal dysfunction ranging from small rises in serum creatinine [5], to the need for renal replacement therapy (RRT). Acute kidney injury in hospital inpatients is associated with significantly higher mortality rates both in the intensive care unit (ICU) [6] and amongst hospital inpatients [7,8]. It is also associated with longer hospital stays, and increased costs [9]. The associated mortality from AKI depends on the degree of kidney injury, but has been estimated as high as 26.3% for those in RIFLE class F [10]. This figure rises to over 50% for patients requiring RRT, and up to 70% for those requiring treatment in intensive care [7,11,12]. There is a distinction to be made between AKI that develops in the community (community acquired AKI; c-AKI) and AKI that sets in during an inpatient episode (hospital acquired AKI; h-AKI). It is well established that h-AKI is associated with significant morbidity and mortality that ⁎ Corresponding author. Tel.: +44 7783921024. E-mail address: [email protected] (C.L. Pang).
cannot be ascribed to comorbid conditions [13]. Despite this, identification of patients who develop AKI post-admission is poor in many centres, with a recent National Confidential Enquiry into Patient Outcome and Death (NCEPOD) study [14] finding that diagnosis was delayed or missed in 43% of admissions. Furthermore, for 20% of these patients, the cause was both predictable and preventable. NCEPOD concluded that the majority of cases had been inadequately assessed on admission for the severity of their illness and pre-existing risk factors, and that referral to nephrology services was delayed or did not happen in 20% of cases. The use of scoring systems to identify patients with AKI is well validated. Two such scoring systems are: the Acute Dialysis Quality Initiative's (ADQI) RIFLE [15] scoring system, and the Acute Kidney Injury Network [16] (AKIN) scoring system. Direct comparison [17] of these two scoring systems shows that they identify slightly different groups of patients. AKIN scoring identified 9% more Stage 1 patients than RIFLE scoring; and conversely, RIFLE scoring identified 26.9% more patients with AKI than AKIN. Nonetheless, both are independently associated with increased mortality rates [6,10]. Only a handful of studies have examined differences in outcome between c-AKI and h-AKI. Most of these studies were carried out in developing countries with the majority concluding that h-AKI is associated with higher mortality [18–21], although two studies from Brazil and Saudi Arabia found that c-AKI was linked with reduced survival
http://dx.doi.org/10.1016/j.ejim.2014.06.005 0953-6205/© 2014 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved.
Please cite this article as: Pang CL, et al, Timing of acute kidney injury — does it matter? A single-centre experience from the United Kingdom, Eur J Intern Med (2014), http://dx.doi.org/10.1016/j.ejim.2014.06.005
2
C.L. Pang et al. / European Journal of Internal Medicine xxx (2014) xxx–xxx
compared to h-AKI [5,22]. A recent North American study showed that patients with c-AKI were more likely to have a shorter length of hospital stay, fewer complications and better overall survival compared to h-AKI [23]. To assess the impact of the timing of acute kidney injury on patient outcome following the NCEPOD report, this study aims to identify and quantify differences in outcomes between patients admitted with c-AKI, and those who develop h-AKI during their hospital stay.
Multivariate analysis: Variables with p value ≤ 0.1 on univariate analysis were included in the multivariate models. Logistic regression was performed for binary outcome variables. Variables were entered in a single step. Odds ratios with confidence intervals and p values are reported. Linear regression was employed for log transformed continuous outcome variables. Variables were entered in a single step. B values with confidence intervals and p values are reported.
3. Results 2. Study cohort and methods 3.1. Patient demographics 2.1. Study cohort and patient selection This was a single centre, retrospective observational study. Our hospital operates across three sites, all of which have telephone access to a duty nephrologist. Two of the sites have an intensive care unit (ICU) with the capacity to provide continuous haemofiltration (CVVH); but only the main site has the facilities for intermittent haemodialysis (IHD) with a dedicated inpatient & outpatient nephrology presence. Using the renal patient database (PROTON), we identified a total of 306 consecutive patients that were admitted via one of our three hospital sites and received intermittent haemodialysis (IHD) at some point during their admission in the context of AKI between 2009 and 2011. We included patients with AKI on a background of chronic kidney disease (CKD). Patients were split into two groups: those who already had AKI on admission to hospital (c-AKI), and those in whom AKI developed 48 hours after admission (h-AKI). We identified AKI and differentiated between c-AKI and h-AKI based on the admission creatinine, available creatinine values prior to admission and serial creatinine measurements post-admission according to the AKIN staging system. We did not define AKI based on urine output as this data was not available for c-AKI patients. Data was collected from the renal unit's records, as well as the electronic patient record database (iCare), both of which are continuously updated and include all investigations and blood results performed at all 3 hospital sites, as well as by most GP practices within the trust catchment area. Data collected comprised basic patient demographics, including: age, gender, ethnicity, modified Charlson comorbidity index, the presence of pre-existing diabetes mellitus, and CKD. The presence of previous CKD was coded based on an eGFR value of b 60 ml/min/ 1.73 m2 within available blood results prior to hospital admission. Renal function was assumed to be normal if there were no prior blood tests available, and the patient gave no history of pre-existing kidney disease on admission. The cause of AKI was coded based on clinical information and biopsy results where available. Patients that suffered a pre-renal insult that might have led to clinical sequelae of acute tubular necrosis (ATN) were coded for their initial renal insult in terms of their AKI cause, i.e. pre-renal. Two primary outcomes were measured: patient survival and renal survival (dialysis independence) at discharge and at 6 months. Secondary outcomes evaluated were: length of hospital stay (LOS), the number of days spent on dialysis, and the rate of admission to a step-up unit (either high-dependency unit (HDU) or ICU).
The characteristics of the study population are shown in Table 1. The mean age of the cohort was 68.2 (+/−13.1) years. 67% of patients were male. The modified Charlson index was used as a measure of comorbidity. The majority of patients (62.7%) had a moderate risk score, with an estimated 10-year mortality of 53% or less [24]. Two hundred and fifty (81.7%) patients were admitted with AKI (c-AKI), whilst fifty-six (18.3%) patients developed AKI at least 48 hours after admission (h-AKI) (Table 1). Patients with c-AKI were most likely to present in Stage 1 AKI. By definition, patients in the h-AKI group had normal renal function, or renal function in keeping with their known baseline on admission. Pre-renal causes accounted for the majority of all AKI (70.6%). Renal and post-renal causes accounted for 21.5% and 7.9% of all AKI episodes respectively. The majority of pre-renal causes were multifactorial, and included conditions associated with effective reduced renal perfusion, such as sepsis associated with hypotension and hypovolaemia (Table 2). Most patients were admitted for an acute medical or surgical reason; however, some were admitted for an elective surgical procedure. Patients with h-AKI were significantly more likely to have sepsis (p = 0.012), and a pre-renal cause of AKI (p = 0.018). They were also more likely to present at the satellite sites (p = 0.021) where no on-site renal services are routinely available. There was no difference in the proportion of patients who had pre-existing diabetes mellitus, or CKD. The Charlson index score was also similar between both groups, indicating a similar co-morbid burden.
3.2. Survival and renal recovery outcomes On univariate analysis, patients in the h-AKI group had significantly higher mortality rates (p b 0.001; Table 3), with just 42.9% surviving to discharge. The initial survival advantage of the c-AKI was short-lived, as no difference in survival was detected amongst the survivors of the two groups at six months post-hospital discharge (p = 0.772). Hospital acquired AKI patients that survived had slightly better renal survival as judged by RRT independence at hospital discharge compared to c-AKI patients (Table 3), although this result did not reach statistical significance (p = 0.072). Nevertheless, h-AKI patients that survived to 6 months post-hospital discharge were still more likely to be dialysisindependent compared to c-AKI patients (Table 3), although again, this result did not reach statistical significance (p = 0.056). There was no difference in the mean eGFR amongst the RRT-independent patients between the two groups either at discharge or at 6 months.
2.2. Statistics
3.3. Secondary outcome analysis
All statistical analysis was carried out using SPSS. A p-value threshold of b0.05 was deemed statistically significant. Univariate analysis: Parametric tests (T-test) were used for transformed continuous variables (length of stay and eGFR at discharge). The chi-square test was used for categorical variables. Non-parametric tests (Mann–Whitney U and Spearman's correlation) were used for continuous variables that were not normally distributed.
Patients in the h-AKI group were significantly more likely to have an extended inpatient stay compared to c-AKI (p b 0.001; Table 3). On average, their length of stay was twenty-five days longer when compared to the c-AKI group. There were no significant differences in the length of time spent on haemodialysis; or on the proportion of patients that required admission to an ICU/HDU setting amongst the two groups (Table 3).
Please cite this article as: Pang CL, et al, Timing of acute kidney injury — does it matter? A single-centre experience from the United Kingdom, Eur J Intern Med (2014), http://dx.doi.org/10.1016/j.ejim.2014.06.005
C.L. Pang et al. / European Journal of Internal Medicine xxx (2014) xxx–xxx
3
Table 1 Characteristics of the study population. c-AKI (n = 250)
h-AKI (n = 56)
p-Value
Age (years +/− SD) Female gender (%) Charlson index score +/− SD Ethnicity (% Caucasian) Presence of CKD (%) Presence of DM (%) Presence of sepsis (%)
67.9 +/− 13.3 87 (34.8) 5.05 +/− 2.3 210 (80.6) 92 (36.8) 102 (40.8) 55 (22.0)
70.0 +/− 11.7 14 (25.0) 4.96 +/− 2.4 42 (75.5) 15 (26.8) 20 (35.7) 21 (38.2)
0.455 0.159 0.867 0.395 0.155 0.482 0.012
Cause of AKI Pre-renal (%) Renal (%) Post-renal (%)
169 (67.5) 60 (24.2) 21 (8.3)
47 (83.6) 6 (10.9) 3 (5.5)
0.018 0.031 0.587a
AKIN on admission Stage 1 (%) Stage 2 (%) Stage 3 (%) Presenting hospital (% that presented at main site) Presenting creatinine (mean)
102 (40.8) 41 (16.4) 107 (42.8) 182 (72.8) 541.2
N/A 32 (57.1) 136.7
0.021 N/A
Bold values indicate significant p-value of pb0.05. a Fisher's exact test.
4. Predictors of primary and secondary outcomes for the whole cohort
4.2. Renal recovery and RRT dependence Underlying CKD was an independent predictor of RRT dependence at both discharge and at 6 months (Table 4). The presence of sepsis predicted greater RRT independence at 6 months.
4.1. Patient survival Using a multivariate logistic regression model, the following variables were independent predictors of survival to discharge: a renal or post-renal cause of AKI as opposed to pre-renal cause, younger age, pre-existing diabetes, c-AKI, and the absence of sepsis (Table 4). The only predictor of survival to 6 months was the severity of the comorbid burden, with those patients who had lower Charlson indices being more likely to survive (Table 4).
4.3. Length of stay and length of RRT Using a multivariate linear regression model, h-AKI, and advancing age were independent predictors of an increased length of stay. Amongst patients that recovered renal independence, the presence of underlying CKD, as well as either a renal or a post-renal cause of AKI was associated with a longer length of RRT (Table 5). 5. Discussion
Table 2 Causes of acute kidney injury. Cause
Number of patients (%)
Pre-renal Sepsis associated with hypotension/hypovolaemia Dehydration/fluid losses Post-operative hypovolaemia/hypotension Diarrhoea and vomiting Prolonged hypotension due to cardiac dysfunction/cardiac arrest Prolonged hypotension due to other causes (e.g. ruptured AAA) Hypovolaemia secondary to blood loss Reduced renal perfusion associated with heart failure
216 (70.6) 74 (24.2) 49 (16) 28 (9.2) 20 (6.5) 20 (6.5) 12 (3.9) 8 (2.6) 5 (1.6)
Renal Glomerulonephritis Nephrotoxins (including contrast induced nephropathy) Cast nephropathy secondary to multiple myeloma Rhabdomyolysis Tubulointerstitial nephritis Biopsy-proven acute tubular necrosis with no clear precipitant Haemolytic uraemic syndrome/microangiopathic haemolytic anaemia
66 (21.5) 20 (6.5) 18 (5.9) 9 (2.9) 7 (2.3) 5 (1.6) 5 (1.6)
Post-renal Ureteric obstruction (e.g. secondary to malignancy) Bladder outflow obstruction
24 (7.9) 13 (4.2) 11 (3.6)
2 (0.7)
Inpatient mortality amongst patients that require RRT for AKI remains very high. Our findings are in keeping with data from other studies, where mortality associated with AKI requiring RRT has been reported to be as high as 70% [10]. Multivariate analysis of the entire cohort revealed that h-AKI, sepsis, a pre-renal cause of AKI and advancing age were all independent predictors of inpatient mortality. AKI has been shown to develop in over half of patients presenting with sepsis or septic shock, and mortality is directly correlated with the severity of the AKI [25]. Not having diabetes mellitus also predicted higher inpatient mortality. Patients with diabetes were slightly younger, less likely to have had sepsis and more likely to have had a pre-renal cause of AKI with a shorter duration of RRT (data not shown), possibly suggesting a somewhat less complicated course compared to patients that were not diabetic in this group of patients. We have previously observed a similar apparent survival benefit associated with the presence of diabetes in a cohort of patients with AKI requiring ICU admission and haemofiltration [8]. At 6 months, only a lower co-morbid burden as calculated by Charlson index was an independent predictor of survival. Focusing on differences between h-AKI and c-AKI, h-AKI was associated with a much higher mortality compared to c-AKI. In addition, h-AKI was a strong predictor of inpatient mortality even after controlling for other confounding variables. Only a handful of studies have compared the outcome of h-AKI to that of c-AKI. Furthermore, to the best of our knowledge our study is the first to examine the outcome of AKI requiring RRT in these two groups. Our findings are in agreement with a recent North American study that found c-AKI to be associated with a survival advantage [22]. They are also in agreement with two
Please cite this article as: Pang CL, et al, Timing of acute kidney injury — does it matter? A single-centre experience from the United Kingdom, Eur J Intern Med (2014), http://dx.doi.org/10.1016/j.ejim.2014.06.005
4
C.L. Pang et al. / European Journal of Internal Medicine xxx (2014) xxx–xxx
Table 3 Primary and secondary outcomes at discharge and at 6 months. c-AKI
h-AKI
p-Value
Primary outcomes Survival at hospital discharge (% survived) Dialysis dependence at hospital discharge (% dialysis dependent) eGFR at discharge +/− SD Survival at 6 months (% survived) Dialysis dependence at 6 months (% dialysis dependent) eGFR at 6 months +/− SD
72.0 35.0 33.1 +/− 24.6 82.8 32.2 44.7 +/− 22.9
42.9 16.7 38.2 +/− 25.8 87.5 12.5 52.3 +/− 22.5
b0.001 0.072⁎ 0.266 0.772a 0.056⁎, a 0.325
Secondary outcomes Length of RRT (days +/− SD) Length of stay (days +/− SD) ICU admission (%)
12.4 +/− 30.3 23.9 +/− 19.7 5.6
6.1 +/− 6.9 48.6 +/− 62.6 1.8
0.589 b0.001 0.321a
Bold values indicate significant p-value of pb0.05. ⁎ Difference that approaches statistical significance. a Fisher's exact test.
descriptive studies from India, which describe mortality amongst h-AKI and c-AKI patients, primarily on medical wards, to be 37.2% [26] and 17.39% [27] respectively. Interestingly, two studies from Brazil and Saudi Arabia found that c-AKI was linked with worse survival compared to h-AKI [5,21]. This may reflect specific population characteristics in these patient groups, as well as different access to secondary healthcare, which could potentially delay the presentation of c-AKI in those countries. Survivors in the h-AKI group were less likely to be dialysisdependent, both at discharge and at 6 months, although this difference did not reach statistical significance. The inclusion of an additional 30 patients overall with the same characteristics as observed in our analysis would render these results statistically significant (data not shown). We hypothesise that the renal survival benefit within the h-AKI group was due to a higher proportion of patients with sepsis and a pre-renal cause of AKI and therefore more of a chance to recover renal function amongst survivors. On the contrary, due to the high proportion of sepsis within these patients, they were more likely to have been more systemically unwell, and to have had higher mortality rates. It is also possible that failure to recover renal function might have contributed to higher inpatient mortality in the h-AKI group. Out of the thirty-two patients (57%) in the h-AKI group that died during their hospital admission, twenty-three (71.9%) died within 7 days of receiving the last dialysis session possibly suggesting that a clinical decision might have been taken to withdraw renal support in these patients although we did not have access to patient notes to confirm this. On the other hand, renal causes such as glomerulonephritides were associated with c-AKI. Therefore, c-AKI was linked with worse renal recovery. The frequency of CKD and diabetes was similar in both groups, and therefore unlikely to account for the worse renal recovery of the c-AKI group. This was confirmed in a post hoc analysis described
below. However, pre-existing CKD led to more overall RRT dependence at discharge and at 6 months. This is not a new finding as it has been noted in several studies including a recent observational study of a CKD disease cohort in Canada [28], where patients with CKD were not only more likely to develop AKI, but also more likely to require dialysis after discharge. Importantly, patients with h-AKI spent significantly longer amounts of time in hospital. This is probably a reflection of the severity of their illness, and may reflect the need for ongoing supportive therapy. With the average inpatient stay extended by almost a month; developing h-AKI has implications on patient experience. Finally, as AKI is estimated to account for up to 10% of all ICU bed days [29], AKI has significant logistic and financial implications for hospitals. The development of h-AKI was also associated with presentation in a hospital with no permanent nephrologist, and may reflect delayed referral to and involvement of the nephrology team. A recent pilot study showed improved outcomes for h-AKI with early nephrologist involvement [30]. Furthermore, delayed nephrology referral is linked to higher mortality and dialysis dependence at discharge in critically ill AKI patients [31]. We also carried out a post-hoc analysis after removing all patients with pre-existing CKD in order to determine differences between h-AKI and c-AKI in ‘pure’ AKI. This did not affect the direction or indeed significance of associations compared to our analysis of the whole group (data not shown). Interestingly, the only discrepancy was that by removing the CKD population, h-AKI patients were much more likely to be RRT independent compared to c-AKI patients at 6 months (p = 0.025). The remainder of the associations remained unchanged. One significant limitation of our study is the relatively low number of patients within the h-AKI group compared to the c-AKI group.
Table 4 Independent predictors of survival & RRT dependence. Odds ratio of survival (95% confidence interval)
p-Value
Survival to discharge Renal or post-renal cause of AKI Younger age Presence of diabetes c-AKI No sepsis
3.745 (1.745–8.000) 1.038 (1.012–1.066) 2.398 (1.300–4.425) 2.531 (1.269–5.048) 2.399 (1.246–4.618)
0.001 0.004 0.005 0.008 0.009
Survival to 6 months Lower Charlson index score
1.211 (1.027–1.429)
0.023
RRT dependence at discharge Presence of CKD RRT dependence at 6 months Presence of CKD No sepsis
Odds ratio of RRT dependence (95% confidence interval)
p-Value
3.247 (1.692–6.211)
b0.001
3.359 (1.423–7.929) 9.328 (1.167–74.546)
0.006 0.035
Please cite this article as: Pang CL, et al, Timing of acute kidney injury — does it matter? A single-centre experience from the United Kingdom, Eur J Intern Med (2014), http://dx.doi.org/10.1016/j.ejim.2014.06.005
C.L. Pang et al. / European Journal of Internal Medicine xxx (2014) xxx–xxx Table 5 Independent predictors of length of stay and length of RRT. B value (95% confidence interval)
p-Value
Increased length of stay h-AKI Advancing age
0.581 (0.271–0.891) 0.009 (0.002–0.017)
b0.001 0.013
Increased length of RRT Presence of CKD Renal or post-renal cause of AKI
0.511 (0.148–0.875) 0.509 (0.132–0.886)
0.006 0.008
Multivariate analysis was therefore only carried out within the entire group of 306 patients. Furthermore, as the data was collected retrospectively the quality of data could have been potentially affected. In order to avoid this we used prospectively collected data within the renal and hospital databases wherever possible. 6. Learning points • We have examined short and long-term outcomes in 306 consecutive patients with AKI that required RRT in the form of intermittent haemodialysis. • Patients with h-AKI had significantly higher mortality rates, compared with patients who had c-AKI. They were also more likely to have an extended hospital stay. However, survivors from the h-AKI group were less likely to be dependent on RRT at discharge, and at six months. • Independent predictors of mortality included sepsis, advancing age, a pre-renal cause of AKI, and h-AKI. • Early recognition and management by physicians, with timely input from nephrologists, are likely to be crucial in improving outcomes for patients with h-AKI. 7. Conclusions Acute kidney injury requiring RRT continues to be associated with high mortality, as well as a need for ongoing RRT, reaffirming the notion that AKI does indeed represent a ‘global health alert’ [32]. As we have shown, timing of AKI does matter, and h-AKI requiring RRT is associated with higher mortality, as well as an increased length of hospital stay. There is an ongoing need to further increase awareness of the multiple problems associated with AKI. Early recognition and management by clinicians is crucial, as the majority of AKI secondary to pre-renal causes is often avoidable, but linked with very high mortality. The early involvement of nephrologists is key, as are efforts to improve our diagnosis and prediction of incipient and established AKI with novel biomarkers [33]. Finally, there is a need for ongoing education to improve awareness and implement local strategies aimed at improving early diagnosis of AKI, prompt nephrology referral and appropriate management. References [1] Hegarty J, Middleton R, Krebs M, Hussain H, Cheung C, Ledson T, et al. Severe acute renal failure. Place of care, incidence, and outcomes. QJM 2005;98:661–6. [2] Stevens PE, Tamimi NA, Al Hasani MK, Mikhail AI, Kearney E, Lapworth R, et al. Nonspecialist management of acute renal failure. QJM 2001;94:533–40. [3] Ali T, Khan I, Simpson W, Prescott G, Townend J, Smith W, et al. Incidence and outcomes in acute kidney injury: a comprehensive population-based study. J Am Soc Nephrol 2007;18:1292–8. [4] Liangos O, Wald R, O'Bell JW, Price L, Pereira BJ, Jaber BL. Epidemiology and outcomes of acute renal failure in hospitalized patients: a national survey. Clin J Am Soc Nephrol 2006;1:43–51 [Epub 2005 Oct 26].
5
[5] Praught ML, Shlipak MG. Are small changes in serum creatinine an important risk factor? Curr Opin Nephrol Hypertens 2005;14:265–70. [6] Daher EF, Silva Junior GB, Santos SQ, R Bezerra CC, Diniz EJ, Lima RS, et al. Differences in community, hospital and intensive care unit-acquired acute kidney injury: observational study in a nephrology service of a developing country. Clin Nephrol 2012;78:449–55. http://dx.doi.org/10.5414/CN107167. [7] Meier P, Bonfils RM, Vogt B, Burnand B, Burnier M. Referral patterns and outcomes in noncritically ill patients with hospital-acquired acute kidney injury. Clin J Am Soc Nephrol 2011;6:2215–25. http://dx.doi.org/10.2215/CJN.01880211. Epub 2011 Aug 4. [8] Ng KP, Chanouzas D, Fallouh B, Baharani J. Short and long-term outcome of patients with severe acute kidney injury requiring renal replacement therapy. QJM 2012;105:33–9. [9] Chertow GM, Burdick E, Honour M, Bonventre JV, Bates DW. Acute kidney injury, mortality, length of stay, and costs in hospitalized patients. J Am Soc Nephrol 2005;16:3365–70. [10] Hoste EA, Clermont G, Kersten A, Venkataraman R, Angus DC, De Bacquer D, et al. RIFLE criteria for acute kidney injury are associated with hospital mortality in critically ill patients: a cohort analysis. Crit Care 2006;10:R73. [11] Liaño F, Junco E, Pascual J, Madero R, Verde E. The spectrum of acute renal failure in the intensive care unit compared with that seen in other settings. The Madrid Acute Renal Failure Study Group. Kidney Int 1998;53:S16–24. [12] Karsou SA, Jaber BL, Pereira BK. Impact of intermittent hemodialysis variables on clinical outcomes in acute renal failure. Am J Kidney Dis 2000;35:980–91. [13] Nolan CR, Anderson RJ. Hospital-acquired acute renal failure. JASN 1998;9:710–8. [14] National Confidential Enquiry into Patient Outcome, Death. Acute kidney injury: adding insult to injury; 2009. [15] Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P. Acute renal failure — definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care 2004;8:R204–12. [16] Mehta RL, Kellum JA, Shah SV, Molitoris BA, Ronco C, Warnock DG, et al. Acute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury. Crit Care 2007;11:R31. [17] Joannidis M, Metnitz B, Bauer P, Schusterschitz N, Moreno R, Druml W, et al. Acute kidney injury in critically ill patients classified by AKIN versus RIFLE using the SAPS 3 database. Intensive Care Med 2009;35:1692–702. [18] Wang Y, Cui Z, Fan M. Hospital-acquired and community-acquired acute renal failure in hospitalized Chinese: a ten-year review. Ren Fail 2007;29:163–8. [19] Kohli HS, Bhat A, Jairam A, Aravindan AN, Sud K, Jha V, et al. Predictors of mortality in acute renal failure in a developing country: a prospective study. Ren Fail 2007;29:463–9. [20] Dharan KS, John GT, Antonisamy B, Kirubakaran MG, Jacob CK. Prediction of mortality in acute renal failure in the tropics. Ren Fail 2005;27:289–96. [21] Sesso R, Roque A, Vicioso B, Stella S. Prognosis of ARF in hospitalized elderly patients. Am J Kidney Dis 2004;44:410–9. [22] Al-Homrany M. Epidemiology of acute renal failure in hospitalized patients: experience from southern Saudi Arabia. East Mediterr Health J Sep-Nov 2003;9:1061–7. [23] Schissler MM, Zaidi S, Kumar H, Deo D, Brier ME, McLeish KR. Characteristics and outcomes in community-acquired versus hospital-acquired acute kidney injury. Nephrology (Carlton) 2013;18:183–7. http://dx.doi.org/10.1111/nep.12036. [24] Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–83. [25] Suh SH, Kim CS, Choi JS, Bae EH, Ma SK, Kim SW. Acute kidney injury in patients with sepsis and septic shock: risk factors and clinical outcomes. Yonsei Med J 2013;54:965–72. [26] Singh TB, Rathore SS, Choudhury TA, Shukla VK, Singh DK, Prakash J. Hospitalacquired acute kidney injury in medical, surgical, and intensive care unit: a comparative study. Indian J Nephrol 2013;23:24–9. [27] Prakash J, Singh TB, Ghosh B, Malhotra V, Rathore SS, Vohra R, et al. Changing epidemiology of community-acquired acute kidney injury in developing countries: analysis of 2405 cases in 26 years from eastern India. Clin Kidney J 2013;6:150–5. [28] Lafrance JP, Djurdjev O, Levin A. Incidence and outcomes of acute kidney injury in a referred chronic kidney disease cohort. Nephrol Dial Transplant 2010;25:2203–9. [29] Intensive-Care National Audit Research Centre. www.icnarc.org; 2005. [30] Balasubramanian G, Al-Aly Z, Moiz A, Rauchman M, Zhang Z, Gopalakrishnan R, et al. Early nephrologist involvement in hospital acquired acute kidney injury: a pilot study. Am J Kidney Dis 2011;57:228–34. [31] Costa e Silva VT, Liano F, Muriel A, Diez R, De Castro I, Yu L. Nephrology referral and outcomes in critically ill acute kidney injury patients. PLoS One 2013;8:e70482. [32] Li PK, Burdmann EA, Mehta RL. World Kidney Day Steering Committee 2013: acute kidney injury: global health alert. 2013;83:372–6. [33] Obermuller N, Geiger H, Weipert C, Urbschart A. Current developments in early diagnosis of acute kidney injury. Int Urol Nephrol 2014;46(1):1–7.
Please cite this article as: Pang CL, et al, Timing of acute kidney injury — does it matter? A single-centre experience from the United Kingdom, Eur J Intern Med (2014), http://dx.doi.org/10.1016/j.ejim.2014.06.005
