|Year : 2019 | Volume
| Issue : 1 | Page : 25-30
Renal angina index in the prediction of acute kidney injury in critically ill children
Suma Sundararaju, Aditi Sinha, Pankaj Hari, Rakesh Lodha, Arvind Bagga
Department of Pediatrics, All India Institute of Medical Sciences, New Delhi, India
|Date of Web Publication||17-May-2019|
Department of Pediatrics, Division of Nephrology, All India Institute of Medical Sciences, New Delhi
Source of Support: None, Conflict of Interest: None
Background: Renal Angina Index (RAI) was recently proposed as a tool to identify patients at high risk for severe acute kidney injury (AKI) by integrating baseline, contextual, and clinical evidence of renal injury and to optimize biomarker utility in intensive care units. Methods: In this unicentric prospective observational study, we estimated RAI at admission in 285 critically ill patients, aged 1 month to 18 years, and evaluated the utility of renal angina (RAI ≥8) in identifying patients with severe or any AKI on day 3 and day 7. The relationship between RAI and need for renal replacement therapy (RRT) and duration of mechanical ventilation and hospital stay was also examined. Results: Renal angina was present in 49.4% of 285 patients. Severe AKI, observed in 29 (10.2%) patients on day 3 and 13.2% of 144 patients followed to day 7, had an incidence of 1.1 (0.8–1.6) episodes per 100 person-days. Thirty-six (12.6%) patients required RRT. RAI satisfactorily identified patients at risk of severe AKI on days 3 (area under the curve [AUC]: 0.82; 95% confidence interval [CI]: 0.73–0.90) and 7 (AUC: 0.73; 95% CI: 0.62–0.84), was more useful than Pediatric Index of Mortality in such discrimination, and correlated with duration of mechanical ventilation and hospital stay. However, RAI thresholds ≥12 or ≥20 had higher specificity, Youden index, and positive predictive value than that of RAI ≥8 in discriminating severe AKI on day 3 or 7 and distinguished between patients with and without need for RRT. Conclusions: RAI usefully predicts the development of subsequent severe AKI on days 3 and 7, and is associated with duration of mechanical ventilation and hospital stay. A higher RAI threshold (≥12 or ≥20) is more discriminatory than RAI ≥8.
Keywords: Acute renal failure, fluid overload, pediatric
|How to cite this article:|
Sundararaju S, Sinha A, Hari P, Lodha R, Bagga A. Renal angina index in the prediction of acute kidney injury in critically ill children. Asian J Pediatr Nephrol 2019;2:25-30
|How to cite this URL:|
Sundararaju S, Sinha A, Hari P, Lodha R, Bagga A. Renal angina index in the prediction of acute kidney injury in critically ill children. Asian J Pediatr Nephrol [serial online] 2019 [cited 2019 Dec 9];2:25-30. Available from: http://www.ajpn-online.org/text.asp?2019/2/1/25/258565
| Introduction|| |
The diagnosis of acute kidney injury (AKI), characterized by rapid loss of renal function, is based on an increase in serum creatinine, decrease in urine output, or both. AKI affects a high proportion of critically ill patients, both pediatric and adult, and is strongly associated with mortality, serious comorbidities, length of hospital stay, and risk of developing chronic kidney disease.,, Given its significance, recent research has focused on the early detection of AKI to enable timely intervention to prevent severe AKI. Serum and/or urinary biomarkers, namely neutrophil gelatinase-associated lipocalin, interleukin-18, kidney injury molecule-1, and liver fatty acid-binding protein are validated as useful screening tools in defined populations, for example, postcardiopulmonary bypass., As the diagnostic utility of these biomarkers in unspecified critically ill patients remains unclear,, the concept of renal angina was introduced to optimize their performance.
Renal Angina Index (RAI), a composite of risk strata and clinical injury signs of AKI [Table 1], helped identify, in pediatric intensive care unit (PICU) admission in multicentric studies, sick children most at risk of developing severe AKI (Kidney Diseases Improving Global Outcomes [KDIGO], Stages 2–3) and outperformed various severity of illness scores in such prediction., Because these studies were performed in predominantly Caucasian patients in developed countries, the present study was planned to validate RAI in predicting AKI in critically ill children admitted at a tertiary care center in India.
| Methods|| |
This prospective observational study enrolled consecutive patients, 1 month to 18 years old, admitted to the PICU at a single tertiary care center between July 2014 and March 2016. Patients with chronic kidney disease (CKD) with estimated glomerular filtration rate (eGFR) below 15 ml/min/1.73 m2 and those who underwent renal transplantation in the last 90 days or cardiac catheterization or surgery in the last 30 days, were excluded from the study. Following ethical approval and written informed parental consent, eligible children were enrolled. Information was collected on demographic variables; diagnosis; various comorbidities; fluid intake and urine output; type, intensity, and duration of mechanical ventilation; vasopressor support and renal replacement therapy (RRT); Pediatric Index of Mortality (PIM-2) score; use of nephrotoxic medications; anthropometry; vital signs; disease severity scores; serum creatinine; urine biochemistry; length of PICU and hospital stay; and outcomes (discharge or death), at admission, and daily during PICU stay until day 7, unless the patient was discharged or died earlier. For patients admitted to the PICU for longer than a week, outcome information was also collected at discharge or death, or on day 28, whichever was earlier. Data were anonymized upon data entry.
Estimated creatinine clearance (eCrCl) was calculated from serum creatinine and body length using the modified Schwartz formula. Baseline serum creatinine was the lowest value available in medical records in the 90 days prior to admission; if no value was available, baseline eCrCl was assumed as 80 ml/min/1.73 m2 for patients aged 1–6 months, 100 ml/min/1.73 m2 for those 6–12 months old, and 120 ml/min/1.73 m2 for children older than 1 year. Change in eCrCl was computed as follows: ([current eCrCl − baseline eCrCl] × 100)/baseline eCrCl. Percentage fluid overload was estimated on the day of admission as follows:([sum of all fluids in PICU – sum of all fluids out of PICU, in L] × 100)/weight at PICU admission, in kg. RAI was calculated on the day of PICU admission as a product of risk and injury scores, as shown in [Table 1]. Renal angina was present if RAI equaled or exceeded 8.
The primary outcome was the diagnostic accuracy (sensitivity, specificity, and Youden index) of RAI to predict the presence of severe AKI, defined as KDIGO AKI 2 or 3, namely increase in serum creatinine to ≥200% from baseline, decrease in estimated CrCl (eCrCl) of ≥50% from baseline, or fall in urine output to <0.5 ml/kg/hr for ≥8 hr, on day 3 of PICU stay. Secondary outcomes were the diagnostic accuracy of RAI in detecting severe AKI on day 7 of PICU stay or any AKI (increase in serum creatinine by ≥0.3 mg/dl from, or to >1.5 times the value at, baseline, or fall in urine output to <0.5 ml/kg/hr for 6 hr) on days 3 and 7, need for RRT, length of PICU stay, and mortality.
Data were analyzed using IBM-SPSS (IBM SPSS Statistics for Windows, version 22.0, released 2011; IBM Corp., Armonk, NY, USA). Categorical data were expressed as percentage, and/or 95% confidence interval (CI) of the estimate, and compared using Chi-square or Fisher's exact test and/or relative risk (RR) (95% CI). Continuous data were expressed as median (interquartile range) or mean ± standard deviation (SD) and compared using parametric or nonparametric tests, as appropriate, and mean difference (95% CI); P ≤ 0.05 was considered statistically significant. The diagnostic utility of RAI in predicting outcomes was assessed at various thresholds including ≥8, using sensitivity, specificity, Youden index, negative and positive predictive values (NPV and PPV), and estimates of area under the curve (AUC). Spearman's rank-order correlation was used to examine associations between continuous variables and occurrence of renal angina.
Sample size was estimated to be 200 patients, based on the assumptions that the proportion of critically ill patients developing AKI would be similar to that (36%) reported from our center previously and that renal angina will have 75% sensitivity and specificity each in detecting severe AKI, as reported previously, allowing for 10% absolute precision and 95% confidence in the estimate.
| Results|| |
Of 356 patients screened, 69 (19.4%) were excluded (including 56 neonates, 11 with known CKD5 and two admitted following cardiac surgery); two children were excluded later for missing outcome data. Therefore, 285 patients (66.3% boys) were enrolled, with median (interquartile range [IQR]) age of 4.4 (0.5–7.7) years. Reasons for PICU admission were septic shock (74.0%), respiratory failure (83.9%), congestive cardiac failure (24.2%), and/or encephalopathy (12.3%). Common underlying systemic conditions included infections (29.1%) and pulmonary (18.6%), hematological (16.8%), hepatic or gastrointestinal (12.6%), and cardiac illnesses (10.8%). Renal angina (RAI ≥8) at PICU admission was present in 141 (49.4%) patients.
Acute kidney injury and related outcomes
On day 3, 62 (21.8%) of the 285 patients had AKI and 29 (10.2%) patients had severe AKI [Table 2]. Risk factors for developing AKI included sepsis (96.1%), multiorgan failure (64.5%), need for inotropic support (54%), mechanical ventilation (64.2%) or both (39.6%), use of nephrotoxic drugs (74.2%), and transplantation (0.4%). Patients with and without severe AKI on day 3 showed similar duration of mechanical ventilation (median [IQR], 7 [5,10] days vs. 7 [4,11] days) and length of PICU stay (7 [5,17] days vs. 6 [3,11] days).
|Table 2: Occurrence of acute kidney injury on days 3 and 7 based on estimated glomerular filtration rate and/or urine output|
Click here to view
Among 144 patients with PICU stay lasting a week or longer, any and severe AKI on day 7 were observed in 20 (13.9%) and 19 (13.2%) patients, respectively. Overall, 67 patients developed any AKI and 34 developed severe AKI over 2961 person-days, based on which the incidence of AKI was estimated at 2.3 (1.8–2.9) episodes per 100 person-days and that of severe AKI at 1.1 (0.8–1.6) episodes per 100 person-days. Thirty-six (12.6%) patients required RRT.
Outcomes in relation to renal angina
AKI was observed on day 3 in 47 (33.3%) patients with renal angina, of whom 24 (17.0%) had severe AKI. As shown in [Table 3], patients with renal angina, compared to those with RAI <8, had significantly increased risk of any AKI (RR: 3.2, 95% CI: 1.9–5.5; P < 0.0001) and severe AKI (RR: 4.9; 95% CI: 1.9–12.5; P = 0.0001). Compared to those with RAI <8, patients with renal angina also showed significantly prolonged duration of PICU stay (P = 0.0001) and mechanical ventilation (P < 0.0001). As shown in [Table 3], the need for RRT and rate of mortality were comparable between patients with and without renal angina. Further, the proportion of patients with any and severe AKI on day 7 was only insignificantly higher in patients with renal angina compared to those with RAI <8 [Table 3].
|Table 3: Comparison of outcomes among patients with or without renal angina (Renal Angina Index ≥8) and Renal Angina Index ≥12|
Click here to view
Predictive accuracy of Renal Angina Index scores
Receiver operator characteristic curves to examine the diagnostic accuracy of RAI score in discriminating patients with severe AKI showed AUC of 0.82 (95% CI: 0.73–0.90) for severe AKI on day 3 and AUC of 0.73 (95% CI: 0.62–0.84) for severe AKI on day 7 [Figure 1]. An RAI score of ≥12 or ≥20 had higher diagnostic utility than RAI ≥8 in predicting severe AKI on day 3 [Figure 1] and [Supplementary Table 1] and day 7 (data not shown). RAI thresholds of 8 as well as 12 or 20 had satisfactory sensitivity (82.8% and 79.3%, respectively) and high NPV (96.5% and 97%, respectively) for the development of severe AKI on day 3.
|Figure 1: Receiver operator characteristic curves for diagnostic accuracy of RAI score in discriminating patients with severe AKI on days 3 and 7. (a) RAI had an AUC of 0.82 (95% CI: 0.73–0.90) in discriminating patients with severe AKI on day 3. Compared to RAI threshold of 12 or 20, RAI score ≥8 had lower Youden index (0.55 vs. 0.37), similar sensitivity (79.3% vs. 82.7%), and lower specificity [75.8% vs. 54.3%; Supplementary Table 1]. (b) AUC of RAI for severe AKI on day 7 was 0.73 (95% CI: 0.62–0.84). Compared to RAI threshold of 12 or 20, RAI score ≥8 had lower Youden index (0.41 vs. 0.34), similar sensitivity (68.4% vs. 73.7%), and lower specificity (75.8% vs. 52.3%). AUC: Area under the curve, CI: Confidence interval, RAI: Renal Angina Index, AKI: Acute kidney injury|
Click here to view
[Table 3] shows that RAI score of ≥12 or ≥20 was associated with significantly increased risk of any and severe AKI on days 3 and 7 and of need for RRT. Patients with RAI ≥12 or ≥20 also had significantly prolonged duration of PICU stay and mechanical ventilation [Table 3]. As compared to RAI of ≥8, the significance of differences in all these parameters was more pronounced for a threshold of ≥12. However, mortality rates were comparable between groups categorized by this threshold.
Performance of Renal Angina Index compared to other scoring systems
RAI score correlated significantly and positively with duration of mechanical ventilation (Spearman's r = 0.232; P = 0.001) and length of ICU stay (r = 0.233, P = 0.001). RAI, at thresholds of 8, 12, or 20, outperformed PIM-2 in predicting severe AKI on day 3; the AUC for the outcome with PIM-2 was 0.48 (95% CI: 0.37–0.59). At either threshold, RAI had higher sensitivity and NPV compared to day 0 KDIGO Stages 1 or 2–3. While both scores had Youden index higher than KDIGO Stage 1, a score of ≥12 or ≥20, but not ≥8, had higher Youden index than KDIGO Stages 2–3.
| Discussion|| |
Findings from this single-center prospective study indicate that almost half of the 285 patients admitted to the PICU over 20-month period fulfilled the criteria for renal angina. One-fifth of the patients admitted to the PICU developed AKI by day 3, half of whom (overall, 10.2%) had severe AKI. Of 144 patients followed to day 7, 13.2% developed severe AKI, all of whom required RRT. The incidence of AKI and severe AKI was estimated at 2.3 and 1.1 episodes per 100 person-days. RAI had satisfactory utility in discriminating patients with severe AKI on days 3 and 7, significantly correlated with length of hospital stay and duration of mechanical ventilation, and outperformed PIM-2 score and day 0 KDIGO stage in identifying patients at risk of severe AKI. However, a threshold of ≥12 or ≥20 was more useful than the traditional definition of renal angina (RAI ≥8), as it was associated with overall higher diagnostic accuracy based on higher specificity, positive likelihood ratio, and PPV and NPV, in discriminating between patients with and without any or severe AKI at day 3 or 7. RAI threshold of 12 or 20, but not 8, was associated with the risk of need for RRT and significant differences in durations of ICU stay and mechanical ventilation.
RAI, a score that combines risk factors with early signs of AKI on the day of admission to identify patients at risk of subsequent severe AKI (KDIGO Stages 2–3), was derived and validated in single-center patient cohorts chiefly comprising Caucasian children., These and subsequent studies, summarized in [Table 4],,,,,, indicate satisfactory diagnostic accuracy of RAI score ≥8 in predicting the development of severe AKI 3 days after admission to the ICU. Further, RAI (RAI ≥ 8) consistently enabled excluding the subgroup of patients not at risk of severe AKI, indicated by its high NPVs (92%–99%) [Table 4], apart from being easier to calculate, and more useful, than illness prediction scores.,,, While our study did not examine the utility of biomarkers, limiting the application of three biomarkers to patients with renal angina improved the diagnostic accuracy for each biomarker in critically ill patients., However, the index is criticized for being a misnomer etymologically (“angina” refers to pain in throat) and for its incorporation of similar features as used to classify AKI, suggesting that it may just be detecting mild renal injury that evolves into severe AKI. Even so, similar to our results, day 0 RAI has had higher diagnostic accuracy than day 0 KDIGO stage 1, and change in serum creatinine from baseline in detecting those at risk of subsequent severe AKI, underscoring the importance of the risk component of the score.
|Table 4: Comparison of performance of Renal Angina Index in predicting severe acute kidney injury on day 3 in various studies|
Click here to view
One-tenth of critically ill children in the present study developed severe AKI, similar to the rate (11.9%) reported by us previously. Severe AKI was reported in 8%–14% patients in four studies on RAI, including one from India [Table 4];,,, higher rates of severe AKI in two other studies may reflect a preponderance of postoperative critically ill patients., Similar to our study, all studies listed in [Table 4] excluded neonates, in whom different criteria are proposed to define AKI, and children with CKD Stage 5, that cannot be discriminated readily from severe AKI. Evaluation for AKI was at 3 days to allow time for serum creatinine to rise and to detect persistent rather than transient fall in urine output; further, day 3 AKI usefully discriminates reversible (functional) from damage-associated AKI. The outcome of interest has been severe, rather than any, AKI as mild AKI has a weak association with adverse outcomes.
The present work examined the utility of RAI in a relatively large and heterogeneous population of pediatric patients admitted for primarily medical indications and found that application of the RAI at a threshold of 12 or 20 was more useful than at a threshold of 8 in terms of its predictive accuracy for severe AKI. Similar to other reports,,, we found that RAI outperformed signs of injury (KDIGO Stage 1 AKI) and severity of illness score (PIM-2) in predicting subsequent severe AKI; however, higher thresholds of 12 and 20 were more discriminatory than the conventional threshold of ≥8. These higher RAI thresholds were more useful than renal angina in predicting the occurrence of any AKI on day 3 and severe or any AKI on day 7. While length of hospital stay and duration of mechanical ventilation differed significantly between patient subgroups defined by any of the three RAI thresholds (8, 12, and 20), thresholds of 12 and 20, but not 8, predicted the need for RRT. The reasons for this discordance between our findings and the threshold confirmed by other studies are unclear. In absence of patients with bone marrow or solid organ transplantation in the present study, we found no patients with RAI scores between 11 and 19 [Table 1]. Hence, a higher RAI threshold might have identified chiefly a subgroup of patients in whom mechanical ventilation and/or inotrope use was indicated, known to be associated with severe AKI.,
| Conclusions|| |
Findings from this prospective study confirm the utility of RAI in predicting the development of severe AKI in critically ill pediatric patients and support its use as a feasible bedside method to enable focused attention of preventive measures on those at highest risk while enabling liberal interventions in those excluded from such risk. While the application of RAI in predicting severe AKI seems valid in developing countries, the threshold for renal angina might differ from that advocated in studies including predominantly Caucasian patients.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Akcan-Arikan A, Zappitelli M, Loftis LL, Washburn KK, Jefferson LS, Goldstein SL, et al.
Modified RIFLE criteria in critically ill children with acute kidney injury. Kidney Int 2007;71:1028-35.
Slater MB, Anand V, Uleryk EM, Parshuram CS. A systematic review of RIFLE criteria in children, and its application and association with measures of mortality and morbidity. Kidney Int 2012;81:791-8.
Hoste EA, Bagshaw SM, Bellomo R, Cely CM, Colman R, Cruz DN, et al.
Epidemiology of acute kidney injury in critically ill patients: The multinational AKI-EPI study. Intensive Care Med 2015;41:1411-23.
Kaddourah A, Basu RK, Bagshaw SM, Goldstein SL; AWARE Investigators. Epidemiology of acute kidney injury in critically ill children and young adults. N
Engl J Med 2017;376:11-20.
Krawczeski CD, Goldstein SL, Woo JG, Wang Y, Piyaphanee N, Ma Q, et al.
Temporal relationship and predictive value of urinary acute kidney injury biomarkers after pediatric cardiopulmonary bypass. J Am Coll Cardiol 2011;58:2301-9.
Kashani K, Al-Khafaji A, Ardiles T, Artigas A, Bagshaw SM, Bell M, et al.
Discovery and validation of cell cycle arrest biomarkers in human acute kidney injury. Crit Care 2013;17:R25.
Bagshaw SM, Zappitelli M, Chawla LS. Novel biomarkers of AKI: The challenges of progress' amid the noise and the haste'. Nephrol Dial Transplant 2013;28:235-8.
Murray PT, Mehta RL, Shaw A, Ronco C, Endre Z, Kellum JA, et al.
Potential use of biomarkers in acute kidney injury: Report and summary of recommendations from the 10th
Acute Dialysis Quality Initiative Consensus Conference. Kidney Int 2014;85:513-21.
Basu RK, Wang Y, Wong HR, Chawla LS, Wheeler DS, Goldstein SL. Incorporation of biomarkers with the Renal Angina Index for prediction of severe AKI in critically ill children. Clin J Am Soc Nephrol 2014;9:654-62.
Kidney Disease Improving Global Outcomes (KDIGO) Working Group on Acute Kidney Injury (AKI). KDIGO Clinical Practice Guideline for AKI. Kidney Int Suppl 2012;2:1-138.
Menon S, Goldstein SL, Mottes T, Fei L, Kaddourah A, Terrell T, et al.
Urinary biomarker incorporation into the Renal Angina Index early in intensive care unit admission optimizes acute kidney injury prediction in critically ill children: A prospective cohort study. Nephrol Dial Transplant 2016;31:586-94.
Basu RK, Zappitelli M, Brunner L, Wang Y, Wong HR, Chawla LS, et al.
Derivation and validation of the Renal Angina Index to improve the prediction of acute kidney injury in critically ill children. Kidney Int 2014;85:659-67.
Schwartz GJ, Muñoz A, Schneider MF, Mak RH, Kaskel F, Warady BA, et al.
New equations to estimate GFR in children with CKD. J Am Soc Nephrol 2009;20:629-37.
Shann F, Pearson G, Slater A, Wilkinson K. Paediatric Index of Mortality (PIM): A mortality prediction model for children in intensive care. Intensive Care Med 1997;23:201-7.
Zappitelli M, Parikh CR, Akcan-Arikan A, Washburn KK, Moffett BS, Goldstein SL, et al.
Ascertainment and epidemiology of acute kidney injury varies with definition interpretation. Clin J Am Soc Nephrol 2008;3:948-54.
Goldstein SL, Currier H, Graf CD, Cosio CC, Brewer ED, Sachdeva R. Outcome in children receiving continuous venovenous hemofiltration. Pediatrics 2001;107:1309-12.
Böhning D, Böhning W, Holling H. Revisiting Youden's index as a useful measure of the misclassification error in meta-analysis of diagnostic studies. Stat Methods Med Res 2008;17:543-54.
Mehta P, Sinha A, Sami A, Hari P, Kalaivani M, Gulati A, et al.
Incidence of acute kidney injury in hospitalized children. Indian Pediatr 2012;49:537-42.
Sethi SK, Raghunathan V, Shah S, Dhaliwal M, Jha P, Kumar M, et al.
Fluid overload and Renal Angina Index at admission are associated with worse outcomes in critically ill children. Front Pediatr 2018;6:118.
Kaur R, Dhooria GS, Pooni PA, Bhat D, Bhargava S, Kakkar S, et al.
Utilization of the Renal Angina Index in PICU of a developing country for prediction of subsequent severe acute kidney injury. Pediatr Nephrol 2018;33:2185-91.
Basu RK, Kaddourah A, Goldstein SL; AWARE Study Investigators. Assessment of a Renal Angina Index for prediction of severe acute kidney injury in critically ill children: A multicentre, multinational, prospective observational study. Lancet Child Adolesc Health 2018;2:112-20.
Palevsky PM. Renal angina: Right concept. Wrong name? Clin J Am Soc Nephrol 2014;9:633-4.
Zappitelli M, Ambalavanan N, Askenazi DJ, Moxey-Mims MM, Kimmel PL, Star RA, et al.
Developing a neonatal acute kidney injury research definition: A report from the NIDDK neonatal AKI workshop. Pediatr Res 2017;82:569-73.
Endre ZH, Kellum JA, Di Somma S, Doi K, Goldstein SL, Koyner JL, et al.
Differential diagnosis of AKI in clinical practice by functional and damage biomarkers: Workgroup statements from the Tenth Acute Dialysis Quality Initiative Consensus Conference. Contrib Nephrol 2013;182:30-44.
[Table 1], [Table 2], [Table 3], [Table 4]