European Journal of Cardiovascular Medicine

Acute Kidney injury (AKI) in the ICU set up has been a rising problem with the past few decades showing a change in the profile of patients getting admitted to the ICU‘s. They are more severely ill as compared to 10-15 years back and often have multiple organ involvement and associated sepsis and other comorbidities.[1,2,3] AKI in ICU has a significant effect on patient outcome and recently has been equated to other two syndromes of Severe sepsis and Acute lung injury that determine patient prognosis in ICU. An understanding of factors affecting renal recovery might improve overall outcome. The current study is undertaken to study the profile of AKI and factors which predict its outcome in the ICU set up. Recently a number of novel biomarkers have not only been shown to predict AKI but also its outcome. Neutrophil gelatinase associated Lipocalin is one such biomarker that has also been called as the Renal Troponin due to its excellent performance in predicting AKI prior to rise of serum creatinine but also in predicting outcome. Few studies from South India have validated the utility of NGAL in AKI. [4,5] Present study also aims to look into the utility of urinary NGAL in predicting AKI outcomes as a Pilot project in view of the costs incurred. The Medical Intensive Care unit in Kurnool Medical College, Kurnool has a monthly average admission rate of 80-90 patients from all medical specialties. Hencethe study is undertaken with an intention to look into the causes of AKI as well as into the factors affecting the renal outcome in the ICU of this tertiary care hospital of South India. Simultaneously the study of urinary NGAL in predicting the outcomes shall be done as a PILOT project.

Aims and Objectives

Aim:

• To study the clinical profile, incidence and prevalence of Acute Kidney Injury (AKI) patients in a Medical ICU at a tertiary care centre.

Objectives:

• To study the Renal outcome and Mortality at day 7 of follow up of these patients.
• To study the utility of urinary Neutrophil Gelatinase associated Lipocalin in predicting the severity and outcome of acute kidney injury.
• To distinguish between the characteristics of Transient and Persistent AKI.
• To study the characteristics of patients requiring RRT in this cohort.

Type of study: Prospective observational cohort study approved by the Institutional review board. Setting and location: Departments of Nephrology, Critical care medicine, Kurnool Medical College.

Study Period: 6 months

Participants:

❖ 103 consecutive patients who had AKI at the time of ICU admission or who developed AKI during ICU stay were recruited in the study over a period of 6 months. A) For studying clinical profile:

Inclusion criteria:

• Incident /prevalent AKI patients > 18 year in ICU as per AKIN criteria.

Exclusion Criteria

• Pregnant women and children.

• Patients in whom baseline serum creatinine cannot be found prospectively or retrospectively.

 • Chronic Kidney Disease patients eGFR < 60 ml/min/m2 BSA prior to current illness.

B) For studying urinary NGAL

Inclusion criteria

Patients with eGFR > 60 ml/min/ m2 BSA

Exclusion criteria

1. Patients in whom baseline serum creatinine cannot be found prospectively or retrospectively

2. Patients with obstructive uropathy.

3. Suspected Urinary tract infection.

4. Patients with CKD i.e .eGFR< 60 ml/min/m2 BSA prior to current illness

5. Pregnant women and children.

Data Sources/measurement:

History and treatment details were noted from the chart records. Baseline creatinine was retrieved from prior records within 3 months. If no previous records were available and if the lowest creatinine attained in the hospital corresponded to GFR >60ml/min/1.73m2, it was considered as baseline creatinine. Blood pressure, height, weight, BMI, age, gender and other demographic details were noted from the chart records. Serum creatinine, urea, total leukocyte counts, differential counts, liver function test, Urine analysis & microscopy, urine osmolarity and other investigations were noted from electronic records. Fractional excretion of Sodium was calculated as per formula – (Urine Sodium X Plasma Creat) / (Urine creat X Plasma Sodium). Patients were followed up daily with blood and urine collection at inclusion and daily record of renal parameters, hemodynamic variables, Intake/Output, SOFA score, requirement of RRT for next 7 days. Final outcome was recorded from discharge summary as well as from electronic medical records. Urinary Neutrophil Gelatinase associated Lipocalin was measured with a commercial ELISA kit (BIOPORTO) at biochemistry lab.

One hundred and three consecutive patients who had AKI at the time of ICU admission or who developed AKI during ICU stay were recruited in the study over a period of 6 months as per the protocol

BASELINE CHARACTERISTICS OF STUDY POPULATION

Table 2: study population comorbidities

As shown in the Table 4, Diabetes (30.1%) followed by hypertension (26.2%) were the most frequent comorbidities of the ICU population with AKI. Eight (7.8%) patients had associated malignancies of which 50% were haematological in origin. Other comorbidities were CAD (10.7%), COPD (8.7%), CLD (8.7%) and CVA (9.7%). 71.8% of patients hadno prior renal dysfunction. Those who were classified as CKD stage 1or 2 were in older age group with GFR corresponding to the age-related decline.

Table 3: Clinical Profile at admission

As seen from Table 5 most patients had high counts at admission, though a few were leukopenic due to presence of aplastic anemia. Mean urea and creatinine at presentation were 79.6±43.6 mg% and 2.3±1.3 mg% respectively. Serum albumin and bicarbonate levels were on the lower side being 2.9±0.9 g% and 16.8±6.8 mmol/L respectively. Significant patients had high lactates at admission averaging to 4.2±3.9 mmol/L. Mean SOFA score at admission was 10.7±4.3. Urinary NGAL was analysed in total of 63 patients with mean value of 4224.8± 4586.1 ng/ml.

Table 4: Baseline comparison between AKIN 1,2 and3

Table 6 shows the comparison of baseline parameters between different AKIN stages using One way ANOVA. Age, haemoglobin, total leukocyte counts, total bilirubin, AST, ALT, Sodium, bicarbonate, chloride and lactate levels were not different between the three groups. However, serum urea and creatinine were found to be significantly different between all three groups in the post hoc analysis of One-way ANOVA. Similarly, SOFA score at admission was significantly different between all three AKIN stages. Platelet count, day 1 urine output, serum protein had significant P value in ANOVA, but the post hoc analysis showed that the difference in their values was significant only between AKIN 1 and AKIN 3 stages, while AKIN 1vs 2 and AKIN 2 vs 3 were not significantly different. Serum Albumin was significantly higher in AKIN 1 patients as compared to both AKIN 2 and AKIN 3 separately.

Table5. Urinary NGAL between AKIN stages

Comparing urinary NGAL (UNGAL) at various AKIN stages. The mean values were very significant when comparing AKIN 1 and AKIN 3 (P=0.001) and considerably different when comparing AKIN stage 1 and stage 2 (P=0.024). However, AKIN 2 and AKIN 3 did not significantly differ from one another. With a two-tailed p value of 0.085, a pearson correlation coefficient of 0.212 was established between UNGAL and serum creatinine.

Table 8 demonstrates that there was no statistically significant difference between the groups with transient and persistent AKI in any of the comorbidities, including diabetes, hypertension, coronary artery disease (CAD), chronic obstructive pulmonary disease (COPD), cerebrovascular disease (CVA), or chronic liver disease (CLD). Notably, a considerably higher number of patients with transient AKI had fractional excretion of sodium (FeNa) that was less than 1, and none of them needed hemodialytic treatment. In addition, 72% (18) of patients with transient AKI were in AKIN 1, compared to only 24.3% (17) of patients at inclusion (P<0.001).

Table 6: Baseline Characteristics of Transient and Persistent AKI

               a: For FeNa N=23 for Transient and 43 for persistent AKI

According to Table 8, FeNa could be utilised on 66 patients in total, including 23 patients with transient AKI and 43 patients with persistent AKI. 65.2% (15) of the patients in the transient AKI group showed FeNa values that were suggestive of pre-renal AKI.

Table 7: Baseline Clinico-laboratory characteristics in Transient and Persistent AKI

The clinical and laboratory data for patients with transient and persistent AKI are shown in Table 9. The groups' mean ages, haemoglobin levels, total bilirubin, AST, ALT, alkaline phosphate levels, sodium levels, and chloride levels did not differ substantially. The length of the ICU stay was also the same. Comparatively, there were significant differences between the groups in total leukocyte count (P=0.028), serum potassium (P=0.031), serum bicarbonate (P=0.025), and fractional excretion of sodium (P=0.008). In comparison to patients with persistent AKI, patients with transient AKI had higher mean serum protein and albumin levels, lower mean urea and creatinine concentrations, and all of these differences had a very significant P value of 0.001 on the independent 't' test. Patients with persistent AKI presented with noticeably more acidosis

(<0.001), higher lactate levels (<0.001), and poorer urine production than control patients (P=0.005). When SOFA scores upon inclusion were examined, individuals with transient AKI had significantly lower mean scores (7.7 4.0) than patients with persistent renal impairment (11.5 3.9; P 0.001).   Mean urine NGAL levels were substantially lower in the group of patients with transient  AKI, 1078.4 ± 1702.5 ng/ml, than in the group of patients with persistent AKI, 5292.9± 4913.5 ng/ml (P <0.001). As a result, in univariate analysis, urine NGAL levels at inclusion were a good indicator of renal recovery.

Table 8: Characteristic of patients with renal recovery vs non recovery at day 7

Table 10 shows that diabetic individuals had a higher likelihood of having persistent renal impairment even after a 7-day follow-up (P=0.021). Patients with improved renal function had normal baseline renal function that was considerably higher, i.e., no underlying CKD (P=0.003), and more patients were in AKIN 1 at presentation than patients who did not improve (P=0.015

Table 9: Clinical and Lab parameters of patients with renal recovery vs non recovery at day 7

According to Table 11, those with persistent renal impairment at day 7 had mean serum urea and creatinine values of 95 ± 48 mg/dl and 2.95 ± 1.59 mg/dl, respectively. These were noticeably greater than individuals who had fully regained their renal function by day 7 (P < 0.01). Furthermore, it can be inferred from the table that patients with incomplete renal recovery had considerably worse acidosis (P=0.012) and more lactate build-up (P=0.011) at presentation than those with full renal recovery. Additionally, they had significantly lower mean serum levels of sodium and chloride but higher levels of potassium. Mean urinary NGAL levels were 1797.6 ± 2589.4 ng/ml in the group whose renal functionimproved, which was considerably lower than the 4180.1 ± 4775.4 ng/ml levels in the group whose renal function did not improve (P=0.037). Additionally, patients who did not regain their renal function by day 7 had considerably higher fractional excretion of sodium.

Table 10: SOFA score and Renal outcome at day 7

Multivariate analysis of Renal out come at Day 7: A model for forecasting renal outcome at day 7 was created using binary logistic regression and the backward conditional technique, and it had a 71% predictive power (Table 13). Diabetes Mellitus, SOFA score at day 2, and the lack of pre-existing CKD are the factors that are significantly associated with predicting renal prognosis. Serum lactate, the need for hemodialysis, serum potassium, and urine NGAL on day 1 Although significant in univariate analysis, urea andcreatinine were not identified in multivariate analysis as independent predictors of renal outcome at day 7.

Table 11: Goodness of Fit of Model

Table 12: Regression model predicting Renal outcome at day 7

Renal replacement Therapy requirement:

There were 29 patients (28.2%) who needed hemodialysis. corresponding comorbidities Diabetes, hypertension, CAD, COPD, CVA, and CLD were not linked to the need for RRT, nor were diabetes, hypertension, or CAD, chronic obstructive pulmonary disease, or chronic liver disease. A major predictor of the need for hemodialysis over the next week was an AKIN stage > 1.

According to Table 15, patients who started receiving hemodialysis had significantly higher mean values for urea (P=0.032) and creatinine (<P0.001) on their first day. Additionally, they presented with substantial lactic and metabolic acidosis as well as significantly increased potassium levels.

Table 13: Clinico-biochemical parameters associated with RRT requirement

Patients who received haemodialysis also had substantially higher mean SOFA scores on Days 1, 2, and 3 with P values of <0.014, 0.002, and 0.02, respectively. When urine NGAL  levels were examined, patients needing hemodialysis had mean values of 6470 ± 5668  ng/ml while those not needing RRT had mean values of 3130 ± 3814 ng/ml. The levels between the two groups were significantly different (P=0.038). Additionally, mean FeNa% was noticeably greater in patients needing RRT.

Table 14: Univariate analysis of Mortality outcome

It was discovered that the presence of chronic liver disease was substantially related to death (P=0.028). Although the difference did not reach statistical significance (p=0.093), 38.2% of patients in the expired group and 22.4% of survivors were started on hemodialysis. The patients who died had mean lactate and pH values that were significantly higher than those of the patients who survived: 5.9 ± 4.9 mmol/l and 7.22 ± 0.14, respectively. The patients who experienced mortality also had considerably lower total protein and platelet levels. Additional SOFA values at all days had a statistically significant relationship with death (p <0.001). Additionally discovered to be substantially

linked with mortality was urinary NGAL.Renal replacement therapy was necessary for 13 members of the deceased cohort (38.2%), as opposed to 15 survivors (22.4%). With a P value of 0.093, the difference between the two was not statistically significant. Therefore, in a univariate study, a requirement for hemodialysis did not have a mortality association.

Multivariate analysis for Mortality predictors: A model for predicting death was created using stepwise backward logistic regression and has a predictive value of 77.6%. The most effective independent predictor of mortality in the model was SOFA score at day 2 (P = 0.006), followed by pH at presentation (P = 0.015), and SOFA score at day 1 (P = 0.026). Although urine NGAL exhibited a borderline relevance in predicting mortality in a multivariate model (P=0.078), it was significantly linked with mortality in a univariate analysis.

Table 15: Predictors of mortality in multivariate analysis

ICU patients' morbidity and mortality are significantly impacted by acute renal damage. It is widely recognised that it lengthens hospital stays and increases the risk of developing chronic kidney disease. The ability of biomarkers to identify and predict outcomes at various stages of AKI has been demonstrated by studies of biomarkers in ICU AKI. We designed the study to examine the clinical characteristics and prognoses of incident and prevalent AKI patients in the medical ICU and analyse urinary NGAL as a prognostic indicator. Study Population characteristics: The study's participants were drawn from the Medical Intensive Care Unit of our hospital, which receives a large number of unwell patients from many medical specialties, including gastroenterology, neurology, internal medicine, nephrology, etc. As a consequence, the findings of this study are relevant to any general medical ICU population. In the research cohort, which included a sizable proportion of older patients, men were more prevalent than women. At the time of inclusion, patients with AKIN stages of 1 - 37%, 2 - 30%, and 3 - 33% were included.Causes of AKI: In our study, sepsis, which affected 58.3% of patients, was the main factor in acute kidney injury. This is consistent with the fact that ARDS and sepsis are the two major clinical disorders seen in the ICU. With regard to Septic AKI patients produced less urine on average than other AKI patients (P=0.015). Additionally, they exhibited noticeably reduced serum albumin levels, which indicated a catabolic condition. Despite the lack of statistical significance, individuals with septic AKI also had higher mortality and longer hospital stays. Sepsis is the most common cause of AKI in this cohort, accounting for an estimated 47.5% of the burden, according to the Beginning and End Supportive Therapy (BEST) kidney investigators, who examined a broad sample of critically ill patients in 54 hospitals across 23 countries. Additionally, it was shown that the septic AKI patients had more oliguria (67% vs 57%, P 0.001), higher inpatient mortality (70.2% vs 51.8%, P < 0.001), and longer hospital stays (37d vs 21d, P 0.001). [6] Drugs, scrub typhus, pigment nephropathy, and cardiorenal syndrome were additional significant causes. Even though scrub typhus is not a particularly well-known AKI cause, it is fairly prevalent in South India and is linked to renal impairment.In fact, Basu et al. studied the several tropical acute febrile diseases that result in AKI. Overall, scrub typhus accounted for 53% (80) of AKI; additional causes included malaria, Leptospirosis, and Dengue [7].The sample included a sizable percentage of post-cardiac arrest AKI patients who needed ongoing inotropic treatment after resuscitation. According to Chua et al., among cardiac arrest survivors, the requirement for inotropes after 24 hours of recovery of spontaneous circulation increases the risk of developing AKI (51.7% versus 6.4%) Predictors of Renal outcome: Diabetes mellitus, the SOFA score at day 2, and baseline renal function were significant predictors in the multivariate analysis of renal outcome. With an area under the ROC curve of 0.71, urinary NGAL was significant in predicting renal outcome in univariate analysis. Notably, SOFA ratings were consistently shown to be considerably high in individuals whose renal function did not improve. The risk factors for developing AKI in the intensive care unit (ICU) and its relationship to multiple organ failure (MOF), which is indicated by a SOFA score of three or higher for two or more organ systems other than the kidney, were explored by Mendonca et al. AKI and multiple organ dysfunction were strongly associated. 241 of the 348 AKI patients had MOF. Of them, 25% of the patients developed before AKI, 10% after AKI, and 65% had AKI concurrently with MOF. From the foregoing explanation, it may be inferred that the coexistence of other organ failure might further hinder renal recovery. This explains why patients who did not regain renal function by day 7 had considerably higher SOFAvalues. It is necessary to delve at the cause-and-effect relationship between MOF and AKI in more detail. Mortality: In our research sample, overall mortality at day 7 was 34%. The death rate for AKI patients has been reported in a variety of ICU studies to range from 10.9% [8] to 36.3% [9] to 47.5% [10]. Most of these investigations were conducted in icus that combined medical and surgical care. In our investigation, the multivariate logistic regression analysis identified pH on the day of inclusion, SOFA score at days 1 and 2, and a minor contribution from urine NGAL (P = 0.078) as the predictors of death. The mortality area under the ROC curve for NGAL was reported to be 0.81. A tested tool for predicting outcomes in critically sick patients is the SOFA score. According to SOFA, mortality rates range from as low as 3.2% in patients with no organ failure to as high as 91.3% in patients with failure of all six organs [11]. Although the P value for our study's analysis was 0.093 and RRT need is recognised to be an independent risk factor for mortality in ICU [8,12], death was not significantly correlated with it. The sample size can be the cause of this.

Urinary NGAL: At the time of study enrolment, the mean urine NGAL levels were 1270.2 ± 1659.5 ng/ml in AKIN stage 1 patients, 4865.9 ± 4992.3 ng/ml in AKIN stage 2 patients, and 6135 ± 4586.1 ng/ml in AKIN stage 3 patients. Our urine NGAL levels were greater than those that de Geus et al. [12] found in their study of AKI patients in the ICU. He discovered that the median urine NGAL levels in RIFLE "R" were 323 ng/ml, RIFLE "I" were 523 ng/ml, and RIFLE "F" were 2013 ng/ml. The large increase in septic patients in our research sample can be blamed for the disparity in observation. The mean urine NGAL levels in our research were 1270.2 ± 1659.5 ng/ml in AKIN stage 1 patients, 4865.9 ± 4992.3 ng/ml in AKIN stage 2 patients, and 6135 ± 4586.1 ng/ml in AKIN stage 3 patients at the time of enrolment. In our research of AKI patients in the ICU, urine NGAL levels were greater than those seen in de Geus et al's study [13]. The median levels of NGAL in the urine were 323 ng/ml in RIFLE R, 523 ng/ml in RIFLE I, and 2013 ng/ml in RIFLE F, according to his research. Due to the much greater number of septic patients in our research sample, the discrepancy in observation can be explained. Only .17% of the patients in de Geus' research had septic AKI, while a sizable portion of the patients had post-surgical and trauma-related AKI. Our trends of urine NGAL appear to follow a similar pattern to the studies mentioned above and were considerably greater inpatients in the AKIN2 and AKIN 3 stages at inclusion than in those in the AKIN 1 stage. In univariate analysis, urinary NGAL was shown to be a strong predictor of outcomes, including renal recovery at days 3 and 7. In actuality, NGAL had a 0.71 area under the ROC for predicting renal outcome and the need for RRT. This result is supported by a research by Singer et al. [14] where urine NGAL predicted a composite outcome as defined by an upgrade in RIFLE class, the start of dialysis, and death with an area under the receiver operating characteristic (ROC) of 0.71. In contrast, NGAL in our analysis had a superior area under the ROC for predicting death, at 0.81.However, simultaneous inclusion of SOFA scores in multivariate analysis reduced the NGAL's predictive potential.

Limitations of Study: Significant limitations apply to our investigation. First, a large amount of sepsis, including severe sepsis, was present in our AKI patients. As previously mentioned, the cause of urinary NGAL can be activated neutrophils in inflammatory conditions like sepsis or damaged tubules. Thus, NGAL in urine in these situations may really originate from sites other than the kidneys. The results of our investigation, where mean urinary NGAL levels were greater than those seen in other studies, appear to support this condition.Additionally, individuals with septic vs. non-septic AKI had higher urinary NGAL levels (p=0.063).

Therefore, it is possible that the elevated urinary NGAL levels identified in our study were caused by sepsis in addition to AKI. Second, because of the limited sample size, several findings, such as the association between RRT and mortality, could not have been statistically significant.

Conflict of Interest: None

Funding Support: Nil

1.       Hoste EAJ, Schurgers M. Epidemiology of acute kidney injury: How big is the problem? Crit Care Med 2008; 36 (4) (Suppl.):146-151

2.       Srisawat N, Kellum JA .Acute kidney injury: definition, epidemiology, and outcome. Current Opinion in Critical Care 2011;17:548–555

3.       Ronco C, Bellomo R, Kellum JA . The Concept of Acute Kidney Injury and the RIFLE Criteria Acute Kidney Injury. Contrib Nephrol. Basel, Karger, 2007, vol 156, pp 10–16

4.       Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P; Acute Dialysis Quality Initiative workgroup: 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–R212.

5.       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;16:3365–3370.

6.       Bagshaw SM, Uchino S, Bellomo R, Morimatsu H, Morgera S, Schetz M, et al. Septic acute kidney injury in critically ill patients: Clinical characteristics and outcomes. Clin J Am Soc Nephrol. 2007;2:431–9.

7.       Basu G, ChrispalA ,BooruguH.Acute kidney injury in tropical acute febrile illness in a tertiary care centre—RIFLE criteria validation, Nephrol Dial Transplant (2010) 1 – 7.

8.       Thakar CV, Christianson A, Freyberg R, et al. Incidence and outcomes of acute kidney injury in intensive care units: a Veterans Administration study. Crit Care Med 2009; 37:2552–2558.

9.       Cruz DN, Bolgan I, Perazella MA, et al. North East Italian Prospective Hospital Renal Outcome Survey on Acute Kidney Injury (NEiPHROS-AKI): targeting the problem with the RIFLE Criteria. Clin J Am Soc Nephrol 2007; 2:418–425.

10.    1Abosaif NY, Tolba YA, Heap M, et al: The outcome of acute renal failure in the intensive care unit according to RIFLE:Model application, sensitivity, and predictability.Am J Kidney Dis 2005; 46:1038 –1048.

11.    Moreno.R,VincentJL,MatosR.The use of maximum SOFA score to quantify organ dysfunction/failure in intensive care.Results of a prospective multicentre study. Intensive Care Medicine 1999;25:686-696.

12.    Piccinni P, Cruz DN, Gramaticopolo S et al .Prospectivemulticenter study on epidemiology of acute kidney injury in the ICU: a critical care nephrology Italian collaborative effort (NEFROINT) Minerva Anestesiol 2011;77:1072-83

13.    Hilde R. H. de Geus, Bakker J, Emmanuel M. E. H. Lesaffre, Jos L. M. L. le Noble. Neutrophil Gelatinase-associated Lipocalin at ICU Admission Predicts for Acute Kidney Injury in Adult Patients. Am J Respir Crit Care Med 2011; 183; 907–914.

14.    Singer.E, Elger A, Elitok S, Kettritz R et al. Urinary neutrophil gelatinase-associated lipocalin distinguishes pre-renal from intrinsic renal failure and predicts outcomes. Kidney International 2011;80: 405–414.