Epidemiological study of hospital acquired acute kidney injury in critically ill and its effect on the survival

Abstract

In the intensive care unit (ICU), acute kidney injury (AKI) is the most common cause of morbidity and mortality. Hospital-acquired acute kidney injury (HAAKI) is AKI developing after 48 h. We aimed to study the development of AKI and its associated risk factors. We conducted a longitudinal observational study. Inclusion criteria were patients > 18 years of age admitted to ICU. The primary outcome was the development of AKI as defined by Kidney Disease Improving Global Outcomes (KDIGO) criteria. A total of 273 patients were included in the study. Out of 273, 44(16.11%) patients developed AKI. The mean age was 45.80(17.39) years, and 60.81% were males. The median acute physiology and chronic health evaluation (APACHE II) and sequential organ failure assessment (SOFA) scores were 12(8–18) and 5(3–7), respectively. Diabetes mellitus (23.44%) and hypertension (23.81%) were predominant comorbidities. The risk factors associated with AKI were serum chloride level, colistin, invasive ventilation, positive end-expiratory pressure (PEEP), and fluid balance. The hospital mortality was significantly higher in patients with AKI (43.18%) as compared with no AKI (14.41%). Among the secondary outcomes, 7 (15.90%) patients required renal replacement therapy (RRT) during hospitalisation. The length of ICU stay was higher in patients with AKI 8(5–13) compared to no AKI 5(3–8). A total of 16.11% developed HAAKI, and mortality was 43.18%. Post 6 months follow-up of AKI patients, mortality was 23%. Among survivors none of the patients were on RRT.Patients admitted with normal kidney function can develop AKI. Hence, careful monitoring of ICU patients is necessary.

Introduction

In the ICU, AKI develops due to a renal cause or as an organ failure secondary to nonrenal insult^1–3. The most common etiology for the development of AKI is sepsis, and it is observed in 50% of the patients with AKI³. AKI is associated with prolonged hospitalisation, need for renal replacement therapy, and has huge cost implications. AKI can progress to acute kidney disease or chronic kidney disease in some patients⁴.

The incidence of AKI ranges from 20 to 57%^1,5,6. In the multinational AKI-EPI study of 1032 ICU patients, > 50% of the patients who developed AKI had higher mortality¹. The types of AKI are community-acquired AKI and hospital-acquired AKI (HAAKI)⁷. Wonnacott A, et al. showed although the incidence of community-acquired AKI was higher (4.3%) and was severe in nature it had a better survival (45% mortality) than HAAKI (62.9% mortality)⁷.

Meta-analysis by Huang, et al., showed mortality and other outcomes were better in community acquired AKI than in hospital acquired AKI⁸. HAAKI is seen in older patients and carries higher mortality. The risk factors associated with HAAKI are nephrotoxic medications and sepsis⁹. Limited studies describe characteristics, risk factors associated with the HAAKI, follow up and outcome of these patients^8–10.

We aimed to evaluate the proportion of ICU patients developing AKI, risk factors associated with it, and survival rates of AKI patients after hospital discharge over six months in a longitudinal study.

Methodology- A prospective observational study was conducted in a tertiary care hospital from 5th September 2018 to 16th October 2023. Institutional ethics committee approval was obtained (IEC205/2018 dated 28th August 2018, CTRI/2018/09/015615, 5th Sept2018). All the procedures were followed in accordance with the ethical standards of the responsible committee on human experimentation (institutional or regional) and with the Helsinki Declaration of 1975. Strengthening the reporting of observational studies in epidemiology (STROBE) guidelines were followed (Strobe checklist). The written informed consent was obtained from the legally acceptable representative (LAR). The sample size 246 was adequate, with 20% incidence of AKI, 95% confidence interval, and 5% relative precision. With 10% missing data, the total sample size required was 271. The inclusion criteria were patient’s ≥ 18 yrs. of age, admitted in the ICU. Patients were enrolled within 24–48 h. of hospitalization. Patients presenting with AKI or had chronic kidney disease were excluded. Pregnant patients and terminally ill patients were excluded. Diagnosis of HAAKI was made based on the admission creatinine and using KDIGO guidelines. Baseline and demographic characteristics were collected. Acute physiology and chronic health evaluation (APACHE II) and sequential organ failure assessment (SOFA) scores were calculated. Laboratory parameters, serum creatinine, blood urea, electrolytes, fluid balance, need for renal replacement therapy, vasopressors, and ventilator parameters (noninvasive or invasive) were collected for the first seven days of ICU admission. Post ICU transfer, patients were followed up till the hospital discharge. At 6 months from discharge telephonic follow up was done. Survival outcome in the AKI patients was collected.

Statistics

Continuous variables presented as mean (standard deviation SD) or median (interquartile range IQR) as applicable. Categorical data presented as percentages (%). Clinical characteristics compared between AKI and no AKI using independent ‘t test’ or ‘Mann Whitney U’ test as applicable. A generalized linear mixed model analysis was performed to assess the clinical factors associated with the development of AKI over 7 days. Age, gender, history of diabetes mellitus, APACHE II score were considered in the regression model for adjusted analysis. Risk factors associated with the hospital mortality were analysed. Estimation of survival probabilities was done using Kaplan-Meier and log-rank test. The p-value < 0.05 was considered statistically significant. Statistical analysis was done using STATA 15.0.

Results

A total of 273 patients were included in the analysis. The mean age was 45.80(standard deviation SD 17.39) yrs., and predominantly were males (60.81%). Out of 273, 44 (16.11%) patients developed AKI. Among 44 patients, 28(63.64%), 11(25%) and 5(11.36%) patients developed stage 1, 2 and 3 AKI respectively. Total 7 patients (15.90%) required renal replacement therapy.

Comparison of baseline characteristics of the patients with AKI and no AKI is presented in Table 1. Mean age of the AKI patients was significantly higher than no AKI group (p < 0.01). Presence of any comorbidities was significantly associated with the development of AKI (p = 0.027). Diabetes mellitus and ischemic heart disease (IHD) were significantly associated with the development of AKI. Patients of AKI had significantly higher median APACHE II 16 (12–23) score, higher requirement of vasopressors on day1, and invasive ventilation. Postoperative status and sepsis were not significantly associated with the development of AKI. Readmissions were significantly associated with the development of AKI.

Table 1.

Baseline characteristics of the patients.

Parameters	All N = 273	AKI N = 44	No AKI N = 229	p value
Age¥	45.80(17.39)	50.61(17.53)	44.88(17.25)	0.045
Gender Male/Female	166/107 (60.81/39.19)	31/13 (70.45/29.55)	135/94 (58.95/41.05)	0.152
Comorbidities	120(43.96)	26(59.09)	94(41.05)	0.027
Diabetes Mellitus	64(23.44)	17(38.64)	47(20.52)	0.009
Ischemic heart disease	11(4.03)	5(11.36)	6(2.62)	0.007
Hypertension	65(23.81)	14(31.82)	51(22.27)	0.173
Chronic liver disease	14(5.13)	3(6.82)	11(4.80)	0.579
Cerebrovascular accident	7(2.56)	3(6.82)	4(1.75)	0.051
Malignancy	7(2.56)	3(6.82)	4(1.75)	0.051
Immunosuppressant	3(1.10)	1(2.27)	2(0.87)	0.415
APACHE II Score*	12(8–18)	16(12–23)	11(8–17)	0.002
SOFA Score*	5(3–7)	5.5(3–8)	5(3–7)	0.063
Vasopressors	45(16.48)	14(31.82)	31(13.54)	0.003
Day1 Invasive ventilation	129(47.60)	28(63.64)	101(44.49)	0.020
Noninvasive ventilation	27(9.96)	4(9.09)	23(10.13)	0.833
Postoperative	52(19.05)	12(27.27)	40(17.47)	0.129
Sepsis	114(41.76)	23(52.27)	91(39.74)	0.123
Readmission	19(6.96)	7(15.91)	12(5.24)	0.011
Admission diagnosis system involved
CNS	79(28.94)	12(27.27)	67(29.26)	0.801
CVS	2(0.73)	0(0)	2(0.87)
PA	33(12.09)	7(15.91)	26(11.35)
RS	65(23.81)	8(18.18)	57(24.89)
Tropical fever	15(5.49)	2(4.55)	13(5.68)
Miscellaneous	79(28.94)	15(34.09	64(27.95)
Secondary outcome
Hospital mortality	52(19.05)	19(43.18)	33(14.41)	< 0.001
ICU Length of stay*	5(4–9)	8(5–13)	5(3–8)	0.001

^¥Mean (Standard Deviation), * Median (Interquartile range).

APACHE II- Acute physiology and chronic health evaluation, SOFA-Sequential organ failure assessment, CNS- Central nervous system, CVS- Cardiovascular system, PA-Per abdomen, RS- Respiratory system, ICU-Intensive care unit.

Primary and secondary outcomes

Total 44(16.11%) patients developed AKI. Overall, the mortality rate was 19.05%. The mortality was significantly higher in AKI (43.18%) as compared to no AKI group (14.41%). The length of ICU stay was significantly higher in AKI group (8 Vs. 5 median days, p = 0.001). Factors associated with the development of AKI are presented in Table 2. Patient with increased serum chloride [1.06 (1.03, 1.08)], requiring invasive ventilation [2.583(1.708, 3.905)], PEEP [1.126 (1.02, 1.24)], received colistin [1.98 (1.22, 3.19)], requiring fentanyl [1.78, (1.28, 2.48)], dexmedetomidine [1.92 (1.02, 3.61)], increased fluid balance [1.05 (1.02, 1.08)] had higher odds for AKI. Low molecular weight heparin (LMWH), furosemide, midazlam and pantoprazole were found to be protective.

Table 2.

Risk factors associated with the development of AKI.

Parameters	Adjusted odds ratio	95%C.I.	p value
Sodium	1.020	0.996, 1.044	0.105
Chloride	1.060	1.032, 1.089	< 0.001
Potassium	0.935	0.742,1.180	0.571
Invasive mechanical ventilation	2.583	1.708,3.905	< 0.001
PEEP (Invasive/Noninvasive)	1.126	1.021, 1.242	0.018
Medications
Steroids	0.731	0.484,1.103	0.136
Colistin	1.980	1.226,3.195	0.005
Piperacillin + Tazobactum	1.120	0.775,1.619	0.547
Vancomycin	0.954	0.603,1.511	0.843
Meropenem	1.119	0.763,1.641	0.563
Acyclovir	1.363	0.624,2.976	0437
Oseltamavir	0.990	0.337,2.908	0.986
Levipril	1.059	0.741,1.514	0.751
Pantoprazole	0.630	0.459,0.864	0.004
Paracetamol	0.887	0.577,1.363	0.585
3%saline	0.923	0.545, 1.564	0.765
Mannitol	0.691	0.431,1.109	0.126
Amikacin	0.462	0.180,1.178	0.106
Heparin	1.178	0.746,1.864	0.481
LMWH	0.523	0.303,0.904	0.020
Sedation	1.708	1.232,2.369	0.001
Fentanyl	1.788	1.289,2.480	< 0.001
Midazolam	0.106	0.013,0.819	0.031
Morphine	3.168	0.946,10.607	0.061
Propofol	0.858	0.287,2.567	0.785
Dexmedetomedine	1.921	1.021,3.612	0.043
Shock	1.113	0.713, 1.737	0.636
Balance	1.050	1.019, 1.082	0.001
Furosemide	0.120	0.016, 0.905	0.040

Adjusted for age, gender, presence of DM and APACHE –II, C.I. Confidence interval.

LMWH- Low molecular weight heparin, MV Mechanical ventilation, NIV Noninvasive ventilation, PEEP Positive end expiratory pressure.

The risk factors associated with hospital mortality were elderly patients, higher APACHE II score, and AKI. Presence of AKI had higher odds [1.626(1.164–2.272)] for hospital mortality (Table 3). Figure 1 shows a comparison of survival plots between AKI and no AKI using Kaplan-Meier survival curve. Significant difference in the survival probability was noted between AKI and no AKI groups showing patients with AKI had significantly lower survival (p = 0.01) (Fig. 1).

Table 3.

Adjusted odds ratio for hospital mortality.

Parameters	Adjusted odds ratio	95% C.I.	p value
Age	1.021	(1.011–1.031)	< 0.001
Gender Male/Female	0.912	(0.691–1.203)	0.516
Diabetes Mellitus	0.774	(0.561–1.069)	0.121
APACHE II Score	1.123	(1.101–1.145)	< 0.001
AKI	1.626	(1.164–2.272)	0.004

APACHE II- Acute physiology and chronic health evaluation, AKI- Acute Kidney Injury, C.I. Confidence Interval.

Fig. 1.

Kaplan-Meier graph showing comparison of survival probabilities between patients with AKI and no AKI.

Among 44 patients of AKI, 25 discharged from the hospital. On follow-up of AKI patients, after 6 months of discharge, data of 22 patients was available. We observed a mortality of (5/22) 23%. Among the patients who were alive none was requiring renal replacement therapy.

Discussion

Among 273 ICU patients, 44 developed acute kidney injury. The mean age was [50.61(SD17.53) yrs] significantly higher in patients who developed AKI in our study. This was similar to the study published by Biradar, et al⁹. Our study had predominantly younger population as against older population seen in sepsis associated AKI¹¹.

We observed 43.18% mortality among AKI patients. Singh, et al., described clinical profile of the patients of HAAKI who were admitted in the medical, surgical ward or intensive care unit showed higher mortality in ICU (73.5%) patients due to multiorgan involvement¹⁰. Multinational, multicentre observational study of AKI also showed mortality of 60.3%¹². Long term follow up of RENAL trial observed 62–63% mortality and 5.1–5.8% patients were on RRT¹³. In our study mortality at the end of six months was 23%. We observed lower mortality in follow up patients possibly due to less severe AKI and none of the patients were requiring RRT.

Presence of one or more comorbidities had higher odds for AKI [2.07(1.07–3.99)]. Among the comorbidities we observed diabetes mellitus and ischemic heart disease were significantly associated with AKI. The study by Hoste, et al. showed higher proportions of patients with diabetes and hypertension developed AKI¹.

We observed administration of intravenous contrast was not associated with HAAKI. The clinical decision about administration of intravenous contrast needs to be individualized after carefully evaluating risk and benefits. Although, antiedema therapy with mannitol and 3% saline was not associated with the development of AKI in our study, higher serum chloride level was associated with AKI. Chloride excess causes vasoconstriction of the afferent glomerulus and decreases glomerular filtration rate¹⁴. This indicates careful monitoring of electrolytes especially chloride in ICU patients. However, the available evidence about hyperchloremia and development of AKI is mixed^15–18. The meta-analysis of 8 studies showed no difference in development of AKI and type of fluid used¹⁸. Surprisingly studies done in pediatric population showed chloride levels associated with the development of AKI¹⁹.

Colistin was associated with higher risk of AKI with the adjusted odds ratio of 1.980(1.226, 3.195). However, study by Rocco et al., did not find association between colistin methane sulphonate (CMS) and development of AKI. The patients included in this study were elderly with median age of 61(43–78) years²⁰. The preparation used in our study was CMS, 1 million equals 80 mg and we used loading dose of 9 million international units (MIU) followed by 4.5 MIU 12 h. Lower dose was used in study by Rocco et al.as compared to our study. Hence while comparing the results emphasis should be given to colistin preparation, concentration per 1 MIU and dose used. The meta-analysis including 34 studies showed colistin is being used as a last therapeutic strategy for multidrug resistant gram-negative bacteria. Hence the development of AKI should not be the limiting factor for its use²¹. Higher dose is necessary to achieve the serum concentration above minimum inhibitory concentration in critically ill patients²².

We did not find any protective effect of steroids in preventing AKI. The study by Rubin et al., in patients of COVID-19 showed patients who received dexamethasone had lesser risk of AKI²³. This observation needs to be studied in future trials. Multicentre study from Belgium in COVID-19 patients found older age, obesity, higher APACHE II and requirement of invasive ventilation were the risk factors associated with AKI²⁴. These were similar to our study.

Interesting observation was patients who were requiring invasive ventilation or positive end expiratory pressure (PEEP) either during invasive or noninvasive ventilation had significantly higher risk of AKI. Similar observation was seen in covid patients and PEEP had shown fivefold increase in the risk of AKI²⁵. The systematic review and meta-analysis of 31 studies showed invasive mechanical ventilation as a risk factor for AKI²⁶. This indicates although PEEP is beneficial in improving oxygenation one need to check its effects on kidney.

Among the sedatives, fentanyl and dexmedetomidine had higher odds for AKI. The post hoc analysis of the randomized controlled trial showed lesser renal impairment was seen with no sedation strategy²⁷. We could not find any evidence showing fentanyl or dexmedetomidine as a possible risk factor for AKI. We observed patients who were on invasive ventilator support were on sedatives. Hence invasive ventilation was a possible risk factor for AKI. Among patients who were on midazolam we observed lesser renal impairment (p = 0.031). The patients who received midazolam were predominantly neurological patients who presented with neurotrauma or had presented with seizures. Antiepileptic effect of midazolam was possibly renal protective. We did not find any association with use of paralytics like vecuronium and AKI.

Positive fluid balance has been shown to be associated with AKI and mortality in various studies^28,29. The fluid overload leads to distension of renal capsule, congestion and causes decrease in glomerular filtration rate. We observed positive fluid balance as a risk factor for AKI and administration of diuretics had a protective effect. Possible reason being diuretics would have helped in managing fluid balance and thereby reducing the development of AKI. The findings of our study are similar to FACTT study, where use of diuretics was not associated with new onset non-pulmonary organ failure^30,31.

We observed lesser risk of AKI with pantoprazole. As per the available literature it is one of the drugs causing interstitial nephritis³². However, there are animal studies reporting protective effect of pantoprazole³³. This needs to be studied in future trials.

Also, LMWH was associated with lesser risk of AKI. We could not find any evidence supporting this finding and possible reason could be formation of microthrombi causing AKI. Similar etiology was found in AKI seen in COVID-19 patients³⁴. This needs to be studied in future trials.

The strengths of our study are this is one of the study that showed the longitudinal data of patients who developed HAAKI and also included follow up of these patients. We observed several risk factors associated with HAAKI. We studied various medications which are routinely used in ICU patients. The findings of this study will help the clinicians in identifying patients at risk of developing HAAKI.

There are certain limitations. AKI was diagnosed based on the serum creatinine only. The measurement of hourly urine output requires intensive monitoring of the patient and is associated with errors. Hence, we used change in the serum creatinine to diagnose AKI. We did not study the effect of antifungals (Amphotericin B) as only 5(1.83%) patients received this drug. The future studies should focus on early diagnosis of AKI using a specific biomarker.

The results of the study showed possible modifiable risk factors like monitoring of serum chloride levels, fluid balance and early weaning thereby reducing the effect of PEEP and sedation. Findings of this study need to be validated in a larger cohort.

Conclusion

The proportion of patients developing AKI was 16.11%. Risk factors associated with AKI were high serum chloride level, colistin, invasive ventilation, PEEP, and fluid balance. Mortality was higher (43.18%) in AKI patients. Among patients of AKI, 7 (15.90%) required RRT. At the end of six-month, mortality was 23%. As AKI is associated with the higher mortality vigilant monitoring for risk factors is necessary in ICU patients.

Author contributions

AAH-Concept, design, data collection, conduct, writing of the final draft, EAC, SBS, MHS, MN - conduct and data collection. SS -Helped in statistical analysis and writing the manuscript. All the authors have approved the final draft.

Data availability

Data will be available on reasonable request from the principal investigator, Dr Amarja Ashok Havaldar.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval and consent to participate

IEC approval was obtained from the respective hospitals. Consent was obtained from the legally acceptable representative (LAR).