Predictive Factors and Risk Scoring System for Acute Kidney Injury (AKI) in Sick Neonates-A Prospective Cohort Study

doi:10.21203/rs.3.rs-4839389/v1

Download PDF

Research Article

Predictive Factors and Risk Scoring System for Acute Kidney Injury (AKI) in Sick Neonates-A Prospective Cohort Study

https://doi.org/10.21203/rs.3.rs-4839389/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background

Neonatal AKI (Acute kidney injury) is an underreported entity in sick neonates associated with significant morbidity and mortality.

Methods

It was a prospective cohort study, to study the incidence, risk factors, and outcomes of AKI among sick neonates. We included sick inborn neonates admitted at a level III neonatal intensive care unit. Neonates with congenital anomalies and who died within 72 hours of life were excluded. AKI was defined and categorized as per KDIGO guidelines.

Results

Two hundred and seventy-six neonates were enrolled over 1 year, of which 115(42%) had AKI. AKI was higher in the first week of life as compared to beyond the first week n = 103/115 (89%) vs. n = 12/115 (7%), p = < 0.01). The incidence of AKI was highest n = 27/38 (71%) among extremely preterm (< 28 weeks) infants. On Cox regression analysis, sepsis, invasive ventilation, acidosis, and perinatal asphyxia were significantly associated with AKI with a hazard ratio (95% CI) of 4 (1.21–13.42), 2.3 (1.32–4.03), 1.9 (1.13–3.36), and 1.5 (1.04–2.31), respectively. The risk prediction model, using the 4 predictors mentioned above, had good diagnostic accuracy (area under the curve- 83.6%) with a sensitivity and specificity of 77% and 80%, respectively. Infants with AKI have significantly higher mortality, compared to those who did not have AKI n = 45/115(39%) vs. n = 5/161 (3%), p < 0.01.

Conclusion

Nearly half of sick neonates admitted to NICU have AKI, and is maximum in extremely preterm infants. Sepsis, invasive ventilation, acidosis, and perinatal asphyxia have good diagnostic accuracy in identifying neonates likely to develop AKI.

Asphyxia, prematurity, sepsis, shock, hypotension, drugs, congenital heart diseases contribute to neonatal AKI.

What this study adds:

Our simple risk prediction model can be used in sick neonates to identify infants who are at risk for developing AKI.

Neonatal AKI is associated with significant mortality and morbidity. Early identification and prevention of severe forms aids to prevent its complications and mortality in sick neonates.

Acute kidney injury (AKI) is common among sick neonates and is a major contributor to neonatal morbidity and mortality. Newborns are more vulnerable to AKI due to their functionally immature kidneys. There is a wide variation in the incidence of AKI, ranging from 8.4 to 63.3% (1), owing to a lack of uniformity in the AKI definition. Furthermore, there are wide variations in the 'normal' creatinine and urine output in neonates, further complicating the issue. Accurate urine output measurement and the staging of AKI based on urine output alone are challenging in neonates. KDIGO guidelines were proposed in 2012, and since then, there has been some uniformity in reporting the incidence and risk factors. Extreme prematurity, neonatal sepsis, perinatal asphyxia, congenital heart disease, need for invasive ventilation, inotropic usage, nephrotoxic drug usage, etc., are major risk factors associated with AKI development (2–5). Neonates with AKI are at much higher risk for mortality than those without (6–12). Although a large proportion of neonates admitted to the intensive care unit develop AKI, literature is scarce on its predictors and risk stratification in neonates (13).

In this prospective observational cohort study, we aimed to assess the incidence, risk factors, and outcomes of AKI in sick neonates and to develop a risk prediction model for identifying neonates likely to develop AKI during hospital stay.

This prospective observational cohort study was conducted in the level III neonatal unit of northern India from March 2022 through May 2023. The institute’s ethics committee approved the study,and infants were enrolled after getting written informed consent from the parents. We enrolled sick infants during first 28 days of life and sickness was defined as any one of the following: if they received either intravenous fluids for at least 72 hours or oxygen $\:\ge\:$24 hours, or developing sepsis or shock or had perinatal asphyxia. Infants with congenital anomalies, severe hydronephrosis (anteroposterior diameter >15 mm), and those who died within 72 hours of life were excluded. These neonates were followed up prospectively, and baseline urea and creatinine were obtained on the day of enrolment. If the renal function test (RFT) was normal at enrolment, it was repeated twice in the first week of admission as per study protocol, following which atleast once in a week or more frequently as per the treating clinician. If abnormal, RFT was done more frequently(24th hourly) till resolution of AKI. Urine output was measured daily for the first 2 weeks of admission in non-AKI cases and till resolution in AKI cases. It was done by diaper weighing in non-catheterised babies. Daily weight, fluid intake and output, blood pressure, usage of nephrotoxic medications, respiratory support, and all body fluid cultures were noted. Information about neonates requiring renal replacement therapy was recorded. Neonates with AKI as per KDIGO classification (14) were managed as per the institutional protocol. Data was collected till discharge/death, whichever was earlier.

Definitions used: Neonatal KDIGO criteria (14) was used for defining and staging AKI. Reference basal creatinine was chosen according to gestation. Median baseline creatinine of 0.88mg/dL for 24–36 weeks gestation, and 0.53 mg/dL for term gestation was used (15, 16). In the initial 2–3 days, as neonate’s creatinine is reflection of maternal creatinine; neonatal AKI was diagnosed considering the serum creatinine values after 48 hours of life and urine output from day 2 of life. Early AKI was defined as AKI within the first 7 days of life. Asphyxia was defined as those cases who required positive pressure ventilation for $\:\ge\:$ 60 seconds at birth.

Primary outcome was incidence of AKI in sick neonates, secondary outcomes were to identify the risk factors, outcomes of AKI in the form of resolution or mortality and building a risk prediction model.

Sample size

Based on a previous study, the estimated incidence of AKI was 20% (8–20%) (17). To have an absolute precision of 5%, we needed to enroll 153 neonates.

Statistical analysis

Data was analyzed with SPPS version 26. Normality distribution was tested using the Kolmogorov-Smirnov test. Descriptive statistics were calculated. Univariate analysis of continuous variables was done using Wilcoxon's rank-sum test and compared by Mann Whitney U test, and for categorical variables, chi-square or Fischer exact test was done. Multivariate analysis was done for those risk factors that were significant on univariate analysis with considering a 1:10 ratio, i.e., for 115 AKI cases, 11 significant risk factors were evaluated in multivariate logistic regression and Cox regression analysis.

To generate an AKI predictive model- The score obtained from the Cox regression coefficient ($\:\beta\:$) was used. The best-fit model was chosen, whose predictive validity was based on standard statistical measures like specificity, sensitivity, negative predictive value, positive predictive value, and area under the ROC curve.

We screened 340 neonates for eligibility, of which 276 met the inclusion criteria and were enrolled. The mean(SD) gestational age was 31.4 (3.6) weeks and median(IQR) birth weight 1315 (1018,1835) grams, respectively. The incidence of AKI was 115/276 (42%) among sick enrolled neonates of which 63/115(54%) had stage 1 AKI and both stage 2 and stage 3 were each observed in 26/115(23%). Renal replacement therapy was done in 2 cases. On subgroup analysis, AKI was highest in the extremely preterm group ($\:\le\:$ 28 weeks, 27/38 (71%), followed by late preterm (34-36weeks, 15/34 (44%) and in very preterm (29-33weeks, 59/166 (36%). AKI in term 14/38(37%) v/s preterm 101/238(42%). Gestational age was inversely proportional to the incidence of AKI. Incidence of AKI was higher in SGA babies 41/90 (46%) compared to AGA babies 70/182(38%) though not statistically significant.

We assessed the risk factors of AKI (Table 1). On multivariate logistic regression, sepsis, asphyxia, invasive ventilation, acidosis, and IVH were statistically significant, and on Cox regression analysis, asphyxia, sepsis, invasive ventilation, and acidosis were significantly associated with AKI. Sepsis, shock, requiring invasive ventilation, hypotension, and acidosis were significant risk factors for AKI in preterm infants.

Table 1

Risk factors of Acute Kidney Injury (AKI) (n = 276)
Sl no	Risk factors	AKI group (N = 115)	Non-AKI group (N = 161)	Univariate analysis Odds ratio (95% CI)	P value	Multivariate analysis odds ratio (95% CI)	P value
1.	Perinatal Asphyxia	45 (39%)	36(23%)	1.5 (1.1–1.9)	< 0.01	2.5 (1.2–5.2)	0.01
2.	Sepsis	112(97%)	127(79%)	9.9(2.9–33.4)	< 0.01	8.1 (1.8–35.3)	< 0.01
3.	Shock	62(54%)	18(11%)	9.9(5–17)	< 0.01	1.4 (0.5–3.7)	0.41
4.	Congenital heart diseases *	47 (41%)	30 (19%)	2.9 (1.7-5)	< 0.01	1.02 (0.4–2.2)	0.95
5.	NEC	14(12%)	4(3%)	5.4 (1.7–16.9	< 0.01	1.2 (0.2–5.5)	0.78
6.	IVH	49(43%)	17(11%)	6.3(3.3–11.7)	< 0.01	2.6 (1.2–5.9)	0.01
7.	Invasive ventilation	84(73%)	32(20%)	10.9(6.2–19)	< 0.01	3.0 (1.4–6.5)	< 0.01
8.	Fluid overload	28(24%)	4(3%)	12.6(4.3–37)	< 0.01	3.9 (0.8–19.2)	0.09
9.	Hypothermia (Moderate-Severe)	17(15%)	9(6%)	2.9(1.2–6.8)	0.01	1.1 (0.3–4.1)	0.80
10.	Hypotension	45 (39%)	4 (3%)	24 (8.5–71)	< 0.01	1.6 (0.3–7.3)	0.50
11.	Acidosis(pH < 7.30)	59 (51%)	11 (7%)	14 (6.8–28)	< 0.01	3.1 (1.1–8.2)	0.02
12.	Nephrotoxic drug usage	103 (89%)	103 (65%)	2.6 (1.6–4.4)	< 0.01	1.0 (0.4–2.3)	0.97
Variables in tables expressed as n(%)
P < 0.05 is significant, NEC-Necrotising enterocolitis, IVH-intraventricular hemorrhage,
*Congenital heart diseases- included hemodynamically significant patent ductus arteriosus(hs-PDA), ventricular septal defect, atrial septal defect

It was observed that asphyxia-related AKI had higher rates of resolution 28/45(63%) and less mortality 16/45(36%) compared to sepsis-related AKI whose resolution was 66/112(59%) and mortality 45/112(40%) p value < 0.01 and shock-related AKI with resolution of 19/62 (31%) and mortality of 43/62 (69%), p < 0.01. Oliguric AKI was observed in 37/115 (31%) cases. Median(IQR) duration of hospitalisation in AKI group was higher than non-AKI group (Table 2).

Table 2

Outcomes of AKI
Sl no	Outcomes	Non-AKI (N = 161)	Total AKI (N = 115)
1.	Resolution of AKI * a) AKI-stage 2(n = 26) b) AKI-stage 3(n = 26)	-	69/115 (60%) 14/26 (54%) 3/26 (12%)
2.	Days taken for resolution (days) -^# a) AKI-Stage 2# b) AKI-Stage 3#	-	9 (1.6) 5 (1.7) 7.3 (4)
3.	Mortality * a) AKI-stage 2* b) AKI-stage 3*	5 (3%)	45/115 (39%) 11/26 (42%) 23/26 (89%)
4.	Duration of hospitalization days ^$	24 (15,41)	30 (16,50)
Variables in tables expressed as ^#mean (SD), ^$Median (IQR) and ^*n(%)

We constructed a risk prediction model with significant risk factors on Cox regression analysis (Table 3). The $\:\beta\:$-coefficient value for each risk factor was multiplied by 2 to obtain a whole number, and the scores were assigned. From this scoring system (Table 3), we calculated the total HR score for all 276 cases (115 AKI cases and 161 non-AKI cases). Then, the ROC curve (Fig. 1) was generated from this scoring to identify the cut-off points at which this scoring model can better predict the risk of developing AKI in the enrolled sick neonates. We could obtain a curve that was above the random classifier diagonal line closer to the top-left corner, with the area under the curve of 0.836 (95%CI: 0.786–0.885), suggesting that the model is capable of effectively predicting the risk of AKI in the enrolled sick newborn population. From the coordinates of the ROC curve, it has been observed that a cut-off of $\:\ge\:$ 4 would be significant for predicting the high risk of AKI, as at this cut-off, there is the highest sensitivity and specificity. We constructed a 2*2 table (Table 4) to see the efficacy of a cut-off score $\:\ge\:$ 4 to predict the AKI risk better. We found that the odds of developing AKI were 13.7 times higher if the score was $\:\ge\:$4, with a sensitivity of 77%, specificity of 80%, and p-value of <0.001. Negative predictive value of 83% and positive predictive value of 74%.

Table 3

AKI risk prediction scoring system from Hazards ratio and Cox regression
Sl no	Risk factors	Hazards ratio (95% CI)	P value	$\:\varvec{\beta\:}$-coefficient	$\:\varvec{\beta\:}$-coefficient *2	Score assigned
1	Sepsis	4 (1.21–13.42)	0.02	1.39	2.78	3
2	Invasive ventilation	2.3 (1.32–4.03)	< 0.01	0.83	1.66	2
3	Asphyxia (PPV $\:\ge\:$ 30 secs)	1.5 (1.04–2.31)	0.03	0.44	0.88	1
4	Acidosis	1.9 (1.13–3.36)	0.01	0.67	1.34	1
TOTAL SCORE =						7

Table 4

Total score cut off $\:\ge\:$ 4 in AKI v/s Non-AKI group
Total HR score $\:\ge\:$ 4	AKI	Non-AKI
YES (N = 121)	89 (74%)	32 (26%)
NO (N = 155)	26 (17%)	129 (83%)
P value < 0.001, Odds ratio (95% CI) = 13.7 (7.6–24.7), Sensitivity = 77%,Specificity = 80%, Negative predictive value = 83%,Positive predictive value = 74%

The weight centile at discharge, it was noted that 44% of non-AKI children weighted 10-90th centile as compared to 11% in the AKI group (p < 0.01)

Nearly half of the enrolled sick neonates had AKI according to KDIGO criteria, with the highest incidence among the extremely preterm group. On Cox regression, asphyxia, sepsis, invasive ventilation, and acidosis were significant risk factors for AKI. Resolution of AKI was seen mostly in asphyxia cases, and mortality was highest in stage 3 AKI.

The incidence of neonatal AKI varied from 8.4–63.3%(1). In the AWAKEN study(2), the incidence of AKI was 48% in extremely preterm and 37% in the very preterm and late preterm combined group. Our results are almost similar to the AWAKEN study.

We found 4 risk factors (asphyxia, sepsis, invasive ventilation and acidosis) were significant in our study. Of these perinatal asphyxia, sepsis, invasive ventilation were consistent significant risk factors in other studies (1, 2, 5, 18, 19). Intraventricular hemorrhage has also been found as significant risk factor of AKI on multivariate analysis in our study, as Algadeep et al (19) noted. Amongst the physiological and lab parameters we found acidosis (pH < 7.30) was notifiable significant risk factor for AKI. This parameter has not been reported in previous studies.

In systematic review(1), there was increased risk by 3 times for congenital heart disease (CHD), 6 times for Necrotizing enterocolitis (NEC), in our study these risk factors were significant on univariate analysis but on multivariate analysis and cox regression analysis they were not statistically significant. Probable reasons being our smaller sample size compared to systematic review, differences in categories included under CHD and NEC, also possibly other variables included in our multivariate analysis have stronger influence on the risk of AKI, masking the effects of CHD and NEC.

Data from multiple studies showed that AKI is related to poor outcomes (2–4, 20). In Jetton et al (21) study, neonates with AKI had 4.6 times higher mortality and 8.8 times longer duration of hospitalization compared to non-AKI group. In our study mortality rate in AKI was 45/115 (39%) compared to non-AKI group which was 5/161 (3%) p < 0.01. Late AKI had higher mortality than early AKI. In our study we uniquely also studied the outcome of AKI as per the cause, and found that AKI resolution rates were better in asphyxia group compared to sepsis and shock group.

We also studied the anthropometric parameters like weight and head circumference centile at discharge, it was noted that discharge weight centile was better in non-AKI group compared to AKI group. Though this is a short term outcome, but might act as baseline data for longterm follow up, as the data regarding these outcomes are limited from available literature.

Recently Wazir et al. (13) proposed a neonatal AKI risk stratification score (STARZ). Variables used were age at NICU admission, need for PPV at labour room, < 28 Weeks of gestation, significant heart disease, sepsis, post admission (within 12 hours) of serum creatinine $\:\ge\:$ 0.98 mg/dl, urine output < 1.32 ml/kg/hour, nephrotoxic drugs, inotropes usage. It predicted the increased risk of AKI at 7 days after admission. The scoring model had 95.6% negative predictive value, 80.5% positive predictive value, 87.4% specificity, and 92.8% sensitivity. Though it had a better sensitivity and specificity, it had many variables and needs validation in larger cohort. Similarly, Qian hu et al (22) developed a predictive nomogram for the early detection and management of VLBW infants at high risk of developing AKI from a retrospective study. Gestational age, red blood cell count within 3 days of birth, serum calcium concentration within 3 days of birth, maternal age $\:\ge\:$ 35 years and pulmonary arterial hypertension or myocardial injury were the predictive factors in their model.The area under the curve(AUC) was 0.794 (95% CI: 0.754-0834), and sensitivity was 75% specificity of 71% here they had followed only the the serum creatinine criteria of KDIGO definition. This study has variables whose association with AKI are less described in literature, requiring further validation. Compared to these, our AKI risk prediction scoring system is easy to use with less easily accessible components and better predictive value.

Strengths of the study

Firstly, this was a prospective cohort study, done following adherence to KDIGO criteria and secondly has a sufficient sample size for a risk prediction model that can be used for early identification of AKI.

Limitations of the study

Though we had used KDIGO criteria for defining AKI, we used the reference baseline creatinine from the mean values of the previously established nomograms. This might be one of the reasons for a higher incidence of AKI in our population. External validation of our risk prediction model to be done.

Nearly half of the enrolled sick neonates had AKI. Significant risk factors for AKI are asphyxia, sepsis, respiratory failure requiring invasive ventilation and acidosis. The risk prediction scoring system can be used in sick neonates to predict AKI, an important contributor to neonatal mortality. Early identification of risk factors, risk prediction scoring, and prevention of severe forms of AKI may prevent complications and mortality in sick neonates.

Author Contribution

RK: Conceptualised and designed the study, collected and analyzes data, and drafted the manucript.KM: Conceptualised and designed the study, supervised data collection, analysis, and critically reviewed the manuscript.JK: Reviewed and scrutinised the protocol and manuscript.All the authors approved the final version and shall be accountable for all aspects.

Financial disclosure: None

Conflict of interest: None

Ethical Permissions: The study was approved by the Institute Ethics Committee (IEC-INT/2022/DM-289) of PGIMER Chandigarh.

Hu Q, Li SJ, Chen QL, Chen H, Li Q, Wang M (2021) Risk Factors for Acute Kidney Injury in Critically Ill Neonates: A Systematic Review and Meta-Analysis. Front Pediatr 9:666507
Jetton JG, Boohaker LJ, Sethi SK, Wazir S, Rohatgi S, Soranno DE et al (2017) Incidence and outcomes of neonatal acute kidney injury (AWAKEN): a multicentre, multinational, observational cohort study. Lancet Child Adolesc Health 1(3):184–194
Lee CC, Chan OW, Lai MY, Hsu KH, Wu TW, Lim WH et al (2017) Incidence and outcomes of acute kidney injury in extremely-low-birth-weight infants. PLoS ONE 12(11):e0187764
Mwamanenge NA, Assenga E, Furia FF (2020) Acute kidney injury among critically ill neonates in a tertiary hospital in Tanzania; Prevalence, risk factors and outcome. PLoS ONE 15(2):e0229074
Gohiya P, Nadkarni J, Mishra M (2022) Study of neonatal acute kidney injury based on KDIGO criteria. Pediatr Neonatol 63(1):66–70
Selewski DT, Jordan BK, Askenazi DJ, Dechert RE, Sarkar S (2013) Acute kidney injury in asphyxiated newborns treated with therapeutic hypothermia. J Pediatr 162(4):725–729e1
Koralkar R, Ambalavanan N, Levitan EB, McGwin G, Goldstein S, Askenazi D (2011) Acute kidney injury reduces survival in very low birth weight infants. Pediatr Res 69(4):354–358
Viswanathan S, Manyam B, Azhibekov T, Mhanna MJ (2012) Risk factors associated with acute kidney injury in extremely low birth weight (ELBW) infants. Pediatr Nephrol Berl Ger 27(2):303–311
Mathur NB, Agarwal HS, Maria A (2006) Acute renal failure in neonatal sepsis. Indian J Pediatr 73(6):499–502
Taylor ML, Carmona F, Thiagarajan RR, Westgate L, Ferguson MA, del Nido PJ et al (2013) Mild postoperative acute kidney injury and outcomes after surgery for congenital heart disease. J Thorac Cardiovasc Surg 146(1):146–152
dos Halal SE, Carvalho MG (2013) Acute kidney injury according to pediatric RIFLE criteria is associated with negative outcomes after heart surgery in children. Pediatr Nephrol Berl Ger 28(8):1307–1314
Aydin SI, Seiden HS, Blaufox AD, Parnell VA, Choudhury T, Punnoose A et al (2012) Acute kidney injury after surgery for congenital heart disease. Ann Thorac Surg 94(5):1589–1595
Wazir S, Sethi SK, Agarwal G, Tibrewal A, Dhir R, Bajaj N et al (2021) Neonatal acute kidney injury risk stratification score: STARZ study. Pediatr Res
Zarbock A, John S, Jorre A, Milles KD (2014) Kidney Disease: Improving Global Outcome New KDIGO guidelines on acute kidney injury. Practical recommendations Anaesthetist 63:578–588
Kastl JT (2017) Renal function in the fetus and neonate – the creatinine enigma. Semin Fetal Neonatal Med 22(2):83–89
Finney H, Newman D, Thakkar H, Fell J, Price C (2000) Reference ranges for plasma cystatin C and creatinine measurements in premature infants, neonates, and older children. Arch Dis Child 82(1):71–75
Yang H, Lin C, Zhuang C, Chen J, Jia Y, Shi H et al (2021) Serum Cystatin C as a predictor of acute kidney injury in neonates: a meta-analysis. J Pediatr. Oct 1
AlGadeeb K, Qaraqei M, Algadeeb R, Faqeehi H, Al-Matary A (2021) Prediction of risk factors and outcomes of neonatal acute kidney injury. J Nephrol 34(5):1659–1668
Elmas AT, Tabel Y, Özdemir R (2018) Risk factors and mortality rate in premature babies with acute kidney injury. J Clin Lab Anal 32(7):e22441
Kapoor K, Jajoo M, Dabas V (2013) Predictors of mortality in out born neonates with acute renal failure; an experience of a single center. Iran J Pediatr 23(3):321–326
Jetton JG, Askenazi DJ (2014) Acute kidney injury in the neonate. Clin Perinatol 41(3):487–502
Hu Q, Shi Y, Hua ZY, Bao L, Li F, Wei H et al (2021) Jan A Prediction Nomogram for Acute Kidney Injury in Very-Low-Birth-Weight Infants: A Retrospective Study. Front Pediatr

No competing interests reported.

Download PDF

Editorial decision: Revision requested
27 Aug, 2024
Reviews received at journal
25 Aug, 2024
Reviewers agreed at journal
08 Aug, 2024
Reviewers invited by journal
08 Aug, 2024
Editor assigned by journal
02 Aug, 2024
Submission checks completed at journal
02 Aug, 2024
First submitted to journal
01 Aug, 2024

You are reading this latest preprint version

Predictive Factors and Risk Scoring System for Acute Kidney Injury (AKI) in Sick Neonates-A Prospective Cohort Study

Status:

Version 1

Abstract

Background

Methods

Results

Conclusion

Figures

What is already known

Key messages

Introduction

Methods

Sample size

Statistical analysis

Results

Discussion

Conclusions

Declarations

Author Contribution

References

Additional Declarations

Status:

Version 1