Creatinine and Cystatin-C Levels in Late Preterm and Small-for-Gestational-Age Neonates – An Observational Study

INTRODUCTION

Renal function assessment is vital for fluid electrolyte management, drug dosage, and monitoring of sick neonates.^[1] Asphyxia, sepsis, cardiac dysfunction, prematurity, nephrotoxic drugs, congenital anomalies, and maternal illness increase the risk of acute kidney injury (AKI) in neonates.^[2] Current definitions of Neonatal AKI as per Kidney Disease: Improving Global Outcomes criteria and risk, injury, failure, loss of function, and end-stage kidney disease criteria, are based on serum creatinine levels and urine output volumes.^[3-6] Inulin, the gold standard for glomerular filtration rate (GFR), is impractical due to cost and complex measurement.^[7]

Creatinine, a byproduct of muscle metabolism, is freely filtered at the glomerulus with some tubular secretion. Serum creatinine is easy to measure, but maternal influence, low muscle mass, and immature kidney function limit its accuracy for neonatal AKI.^[8] Despite limitations, creatinine is used for neonatal GFR estimation with special formulas.^[7,8] Cystatin-C, a low-molecular-weight inhibitor of proteinase enzymes, is synthesized consistently by all nucleated cells and readily filtered by the glomerulus and subsequently reabsorbed by proximal tubules.^[9] Unlike creatinine, cystatin-C remains unaffected by muscle mass, age, sex, or placental transfer and could accurately measure neonatal GFR.^[10-12]

Cystatin-C could be superior to creatinine for neonatal AKI, due to its stability and minimal maternal influence. In a large cohort study, Xin Xu et al., found cystatin-C detected more cases of neonatal AKI and mortality than creatinine.^[12,13] Neonatal cystatin-C levels differ from maternal levels at birth and levels stabilize by the 3^rd day of life, making it a reliable marker for renal function in healthy term newborns.^[14]

Cystatin-C was initially measured using radioimmunoassay, but nowadays, feasible, accurate, enzymatic immunoassays such as particle-enhanced turbidimetric immunoassays (PETIA) or particle-enhanced nephelometric immunoassay (PENIA) are used.^[9] Both techniques provide reliable reference ranges, independent of gender, rendering them appropriate for clinical application in neonates.^[10,11] Cystatin-C levels are higher in preterm neonates with values of 1.5–1.9 mg/L for those below 28 weeks, 1.4–1.9 mg/L for 28–33 weeks, 1.2–1.9 mg/L for 34–36 weeks of gestation, and 0.97–1.8 mg/L for term neonates.^[11] Cystatin-C levels gradually decline in the neonatal period and infancy, to around 0.80 mg/L at 1^st year of life, and values remain stable into adulthood.^[15]

There is a lack of Indian studies and a need to establish cystatin-C cut-off values in different groups of neonates to measure neonatal renal function. The primary objective of this study was to compare creatinine and cystatin-C levels in well-appearing term small-for-gestational-age (SGA) and late preterm appropriate-for-gestational-age (AGA) neonates, with term appropriate-for-age neonates as controls in the ^1st week of life.

MATERIAL AND METHODS

RESULTS

During the study period from June to October 2023, 1,285 neonates were born at 35 completed weeks to 41 6/7 weeks of gestation. We screened 76 late preterm AGA neonates for Group 1 and 145 term SGA neonates for Group 2, and 55 were enrolled in both the groups. For Control Group 3, 124 term AGA neonates delivered after group enrolment were assessed and 110 neonates who met the inclusion criteria were enrolled as controls. A summary of the enrolment procedure with reasons for exclusion is shown in Figure 1.

Close

Figure 1:

Enrollment details. AGA: Appropriate for gestational age, SGA: Small for gestational age,

Export to PPT

Baseline maternal and neonatal data, including anthropometry, are represented in Table 1. A majority of neonates (n = 210, 95.5%) were on exclusive breastfeeding, while a few neonates (n = 10, 4.5%) required additional formula feeding. Some neonates required NICU admission within first 72 hours of life for the management of hypoglycemia (n = 7) or neonatal hyperbilirubinemia (n = 2), and this rate of early NICU admissions was significantly higher in Group 1 (n = 5, 9%) and Group 2 (n = 3, 5%) compared to the control group, Group 3 (n = 1, 0.9%; P = 0.037).

The mean ± SD maternal serum creatinine levels and cystatin-C levels in the study population (n = 220) and corresponding cord values were 0.63 ± 0.18 mg/dL and 0.91 ± 0.31 mg/L; 0.63 ± 0.2 mg/dL and 1.582 ± 0.42 mg/L, respectively, and no statistically significant differences were observed among the three study groups [Table 2].

A significant positive correlation seen between maternal and cord creatinine levels (Pearson’s r = 0.832, P < 0.001) [Figure 2]. Spearman’s rho correlation coefficient for maternal and cord blood creatinine among late preterm AGA, term AGA, and term SGA were 0.826, P < 0.001; 0.861, P < 0.001; and 0.75, P < 0.001, respectively. However, no significant correlation was found between maternal and cord serum cystatin-C (Pearson’s r = 0.019, P = 0.78).

Close

Figure 2:

(a) Correlation between maternal cystatin C and cord cystatin C levels (b) Correlation between maternal creatinine and cord creatinine levels. AGA: Appropriate for gestational age, SGA: Small for gestational age

Export to PPT

The mean ± SD serum creatinine and cystatin-C levels among the neonates (n = 220) at day 3 of life were 0.65 ± 0.17 mg/dL and 1.46 ± 0.41 mg/L, respectively. A significant positive correlation was seen between day 3 and cord blood cystatin-C levels (Pearson’s r = 0.982, P < 0.001) [Figure 3]. However, no correlation was observed between day 3 creatinine and cord blood creatinine levels (r = 0.069, P = 0.308). Multivariate linear regression analysis to find predictors of day 3 serum cystatin-C levels after adjusting for gestation, birth-weight, size-at-birth classification (AGA vs. SGA), maternal cystatin-C levels, and day 3 creatinine levels revealed that cord blood cystatin-C level was a significant independent predictor of day 3 cystatin-C (R² = 0.963, P < 0.001). Regression analysis performed to identify predictors of day 3 serum creatinine after adjusting for gestational age at birth, birth-weight, size-at-birth classification (AGA vs. SGA), maternal serum creatinine, cord blood creatinine, and day 3 cystatin-C levels did not reveal cord blood creatinine as a significant predictor of day 3 serum creatinine (R² = 0.023, P = 0.18). Regression analysis coefficient-beta and significance for the determinants of day 3 creatinine and day 3 cystatin-C are shown in Supplementary Table 1.

Close

Figure 3:

(a) Correlation between day-3 cystatin C and cord cystatin C levels (b) Correlation between day-3 creatinine and cord creatinine levels. AGA: Appropriate for gestational age, SGA: Small for gestational age.

Export to PPT

Cord and day 3 cystatin-C values were significantly higher in Group 1 (late preterm AGA) compared to Group 2 and Group 3 (P = 0.011 and P = 0.021, respectively) [Table 2]. Further, Group 2 (term SGA neonates) had a significantly lower day 3 creatinine level compared to Group 3 (term AGA controls) (0.61 ± 0.19 vs. 0.67 ± 0.14 mg/dL; P = 0.04). Cystatin-C percentile charts among different groups are represented in Table 3.

DISCUSSION

In this prospective cohort study, we analyzed serum cystatin-C and creatinine in assessing renal function among term and late preterm neonates across different intrauterine growth groups. We found higher cord cystatin-C levels in late-preterm AGA, as compared to term AGA neonates. Cord cystatin-C showed a strong positive correlation with day 3 cystatin-C levels. Regression analysis confirmed that cord cystatin-C predicts cystatin-C on day 3 (R² = 0.963, P < 0.001). Cord creatinine had a strong positive correlation with maternal levels possibly due to placental transfer. However, day 3 creatinine was not influenced by cord creatinine values in adjusted models (R² = 0.023, P = 0.18). These findings suggest that cystatin-C is a stable marker to assess neonatal renal function in the 1^st week of life, unaffected by maternal values.

High creatinine levels in the early neonatal period reflect maternal creatinine and immature renal handling of creatinine. Hence, the definitions of neonatal AKI based on serum creatinine levels may be insensitive. Gupta et al., suggested that failure of serum creatinine level to decrease in the 1^st week of life would be a sensitive marker of renal dysfunction, with improved prediction for perinatal asphyxia.^[3,18,19] Neonatal estimated-GFR using the modified Schwartz formula performed poorly, and complex formulas are needed.^[7,8,19] The rise of serum creatinine by 0.3 mg/dL could be normal in preterm neonates during the initial days of life as identified by Allegaert et al., in their observational study.^[3,4,20] This highlights the need for evaluating novel biomarkers in neonatal renal function, which is independent of maternal levels.

The median (IQR) cord-to-maternal creatinine ratio in our study, 0.97 (0.82–1.12), is comparable to the ratio reported by Go et al.^[21] Maternal and cord creatinine had a strong Spearman correlation in late-preterm AGA and term AGA neonates, ρ = 0.826 and 0.861, respectively; whereas correlation was lower in term SGA neonates, ρ = 0.75. However, Go et al.^[21], reported that cord and maternal creatinine correlation was stronger in term neonates than preterm, and this difference could be attributed to differences in the study population. Sick neonates in NICU were enrolled by Go et al., whereas we enrolled healthy-appearing neonates.^[21] In contrast, no significant correlation was observed between cord and maternal cystatin levels in our study, which is consistent with previous studies involving term neonates in uncomplicated pregnancies reporting no correlation.^[22,23]

Automated immunoassays by PENIA and PETIA are highly sensitive cystatin-C estimation methods with less interference from glucose, bilirubin, and protein.^[9,24] Serum or plasma cystatin-C measured by PETIA and PENIA has excellent correlation and is widely used in neonates (r = 0.97). The PETIA method may slightly overestimate cystatin-C in neonates.^[11,24,25] Cystatin-C levels are high at birth, rapidly fall in the 1^st month, then gradually decrease till 1–2 years subsequently without much changes in both term and preterm neonates.^{[10,11,15,17,26]} In our study, serum cystatin-C was measured by the automated immunoassay PETIA method, which was compatible with our laboratory Beckman-Coulter AU480 autoanalyzer.

Nephrogenesis with an increase in nephron mass continues till 36 weeks, hence preterm neonates have higher cystatin-C than term infants. In a systematic review of 15 studies, Renganathan et al., reported cystatin-C levels in cord and day 3 in late-preterm neonates ranged from 1.22 to 1.96 mg/L and 1.24 to 1.85 mg/L, respectively.^[11] These values are similar to our study, a prospective study by Elmas et al., and Harmoinen et al.^[27,17] However, lower cystatin-C levels were reported by Ibrahim et al., 1.1 ± 0.19 mg/L in late-preterm neonates, which could be due to the enzyme-linked immunosorbent method utilized by them for cystatin-C measurement.^[28]

We found no significant difference between cystatin-C levels in term SGA and term AGA neonates, both at birth and on day 3 [Table 2]. Our findings are similar to Treiber et al., who also reported no significant difference in cystatin-C between term AGA and term SGA neonates, and they noticed that a 10% decrease in three-dimensional kidney volume was associated with 9.3% increment in cord cystatin level.^[29] Whereas in their longitudinal study involving 166 neonates, Iyengar et al., found low birth weight and SGA neonates had lower cystatinC-based GFR when compared to term AGA neonates, with comparable values observed by 2 years of age.^[30]

Inulin-based GFR measurement is the gold standard, but bedside application is limited by feasibility, cumbersome procedure, and cost. Yang et al., in their meta-analysis including 12 studies and 1572 samples, reported cystatin-C had a pooled sensitivity and specificity of 84% (74–91) and 81% (75–86) for detecting neonatal AKI. However, heterogenicity was seen and confirmatory tests were inconsistent.^[10] Smeets et al., reported that neither cystatin-based nor creatinine-based formulae correctly estimated iohexol-based GFR measurement and poor correlation seen in sick neonates and children.^[19] Cystatin-C models did not outperform creatinine models for prediction of Inulin GFR in preterm neonates.^[31] In a multicentric retrospective cohort study from China among 52,333 sick neonates cystatin-C levels >2.2 mg/dL or a 25% raise in levels for defining AKI, Xu et al., estimated cystatin-based mortality detected 6.5 times higher mortality than creatinine-based models.^[13] The 95^th centile cystatin-C in our study on day 3, 2.198 mg/L is similar to this cut-off. Thus, cystatin-C could be a more sensitive indicator for renal function and AKI in neonates.

The merits of study are prospective design, enrolment of healthy-appearing neonates, centile charts, and automated immunoassay by the PETIA method for cystatin-C measurements. The cystatin values in our study could guide further studies to estimate AKI in sick neonates. However, accurate GFR estimation by inulin studies could not be done for accurate evaluation of cystatin-C and the sample size for percentile chart is small. The cost of cystatin-C measurement is higher than that of creatinine measurement.

CONCLUSION

In conclusion, we found that serum cystatin-C levels in neonates during 1^st week of life were independent of maternal values as compared to creatinine. Cord and day 3 cystatin-C levels had a significant positive correlation. Cystatin-C levels were higher in late-preterm AGA neonates when compared to term AGA neonates. Cystatin-C levels were comparable between term AGA and term SGA neonates. The cystatin-C percentile chart of our study would guide subsequent application of cystatin-C in sick neonates for assessment of AKI.