Introduction

Infants born low birth weight at term gestation are at increased risk for “adult onset” adiposity, hypertension, and renal dysfunction.1,2 This may be due to intrauterine stresses associated with nutrient deficiency that modify developmental programming and contribute to the occurrence of chronic disease states in adulthood.3,4,5 Preterm birth also has been associated with the occurrence of obesity, hypertension, and renal disease in young adults6,7,8,9,10,11,12 and the incidence of each increases with decreasing gestational age (GA) at birth.13,14,15,16,17 Duncan et al.18 reported that infants born very preterm (≤32 weeks GA) and very low birth weight (≤1500 g; PT-VLBW) had elevated weight-for-length ratios and systolic blood pressure (SBP) as early as 6 months to 1 year corrected age. Frankfurt et al.19 observed that the disproportionate growth and elevated SBP in this population were associated with evidence of renal dysfunction; however, the relationships between obesity, elevated SBP, and renal dysfunction were unclear. Notably, excessive weight gain and the velocity of weight gain in preterm and term infants in the first 3 years of life is also associated with later elevations in SBP.13,20,21,22,23 While it is clear that preterm birth is associated with subsequent obesity and hypertension in young adults, the time course in childhood remains unclear and requires further study.

Low birth weight in term infants and intrauterine growth restriction in term fetal sheep are also associated with decreased nephrogenesis and subsequent hyperfiltration, glomerulosclerosis, and chronic renal disease with hypertension.23,24,25,26 Similar observations have been reported in adolescents and young adults born preterm.27,28,29,30 Birth <35 weeks GA is associated with incomplete nephrogenesis, and it has been suggested that this also results in subsequent hyperfiltration, glomerulosclerosis, and chronic renal disease with hypertension.31,32 This is supported by autopsy data that revealed morphologically abnormal glomeruli in preterm neonates.33 Frankfurt et al.19 reported loss of renal autoregulation and increased perfusion in 1-year-old PT-VLBW infants and low estimated glomerular filtration rates (eGFR) in a cohort of 3 year olds born PT-VLBW. Thus, birth <32 weeks GA is associated with incomplete nephrogenesis, increased risk for hyperfiltration in infancy and renal dysfunction, and elevations in SBP later in childhood. However, there are no serial measurements examining renal function and blood pressure from infancy to preadolescence in PT-VLBW.

In the present study, PT-VLBW infants originally seen at 1, 2, or 3 years of age and consented for follow-up were reenrolled 10 years later in order to determine (1) whether the obesity, high SBP, and renal dysfunction observed at 1–3 years of age persists at preadolescence18,19 and (2) whether the occurrence of obesity, high SBP, and renal dysfunction in the reenrolled PT-VLBW Cases differs from Term-matched Controls identified in the National Health and Nutrition Examination Survey (NHANES) dataset for 2015–2016.34

Methods

Study population and recruitment

PT-VLBW infants initially seen and recruited between May and September 2007 in a prospective cohort study and consented to subsequent follow-up were contacted in 2017 in order to conduct a 10-year follow-up.18,19 These PT-VLBW Cases were born at Parkland Memorial Hospital after January 1, 2004, had a GA ≤ 32 weeks, birth weight ≤1500 g, and received primary care in the Low Birth Weight Clinic at Children’s Medical Center, Dallas, TX at 1, 2, or 3 years corrected age. Race/ethnicity was verbally obtained at the initial recruitment and is self-reported. We excluded infants with congenital anomalies or discharged before 35 weeks postmenstrual age. We identified all of the PT-VLBW Cases (n = 120) in the electronic health records of Children’s Health and/or Parkland Health and Hospital System and successfully contacted 86 (72%) by telephone; 68 agreed to a clinic visit and 62 (52%) were seen, consented, and reenrolled at 10–13 years of age. A health survey was completed by V.B. for all enrolled patients to examine medications and health care.

Blood pressure measurements

We obtained standardized blood pressure measurements from consented PT-VLBW Cases during the clinic visit.35 After the patient rested at least 5 min while seated, upright and feet flat on the floor, two manual blood pressure measurements, 5 min apart, were obtained by one investigator (V.B.) from the right upper arm using an appropriately sized cuff and a 9002 digital ADC® e-sphygTM 2 sphygmomanometer (American Diagnostic Corp., Hauppauge, NY). The cuff had a bladder length 80–100% of arm circumference and width at least 40%. If the two sequential blood pressure measurements were within the range of error for the sphygmomanometer, ±3 mmHg, and <90th centile for age using the 2004 guidelines available at that time for the diagnosis, evaluation, and treatment of high blood pressure in children and adolescents,36 we used the average value for subsequent analyses. If either SBP was ≥90th centile, we obtained a third measurement before concluding the visit and used the average of the two lowest measurements for analyses. Parents of children with multiple high blood pressure measurements received a letter from V.B. to give their primary care physician for follow-up assessment. The updated 2017 AAP Clinical Practice Guidelines for blood pressure diagnosis and management were subsequently used to reclassify and analyze SBP in PT-VLBW Cases at 1–3 years, 10–13 years of age, and in Term-matched Controls.37 Measurements of SBP were categorized as normal (<90th centile), elevated (90–95th centile), and hypertensive (>95th centile). None of the study patients was receiving an antihypertensive medication.

Anthropometric measurements

One investigator (V.B.) measured weight (g) and height (cm) twice, weight on a digital scale (accurate to the nearest 100 g) and length by stadiometer (accurate to the nearest millimeter). We used the average of the two measurements for each analysis. Body mass index (BMI) was calculated as weight (kg)/height (m)2 and classified using the sex-specific reference ranges from the Centers for Disease Control and Prevention.38 We categorized values as normal (<85th centile), overweight (85–95th centile), or obese (>95th centile).

Blood samples

We collected 3.0 mL of blood from all consented Cases, n = 62. Samples were allowed to clot, centrifuged at 10,000 rpm for 10 min, and serum divided into three aliquots that were stored at −80 °C until analysis. Serum creatinine (Cr) was measured using the modified Jaffe method and cystatin C (cysC) using the particle enhanced turbidimetric immunoassay. All samples (n = 62) were measured in a single assay at the Children’s Medical Center Dallas laboratory. We calculated eGFR using currently recommended equations by the National Kidney Foundation for children aged 10–13 years: the bedside Schwartz for creatinine (eGFRCr = 0.413 × L (cm)/Cr (mg/dL)) and the equation for cystatin C (eGFRcysC = 70.69 × [cystatin C (mg/L)]−0.931.39

Control population

Term-matched Controls were selected from the 2015–2016 NHANES dataset and matched two Controls for each PT-VLBW Case using age at reenrollment or date of birth if available, sex, and race/ethnicity.34 Since GA at birth was unavailable, we used birth weight >2000 g as a proxy for term birth. This cutoff allowed us to consider a larger control population that might have had a serum Cr measured and still exclude fetal growth restriction. Serum Cr measurements were only available in a subset of 12–13-year-old Controls; thus, we matched them with a corresponding age-matched subset of Cases in order to compare eGFRCr; serum Cr was not available in three of the 12–13-year-old matched Controls. A sample size of 124 Controls matched to 62 Cases achieved a 90% power to detect a 0.25 between-group difference in proportions with SBP ≥ 90th centile, assuming the proportion in the Cases with elevated SBP was 0.25 under the null hypothesis and 0.50 under the alternative hypothesis. We targeted a significance level of 0.05.

Height (cm) and weight (Kg) were measured using the NHANES mobile examination center protocol.34 In this cohort, blood pressure was measured using the American Heart Association guidelines: after resting 5 min, three measurements were manually obtained with a Baumanometer® in the right upper arm with an appropriately sized cuff.37 A fourth measurement was obtained if any of the first three were interrupted or incomplete. We used the average of the three SBP measurements for analyses.

Statistical analyses

Demographic characteristics, e.g., ethnicity, birth weight, and GA, were matched and compared in PT-VLBW Cases at 1, 2, and 3 years and at 10, 11, 12, and 13 years. We also compared demographics between PT-VLBW Cases and Term-matched Controls at 10–13 years. We analyzed categorical variables (e.g., race, obesity) by χ2 analysis and continuous variables (e.g., birth weight, GA, eGFR) by t test or Mann–Whitney U-test dependent on data distribution. The alpha level was 0.05 for all statistical tests. We used linear regression analysis to determine correlations between continuous variables, e.g., GA and eGFR at 12–13 years. Analysis of variance (ANOVA) for multiple groups followed by the correct post hoc test and pair-wise analysis, e.g., Bonferroni correction, were used to compare multiple groups. Data are means ± SD or medians with 25th and 75th centiles unless noted otherwise. We also performed exploratory analyses using multivariate logistic regression to determine which clinical variables are associated with development of an elevated SBP at 10–13 years in the Cases and Controls compared to a normotensive group in the Cases and Controls. Data are the adjusted odds ratio (aOR) of each variable in the model with the 95th confidence limits.

The study was approved by the Institutional Review Boards of the University of Texas Southwestern Medical Center, Children’s Medical Center Dallas, and Parkland Health and Hospital Systems. Informed consent was obtained from parents or guardians of study patients at the clinic visit by V.B. and assent was obtained from all study patients >10 years. PT-VLBW Cases were also consented for later follow-up.

Results

Description of PT-VLBW cases

Sixty two (52%) of 120 patients originally studied at 1, 2, or 3 years were reenrolled at 10–13 years of age. They were evenly distributed between the three original cohorts (18/40, 21/40, and 23/40, respectively) and only differed from the overall original cohort due to fewer Black non-Hispanics (P < 0.03, Table 1). This difference also existed between reenrolled and nonenrolled Cases as well as 1 week lower GA at birth (P ≤ 0.03, Table 1). There is no significant difference between groups in exposure to maternal antenatal steroids (ANS) or postnatal therapeutic indomethacin. We also assessed the demographics of the reenrolled Cases as separate cohorts originally seen at 1, 2, and 3 year of age. The only significant differences are a lower GA (27 vs. 28.6 and 28.4 weeks) and birth weight (970 g vs. 1165 and 1106 g) in Cases originally in cohort 3 vs. cohorts 1 and 2, respectively (P < 0.05, ANOVA).

Table 1 Neonatal characteristics and distribution of cystatin C-derived estimated glomerular filtration rate (eGFRcysC) at 1–3 years in all PT-VLBW Cases originally enrolled at 1–3 years (n = 120) and a comparison of those reenrolled (n = 62) and not reenrolled (n = 58) at 10–13 years.
Full size table

When the characteristics of the Cases reenrolled (n = 62) are examined as a single cohort at 10–13 years (Table 2), the average age at reenrollment is 11 years and 86% are Hispanic, reflecting the delivery population at Parkland Hospital. Notably, the mean weight Z-score at 10–13 years exceeds 0 while the height Z-score is <0; thus, 45% of Cases at 10–13 years exhibit disproportionate weight-for-length growth and are classified as overweight or obese with a BMI Z-score ≥85th centile.38 This is similar within all three cohorts (P > 0.1, ANOVA; data not shown). There is no difference in the velocity of weight gain from birth to 10–13 years between Cases from different cohorts (P > 0.1; data not shown); thus, these data are expressed as the mean value for all reenrolled Cases (Table 2). Although the mean SBP at 10–13 years is 115 ± 11 mmHg, classification of blood pressure measurements using the 2017 AAP Guidelines37 reveals that 48% (30/62) have a SBP ≥ 90th centile and 77% of these have a SBP > 95th centile and are considered hypertensive (Table 2). Four had Stage 2 hypertension. When we compare SBP classification at 1–3 years and 10–13 years using 2017 AAP Guidelines (Fig. 1), the number of Cases <90th centile decreases 36% between 1–3 and 10–13 years while the number ≥90th centile increases 2.5-fold. Notably, the Cases >95th centile increases from 6 to 23 (P = 0.07), while 17/50 Cases originally <90th centile have a SBP > 95th centile at 10–13 years (P = 0.9, χ2 analysis).

Table 2 Patient characteristics in PT-VLBW Cases reenrolled at 10–13 years (n = 62) and Term-matched Controls (n = 124) selected from 2015–2016 NHANES database.34
Full size table
Fig. 1: Comparison of the systolic blood pressure classification in PT-VLBW Cases (n = 62) originally seen at 1–3 years and subsequently at 10–13 years of age.
figure 1

Values within bars are number of patients. Blood pressure is classified using the 2017 AAP Guidelines:37 Normal is <90th centile (white), elevated is 90th–95th centile (light gray), and hypertensive >95th centile (dark gray). P = 0.07 by Fisher exact test comparing patients >95th centile at 1–3 years and 10–13 years.

Full size image

We measured serum Cr and cysC in each of the PT-VLBW Cases at 10–13 years and assessed renal function in the total cohort. The overall mean serum Cr concentration is 0.65 ± 0.16 mg/dL (n = 62; Table 2) and the overall mean eGFRCr is considered normal. However, classification of eGFRCr revealed that 34% are <90 mL min−1 [1.73 m2]−1 and considered low (Table 2). Since differences may exist between eGFRCr and eGFRcysC as well as between females and males, we examined both (Table 3). The analysis of paired values for eGFRCr and eGFRcysC show no significant difference in the overall mean eGFR, 96.2 ± 13 vs. 99.7 ± 21 mg/dL, respectively (P = 0.1, paired t test). There also are no differences in the mean eGFR or classification of eGFR between females and males using either estimate of GFR (P ≥ 0.5, Table 3). However, a third of all Cases at 10–13 years have an eGFRCr or eGFRcysC < 90 mL min−1 [1.73 m2]−1. Birth weight and GA are linearly related to both eGFRCr and eGFRcysC (Fig. 2), but the relationship is significantly greater for birth weight than GA (P < 0.01) and best defined by eGFRcysC (R = 0.30 and P = 0.02 vs. R = 0.49 and P < 0.01, respectively). The correlation between eGFR and GA is also best defined using eGFRcysC (eGFRcysC: R = 0.49 and P < 0.01 vs. eGFRCr: R = 0.11 and P = 0.4, respectively). When we examine the relationship between prenatal ANS exposure and the subsequent occurrence of an abnormal eGFR,40 there is no correlation with either eGFRCr or eGFRcysC at 10–13 years (P = 0.9). However, birth weight in Cases exposed to ANS remains highly correlated to subsequent eGFR at 10–13 years, P = 0.002.

Table 3 Comparison of estimated glomerular filtration rate (eGFR) derived from measurements of serum cystatin C (cysC) and creatinine (Cr) in all female and male PT-VLBW patients reenrolled at 10–13 years of age (n = 62).
Full size table
Fig. 2: Birth Weight Predicts eGFR in Preadolescence.
figure 2

Relationships between cystatin C-derived estimated glomerular filtration rate (eGFRcysC), birth weight (■–■; R = 0.49, n = 62, P < 0.01) and gestational age (♦–♦; R = 0.29, n = 62, P = 0.02) in PT-VLBW Cases at 10–13 years (n = 62).

Full size image

Comparison of PT-VLBW Cases at 10–13 years and Term-matched Control

To determine whether there were differences in blood pressure and renal function between preterm and term births, we matched 10–13-year-old PT-VLBW Cases with Controls (62:124) selected from the 2015–2016 NHANES dataset.34 Clinical characteristics do not differ between groups except for birth weight (Table 2). Notably, the proportion with a BMI Z-score classified overweight or obese is similar in Cases and Controls, 45% and 48%, respectively, as is the mean velocity of weight gain from birth to preadolescence. There is, however, a higher overall SBP in PT-VLBW Cases (P < 0.01), who are sevenfold more likely to be classified as hypertensive with a SBP > 95th centile (P < 0.01). We also examined the relationship between SBP at 10–13 years with the velocity of weight gain from birth to 10–13 years in Cases and Controls (Fig. 3). While both demonstrate significant positive linear relationships (R = 0.74, P < 0.001 vs. R = 0.41, P < 0.001, respectively), the slopes and intercepts differ (P < 0.001) and Cases have a greater rise in SBP with increasing velocity of growth.

Fig. 3: Velocity of Weight Gain Contributes to Blood Pressure in Preadolescence.
figure 3

Relationships between systolic blood pressure (mmHg) at 10–13 years and the velocity of weight gain (g month−1) from birth to 10–13 years in PT-VLBW Cases (○; R = 0.74, n = 62, P < 0.001) and Term-matched Controls (●; R = 0.41, n = 124, P < 0.001). The slopes and intercepts are significantly different, P < 0.001.

Full size image

We19 previously reported abnormalities of renal function at 1–3 years in PT-VLBW Cases; thus, we compared renal function in Cases (n = 62) with matched Controls at 12–13 years (n = 53) using serum Cr measurements to calculate eGFRCr (Table 2). Serum Cr is significantly higher in Cases vs. Controls and overall eGFRCr lower (P < 0.01). Moreover, Cases are fivefold more likely to have an eGFRCr < 90 mL min−1 [1.73 m2]−1, whereas Controls are ~3.5-fold more likely to have an eGFRCr > 140 mL min−1 [1.73 m2]−1 (P = 0.03, χ2 test). Since there may be sex-specific differences in renal function, we compared sex-related characteristics and eGFRCr in Cases and matched Controls (Table 4). There are no sex-related differences in race/ethnicity, weight, height, or distribution of BMI Z-score across groups. However, serum Cr is ~24% higher in male Cases vs. all Controls and male eGFRCr is four times more likely to be <90 mL min−1 [1.73 m2]−1 than either male or female Term-matched Controls (P < 0.01). No PT-VLBW Cases have an eGFRCr > 140 mL min−1 [1.73 m2]−1.

Table 4 Comparison of sex-stratified patient characteristics and serum creatinine-derived estimated glomerular filtration rate (eGFRCr) in PT-VLBW Cases at 12–13 years (n = 28) and Term-matched Controls (2:1; n = 53).
Full size table

Prediction of SBP and eGFR at 10–13 years

We performed exploratory analyses using multivariate logistic regression to determine which clinical variables might be associated with development of an elevated SBP at 10–13 years. The clinical variables included in the SBP classification model are race/ethnicity, birth weight, overweight, obesity, eGFR classification (<90 and ≥140 mL min−1 [1.73 m2]−1), and Case (PT-VLBW) vs. Control (Term). Race/ethnicity, birth weight, eGFR classification, and Case vs. Control are not statistically significant in the SBP classification model. However, overweight and obesity are significant variables with an aOR 4.1 and 5.9, respectively (P ≤ 0.04). Notably, the aOR increases further when we use a SBP > 95th centile, 9.8 (P = 0.003) and 11.2 (P = 0.001), respectively. These observations are exploratory and require further investigation.

We performed a similar multivariate logistic regression to evaluate which clinical variables are associated with eGFR classification at 10–13 years with Cases and Controls in the model. There is no significant variable detected when race/ethnicity, birth weight, overweight, obesity, SBP classification, and Cases (PT-VLBW) vs. Controls are in the model. When we use neonatal characteristics in a model restricted to Cases (n = 62) to assess the roles of birth weight, GA, and ANS, birth weight is significant, P < 0.015, and a 1.0 g rise in birth weight is associated with a 0.56 rise in eGFRcysC. Importantly, 57% of PT-VLBW Cases with a low eGFRcysC at 1–3 years have a persistent low eGFRcysC at 10–13 years, P = 0.03 (Supplementary Table 1).

Discussion

Infants born PT-VLBW are at increased risk for elevated SBP in childhood, adolescence, and early adulthood.6,9,11,41,42,43,44,45,46 These observations are derived from population-based cohort studies that generally do not include parallel serial measurements of blood pressure and/or renal function beginning in infancy and continuing to preadolescence, which might improve our understanding of the presentation, progression, and pathogenesis of disease. To address this, we restudied children at 10–13 years who were born PT-VLBW and originally seen at 1, 2, or 3 years of age.18,19,47 At 10–13 years, nearly half are overweight or obese and have an elevated SBP, three quarters of whom are hypertensive. Moreover, a third have a reduced eGFR that persisted from infancy, while nearly 20% not only have an elevated SBP but also a reduced eGFR. When compared with Term-matched Controls, their BMI Z-score and prevalence of overweight/obesity do not differ; however, the PT-VLBW Cases are more likely to have an elevated SBP with hypertension and low eGFR. Thus, the occurrence of an elevated SBP and low renal function early in infancy in PT-VLBW infants persists at 10–13 years of age, demonstrating the need to not only study this population more intensely across childhood but also provide improved screening for blood pressure, renal function, and obesity starting early in infancy.

We reenrolled and consented 52% of the original PT-VLBW 10 years after their initial enrollment. Due to the relatively small study population, we sought to insure there is no bias. There are no substantive differences within the population studied other than minor differences in race/ethnicity and GA. Moreover, there are no significant differences in other neonatal variables between reenrolled and not reenrolled Cases. Furthermore, we reenrolled approximately 50% of each of the three cohorts originally seen at 1–3 years and observed no differences in neonatal characteristics between these cohorts at 10–13 years. Thus, we believe the reenrolled Cases are representative of the initial 120 infants studied.

A primary objective of our study was to determine whether there was persistence or progression of an abnormal SBP during the first 10–13 years of life in PT-VLBW neonates. Several investigators using variable sized population cohorts have reported an increase prevalence of an elevated SBP during childhood, including adolescence, but only a few included serial measurements within the same population.22,23,44 In our earlier study, SBP and SBP Z-scores at 1–3 years of age exceeded that expected for age and ~20% were ≥90th centile.18 The number with a normal SBP at 1–3 years decreased ~50% at preadolescence while those subsequently classified hypertensive rose fourfold. This redistribution, however, is not significant, likely reflecting the relatively small number of patients studied and/or the fact that only 52% of the original cohort was reenrolled. Furthermore, we excluded SBP measurements at 1–3 years in three reenrolled Cases since they were obtained while the infant was in an unstable state, which falsely elevates the SBP;48 their SBP is normal at 10–13 years. Thus, the prevalence of an elevated SBP increases between infancy and preadolescence and is associated with an increase in the occurrence of hypertension. This should be examined further in larger population-based studies that include serial measurements.

Infants born PT-VLBW exhibit disproportionate weight-for-length growth at neonatal intensive care unit (NICU) discharge and at 6 months and 1–2 years corrected age, which is associated with elevations in SBP at follow-up.18,19,49 At 3 years, the indices of overweight/obesity and an elevated SBP Z-scores were no longer evident, suggesting a relationship between being overweight and having an elevated SBP in infancy.18 At 10–13 years, nearly half of our Cases are overweight/obese; surprisingly, this does not differ from Term-matched Controls. In addition, the overall mean velocity of weight gain from birth to preadolescence does not differ. Nonetheless, the Cases are more likely to have an elevated SBP and be hypertensive. Vohr et al.23 also observed no difference in overall BMI between preterm and term born adolescents at 16 years, yet SBP was higher in the preterm. Of note, in comparison with Vohr’s adolescent population, the incidence of a BMI ≥ 85th centile in our preadolescent population was 40% greater and the incidence of a SBP ≥ 90th centile was three times greater, i.e., 48 vs. 17.6%. This may reflect differences in race/ethnicity, i.e., Vohr’s population is described as being 75% White non-Hispanic and 8% Hispanic as compared to 0% White non-Hispanic and 86% White Hispanic in the present study. This suggests involvement of unique ethnic and/or environmental influences on weight gain, e.g., diet, and the risk of an elevated SBP after preterm birth. This novel observation illustrates the problem of making broad conclusions about the outcomes of preterm births without detailed consideration of racial/ethnic makeup. Nonetheless, the present study confirms that preterm birth plus being overweight or obese (BMI ≥ 85th centile) at preadolescence contributes to an increased likelihood of having an elevated SBP and hypertension and this is more likely in Hispanic children born PT-VLBW.

Comparing serial measurements of blood pressure within the PT-VLBW Cases revealed a nearly fourfold rise in values ≥95th centile between 1–3 years and 10–13 years. When we examined potential factors associated with this change, SBP significantly correlated with the velocity of weight gain from birth to adolescence, suggesting that catch-up growth might contribute. Multivariate regression analysis revealed that only overweight and/or obesity contribute to a SBP > 90th centile. When we assessed SBP ≥ 95th centile at preadolescence, the aOR for a BMI > 85th centile doubled. Notably, neither birth weight nor GA contributed, consistent with a recent analysis of nine cohort studies assessing young adults born VLBW.11 In a retrospective cohort study of extremely preterm low-birth-weight infants followed 7–10 years, Mhanna et al.22 also observed an association between obesity and hypertension and reported that BMI and weight gain velocity from birth to 3 years (P = 0.008) predicted an elevated SBP. Edstedt-Bonamy et al.42 observed similar relationships in a Swedish population-based cohort of extremely preterm infants compared to term-matched controls at 6 years. Vohr et al.44 also examined weight gain velocity in a population of extremely preterm infants between 18 months and school age and noted that it predicted an elevated SBP at 6–7 years with a relative risk 1.36. In addition, in a retrospective review of 160 children born preterm and with at least three blood pressure measurements between 10 and 15 years, the odds of having a SBP ≥95th percentile was 1.6 times greater for every one standard deviation increase in BMI.50 The current study is consistent with these reports but differs due to ethnic/racial differences noted earlier. Moreover, it supports the conclusion that preterm children who are overweight and/or obese in early childhood are at increased risk for elevated blood pressure in preadolescence.

Renal dysfunction in the PT-VLBW population is related to the occurrence of chronic renal disease in adolescence and adults.27,28,29,30,31,32,33,46,51,52,53,54 However, few investigators have examined the age of onset and none obtained serial measurements of renal function between early infancy and later childhood.55 We19 previously reported that renal autoregulation appeared dysfunctional in PT-VLBW at 1 year corrected age, suggesting hyperfiltration. In addition, the 3-year-old cohort had a decreased eGFRcysC, which correlated with decreasing birth weight and GA. We now report that these PT-VLBW Cases have a greater prevalence of low eGFR at 10–13 years than Term-matched Controls and again observe that eGFRcysC decreases with decreasing birth weight and GA; however, the correlation with birth weight is greater than GA. This is consistent with other reports, but it is surprising since GA is generally considered a better marker for organ development than birth weight.32 This observation could reflect the contribution of adverse events occurring before and/or after birth, e.g., intrauterine nutritional deficiency, which occurred in 20% of our cohort, or adverse postnatal events, such as acute kidney injury.31,32 It is notable that none of the PT-VLBW Cases have an eGFRcysC > 140 mL min−1 [1.73 m2]−1 and evidence of hyperperfusion, whereas 17% of Term-matched Controls exhibit hyperperfusion at preadolescence, which is consistent with recent observations.56 It is possible that the development of obesity after birth and its consequences may contribute, suggesting that the low renal function may be due several variables. Future studies should consider modifying the disproportionate weight gain that frequently occurs in the NICU and after discharge as well as the rate of weight gain in childhood. Nonetheless, it appears that decreases in eGFR in PT-VLBW infants persist from infancy to preadolescence, but the pathophysiology and age it evolves into chronic renal disease remain unclear.

Limitations to this study include small sample size, predominant Hispanic population, and lack of concurrently collected data from a term-born control population. We addressed the latter by using the most recent NHANES dataset to match term-delivered controls 2:1. Furthermore, a subset of the Control population at 12 years had serum Cr measurements, which allowed us to compare eGFRCr at 12–13 years. We believe there was no selection bias during recruitment, which also could alter conclusions. A strength of this study is the ability to identify all of the earlier participants within our health care system and the recruitment of >50% 10 years later. Moreover, we collected blood samples from 100% and studied renal function between infancy and preadolescence. It is obvious that a more diverse PT-VLBW population should be studied in the future.

Conclusions

We have shown in this prospective cohort study of PT-VLBW infants originally seen at 1–3 years and again at 10–13 years that overweight/obesity, elevated SBP, and low renal function occur as early as 6 months corrected age and persist at 10–13 years of age. Moreover, while the occurrence of overweight/obesity is similar in Term-matched controls with a similar ethnicity at 10–13 years, the occurrence of an elevated SBP, hypertension, and low eGFR are significantly greater in the PT-VLBW population. Furthermore, nearly 20% of Cases have an elevated SBP plus low eGFR. How this will progress during adolescence and in early adulthood requires further follow-up. It also remains unclear whether measurements at 1–3 years and/or 10–13 years will predict the severity and/or occurrence of adult disease. Nonetheless, it is important that PT-VLBW neonates receive improved comprehensive long-term care that includes serial assessment of renal function, blood pressure, and obesity, which might permit early recognition of renal disease and hypertension and the institution of appropriate therapies.57,58