This Article Abstract Full Text (PDF) Submit a related Letter to the Editor Purchase Article View Shopping Cart Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Jett, P. L. Articles by Hanna, C. E. Search for Related Content PubMed PubMed Citation Articles by Jett, P. L. Articles by Hanna, C. E.

Variability of Plasma Cortisol Levels in Extremely Low Birth Weight Infants¹

Departments of Pediatrics (P.L.J., G.I.B., S.H.L., J.W.R., C.E.H.) and Medicine (M.H.S., P.A.M.), Divisions of Endocrinology and Neonatal Medicine, Oregon Health Sciences University, Portland, Oregon 97201

Address all correspondence and requests for reprints to: Patricia L. Jett, M.D., Neonatology Office, Rogue Valley Medical Center, 2825 East Barnett Road, Medford, Oregon 97504.

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
Cortisol is secreted by children and adults in a pulsatile pattern of 15–30 peaks and nadirs each day with a circadian rhythm. Newborns are known to lack the circadian pattern, leading to uncertainty about the appropriate time for blood sampling for assessment of adrenal function. Because extremely low birth weight (ELBW) infants may manifest signs of adrenal insufficiency, knowledge of the pattern of cortisol levels is necessary to guide the appropriate timing of blood sampling. To define the pattern of plasma cortisol levels in 14 ELBW infants, we obtained blood specimens every 20 min over a 6-h period at 4–6 days of life. Although cortisol levels in the 14 infants ranged from 2.0–54.5 µg/dL, each infant’s cortisol levels varied little from his or her own mean cortisol level. The SDs calculated from each infant’s mean cortisol level were small, ranging from 0.37–4.12 µg/dL. Cluster analysis was applied to the data; only 0.6 cortisol pulses/infant·6-h period were detected. Each infant’s plasma cortisol levels were plotted against time, and regression analysis was performed. The slopes of the resulting lines of regression ranged from -0.0284 to 0.0221. Our data indicate that ELBW infants show little variability in their plasma cortisol levels over time; therefore, a single random measurement provides an adequate reflection of the adrenal status of the ELBW infant.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
CORTISOL is secreted by children and adults in a pulsatile pattern of 15–30 short discrete bursts of varying magnitude each day (1). Cortisol secretion also follows a circadian pattern, with the highest peak seen approximately 1 h after arising in the morning, making peak cortisol secretion easily assessable. This circadian rhythm does not appear in the newborn infant, however, until approximately 3 months of age (2). The lack of defined periodicity of cortisol secretion in this population makes the assessment of adrenal function difficult with a single cortisol measurement.

A study by Metzger et al. (1) of the pulsatility of cortisol secretion in neonates detected a difference in the shape and amplitude of the secretory bursts when comparing premature to term newborns. The premature infants (three at 34 weeks and one each at 24 and 25 weeks gestational age) exhibited secretory bursts of longer duration and lower amplitude than those of the term infants, indicating that the pattern of pulsatility of cortisol secretion may be related to the infant’s degree of maturity.

Extremely low birth weight (ELBW) infants, weighing under 1000 g and less than 28 weeks gestational age, sometimes manifest a constellation of signs indicative of cortisol deficiency: hypotension that is unresponsive to inotropic support, oliguria with edema, hyponatremia, and hyperkalemia. Many ELBW infants, unlike adults and older children, have plasma cortisol levels that have no correlation to their degree of illness (3, 4). Infants with signs of adrenal insufficiency have been reported to have cortisol levels, measured in random blood samples, that are inappropriately low for their severity of illness (5, 6). When treated with stress doses of hydrocortisone, their signs of cortisol deficiency resolved within 2 days of beginning treatment (5, 7). Watterberg and Scott have found a correlation between a blunted response to ACTH stimulation in some very low birth weight infants and their subsequent development of bronchopulmonary dysplasia, a well known sequela of very premature birth (8). They have gone on to speculate that early corticosteroid replacement therapy might decrease the frequency of bronchopulmonary dysplasia. This speculation is in agreement with Korte et al., whose work suggests that morbidity outcomes in ELBW infants are correlated to their glucocorticoid production (9). In an effort to define the pattern of cortisol secretion in ELBW infants and to improve our ability to detect cortisol deficiency, we studied the pattern of cortisol concentrations measured serially over a 6-h period.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
Between November 1994 and December 1995, 38 infants who met the initial criteria of birth weight of less than 1000 g, estimated gestational age of less than 28 completed weeks, and a physical examination without anomalies were admitted to the Doernbecher Neonatal Care Center. Gestational age was determined by the admitting neonatologist using the Ballard score (10) and obstetrical dating. Fourteen of those infants who survived the first days of life met the additional criteria of having an indwelling umbilical catheter, having received no postnatal steroid treatment, and having informed parental consent given on the day of testing. These 14 infants were similar in clinical features to the entire population who were eligible for the study at the time of admission in terms of birth weight, gestational age, and severity of illness.

Infants were studied at 4–6 days of age to allow placental and maternal hormones to be metabolized and yet assure a high percentage of infants still having umbilical catheters. Blood samples were collected between 1800–2400 h to standardize the environmental stimuli of the infants at a time when noise and activity are usually at low levels in the intensive care unit. Clinical care interventions were kept to a minimum; some infants underwent endotracheal suctioning during the study. Blood was sampled every 20 min for 6 h, resulting in 19 specimens/infant. The specimens were heparinized and centrifuged, and the resulting plasma was frozen at -20 C until assayed.

During the first day of life, the infants were assessed for severity of illness using the Score for Neonatal Acute Physiology-Perinatal Extension (SNAP-PE) (11, 12). The SNAP-PE, when applied during the first 24 h of life, has been validated to be highly predictive of mortality; infants with a SNAP-PE equal to or greater than 30 have a percent mortality risk roughly equal to their SNAP-PE. The SNAP-PE was repeated on the day of the study as an informal assessment of the severity of illness at that time.

The study was approved by the institutional review board/committee on human research of the Oregon Health Sciences University.

Measurement of total plasma cortisol was performed by RIA (13) using tritiated cortisol and rabbit anticortisol antibody. The antibody was assayed for cross-reactivity with other glucocorticoid metabolites with good results; a 19% cross-reactivity with prednisolone was found, and all others were less than or equal to 10%, including a cross-reactivity with cortisone of 4%. The functional sensitivity of the assay was 1.0 µg/dL. The intra- and interassay coefficients of variation for this RIA were 10% and 5%, respectively. Standards and quality controls were run in duplicate. Each infant specimen was prepared in duplicate, and each preparation was then run in duplicate. All specimens from a single infant were run in the same assay.

For each individual infant, the mean cortisol level was calculated. The SD was then examined to assess the amount of variability around each infant’s mean. Linear regression was performed on each infant’s cortisol levels as well, with the 19 cortisol levels plotted against time.

Cortisol pulses were located by cluster analysis (14), using dose-dependent coefficients of variation calculated from sample replicates in each infant’s hormone series. Cluster parameters were two points for test nadirs and one point for test peaks. The t statistics were 3.0 for up- and down-strokes. The test nadir and peak parameters were chosen based on published studies of pulsatile cortisol secretion in humans (15, 16). The t statistics were chosen to constrain false positive peak detection rates to less than 10%, as determined by analysis of pooled serum samples run in the same cortisol assay. Estimated false negative rates for peak detection were less than 20%, given the observed sample variability and the known half-life of cortisol. Deconvolution analysis, which has been applied to cortisol profiles in premature neonates in past studies, was not attempted in the current study due to low cortisol variability and poor suitability of the data for the deconvolution process (Veldhuis, J. D., unpublished observations).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
The estimated gestational age of the 14 study infants was 25.6 ± 1.3 weeks (mean ± SD), with a range of 24–28 weeks (Table 1). The mean birth weight was 771 ± 175 g, ranging from 478-1009 g. All infants had birth weights appropriate for their gestational ages with one exception, an infant who had suffered intrauterine growth retardation (birth weight, 478 g at 27 weeks gestation) that was believed to be due to chronic maternal hypertension.

The mothers of 11 of the 14 study infants had received betamethasone before delivery; the first dose was given 2 h to 2 weeks before delivery. Nine of these 11 mothers also received TRH.

The infants demonstrated little variability in their plasma cortisol levels either between consecutive samples or over time. Although plasma cortisol levels in the 14 infants ranged from 2.0–54.5 µg/dL, each individual infant’s 19 cortisol levels varied little from his own mean cortisol level (Table 1). When the SD was calculated for each infant’s cortisol levels, the 14 SDs ranged from 0.37–4.12 µg/dL, with a mean of 1.8 ± 1.0 (±SD) µg/dL. No correlation could be found between the mean cortisol level or the amount of variation in cortisol levels and gestational age or birth weight. The infant with intrauterine growth retardation (no. 7) had similar results. Mean cortisol levels were plotted against the severity of illness; no significant correlation was detected (r = 0.46; P > 0.1).

Employing cluster analysis, 8 of the 14 infants had detectable pulses in their cortisol time series. For the overall patient group, the frequency of pulses was 0.6 pulse/infant·6-h period. Pulses were detected in infants regardless of the mean cortisol value, SNAP-PE, birth weight, or gestational age.

When each infant’s 19 cortisol levels were plotted against time, and regression analysis was performed, the slopes of the individual lines of regression ranged from -0.0284 to 0.0221 with a mean of -0.0057 ± 0.0118 (±SD). Although the most positive slope was seen in the infant with the highest SNAP-PE, no other correlation was noted between severity of illness and slope. There was also no correlation between slope and gestational age or birth weight.

The figure shows the graphs of three representative study infants with cortisol levels, lines of regression, and cluster analysis.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

 
Although the pulsatility of cortisol secretion in term newborns and late gestation premature infants has been documented by Metzger (1), no studies, to our knowledge, have addressed the temporal pattern of plasma cortisol levels in ELBW infants. Yet, these are the very infants who are often found to have low levels of cortisol and most frequently exhibit clinical signs of cortisol deficiency (5, 6). Recent work by Scott and Watterberg suggests that infants with lower cortisol levels and more severe disease have a higher risk of mortality and chronic morbidity (17).

We have demonstrated that ELBW infants (<1000 g and <28 weeks gestational age) show little pulsatility in their plasma cortisol levels over time.

The infrequency of the pulses detected by cluster analysis suggests that a single cortisol level would be representative of the cortisol levels over the time period studied. Given our assay performance characteristics and our pulse analysis parameters, it is possible that we missed very small amplitude pulses; however, the small SD of each infant’s cortisol levels around his mean indicates that those pulses that do occur are of a small enough amplitude to make them insignificant in the interpretation of the infant’s cortisol level (Fig. 1). No correlation was found between the presence of pulse(s) in cortisol levels, located by cluster analysis, and the mean cortisol level. Neither was a correlation seen between the presence of a pulse(s) and the degree of illness, as defined by both the SNAP-PE on the first day of life and the informal assessment on the day of the study. The slopes found on the graphs of cortisol level vs. time suggest that no clinically relevant upward or downward trends existed during the 6 h of the study, either inherent or study induced, and that a single cortisol level would adequately represent the cortisol levels throughout the study period.

View larger version (10K): [in this window] [in a new window]   Figure 1. A, Cortisol levels of an 839-g female delivered at 25.7 weeks EGA due to chorioamnionitis and preterm labor. Both steroids and TRH were given before delivery. She was studied at 4 days of age. Her initial SNAP-PE was 21; on the day of study, it was 18. She had been treated with indomethacin for a patent ductus arteriosus before the study. She required little ventilatory support at the time of the study and was receiving insulin for hyperglycemia. Others medications included ampicillin, cefotaxime, aminophylline (to enhance respiratory drive), and pentobarbital for sedation. Cluster analysis detected one short pulse in cortisol levels during the study. B, The single prolonged pulse found in the cortisol levels of a 26.5-week gestation, 891-g male who was delivered after the development of chorioamnionitis. Both steroids and TRH were given during labor. The initial SNAP-PE was 19; when studied at 3 days of age, it was 15. At the time of study, the infant was breathing room air; he had never required intubation. His only medication was aminophylline. C, The cortisol levels of a 4-day-old female delivered at 25 weeks EGA with a body weight of 640 g. Premature delivery was due to failed tocolysis for preterm labor. Antenatal steroids and TRH were not given. Her initial SNAP-PE was 50; at the time of study, it was 58. She had undergone patent ductus arteriosus ligation on the day before the study. She was receiving dopamine and insulin for hypotension and hyperglycemia, respectively. Other medications included ampicillin, cefotaxime, and fentanyl for postoperative pain control. Her need for ventilatory support was high. She had suffered a bilateral grade II intraventricular hemorrhage. No discrete pulses were detected by cluster analysis.

The issue of pulsatility of cortisol secretion in ill premature and term newborns has also been addressed in two recent studies by Arnold et al. using deconvolution analysis. No significant differences in pulsatility were detected between groups with and without clinical interventions such as endotracheal suctioning, iv catheter insertion, and reintubation (18), and there were no effects on the characteristics of cortisol pulsatility when antenatal steroids were given less than 7 days before birth (19). Although the mass of cortisol secreted did differ between the groups in Arnold’s antenatal steroids study, recent work by Scott and Watterberg found no alteration by antenatal steroids of the pattern of baseline or ACTH-stimulated cortisol levels in premature infants (gestational age, 24–36 weeks) (20).

Like all other organ systems, the endocrine axes progress through stages of maturity during gestation. We propose that the ELBW infants in our study, as a group, represent an earlier stage in the maturation spectrum of the hypothalamic-pituitary-adrenal (HPA) axis than do the premature infants in Metzger’s and Arnold’s works (1, 18). A poor response to stress by the HPA axis secondary to immaturity may be an important factor in defining the limit of viability in the ELBW infant.

Early in development, the fetus relies on transplacental passage of maternal cortisol; this dependence continues to a varying degree as its own HPA axis matures (21). Premature birth may occur before the HPA axis is adequately functional. Studies using pharmacological doses of ACTH and ovine CRH demonstrate appropriate responses of the adrenal gland and the pituitary gland, respectively, in ELBW infants, leading to speculation that the defect lies in the hypothalamus (22). Due to immaturity, that region of the brain may be unable either to secrete appropriate amounts of CRH or to recognize the need for elevated cortisol levels at times of stress. Others have found a poor adrenal response to more physiological doses of ACTH (9). In a study by Hingre et al. of very low birth weight preterm infants (gestational age, <30 weeks), inappropriately low levels of cortisol were found in both the basal and ACTH-stimulated states (23). Levels of several cortisol precursors, however, were dramatically elevated, suggesting one or more enzymatic deficiencies at the distal end of the HPA axis. The placenta is also known to produce a number of steroids and modulators that may act on the fetal HPA axis as both trophic factors and maturation inhibitors (24, 25, 26). It may well be that varying degrees of immaturity exist at numerous points along this endocrine axis. Finally, there are multiple causes of premature delivery, some of which may prompt early maturation of the HPA axis.

Because some ELBW infants appear to be cortisol deficient within the first few days of life, a method of easily assessing their adrenal status is of utmost importance in determining whether hormone replacement therapy is needed. We have shown that a single random plasma cortisol level is representative of the plasma cortisol levels over a prolonged period of time and, thus, can provide an assessment of the functional status of the adrenal cortex of the ELBW infant. The use in acutely ill adult patients of a total plasma cortisol level of less than 5 µg/dL as indicative of adrenal insufficiency (27) appears also to be appropriate in ill ELBW infants (4). Testing of the response to ACTH may be clinically useful as an additional measure of adrenal functional status. If carried out, the use of a low dose of ACTH-(1–24) (0.1 µg/kg), as recommended by Korte et al. (9), is desirable. A recommendation of an appropriate daily physiological replacement dose of hydrocortisone for an acutely ill ELBW infant is beyond the scope of this paper. The issue deserves further clinical study.

	Footnotes

 
¹ This work was supported by grants from the Wyeth Pediatrics Neonatology Research Fund and the General Clinical Research Center Program of the National Center for Research Resources (5-MOI-RR0034). Presented at the 65th Annual Meeting of the Society for Pediatric Research, May 1996, and at the 10th International Congress of Endocrinology, June 1996.

Received January 6, 1997.

Revised May 16, 1997.

Accepted May 27, 1997.

	References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Metzger DL, Wright NM, Veldhuis JD, Rogol AD, Kerrigan JR. 1993 Characterization of pulsatile secretion and clearance of plasma cortisol in premature and term neonates using deconvolution analysis. J Clin Endocrinol Metab. 77:458–463.[Abstract]
2. Vermes I, Dohanics J, Toth G, Pongracz J. 1980 Maturation of the circadian rhythm of the adrenocortical functions in human neonates and infants. Horm Res. 12:237–244.[Medline]
3. Lee MM, Rajagopalan L, Berg GJ, Moshang T Jr. 1989 Serum adrenal steroid concentrations in premature infants. J Clin Endocrinol Metab. 69:1133–1136.[Abstract]
4. Hanna CE, Jett PL, Laird MR, Mandel SH, LaFranchi SH, Reynolds JW. 1997 Corticosteroid binding globulin, total serum cortisol, and stress in extremely low birth weight infants. Am J Perinatol. 14:201–204.[Medline]
5. Colasurdo MA, Hanna CE, Gilhooly JT, Reynolds JW. 1989 Hydrocortisone replacement in extremely premature infants with cortisol insufficiency [Abstract]. Clin Res. 37:180A.
6. Ward RM, Kimura RE, Rich-Denson C. 1991 Addisonian crisis in extremely premature neonates [Abstract]. Clin Res. 39:11A.
7. Helbock HJ, Insoft RM, Conte FA. 1993 Glucocorticoid-responsive hypotension in extremely low birth weight newborns. Pediatrics. 92:715–717.[Abstract/Free Full Text]
8. Watterberg KL, Scott SM. 1995 Evidence of early adrenal insufficiency in babies who develop bronchopulmonary dysplasia. Pediatrics. 95:120–125.[Abstract/Free Full Text]
9. Korte C, Styne D, Merritt TA, Mayes D, Wertz A, Helbock HJ. 1996 Adrenocortical function in the very low birth weight infant: improved testing sensitivity and association with neonatal outcome. J Pediatr. 128:257–263.[CrossRef][Medline]
10. Ballard JL, Khoury JC, Wedig K, Wang L, Ellers-Walsman BL, Lipp R. 1991 New Ballard score, expanded to include extremely premature infants. J Pediatr. 119:417–423.[CrossRef][Medline]
11. Richardson DK, Gray JE, McCormick MC, Workman K, Goldmann DA. 1993 Score for neonatal acute physiology (SNAP): a physiology-based severity of illness index. Pediatrics. 91:617–623.[Abstract/Free Full Text]
12. Richardson DK, Phibbs CS, Gray JE, McCormick MC, Workman-Daniels K, Goldmann DA. 1993 Birth weight and illness severity: independent predictors of neonatal mortality. Pediatrics. 91:969–975.[Abstract/Free Full Text]
13. Krey LC, Lu K-H, Butler WR, Hotchkiss J, Piva R, Knobil E. 1975 Surgical disconnection of the medial basal hypothalamus and pituitary function in the rhesus monkey. Endocrinology. 96:1088–1093.[Abstract]
14. Veldhuis JD, Johnson ML. 1986 Cluster analysis: a simple, versatile, and robust algorithm for endocrine pulse detection. Am J Physiol 250:E486–E493.
15. Iranmanesh A, Lizarralde G, Johnson ML, Veldhuis JD. 1990 Dynamics of 24-h endogenous cortisol secretion and clearance in primary hypothyroidism assessed before and after partial thyroid hormone replacement. J Clin Endocrinol Metab. 70:155–161.[Abstract]
16. Mendelson JH, Mello NK, Teoh SK, Ellingboe J, Cochin J. 1989 Cocaine effects on pulsatile secretion of anterior pituitary, gonadal, and adrenal hormones. J Clin Endocrinol Metab. 69:1256–1260.[Abstract]
17. Scott SM, Watterberg KL. 1996 Low cortisol correlates to mortality in preterm infants. Pediatr Res. 39:224A.
18. Arnold JD, Bonacruz G. Leslie GI, Veldhuis JD, Bowen JR, Silink M. 1996 Effect of clinical state and clinical interventions on characteristics of plasma cortisol secretion in neonates using deconvolution analysis. Pediatr Res. 39:193A.
19. Arnold JD, Bonacruz G, Leslie GI, Veldhuis JD, Bowen JR, Silink M. 1996 Effect of antenatal steroids on pulsatile secretion of plasma cortisol in premature neonates using deconvolution analysis. Pediatr Res. 39:193A.
20. Scott SM, Watterberg KL. 1996 Lack of effect of prenatal steroids on baseline or stimulated cortisol values. Pediatr Res. 39:244A.
21. Pepe GJ, Albrecht ED. 1990 Regulation of the primate fetal adrenal cortex. Endocr Rev. 11:151–176.[Abstract]
22. Hanna CE, Keith LD, Colasurdo MA, et al. 1993 Hypothalamic pituitary adrenal function in the extremely low birth weight infant. J Clin Endocrinol Metab. 76:384–387.[Abstract]
23. Hingre RV, Gross SJ, Hingre KS, Mayes DM, Richman RA. 1994 Adrenal steroidogenesis in very low birth weight preterm infants. J Clin Endocrinol Metab. 78:266–270.[Abstract]
24. Partsch C-J, Sippell WG, Mackenzie IZ, Aynsley-Green A. 1991 The steroid hormonal milieu of the undisturbed human fetus and mother at 16–20 wks gestation. J Clin Endocrinol Metab. 73:969–974.[Abstract]
25. Tropper PJ, Warren WB, Jozak SM, Conwell IM, Stark RI, Goland RS. 1992 Corticotropin releasing hormone concentrations in umbilical cord blood of preterm fetuses. J Dev Physiol. 18:81–85.[Medline]
26. Honour JH, Wickramaratne K, Valman HB. 1992 Adrenal function in preterm infants. Biol Neonate. 61:214–221.[Medline]
27. Grinspoon SK, Biller BMK. 1994 Laboratory assessment of adrenal insufficiency. J Clin Endocrinol Metab. 79:923–931.[CrossRef][Medline]

This article has been cited by other articles:

				C. J. Herrington, I. N. Olomu, and S. M. Geller Salivary Cortisol As Indicators of Pain in Preterm Infants: A Pilot Study Clin Nurs Res, February 1, 2004; 13(1): 53 - 68. [Abstract] [PDF]

				S. L. King and K. M. Hegadoren Stress Hormones: How Do They Measure Up? Biol Res Nurs, October 1, 2002; 4(2): 92 - 103. [Abstract] [PDF]

				K. L. Watterberg, S. M. Scott, C. Backstrom, K. L. Gifford, and K. L. Cook Links Between Early Adrenal Function and Respiratory Outcome in Preterm Infants: Airway Inflammation and Patent Ductus Arteriosus Pediatrics, February 1, 2000; 105(2): 320 - 324. [Abstract] [Full Text]

This Article Abstract Full Text (PDF) Submit a related Letter to the Editor Purchase Article View Shopping Cart Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Jett, P. L. Articles by Hanna, C. E. Search for Related Content PubMed PubMed Citation Articles by Jett, P. L. Articles by Hanna, C. E.