Main

Evaluation of adrenocortical function in a preterm neonate is important when the infant has received antenatal or postnatal glucocorticoid treatment for an extended period. These babies may have adrenal suppression leading to failure to thrive, therefore the need for glucocorticoid replacement therapy has to be reliably determined(1–3). Before discontinuation of postnatal dexamethasone treatment, evaluation of the HPA axis is essential to ensure that the adrenal secretory response is fully functional(4).

The standard ACTH stimulation test has long been in clinical use to assess adrenocortical function in adults(5,6). In the standard ACTH test, 250 µg/1.73 m2 of synthetic ACTH is given intramuscularly or i.v. as a bolus, and the serum cortisol response is usually evaluated at 60 min after dosing. The standard test has also been called the pharmacologic or supraphysiologic ACTH test because the dose that is given is large enough to lead to a maximal cortisol response, which may sometimes provide misleading results. The low-dose (physiologic) ACTH test has been proposed to be more sensitive and may reveal mild but clinically significant suppression better than the standard ACTH test(7–9), but its use has not yet been established in preterm infants. In adults, the amount of synthetic ACTH that is administered in the low-dose test is usually 1 µg/1.73 m2, and the cortisol concentration is measured 30 min after dosing.

The purpose of this study was to find out whether the same time points that are used in adults could be used to measure plasma cortisol response to exogenous ACTH in preterm infants. We performed the standard- and the low-dose ACTH tests in neonates and compared the timing of the peak values because the peak values in ACTH stimulation tests are generally accepted as markers of adrenal responsiveness(7). We expected the peak value in the low-dose ACTH test to be reached earlier than in the standard ACTH test, as reported in adults(7,10).

METHODS

Eight preterm and two full-term infants were studied, and the same subjects were examined in both tests (Table 1). All subjects were or had been inpatients in the neonatal intensive care unit in Turku University Hospital. The mean birth weight of the infants was 1327.5 g (SD 850.8 g), and the median gestational age was 28 wk 4 d (range 26 wk 4 d to 40 wk 1 d). Six infants had respiratory distress syndrome. In the low-dose ACTH test the subjects were given 1 µg/1.73 m2 and in the standard-dose test 250 µg/1.73 m2 synthetic ACTH (Synacthen®; Ciba-Geigy, Stein, Switzerland) as an i.v. bolus. In both tests blood samples were taken before dosing and at 30, 40, 60, and 120 min after dosing for the measurement of serum cortisol. Timing of the peak values between the two tests was compared.

Table 1 Clinical data of the study patients
Full size table

The serum cortisol concentration was determined by RIA using a commercial kit (Technicon Immuno 1; Bayer, Middletown, VA) in both tests. The interassay coefficients of variation were 9.8% and 8.3% at serum concentrations of 80 nmol/L and 477 nmol/L, respectively. The intraassay coefficients of variation at the same cortisol concentrations were 2.3% and 2.3%, respectively. The lower limit of detection was 20 nmol/L.

The two tests were performed on separate occasions, approximately 1 mo apart, during the children's stay in the ward or during their regular control visits to the hospital. Nine patients received glucocorticoid treatment during this study (dexamethasone, i.v. or oral hydrocortisone, budesonide), and seven patients had been exposed to dexamethasone antenatally. The tests were performed in the morning, and the infants did not receive glucocorticoids for the preceding 12 h. The study was approved by the local ethics committee, and the children were enrolled in the study after informed consent from parents was obtained.

RESULTS

In the physiologic ACTH test (1 µg/1.73 m2) the peak value was reached at 30 (6/10 cases) or at 40 min (3/10 cases) (Fig. 1A). In one patient the peak value was reached at 60 min, whereas in the others the serum cortisol values had significantly decreased from their peak values at 60 min but were still above the baseline. At 120 min the cortisol values had decreased near or below the basal levels in eight cases. During this test two values at 30 min were missed. In the other case, however, the peak value was reached at 60 min.

Figure 1
figure 1

Serum cortisol responses in low-dose (1 µg/1.73 m2) ACTH test (A) and standard (250 µg/1.73 m2) ACTH test (B) in 10 prematurely born infants.

Full size image

In the pharmacologic test (250 µg/1.73 m2) the peak value occurred at 60 min (5/10) or at 120 min (3/10) and in two cases at 40 min (Fig. 1B). The peak values tended to be higher in the standard ACTH test than in the low-dose test, and the serum cortisol concentration was often already over the limit of 400 nmol/L at 30 min in the standard-dose test.

DISCUSSION

The short ACTH stimulation test (the standard- or low-dose ACTH test) is an alternative to the traditional IIH and the metyrapone tests for evaluating the function of the HPA axis(11,12). The ACTH test is widely used in children and in infants because of safety concerns associated with the IIH and metyrapone tests. IIH especially carries many risks; deaths have even been reported in children(13). Recently there has been concern about the reliability of the standard ACTH test because it may overestimate the adrenal responsiveness. A low-dose ACTH test has therefore been introduced for adults(9,12,14,15), although a recent paper showed an equal sensitivity between standard- and low-dose ACTH tests when detecting pituitary disorders in adults(16). The underdiagnosis of HPA impairment in neonates and preterm infants is a potential risk factor, and it may have severe consequences because these small infants cope poorly with stress. Therefore, the low-dose ACTH test may replace the standard-dose test in the assessment of adrenal function in infants. Before its wide use in infants, the profile of cortisol secretion after a low ACTH dose should be evaluated because the peak value may not appear at the same time in infants as in adults.

There has been some question about the correct lower limit of cortisol concentration that constitutes a normal response in a preterm infant after the ACTH stimulation test. The limit of 550 nmol/L that is used in adults and children may be too high for neonates because basal cortisol secretion of neonates is lower than in older children and adults(17,18). Wilson et al.(19) suggested that ACTH-stimulated cortisol concentration of more than 360 nmol/L indicates a normal adrenal response. Our aim in this study was to find the correct sampling times for neonates in both tests. We did not compare the sensitivities of the two tests, although we used the limit of 400 nmol/L to reflect proper cortisol response to ACTH and clinical well-being of the neonates.

The blood volume taken from the premature infants had to be minimized in this study, and time points of before dosing and 30, 40, and 60 min after dosing of ACTH were chosen to most effectively detect the peak cortisol concentrations. In previous studies with ACTH doses of less than 1 µg/1.73 m2, the peak value has been achieved before 30 min(8). As the peak cortisol response appears later with higher doses of ACTH(10), 30 min was chosen as the first sampling point after dosing in this study. In adult patients, the peak values have been attained between 30 and 40 min after administration of 1 µg of ACTH(10). Pretreatment with dexamethasone (1 to 2 h before the test) has been used to inhibit fluctuations in basal cortisol secretion both in pharmacologic and low-dose ACTH tests(20). In this study the infants did not receive glucocorticoids for the preceding 12 h before the ACTH tests because there is evidence that this kind of dexamethasone pretreatment blunts the cortisol increase in the low-dose test(21). Timing of the peak value is not affected by pretreatment with dexamethasone(21).

The profile of cortisol secretion after ACTH tests is similar in the heterogenic population of infants in the study. The infants who were chosen for this study were of different gestational and postpartum ages, and their total dose of dexamethasone varied from 0 to 4.5 mg. In addition, eight of the neonates had received antenatal dexamethasone. This suggests that the profile of cortisol secretion after ACTH tests is similar in all neonates regardless of the age of the neonate or the stage of dexamethasone treatment; therefore, the same time point for the measurement of serum cortisol concentration after ACTH tests can be used in all neonates.

We conclude that the timing of the peak values in neonates in the low-dose and the standard ACTH tests is similar to that in adults, and therefore the sampling can be performed similarly. Serum cortisol concentration should be measured in the standard test at 60 min after dosing. However, because the cortisol values in the standard-dose test appeared to be relatively stable at all time points between 30 and 120 min, any of these time points can be selected, as long as it is kept constant. In adults, the cortisol value at 30 min is widely used(6). The low-dose test can easily be performed in the wrong way if the correct sampling times are not followed, and this can lead to attaining false cortisol concentrations. According to our results, in the low-dose ACTH test the sampling for the measurement of serum cortisol concentration should be performed before the test and between 30 and 40 min after dosing in infants.