| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Departments of Medicine (D.S., R.E.W.), Obstetrics and Gynecology (J.M., J.U.H.), and Health Studies (K.K.), and Committees on Molecular Medicine and Nutritional Biology and General Clinical Research Center (R.E.W.), The University of Chicago, Chicago, Illinois 60637
Address all correspondence and requests for reprints to: Roy E. Weiss, M.D., Ph.D., Thyroid Study Unit, Department of Medicine, University of Chicago, 5841 South Maryland Avenue, Mail Code 3090, Chicago, Illinois 60637. E-mail: rweiss{at}medicine.bsd.uchicago.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objective: We sought to identify a normative, minimum-response threshold for the ACTH stimulation test in pregnancy. We hypothesized that salivary free cortisol (SaFC) would prove a more physiological and less variable measure of adrenal reserve in pregnancy than serum cortisol (SC).
Design: This is a prospective study of normal controls.
Setting: The study was conducted in an obstetrical clinic in a tertiary care hospital.
Patients: Patients included 36 healthy ambulatory pregnant women (aged 18–37 yr) with singleton pregnancies.
Intervention: The 250-µg ACTH stimulation test was performed in the healthy pregnant volunteers. Based on their gestational age at the time of recruitment, women were studied in one of the trimesters and were restudied at 11–14 wk postpartum.
Main Outcome Measures: Total SC, aldosterone, and SaFC concentrations were measured before and after ACTH. The response in pregnancy was compared with postpartum values.
Results: Basal SC (P = 0.01), aldosterone (P = 0.001), and SaFC (P = 0.01) values progressively increased during the trimesters of pregnancy and decreased postpartum, confirming that pregnant women have increased basal glucocorticoid and mineralocorticoid production. There was enhanced responsiveness of the maternal adrenal glands to ACTH stimulation as pregnancy progressed, as measured by peak stimulated SaFC (P = 0.009) and aldosterone (P = 0.01). In the milieu of altered binding globulins, SaFC is a more consistent, binding-globulin-independent measure of stimulated adrenal function than total SC. Minimum criteria for the normal SaFC response to ACTH stimulation in the second and third trimesters of pregnancy and postpartum have been generated based on a predominantly African-American group of subjects.
Conclusions: Reliable data are available for the evaluation of the adrenal axis in pregnancy with a noninvasive, outpatient measure of SaFC. Glucocorticoid therapy in pregnancy should take into account that adrenal reserve increases as pregnancy progresses.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The above-described changes underlie the diagnostic dilemmas caused by pregnancy, for although it is widely agreed that one should evaluate the hypothalamic pituitary axis (HPA) of gravidas with symptoms that suggest adrenal insufficiency, standard diagnostic criteria do not seem to exist. In addition, clinicians may ignore signs and symptoms such as emesis, fatigue, and mild hyponatremia because they occur in normal pregnancy. Furthermore, an adrenal-deficient gravida often presents with values of total cortisol that are within the normal laboratory range for nonpregnant women, reflecting substantial hormone bound to the high levels of CBG. Diagnostic testing is further limited in pregnancy because the insulin tolerance and metyrapone tests are both contraindicated, whereas cortical-releasing hormone testing is not likely useful due to its robust placental production (17). The 250-µg standard ACTH stimulation test (SCT) thus remains the most widely available dynamic test of adrenocortical function in pregnancy (class C) (13, 14), but, as noted, there are little data available to determine specific cutoff values when attempting to diagnose adrenal insufficiency in pregnant women.
We therefore designed a study to evaluate adrenal function throughout normal gestation with the SCT, measuring both total circulating and salivary cortisol, hypothesizing that the latter would be less variable and would prove a more reliable HPA test. A second aim was to establish norms of cortisol and aldosterone responses to aid in the diagnosis of adrenal insufficiency in gestation.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Women were studied shortly after recruitment either in the first trimester (11–13 wk, n = 4), second trimester (20–24 wk, n = 16), or third trimester (32–36 wk, n = 16), and 32 subjects were then restudied at 11–14 wk postpartum. Subjects were not taking oral estrogens or contraceptives at the time of postpartum testing. Classification into trimesters was based on the time of testing. The study was approved by the University’s institutional review board. The volunteers, who gave written and informed consent, were compensated.
Protocol
Volunteers reported to the outpatient General Clinical Research Center for both the pre- and postpartum visits between 0800 and 0900 h. An iv catheter was placed for rapid phlebotomy, avoiding potential increases in cortisol secretion due to pain from needle sticks. With the volunteer supine for the duration of testing, 250 µg 1–24 synthetic ACTH (Cortrosyn, cosyntropin; Amphastar, Rancho Cucamonga, CA) was administered between 0900 and 1000 h, and serum total cortisol, serum aldosterone, and salivary cortisol concentration were measured before and 30 and 60 min after cosyntropin stimulation. Plasma ACTH and CBG were measured before cosyntropin administration.
The most widely used criteria for interpreting the SCT specify that in normal nonpregnant adults, the peak cortisol value after stimulation should reach a maximal level of more than 18–20 µg/dl (500–550 nmol/liter) (18, 19, 20). Although the 30-min cutoff point is considered the more consistent measure for diagnosis, the 60-min value is almost invariably higher and is the value most widely used for diagnosis (21). The incremental increase in cortisol in response to cosyntropin has not been adopted as a valid diagnostic criteria for adrenal insufficiency (21). We defined the maximal response to cosyntropin as the peak steroidogenic response measured at either 30 or 60 min after cosyntropin administration. Peak aldosterone response was defined as the maximal aldosterone response to cosyntropin at 30 or 60 min.
Laboratory analysis
Plasma cortisol was measured by a competitive chemiluminescent enzyme immunoassay (Immunolite; Diagnostic Products Corp., Los Angeles, CA); 0.2 µg/dl, 5.8–8.8%, and 6.3–10.5% were the lower limit of detection, intraassay coefficient of variation, and interassay coefficient of variation, respectively. Serum aldosterone was measured by RIA (Diagnostic Products Corp.), with a lower limit of detection of 1.6 ng/dl. Plasma CBG was measured by RIA (Nichols Institute, San Juan Capistrano, CA), with an intraassay coefficient of variation of 6% and an interassay coefficient of variation of 9%.
Salivary cortisol was measured by an enzyme-linked immunoassay (Salimetrics, State College, PA). This assay has a lower limit of detection of 0.4 mmol/liter, an intraassay coefficient of variation of 3.8%, and an interassay coefficient of variation of 6.7%. A standard device, composed of a cotton tube and two larger plastic tubes, was used to collect and transport the saliva. Saliva was obtained by chewing on the cotton tube for 1 min. The cotton tube with absorbed saliva was then inserted inside the plastic tube, which was then capped (22).
Statistical analysis
Data are presented as means ± SD as well as medians and ranges. The between-trimester patient data were first analyzed with the use of the Kruskal-Wallis test as a nonparametric alternative to the ANOVA. If the Kruskal-Wallis test was significant, then pair-wise comparisons between trimesters were performed with the use of the Mann-Whitney test. Comparisons between pregnant and postpartum values were performed using the Wilcoxon signed-rank test as a nonparametric alternative to the paired t test. The Pearson correlation coefficient was calculated to quantify the strength of the association between two variables. Linear regression models were fit using standard regression techniques.
Because there were three primary end points [SC, salivary free cortisol (SaFC), and aldosterone], a Bonferroni-adjusted P value of 0.017 or less (i.e. 0.05/3) was used to indicate statistical significance for the Kruskal-Wallis test. Because the pair-wise comparisons between trimesters were only performed if the Kruskal-Wallis test was significant (Fisher’s protected least significant difference method), no further adjustments were made. Additionally, for the signed-rank tests and correlation coefficient analysis, a P value of 0.01 or less was designated as statistically significant.
The distribution of baseline and stimulated SC, serum aldosterone, and SaFC demonstrated a non-Gaussian distribution; thus, the natural log transformation was used for determining normal cutoff limits. Rankit plots of the log-transformed data indicated more normal distributions. The data were analyzed using the Minitab statistical program (version 13; Minitab Inc., State College, PA) and Stata version 9 (Stata Corp., College Station, TX).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
). All patients delivered full-term (
36 wk) healthy infants without significant obstetrical complications.
|
Baseline morning SC rose progressively and significantly from 9.3 ± 2.2 µg/dl in the first trimester to 14.5 ± 4.3 µg/dl in the second trimester and 16.6 ± 4.2 µg/dl in the third trimester and was 9.1 ± 4.8 µg/dl postpartum (Fig. 1 and Table 2). The latter value did not differ significantly from the first trimester but was lower than those measured in the second and third trimesters (P = 0.003 and 0.002, respectively, Table 2). Although the peak SC response to cosyntropin appeared to rise as pregnancy advanced, the increment was not significant (Fig. 1). There was, however, significantly increased responsiveness to ACTH in late gestation compared with the nonpregnant state (cosyntropin stimulated peak cortisol, 37.9 ± 9.0 µg/dl in the second and 34.7 ± 7.5 µg/dl in the third trimester, compared with 27.2 ± 5.6 µg/dl postpartum; P = 0.003 and 0.02, respectively).
|
|
Baseline morning SaFC rose progressively and significantly (0.21 ± 0.10 µg/dl, first trimester; 0.36 ± 0.17 µg/dl, second trimester; and 0.47 ± 0.18 µg/dl, third trimester; Fig. 1
, Table 2). Values in the last two trimesters were each significantly higher than in the first (P = 0.05; P = 0.01), but the difference between trimesters two and three failed to achieve statistical significance (P = 0.06). Baseline morning SaFC in the nonpregnant state, 0.22 ± 0.17 µg/dl, was lower than in the second and third trimesters of pregnancy (P = 0.01; P = 0.008). There was an increased responsiveness of SaFC to cosyntropin as pregnancy advanced (Fig. 1), with the peak salivary cortisol responses being 1.79 ± 0.76 µg/dl, first trimester; 1.70 ± 0.41 µg/dl, second trimester; and 2.37 ± 0.99 µg/dl in the third trimester (P = 0.009). The SaFC response was significantly greater in the third compared with that in the second trimester (P = 0.002), whereas levels during both trimesters were above those obtained postpartum (1.16 ± 0.23 µg/dl; P = 0.008; P = 0.001).
Serum aldosterone measurements
Baseline serum aldosterone rose from 8.0 ± 2.3 to 13.8 ± 5.3 to 25.3 ± 10.8 ng/dl during the first, second, and third trimesters of pregnancy, respectively (P = 0.001; Fig. 1 and Table 2), with values higher in the third trimester compared with those in the first and second trimesters, respectively (P = 0.005 and 0.003), and levels in the second higher than those in the initial trimester (P = 0.02). Aldosterone responsiveness to ACTH was also enhanced in pregnancy; the peak serum response to cosyntropin was 38.0 ± 15.5 ng/dl in the first trimester, 49.5 ± 24.9 ng/dl in the second trimester, and 73.9 ± 30.0 ng/dl in the third trimester (P = 0.01; Table 2). Peak responses in the last two trimesters were greater than those in the nonpregnant state (14.8 ± 8.6 ng/dl; P = 0.001; P = 0.001).
Diagnostic criteria
The antilogarithm of the 2.5th percentile (mean – 2 SD) of the natural-log-transformed data was defined as the minimal steroid response in the second and third trimesters (Table 3). Sample size was too small to provide data for the first trimester of pregnancy. Assuming our sample is representative of healthy gravidas in general permits the conclusion that basal morning salivary cortisol will be at least 0.13 µg/dl in 97.5% of normal second-trimester women. Values below this in a second-trimester pregnant woman, with the appropriate clinical signs and symptoms, should raise the specter of adrenal insufficiency. Also note that, because stimulated SC values were at least 20 µg/dl in all second- and third-trimester volunteers, use of traditional nonpregnant cutoff values (18- 20 µg/dl) would potentially lead to the underdiagnosis of adrenal insufficiency in pregnancy.
|
) (r = 0.79, P < 0.001), but the cosyntropin-stimulated peak values of each did not (r = 0.01, P = 0.96). As expected, baseline SaFC did not correlate with serum aldosterone (Fig. 2B) (r = 0.21, P = 0.22) because activation of the renin-angiotensin system and serum potassium balance are the major regulators of the latter’s secretion. However, cosyntropin-stimulated SaFC did correlate somewhat with serum aldosterone (r = 0.39, P = 0.02), whereas neither baseline nor stimulated SC correlated with serum aldosterone levels (Fig. 2C) (r = 0.27, P = 0.12; r = 0.15, P = 0.39).
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
It is currently difficult to detect adrenal insufficiency during pregnancy, in part because we lack clear biochemical criteria to help establish a diagnosis or to guide corticoid replacement. This is unfortunate given the substantial maternal and fetal morbidity associated with adrenal insufficiency, which is easily preventable with appropriate replacement therapy, particularly when diagnosis occurs early in gestation. A need for appropriate test criteria is further underscored because of the increased prevalence of autoimmune adrenal insufficiency as well as the rates of secondary adrenal insufficiency (3, 18, 23) (e.g. the more widespread use of oral and inhaled steroids for asthma and in preterm labor for fetal lung maturation and the relative adrenal insufficiency of critical illness). Therefore, improved diagnostic tools for identifying adrenal insufficiency in pregnancy are necessary.
The diagnosis of adrenal insufficiency in pregnancy, however, remains challenging. Our data, and numerous other published reports (summarized in Ref. 24), indicate that measuring baseline serum total cortisol concentration in late gestation can be misleading, especially if the criteria used for diagnosis are based on norms determined in nonpregnant populations (who have markedly lower CBG levels). In addition, there appear to be no criteria for interpreting SCT values in pregnancy, a more cost-effective and reliable dynamic test of the HPA in pregnant women.
Our data demonstrate that pregnant women have an activated HPA characterized by nearly 2-fold elevations in serum total cortisol and SaFC and 8-fold elevations in serum aldosterone concentrations in late gestation as compared with the nonpregnant state. We also show an enhanced responsiveness of the maternal adrenal gland to ACTH as pregnancy progresses. Our results are in agreement with previously published studies that have investigated the physiology of the HPA in pregnancy (6, 10) and responses to synthetic ACTH administration in pregnancy (11, 12, 13, 14, 24). However, to the best of our knowledge, this is the first study to evaluate cortisol response to the SCT in a cohort of normal healthy pregnant women, generating specific cutoff points for the interpretation of this test in pregnancy. Also, we could locate no published data of SaFC responses to 250-µg cosyntropin stimulation throughout the three trimesters of pregnancy.
Our data also demonstrate a lack of correlation between cosyntropin-stimulated total SC values and SaFC in our healthy pregnant cohort. Additionally, cosyntropin-stimulated serum aldosterone secretion, a binding-protein-independent marker of adrenal function similar to free cortisol, is correlated with SaFC but not with SC. This is concordant with other published reports that indicate that the correlation between serum total cortisol and SaFC is significantly altered in situations with elevated serum total cortisol concentrations more than 600 nmol/liter (25), the range commonly seen in stimulated adrenal function in the second and third trimesters of pregnancy. Additionally, because CBG has low capacity but high affinity binding for cortisol, CBG binding can become rapidly saturated even in situations of elevated CBG levels with supraphysiological cortisol concentrations (25, 26, 27). Therefore, in pregnancy, with its resultant high degree of interindividual variation in magnitude and timing of CBG elevations, a pregnant woman’s serum total cortisol is significantly affected by her pattern of CBG elevation, as opposed to her actual cortisol production and biologically active free cortisol level. It is for these reasons that we propose that SaFC is the preferred measurement of assessing adrenal responsiveness to ACTH in pregnancy.
Our data demonstrate an increasing adrenal responsiveness to ACTH in normal pregnancy, reflected by a 2- to 5-fold increase in adrenal steroidogenic response to ACTH as pregnancy progresses, despite elevated baseline cortisol and aldosterone levels. This suggests the need for temporally increasing glucocorticoid and mineralocorticoid replacement when treating adrenal insufficiency through pregnancy. However, no widespread consensus has been reached on whether mineralocorticoid and glucocorticoid therapy needs to be adjusted during the course of pregnancy (3, 6, 28). Attempting to decrease maternal steroid replacement for adrenal insufficiency during pregnancy is not recommended in view of the finding that, in normal pregnancy, maternal adrenal steroidogenic response increases as pregnancy progresses.
The primary limitation of our study is that only normal pregnant women have been investigated. Because we have not yet studied the response to the SCT in gravidas with adrenal insufficiency, our minimum response criteria have been derived using a fundamentally statistical definition of "normality," rather than one that is clinical. Therefore, we cannot definitively state that our proposed lower cutoff limits for the SCT will differentiate all normal pregnant patients from those with adrenal insufficiency. Our sample sizes are also rather limited for establishing normal cutoff values even on a statistical basis. Additionally, the accuracy of our diagnostic tool (SCT) was not tested against the insulin tolerance test, a reference standard, as this test is contraindicated in pregnancy. Finally, because the vast majority of the patients studied were African-American, representative of the racial composition of The University of Chicago obstetrics population, the generalizability of our results to all gravidas may be limited. A longitudinal study is needed to confirm the changes observed among trimesters.
In summary, our study demonstrates that total SC, SaFC, and serum aldosterone levels are elevated in normal pregnancy and increase as gestation progresses in African-American women. We have also determined that adrenal reserve increases as pregnancy advances. Our results indicate that SaFC is a more consistent, generalizable, and physiologically rational measure of adrenal function in pregnancy rather than serum total cortisol. Measuring SaFC response to the SCT in pregnancy provides the clinician with a minimally invasive test that can be performed in the outpatient setting and does not require special skills, laboratory facilities, or precise timing. Finally, we have generated trimester-specific minimum-response criteria for interpreting the SCT in normal pregnancy, although further studies are needed to determine more accurately this threshold.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
The authors have nothing to declare.
First Published Online August 8, 2006
Abbreviations: CBG, Cortisol binding globulin; HPA, hypothalamic pituitary axis; LMP, last menstrual period; SaFC, salivary free cortisol; SC, serum cortisol; SCT, 250-µg standard ACTH stimulation test.
Received May 15, 2006.
Accepted July 31, 2006.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. D. Lindheimer, A. Khurana, N. Vinayek, C. A. Stratakis, A. Klibanski, and M. F. Greene Case 36-2006: A Pregnant Woman with New Hypertension N. Engl. J. Med., March 1, 2007; 356(9): 966 - 968. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
represents the mean incremental increase in response to cosyntropin administration. The error bars represent the SD.