| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Endocrinology (K.H.D., R.D.M., H.K.G., S.M.S.), Christie Hospital, Manchester M20 4BX, United Kingdom; and the Department of Medicine (S.S.P., M.O.T.), University of Virginia Health Science Center, Charlottesville, Virginia 22908
Address all correspondence and requests for reprints to Professor S. M. Shalet, Department of Endocrinology, Christie Hospital, Wilmslow Road, Manchester M20 4BX, United Kingdom. E-mail: stephen.m.shalet{at}man.ac.uk.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Study Subjects and Design: Twenty-four-hour GH profiles at 20-min intervals were undertaken in the fed state and in the last 24 h of a 33-h fast in eight young adult cancer survivors (two women) with severe GH deficiency after cranial irradiation for nonpituitary brain tumors in childhood and 14 matched normal controls (three women). A sensitive chemiluminescence GH assay was used with cluster analysis.
Results: Fasting induced a significant (P < 0.05) rise in all amplitude-dependent measures (absolute GH peak and nadir, profile mean GH, and mean pulse amplitude and area) in both groups. Pulse frequency was nonsignificantly increased (by 10%) in normals but significantly increased (by 20%) in the patients. The average increase in the individual fasting profile mean GH concentration was 3.7-fold (range 1.5–8.3) in normals, compared with 2.7-fold (range 1–4.7) in the patients (P > 0.05). Fasting amplified amplitude-related differences between patients and controls, and thus, unlike in the fed state, the day (0900–2040 h) mean GH completely demarcated patients from normals. An absolute GH peak level of 2 and 4 µg/liter and a profile mean GH level of 0.25 and 0.65 µg/liter completely separated patients from normals in the fed and the fasting states, respectively. Overall, fasting seems to induce a feminized pattern of GH secretion with relatively higher interpeak levels, preserved but diminished diurnal variation, and increased secretory disorderliness (increased approximate entropy scores).
Conclusion: The overall pulsatile pattern of GH secretion during fasting in patients with radiation-induced GH deficiency and the relative augmentation in GH release are similar to that seen in normals emphasizing that GH neuroregulation is preserved in these patients even when the hypothalamic-pituitary axis is under physiological stress.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
However, that study examined spontaneous GH secretion under normal physiological circumstances, and it remained unclear whether stressing the h-p axis might unmask neurosecretory dysregulation that is not seen under basal conditions.
Fasting induces well-defined neuroendocrine and metabolic changes that are believed to contribute to the human adaptation to nutrient deprivation (2, 3, 4, 5). GH secretion during fasting is enhanced; this reflects an intact h-p axis response to the stress of fasting and the associated metabolic and hormonal changes (2, 6). Previous studies in normals after a 2-d fast revealed a 3-fold increase in circulating GH concentration during 24-h profiling and a 5-fold increase in GH production rate without any change in the calculated endogenous GH half-life (5, 6).
GH secretion in GH-deficient patients in the fasting state has rarely been studied. Two studies of this nature analyzing diurnal GH secretion at half-hour intervals in the last 8 h of a 44-h fast revealed little change in the mean GH concentration in nine severely GH-deficient patients and nine obese individuals with functional GH deficiency, compared with many-fold increase in mean GH concentration in matched normal controls (7, 8). These two studies, however, failed to examine the whole 24-h profile and identify any changes in the pulsatile pattern or the diurnal variation. In addition, the use of a less sensitive GH assay, with many samples possibly having a GH concentration below the detection limit of the assay used, could have influenced the detection of the actual change in mean GH concentration in the GH-deficient subjects (7, 8).
We therefore followed up our initial study of spontaneous GH secretion in fed patients with severe radiation-induced GH deficiency by examining the impact of fasting, a mild physiological stimulus, on quantitative and qualitative aspects of GH secretion and neuroregulation, with the added purpose of generating robust diagnostic thresholds that can be used in future research to determine whether GH neurosecretory dysfunction exists in cranially irradiated adult survivors with normal GH responses to provocative tests.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Eight (two women) of 10 young adult patients with severe radiation-induced GH deficiency and 14 (three women) of the 30 normal controls who previously took part in 24-h GH profiling in the fed state (1) successfully completed a fasting profile and were suitable for this study. The diagnoses (five medulloblastoma, two ependymoma, and one T cell lymphoma) and clinical characteristics of the patients were described in the original study (1). Apart from GH deficiency, none of the patients had any other pituitary hormone deficiencies. All patients had normal gonadal function except for three male patients who had primary testicular failure attributable to chemotherapy and/or radiotherapy and were receiving testosterone replacement therapy; their circulating testosterone levels on replacement therapy had been normal for at least 3 months before testing. Of these three patients, one was also receiving T4 for primary hypothyroidism. Patients and normals were age and body mass index (BMI) matched (24.6 ± 2.3 and 25.8 ± 1.7 yr and 25.3 ± 0.9 and 22.4 ± 0.5 kg/m2, respectively; P > 0.05).
Study protocol
The study was approved by the South Manchester Local Research Ethics Committee, and informed consent was obtained from all subjects before testing.
Twenty-four-hour profiling at 20-min intervals in the fed state was performed as reported previously (1). Fasting 24-h profiles were performed 1–4 wk after the fed profiles; patients were asked to start fasting from 2400 h (midnight) before their admission at 0800 h. The fasting was strictly supervised on the study unit until 0900 h the next morning. Sampling was undertaken every 20 min, as described in the fed state (1). Basal serum and fluoride oxalate samples were taken at the start of the profiling and at the end (after fasting) for analysis of IGF-I, free fatty acid (FFA), 3-hydroxybutyrate (3-HOB), bilirubin, and glucose. Sera were separated and immediately frozen at –80 C until the assays were performed.
Women were profiled in the first half of their menstrual cycle and none had taken any oral contraception for at least 6 months before the study.
Assays
Serum samples from the fed and the fasting profiles for each subject and their GH stimulation tests were analyzed in duplicates in the same assay run using the modified Luma Tag hGH chemiluminescence immunometric assay (Nichols Institute, Inc., San Juan Capistrano, CA) (1). The GH standards used in this assay are calibrated against the first IS 80/505 International Standard. The sensitivity of this assay was 0.003 µg/liter. The intraassay coefficients of variation were 11.3, 9.8, and 11.7% for GH concentrations of 8.22, 0.293, and 0.027 µg/liter, respectively. The interassay coefficients of variation at the same GH concentrations were 6.6, 7.7, and 10.4%, respectively. Serum IGF-I was determined by a two-site chemiluminescent immunometric assay (1). The FFA and 3-HOB were analyzed as described previously (9, 10).
Analysis of GH concentration profiles
Cluster algorithm for pulsatility analysis (11), Cosinor analysis for appraisal of diurnal variation (12), and approximate entropy statistic (ApEn) to appraise secretory orderliness (13) were performed as described in the initial study of GH secretion in the fed state (1). The day (0900–2040 h) and night (2100–0840 h) mean GH concentrations and the ratio of the night area under the curve for GH (AUCGH) to total AUCGH were noted to measure the nocturnal increase in GH production.
Cumulative profile GH mean and cumulative AUCGH
The cumulative GH mean (or cumulative AUCGH) in the profile is plotted against the time in the profile to identify the earliest time at which maximum separation occurs between the patients and normals. The cumulative GH mean at a particular time point in the profile is the product of the sum of all GH measurements divided by the number of measurements between the start of the profile and that time point. The cumulative AUCGH is the product of the cumulative mean multiplied by the duration (in minutes) to that time point. It is to be noted that whereas AUCGH continues to accrue with more time into the profile, the cumulative GH mean can drift up and down in line with the pulsatile nature of GH secretion.
Statistical analysis
The data (normally distributed) were expressed as mean ± SEM; ranges are shown to give a better idea about the actual spread of the data. Unpaired t test was used to examine differences between normally distributed groups, whereas paired t test was used to compare normally distributed paired data (before and after fasting). Simple correlations to examine the relationship between variables were made using the Spearman rank order correlation test. Statistical significance was accepted at P < 0.05.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
The dynamics of GH secretion in the fed state, IGF-I data, and the effects of gender and BMI on GH secretion are extensively covered in our initial study (1). In brief, GH was detected in all samples and pulsatility was evident in all subjects. All amplitude-related measurements were significantly reduced in the patients, whereas frequency-related properties were similar in both groups and were not affected by gender or BMI.
The absolute peak GH (range 0.04–1.62 µg/liter) and mean GH levels (range 0.02–0.25 µg/liter) in the patients (n = 8) were clearly demarcated from that seen in normals (n = 14; range 2.7–15.2 and 0.28–2 µg/liter, respectively). Considerable overlap was, however, seen in the nadir levels (range 0.01–0.03 µg/liter in patients and 0.018–0.071 µg/liter in normals).
GH secretion in the fasting state
Fasting was confirmed objectively in each individual study subject and the whole group by a significant fall in blood glucose in conjunction with a rise in serum bilirubin, FFA, and 3-HOB (Table 1
).
|
and 2), which were relatively higher in men than women (higher ratio of the profile mean GH in the fasting/fed states). In addition, the augmentation in the average mean valley GH levels (interpeak levels) with fasting was manyfold higher in men, compared with women (Table 2). Similarly, the daytime GH secretion with fasting was relatively higher (reduced ratio of the nighttime AUCGH to total AUCGH) in men only, leading to a significant reduction in the relative amplitude of the diurnal variation. Thus, the greater amplitude-related measurements and the reduced amplitude of diurnal variation (reduced nighttime AUCGH to total AUCGH) seen in normal women, compared with normal men, in the fed state completely disappeared in the fasting state. Fasting also caused a slight advancement in the acrophase that failed to reach statistical significance. The increase in GH secretion is predominantly related to the increase in the GH pulse area with no significant change in pulse frequency (Table 2).
|
|
|
).
|
).
The responses to fasting did not correlate with age, BMI, or IGF-I levels (P > 0.05) in either group. The average increase in the individual mean GH concentration with fasting (profile mean GH in the fasting/fed states) was 3.7-fold (range 1.5–8.3) in normals, compared with a less profound increase of 2.7-fold (range 1–4.7) in the patients; however, the difference was insignificant (Table 3). In addition, the responses to fasting were more heterogeneous; four of the eight patients (50%) failed to show an increase in the absolute and/or mean GH levels, compared with only two controls (14%) (Fig. 2). Failure to respond to fasting was unrelated to the severity of GH deficiency (as measured by the profile mean GH level), biological effective dose or irradiation, age at irradiation, or postirradiation follow-up interval.
The first separation in the cumulative AUCGH between the patients and normals occurred only toward the last few hours of the profile in the fed state (around 0500 h) but much earlier (around 2100 h) in the fasting state (Fig. 3). Thus, unlike the night GH mean, which demarcated patients from normals in both the fed and fasting states, the day GH mean did so in the fasting state only (Fig. 4).
|
|
); however, the overlap in the nadir level persisted in the fasting state (range 0.018–0.085 µg/liter in the patients and 0.025–0.145 µg/liter in normals).
In addition, fasting-induced GH hypersecretion is most exuberant after midday until the first few hours after midnight in such a way that the mean GH level of only a 6-h sampling period starting at any time between 1600 and 2000 h (and finishing at 2200–0200 h) provided excellent separation with a highest value of 0.75 µg/liter in the patients and a lowest value of 1.2 µg/liter in normals (Fig. 5). This provided the earliest and shortest possible period of sampling after commencing the fast at 2400 h that can provide a diagnostic threshold as robust as that offered by a complete 24-h profiling.
|
IGF-I levels were significantly reduced in the patients, and fasting resulted in a mild but significant fall in IGF-I levels in all subjects (Tables 2 and 3).
Secretory disorderliness (ApEn)
Fasting induced a relatively higher increase in the ApEn values in normals, compared with the patients who had considerably higher ApEn values in the fed state, and more so in men, compared with women. Thus, the ApEn scores in normals during fasting were similar to those seen in GH-deficient patients in the fed state. Consequently, the differences in ApEn values between normal men and women and between the patients and normals were attenuated in the fasting state (Tables 2 and 3).
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
In our study, the overall pattern of changes in GH concentration pulse profile is similar to that reported by Bergendahl et al. (5) with a change in pulse frequency of about 10–20% and a significant rise in the absolute and mean peak height (amplitude). Although patients seem to have significantly increased pulse frequency during fasting, this is unlikely to be of biological significance, and it is not possible to conclude that patients are different from normals in this respect without proper estimation of the underlying secretory pulses with deconvolution analysis (15, 17). It is likely that with much higher secretory pulse amplitudes in normals, the chance of two or more pulses becoming superimposed on each other before the GH concentration declines to a plateau level is more frequent, resulting in apparently less frequent (but broader) GH concentration pulses.
The 3-fold increase in GH concentration in our study was similar to that seen in other studies on d 1 of fasting (2) and after more prolonged fasting periods of 36–120 h (2, 5, 6, 7, 8, 16, 18, 19, 20). It is interesting to note that fasting can reach its maximum impact on GH secretion as soon as the first two meals in the day are missed. In addition, the relative increase of GH secretion during the day was even higher than that seen during the night. Thus, complete separation between the patients and normals was achievable by the end of the first 12 h of sampling (day mean). In contrast, there was substantial overlap in the day mean in the fed state and complete separation between the cumulative means only occurred in the last few hours of the profile.
Perhaps the most important finding in this study relates to the observation that the somatotrophic axis in some of the cranially irradiated GH-deficient patients is still capable of mounting an enhanced response to short-term fasting with a relative increase in mean GH concentration that was insignificantly lower (2.7-fold), compared with normals (3.7-fold). These findings suggest that hypothalamic (and suprahypothalamic) neurosecretory mechanisms responsible for fasting-induced augmentation in GH secretion are not significantly damaged by radiation therapy. However, the individual responses were quite heterogeneous, with about 50% of patients failing to have an increase in the absolute peak GH and/or mean GH levels, compared with only 14% in the control group. It is unclear why some patients failed to respond to fasting; it is possible that in those patients the somatotrophic axis is already functioning at its maximum capacity under basal conditions, and thus, further stimulation does not occur with the stress of fasting.
A much lower overall increase in GH secretion was previously reported in nonirradiated severely GH-deficient patients tested in the last 8 h of a 44-h fast (7). Similarly, studies of fasting GH secretion in obese individuals with functional GH deficiency showed very limited GH responses to fasting (8, 18). In contrast, hyposomatotropism associated with aging can be partially reversed by fasting (16).
Fasting does not appear to abolish the sleep-entrained nocturnal increase in GH secretion despite a relatively higher increase in daytime GH secretion. The latter, however, may have contributed to the advancement of the acrophase (by about 1–2 h) in both the patients and controls and a reduction in the amplitude of the diurnal variation. A previous study of a 2.5-d fast in normal women also showed an advancement of the acrophase by about 2 h that failed to achieve statistical significance (5). Overall, fasting seems to induce a feminized pattern of GH secretion with diminished amplitude of diurnal variation and increased secretory disorderliness in normals and to a less extent in the patients. These changes are similar to those seen in GH-deficient patients (1), hence, the absence of differences in these attributes between patients in the fed or fasting state and normals in the fasting state (Table 3).
It has been suggested that fasting can be used to enhance the diagnostic usefulness of GH profiling by eliminating the overlap in integrated GH levels seen between the patients and normals in the fed state (7). This is because the fold (relative) increase in GH levels with fasting in the patients is less than that seen in normal controls. It is to be noted, however, that this overlap in the fed state reported in many studies (21, 22, 23) mostly relates to a variety of methodological factors and does not necessarily reflect a genuine phenomenon. Perhaps the most critical of these factors is the use of less sensitive GH assays that impeded accurate calculation of the mean GH levels due to a large proportion of samples with undetectable levels of GH (21, 24). The use of sensitive GH assays has enhanced the separation between patients and normals but not completely eliminated the overlap (22, 23, 24). This partial failure may be attributed to the inclusion of patients with less severe degrees of GH deficiency (22) as well as the inclusion of heterogeneous study groups with mixed sexes and wide ranges of age and BMI (22, 23, 24). In this study in which all normals and patients are young and not obese, no such overlap was seen in the fed state and only a very minimal overlap was seen in our original study, which included a larger number of patients and controls with a wider range of BMI (1).
Thus, based on the data generated in this and the previous study (1), it is possible to conclude that the combined use of an absolute peak GH level of less than 2 µg/liter and a mean GH concentration of less than 0.25 µg/liter (AUCGH = 360 µg/min·liter) in the fed state and/or an absolute peak GH of less than 4 µg/liter and a mean GH concentration of less than 0.65 µg/liter (AUCGH = 930 µg/min·liter) in the fasting state can identify severely GH-deficient patients with an almost 100% accuracy. These thresholds may be used in future studies to define GH status in patients with discordant results on provocative testing or those suspected to have GH neurosecretory dysfunction.
Robust diagnostic thresholds can also be generated from a limited sampling period as short as 6 h at times when fasting-induced GH hypersecretion is most exuberant. This occurred any time between 1600 and 2000 h after commencing the fast at 2400 h and finished between 2200 and 0200 h. The highest mean GH level in this 6-h period in the patients (0.75 µg/liter) was much lower than that achieved in normals (1.2 µg/liter). This approach is therefore as useful as that previously suggested, which involved 8-h sampling after a 36-h fast (7), with the added benefit of being more practical.
The results of the current study considered in isolation could be explained mechanistically by either a hypothalamic or pituitary site of damage; there is no implicit contradiction with the existing belief that the hypothalamus is the primary site responsible for radiation-induced GH deficiency with somatotroph atrophy occurring secondarily (25).
In conclusion, fasting for 33 h results in a 2.7-fold increase in mean GH concentration in cranially irradiated severely GH-deficient patients, compared with a 3.7-fold increase in normal controls. Fasting amplified the differences and thus, unlike in the fed state, the day (0900–2040 h) mean GH completely demarcated patients from normals. The increase in GH secretion is mostly amplitude dependent, with mild changes in the concentration pulse frequency. Overall, fasting seems to induce a feminized pattern of GH secretion with relatively higher interpeak levels, preserved but diminished diurnal variation, and increased secretory disorderliness (ApEn). The overall pulsatile pattern of GH secretion during fasting and the relative augmentation in GH release are not altered in the irradiated patients, providing further evidence that GH neuroregulation is fully preserved in these patients, even when the h-p axis is under physiological stress.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
First Published Online December 29, 2005
Abbreviations: ApEn, Approximate entropy; AUCGH, area under the curve for GH; BMI, body mass index; FFA, free fatty acid; 3-HOB, 3-hydroxybutyrate; h-p, hypothalamic-pituitary.
Received September 27, 2005.
Accepted December 16, 2005.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. L. Chan, C. J. Williams, P. Raciti, J. Blakeman, T. Kelesidis, I. Kelesidis, M. L. Johnson, M. O. Thorner, and C. S. Mantzoros Leptin Does Not Mediate Short-Term Fasting-Induced Changes in Growth Hormone Pulsatility but Increases IGF-I in Leptin Deficiency States J. Clin. Endocrinol. Metab., July 1, 2008; 93(7): 2819 - 2827. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. H. Darzy, S. S. Pezzoli, M. O. Thorner, and S. M. Shalet Cranial Irradiation and Growth Hormone Neurosecretory Dysfunction: A Critical Appraisal J. Clin. Endocrinol. Metab., May 1, 2007; 92(5): 1666 - 1672. [Abstract] [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 |
) in a severely GH-deficient patient (A) and a normal individual (B). Log scale is used to reveal pulsatility at very low concentrations. Note the relatively greater increase in daytime GH secretion during fasting leading to dampening of the diurnal rhythm and earlier acrophase.
