| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Articles |
Department of Intensive Care Medicine (G.V.d.B., F.W., P.W.) and Laboratory for Experimental Medicine and Endocrinology (R.B.), University Hospital Gasthuisberg, University of Leuven, B-3000 Leuven, Belgium; Kolling Institute of Medical Research, Royal North Shore Hospital, Sydney University (R.C.B.), St. Leonards, New South Wales 2065, Australia; Endocrine Section, Medical Service, Salem Veterans Affairs Medical Center (A.I.), Salem, Virginia 24153; and General Clinical Research Center, Department of Medicine, Division of Endocrinology, University of Virginia Health Sciences Center (J.D.V.), Charlottesville, Virginia 29908
Address all correspondence and requests for reprints to: Greet Van den Berghe, M.D., Ph.D., Department of Intensive Care Medicine, University Hospital Gasthuisberg, University of Leuven, B-3000 Leuven, Belgium. E-mail: greta.vandenberghe{at}med.kuleuven.ac.be
Abstract
Central hyposomatotropism and hypothyroidism have been inferred in long-stay intensive care patients. Pronounced hypoandrogenism presumably also contributes to the catabolic state of critical illness. Accordingly, the present study appraises the mechanism(s) of failure of the gonadotropic axis in prolonged critically ill men by assessing the effects of pulsatile GnRH treatment in this unique clinical context.
To this end, 15 critically ill men (mean ± SD age, 67 ± 12 yr; intensive care unit stay, 25 ± 9 days) participated, with baseline values compared with those of 50 age- and BMI-matched healthy men. Subjects were randomly allocated to 5 days of placebo or pulsatile iv GnRH administration (0.1 µg/kg every 90 min). LH, GH, and TSH secretion was quantified by deconvolution analysis of serum hormone concentration-time series obtained by sampling every 20 min from 2100–0600 h at baseline and on nights 1 and 5 of treatment. Serum concentrations of gonadal and adrenal steroids, T4, T3, insulin-like growth factor I (IGF), and IGF-binding proteins as well as circulating levels of cytokines and selected metabolic markers were measured.
During prolonged critical illness, pulsatile LH secretion and
mean LH concentrations (1.8 ± 2.2 vs. 6.0 ±
2.2 IU/L) were low in the face of extremely low circulating total
testosterone (0.27 ± 0.18 vs. 12.7 ± 4.07
nmol/L; P < 0.0001) and relatively low estradiol
(E2; 58.3 ± 51.9 vs. 85.7 ± 18.6
pmol/L; P = 0.009) and sex hormone-binding globulin
(39.1 ± 11.7 vs. 48.6 ± 27.8 nmol/L;
P = 0.01). The molar ratio of E2/T was
elevated 37-fold in ill men (P < 0.0001) and
correlated negatively with the mean serum LH concentrations (r =
-0.82; P = 0.0002). Pulsatile GH and TSH secretion
were suppressed (P 
0.0004), as were mean serum
IGF-I, IGF-binding protein-3, and acid-labile subunit concentrations;
thyroid hormone levels; and dehydroepiandrosterone sulfate. Morning
cortisol was within the normal range. Serum interleukin-1ß
concentrations were normal, whereas interleukin-6 and tumor necrosis
factor-
were elevated. Serum tumor necrosis factor- was
positively correlated with the molar E2/testosterone ratio
and with type 1 procollagen; the latter was elevated, whereas
osteocalcin was decreased. Ureagenesis and breakdown of bone were
increased. C-Reactive protein and white blood cell counts were
elevated; serum lactate levels were normal.
Intermittent iv GnRH administration increased pulsatile LH secretion compared with placebo by an increment of +8.1 ± 8.1 IU/L at 24 h (P = 0.001). This increase was only partially maintained after 5 days of treatment. GnRH pulses transiently increased serum testosterone by +174% on day 2 (P = 0.05), whereas all other endocrine parameters remained unaltered. GnRH tended to increase type 1 procollagen (P = 0.06), but did not change serum osteocalcin levels or bone breakdown. Ureagenesis was suppressed (P < 0.0001), and white blood cell count (P = 0.0001), C-reactive protein (P = 0.03), and lactate level (P = 0.01) were increased by GnRH compared with placebo infusions.
In conclusion, hypogonadotropic hypogonadism in prolonged critically ill men is only partially overcome with exogenous iv GnRH pulses, pointing to combined hypothalamic-pituitary-gonadal origins of the profound hypoandrogenism evident in this context. In view of concomitant central hyposomatotropism and hypothyroidism, evaluating the effectiveness of pulsatile GnRH intervention together with GH and TSH secretagogues will be important.
THE CATABOLIC STATE of critical illness is accompanied by a biphasic neuroendocrine response, consisting of acute and prolonged adaptations. The prolonged phase of critical illness is marked by impaired pulsatile secretion of GH and TSH, due, at least in part, to reduced hypothalamic stimulation (1, 2, 3, 4, 5, 6). The resultant hyposomatotropism and hypothyrotropism bring about low circulating levels of insulin-like growth factor I (IGF-I), GH-dependent IGF-binding proteins (IGFBPs), and thyroid hormones. This condition is different from the acute phase of illness, during which GH resistance and altered peripheral metabolism and binding of thyroid hormone predominate. Continuous infusion of GH and TSH secretagogues in prolonged critical illness (e.g. GH-releasing peptides, GHRH, and TRH) can reactivate blunted pulsatile GH and TSH secretion, followed by a feedback-controlled, and thus self-limited, increase in serum IGF-I, GH-dependent IGFBPs, and serum concentrations of T4 and T3 without altering rT3 (4). Restoring normal levels of these hormones in prolonged critical illness is followed by an immediately measurable anabolic response (4). Beyond the failure of the somatotropic and thyrotropic axes, hypoandrogenism has been documented consistently in critical illness and has been inferred to contribute to the catabolic state (5, 6, 7, 8, 9). A primary Leydig cell defect of unknown causes appears to predominate immediately after the onset of illness or trauma. In the chronic phase of critical illness, suppressed LH secretion presumably further reduces serum testosterone (T) concentrations in this condition (6).
The disturbances in LH secretion in prolonged critical illness have been shown to include reduced LH pulse amplitude, pointing to impaired compensatory hypersecretion of GnRH and LH in response to reduced T negative feedback (6). Suppressed LH pulse amplitude can be caused either by impaired pituitary responsiveness to GnRH or by attenuated hypothalamic GnRH pulses reaching the pituitary. Determining the effect of exogenous administration of GnRH to critically ill patients would help to address this pathophysiological distinction. Indeed, in analogy with previous findings of hypothalamic hyposomatotropism and hypothyroidism in prolonged critical illness, we postulated that hypogonadotropism in this condition reflects some measure of hypothalamic insufficiency. To test this hypothesis, we here investigate the effects of 5 days of pulsatile iv GnRH administration on LH secretion and sex steroid hormone production in prolonged critically ill men. Changes within the PRL, GH, TSH, and adrenocorticotropic axes are assessed concomitantly to monitor the specificity, and selected metabolic markers are evaluated to document the responsiveness of peripheral target tissues.
Subjects and Methods
Patients and concomitant treatment
Fifteen prolonged critically ill men were studied as part of a larger study of the effects of hypothalamic releasing factors in critical illness. At baseline, patients were matched to a group of healthy community-living volunteers and were subsequently randomized to receive 5 days of placebo or pulsatile iv GnRH administration.
Prolonged critically ill men were considered eligible for inclusion in the study when they were dependent on intensive medical care (including mechanical ventilatory support) for at least 2 weeks with an expected stay in the intensive care unit (ICU) of at least another 2 weeks (5). Patients were suffering from a complicated postoperative course after cardiovascular, pulmonary, or abdominal surgery or after polytrauma. Further inclusion criteria were a stable condition without dopamine treatment for at least 72 h in view of the pronounced suppressive effect that dopamine exerts on pituitary function in this setting (6, 10). Exclusion criteria were age less than 18 yr; preexisting neurologic, psychiatric, metabolic or endocrine disease; intracranial lesions; clinically significant liver failure (spontaneous prothrombin time, <30%); renal failure requiring dialysis or hemofiltration; concomitant treatment with steroids, somatostatin, thyroid hormones, Ca2+ entry blockers, clonidine, amiodarone, etomidate, dopamine agonists, or antagonists; and the use of iodine in antiseptic dressings or iv contrast agents.
Baseline demographic characteristics and comparisons with the matched
control group are shown in Tables 1–3
. The Apache II score on
admission (an indicator of severity of illness, with higher levels
reflecting more severe conditions) was 16 ± 3 (11).
ICU mortality was 40%, which was not different from the expected
mortality of 37% in an historical (1996–1998) matched control group
(12). In those six patients who did not survive, death
occurred a mean 37 ± 22 (range, 16–76) days after study entry.
Concomitant ICU treatment included standardized, continuously
administered feeding: total parenteral nutrition (n = 8), combined
parenteral and enteral nutrition (n = 2), or full enteral feeding
(n = 5). Caloric intake was 25 ± 6 nonprotein Cal/kg·day
(range, 18–39 Cal/kg·day), and composition was 0.8–1.6 g amino
acids/kg·day, 2.8–4.0 g glucose/kg·day, and 1–1.5 g fat/kg·day,
accounting for 25–40% of nonprotein calories (13). Other
concomitant therapies consisted of standardized vitamin supplements,
including 220 IU cholecalciferol/day (n = 15), inotropic support
with exogenous nondopaminergic catecholamines (n = 8), antibiotics
(n = 8), analgesia and sedation with continuously infused opioids
(n = 12), and/or benzodiazepines (n = 7). Blood glucose
levels were monitored every 4 h and were clamped between 5–9
mmol/L with continuous insulin infusion if necessary (n = 9)
(14). The mean ± SD dose of insulin in
those patients (n = 9) was 61 ± 37 (range, 24–152) U/day at
study entry. Continuous hemodynamic monitoring always included
electrocardiogram, intraarterial blood pressure, central venous
pressure, fluid balance, and core and peripheral temperatures. During
the study period, the ICU therapy and feeding regimen remained
virtually unaltered.
|
|
|
The study was approved by the institutional review board of the University of Leuven School of Medicine. Informed consent was obtained from the primary relative before patient inclusion.
Study design and peptide administration
At baseline, the study group was compared with an age-matched
sample of 50 community-living control men, as shown in Table 1
.
Patients were studied during a total time span of 6 days, beginning with overnight (2100–0600 h) sampling (every 20 min) on night 0 and again on nights 1 and 5. Additional blood samples were taken daily. Daily complete 24-h urine collections in HCl were obtained from all patients.
Patients were randomly allocated to receive 5 days of placebo (2 mL
0.9% NaCl /h) or pulsatile iv GnRH administration (0.1 µg/kg every
90 min; Lutrelef, Ferring Pharmaceuticals Ltd., Kiel,
Germany) via a separate lumen of the central venous line (Fig. 1), which had been inserted for clinical
purposes, independently of the study protocol.
|
(TNF), interleukin-1ß (IL-1ß), and IL-6 were
determined daily at 0600 h. Serum concentrations of urea and
creatinine and routine blood chemistry, coagulation profiles, serum
triglyceride levels, blood cell counts, and C-reactive protein level
were determined likewise. Urinary excretion of collagen cross-links [pyridinoline and deoxypyridinoline (PYD and DPD) normalized for urinary creatinine in an acidified (HCl) 24-h urine collection] was determined as previously reported (4).
Blood sampling
All blood samples were collected through an arterial line inserted for clinical purposes independently of the study protocol. The Edwards VAMP system (Baxter Healthcare Corp., Irvine, CA) was used, permitting withdrawal of undiluted blood samples from an indwelling catheter without undue blood loss. The total amount of blood sampled per patient over 6 days was 100 mL. Blood was collected into Vacutainer tubes. After clotting and centrifugation, sera were frozen initially at -20 C and subsequently at -70 C until assay.
Assays
All samples from each patient were processed in the same assay run. All samples had detectable values using the assays described below, with serum IGFBP-1 being a notable exception.
Serum LH concentrations were assayed by high sensitivity chemiluminescence-based assay, as previously described (15, 16). The LH standard was the Second International Reference Preparation. Assay sensitivity was 0.05 IU/L, with corresponding inter- and intraassay coefficients of variation (CVs) of 6.5% and 0.3%.
Total serum T concentrations were measured by RIA (Biosource Technologies, Inc.-Europe SA, Nivelles, Belgium). Within- and between-assay CVs were 4% and 8.3%, respectively. Total serum concentrations of E2 were determined by RIA (Estradiol-2, Clinical Assays, Cambridge, MA). Within- and between-assay CVs were 4.2% and 4.9%, respectively. Serum concentrations of sex hormone-binding globulin (SHBG) were determined using ammonium sulfate precipitation, as previously described (17). Within- and between-assay CVs were 3.7% and 3%, respectively. Unbound serum T and E2 were roughly estimated by the molar ratio of their total serum concentrations and the levels of SHBG. The serum concentrations of DHEAS were measured by RIA (kit from Diagnostics Systems Laboratories, Inc., Webster, TX). The intraassay CV was 5.1% at 0.6 µmol/L and 4.1% at 6.1 µmol/L. Normal values in men range between 2–10 µmol/L.
Serum concentrations of cortisol were assayed by RIA after extraction with dichloromethane. The intraassay CV was 3.1% at 417 nmol/L. Normal morning values are 200–700 nmol/L.
GH was measured by RIA, using the Nichols Institute Diagnostics HGH immunoassay 100T kit 40–2155(40–2155, San Juan Capistrano, CA). The intraassay coefficient of variation was 4.2% at 1.4 µg/L and 2.8% at 12.2 µg/L. Total IGF-I was quantified by RIA after acid-ethanol extraction (12). Serum concentrations of IGFBP-3, ALS, IGFBP-1 (18, 19, 20), and IGFBP-2 (21) were measured by RIA. Serum insulin levels were determined by immunoradiometric assay (Medgenix INS-IRMA, BioSource, Fleurus, Belgium). The intraassay CV was 4.5% at 6.6 µIU/mL and 2.1% at 53 µIU/mL.
Serum concentrations of TSH were measured by a third generation sandwich immunoassay using the Immuno 1 System (Bayer Corp., Tarrytown, NY). The within-run and total CVs were 2.4% and 2.5%, respectively. The detection limit was 0.014 mIU/L.
The serum concentrations of T4 and T3 were measured by homogeneous latex agglutination using the Technicon Immuno 1 System (Bayer Corp.). The within-assay CVs were 3.1% and 3.9% for T4 and T3, respectively. The between-assay CVs were 3.6% and 6% for T4 and T3, respectively. The detection limit was 5.2 nmol/L for T4 and 0.09 nmol/L for T3.
The serum concentrations of rT3 were measured by RIA using the rT3 Kit (Techland SA, Liege, Belgium). The intraassay CV was 10.5% at 0.8 nmol/L and 10.3% at 4 nmol/L.
Serum OC was measured by a homologous human osteocalcin RIA, as previously described (4, 12, 22). The within- and between-assay CVs were 5% and 7%, respectively. Serum levels of PICP were measured by RIA (Orion Diagnostica, Espoo, Finland). The between-assay CV was 6.6% at 216 µg/L and 4.0% at 435 µg/L. The within-assay CV was 2.7% at 214 µg/L and 3.2% at 451 µg/L. Serum skeletal ALP was determined by immunoradiometric assay (Tandem-R Ostase, Hybritech, Inc., San Diego, CA). The intraassay CV was 6.7% at 13.2 µg/L, and the interassay CV was 8.1% at 11.7 µg/L. Urinary pyridinolines (PYD and DPD) were measured in 24-h urine collections acidified with hydrochloric acid, as previously described (4, 12, 23). Values are expressed as nanomoles per mmol creatinine. Urinary creatinine was measured colorimetrically. Interassay CVs were 11.5% and 13.3% for PYD and DPD, respectively (n = 12), and within-assay CVs were 10.2% and 12.5% (n = 9). Serum 25OHD was measured by competitive binding assay as previously reported (24). Serum intact PTH was measured by a two-step immunochemical method, involving an amino-terminal capture and a midregional detecting antibody, as described previously (25).
Serum concentrations of TNF were measured by a solid phase Enzyme
Amplified Sensitivity Immunoassay (EASIA; Biosource,
Fleurns, Belgium). The normal mean is 6 ± 4 pg/mL, ranging from
0–20 pg/mL. The within-assay CV was 5.2%, and the between-assay CV
was 8% (n = 20).
Serum concentrations of IL-1ß were measured by EASIA (Biosource). The normal mean is 5 ± 8 pg/mL, ranging from 0–15 pg/mL. The within-assay CV was 3.4%, and the between-assay CV was 4.4% (n = 20).
Serum concentrations of IL-6 were measured by EASIA (Biosource). Normal values are usually undetectable, but range up to 72 pg/mL. At 205 pg/mL, the within-assay CV was 4.7%, and the between-assay CV was 7.5% (n = 22).
The urea concentration was measured in serum on a routine clinical basis using a kinetic UV test (Roche Molecular Biochemicals, Mannheim, Germany; Hitachi 737). Imprecision averaged 3.15% at 50 mg/dL and 1.76% at 158 mg/dL. Serum creatinine was measured using the Jaffé method (Roche Molecular Biochemicals, Hitachi 737). The total inaccuracy was 3% at 1.9 mg/dL and 1.5% at 4.1 mg/dL.
Data analysis
The time series of serum LH, GH, and TSH concentrations sampled repetitively for 9 h overnight was transformed into pituitary secretion profiles by adjusting for each endogenous hormone half-life by multiple parameter deconvolution analysis (26). This method is designed to compute hormonal half-life and the number, amplitude, and mass of underlying pituitary secretory bursts and to estimate tonic (basal or nonpulsatile) secretion, if any. The following parameters were calculated: half-life, basal secretion rate (estimated as the amount of hormone continuously released from the pituitary to achieve serum concentrations approximating the mean of the lowest 5% of all values observed [micrograms per Lv (liter of distribution volume) and per min], amplitudes (maximal secretory rate (micrograms per Lv and per min) and temporal positions of all secretory bursts, the mass of hormone secreted per burst [calculated as the area of the resolved secretion burst (micrograms per Lv)], and mean pulsatile production [estimated as the product of the number of secretory bursts and the mean secretory burst mass over the time interval considered (micrograms per Lv over 9 h)]. Total nocturnal production (micrograms per Lv) was calculated as the sum of pulsatile production (micrograms per Lv) over 9 h and the nonpulsatile secretion over 9 h [basal secretion rate (micrograms per L/min) x 540 min].
Irregularity of LH, GH, and TSH secretory profiles was determined using the approximate entropy statistic (ApEn). ApEn is a model-independent statistic for assessing the regularity of time series that is complementary to deconvolution analysis (27). This metric assigns a single nonnegative number to a time series, with larger values corresponding to greater apparent process randomness. ApEn has been demonstrated to be stable to small changes in noise characteristics and infrequent and significant artifacts, as previously validated in the present setting (1, 2, 3, 4, 12).
We used changes in the serum urea/creatinine ratio to monitor alterations in protein breakdown, mainly from muscle (28), which is appropriate in the absence of prerenal kidney failure and with normal and constant protein and fluid intake, as present in this study (4, 12, 29).
At inclusion, groups were compared using a two-tailed unpaired
t test, Mann-Whitney U test, and 
2
test. Comparison with literature reference values for pulsatile
fraction of GH and TSH release (30, 31) was performed
using the two-tailed, one-sample t test. The effect of
intervention was analyzed using ANOVA, unpaired Student’s t
test, Mann-Whitney U test, or Wilcoxon signed rank test, as
appropriate. Results are expressed as the mean ±
SD unless indicated otherwise. P
0.05 was construed as significant.
Results
Prolonged critically ill men were matched with controls for age
and body mass index (Table 1). Compared with healthy controls, patients
exhibited lower serum LH, total T, total E2,
SHBG, IGF-I, IGFBP-3, ALS, T4, and
T3 concentrations; higher IGFBP-2, TNF, and
IL-6 levels; and an extremely elevated E2/T molar
ratio (Table 1). In addition, patients had higher urea/creatinine
ratios, elevated urinary excretion of collagen cross-links (Pyd and
DPD), higher serum PICP and sALP levels, and lower 25OHD and OC
concentrations than controls.
Patients randomized to receive placebo and GnRH (Table 2) were comparable at baseline with
regard to demographic, clinical, and endocrine characteristics.
LH secretion and sex steroids
In healthy elderly men, the serum LH concentration was 6.04
± 3.98 IU/L (10th percentile, 2.46 IU/L; 90th percentile =
10.23 IU/L), and the fraction of LH released in a pulsatile fashion is
known to exceed 85%, with four to nine pulses per 9 h
(32). In the prolonged critically ill men, serum LH
concentrations measured in a single sample were low compared with
values in matched controls (Table 1). Furthermore, the baseline LH
secretory pattern was characterized by a relatively normal pulse
frequency (five to eight pulses per 9 h), but a severely reduced
amount of LH secreted per pulse (Table 3
and Fig. 1) (32), resulting in low mean nocturnal LH serum
concentrations in the face of extremely low circulating total T and
moderately low E2 and SHBG concentrations (Tables 1 and 3). This resulted in dramatically suppressed estimated free T
levels, whereas those of E2 were normal (Table 1). The molar ratio of E2/T, an index of
aromatase activity and sex steroid metabolism (33), was
37-fold increased in the critically ill men compared with that in
healthy controls, and the log (molar ratio E2/T)
correlated inversely with the mean nocturnal LH (r = -0.82;
P = 0.0002), LH pulse amplitude (r = -0.76;
P = 0.0009), and LH pulse mass (r = -0.84;
P < 0.0001). These correlations were entirely
attributable to the effect of E2, whereas serum T
in itself appeared unrelated to LH secretion. LH secretion was also
unrelated to any of the other studied hormones, including cortisol, and
was unrelated to the exogenous administration of opioids or
benzodiazepines.
Administration of iv GnRH at 90-min intervals increased mean LH
concentrations (a mean increment of +8.1 ± 8.1 IU/L; a median
10-fold rise; Figs. 1 and 2 and Table 3)
and pulsatile LH secretion (increased LH pulse amplitude and mass
without an altered pulse frequency) 8- to 10-fold after 24 h
(Table 3). This stimulatory response
waned, but remained detectable after 5 days of treatment. The
irregularity within the LH secretory pattern, as assessed by the ApEn
score, remained unaltered in both groups (Table 3).
|
). The molar
ratio of E2/T decreased on day 2 of GnRH
intervention compared with that in the placebo group (P
= 0.004), but returned to baseline after 5 days. SHBG levels, low at
study entry, remained unaltered (Fig. 2). GH and TSH axes, insulin, DHEAS, and cortisol
Single sample serum concentrations of GH and TSH were normal
(Table 1), but pulsatility of the GH and TSH secretory patterns at
baseline was reduced compared with that in healthy controls [pulsatile
GH fraction of 60 ± 23% vs. normal of >90%
(P = 0.0002) and pulsatile TSH fraction of 41 ±
19% vs. normal of >65% (P = 0.0004)]
(30, 31), as previously reported (1, 2, 3, 4, 12).
Reduced pulsatile GH and TSH secretion coincided with low IGF-I,
IGFBP-3, ALS, T4, and T3
levels and elevated IGFBP-2 and rT3
concentrations (Table 1). GnRH infusions did not alter GH or TSH
secretory patterns or the low circulating levels of IGF-I, IGFBPs,
thyroid hormones, or baseline insulin concentrations (41 ± 32
µIU/mL; range, 5–113 µIU/mL; data not shown).
Baseline DHEAS levels were reduced by 30-fold in the sick men (0.9 ± 0.8 µmol/L; range, 0.03–2.7 µmol/L; normal range, 2–10 µmol/L) and remained low in both study groups (data not shown).
Morning serum cortisol levels were normal at baseline (444 ± 157 nmol/L; range, 270–781 nmol/L; normal range, 200–700 nmol/L for a morning sample) and during the 6 study days (data not shown).
TNF, IL-1ß, and IL-6
Baseline serum TNF and IL-6 concentrations were elevated,
whereas IL-1ß levels were normal (Table 1). Serum concentrations of
TNF correlated positively with the E2/T molar
ratio (r = 0.68; P = 0.0049) and with PICP (r
= 0.76; P = 0.001). IL-1ß and IL-6 levels were
unrelated to the studied endocrine/metabolic markers. Circulating
levels of these proinflammatory cytokines were not significantly
altered by GnRH administration (data not shown).
Metabolic markers and clinical chemistry/hematology
The serum urea/creatinine ratio was substantially elevated at
baseline (Table 1), but decreased significantly with 5 days of GnRH
administration compared with that in the placebo group (Fig. 3). Concomitantly, serum lactate,
C-reactive protein levels, and white blood cell count increased in
the treatment group (Fig. 3).
|
(r = 0.76;
P = 0.001). There was a trend toward a further increase
in PICP (P = 0.06; Fig. 4) with GnRH administration, whereas
serum concentrations of sALP and OC and the urinary excretion of
collagen cross-links (PYD and DPD) remained unaltered (Fig. 4). Serum
concentrations of PTH remained constant (change in PTH, -2.4 ±
9.2 pg/mL with GnRH vs. -4.8 ± 8.5 pg/mL with
placebo; P = 0.6), whereas serum 25OHD levels slightly
decreased during GnRH infusion (change in 25OHD, -1.6 ± 2.0
µg/L with GnRH vs. +0.5 ± 0.1 µg/L with
placebo; P = 0.03). All other studied variables
remained unchanged.
|
Impoverished LH secretion associated with blunted pulsatility is a consistent feature in prolonged critically ill men along with reduced serum concentrations of androgens, consistent with a clinical inference of acquired hypogonadotropic hypogonadism (6, 8, 9). The pathophysiology of this syndrome in extended life-threatening illness is not yet known. The present interventional study using pulsatile iv GnRH administration for 5 days unmasks a combined hypothalamic-pituitary origin of hypogonadotropism in prolonged critically ill men, as documented by an acute 8.1 IU/L rise in mean LH (P = 0.001) levels at 24 h and a 174% increase in mean T (P = 0.05) concentrations at 48 h, followed by a waning of both responses after 3–5 days of continued iv GnRH drive. Normal estimated free E2 levels in the presence of dramatically low serum concentrations of its precursors T and DHEA indicate increased aromatase activity, which at the pituitary level may contribute to the impairment of pulsatile LH secretion (27) and to desensitization of the gonadotropes despite sustained exogenous pulsatile GnRH stimulation.
In addition to evident hyposomatotropism and central hypothyroidism, as
previously described in prolonged critical illness (1, 2, 3, 4, 5, 6),
critically ill men manifested with reduced pulsatile LH secretion,
comprising relatively normal pulse frequency, but severely reduced
pulse amplitude. In addition, very low serum T (1/40th of normal) and
moderately suppressed serum E2 (two thirds of
normal) and SHBG (two thirds of normal) were present in these sick men.
This constellation is compatible with dramatically reduced bioavailable
T, whereas the free fraction of E2 appears
normal. The molar ratio of E2/T, an aromatization
index (33) that reflects glandular and/or extraglandular
net aromatase activity, was extremely elevated and inversely correlated
with LH secretion. The site and mechanisms of this dramatic activation
of aromatase in critical illness are unknown, but fat storage with
feeding and the effects of catecholamines, endotoxin, hyperinsulinism,
hypercortisolism, and hyperprolactinemia could play a role
(34, 35, 36, 37, 38, 39). The observed positive correlation between TNF
and the molar ratio of E2/T points to a role for
this proinflammatory cytokine.
In the healthy human male, both T and E2 exert dual negative feedback actions, viz. at the hypothalamus to decrease GnRH pulse frequency and amplitude and at the pituitary to decrease responsiveness to GnRH (27, 40, 41, 42, 43, 44, 45). The direct inhibitory effect of T on gonadotropes in men, but possibly also its hypothalamic effect on the GnRH pulse amplitude, appear to be mediated by the aromatization of T to E2 (27, 40, 46). Furthermore, on a molar basis, E2 is 200-fold more effective than T in suppressing GnRH-driven gonadotropin secretion (47). Hence, the lack of a compensatory rise in LH secretion given low T feedback might be explained by preserved bioactive E2 levels at the hypothalamus/pituitary level. Analogous, but less striking, findings have been reported in aging healthy men and patients undergoing hemodialysis (48, 49, 50, 51). Normal LH pulse frequency in critically ill men indicates that the GnRH pulse generator is active, consistent with data in pharmacological states of hyperestrogenism in the male (52). The reduced LH pulse amplitude suggests that either a reduced amount of GnRH is being released from GnRH neurons or pituitary responsiveness to GnRH pulses is impaired. These two possibilities can be distinguished at least in part by the LH responses observed after the administration of GnRH pulses using a GnRH dose (0.1 µg/kg) and administration interval (90 min) that have been shown to normalize hypothalamic hypogonadotropism such as occurs in starvation (53) and healthy aging (49). Pulsatile LH secretion in the critically ill men was clearly reactivated after 24 h of GnRH pulses, to an extent that is more or less comparable to that observed in healthy age-matched men (49, 54). However, the degree of rise in LH secretion was inferentially limited in the face of low T feedback inhibition and was variable among subjects, suggesting variably reduced gonadotrope sensitivity to GnRH. Previous studies suggest, but do not prove, that putatively impaired GnRH actions could reflect some increase in hypothalamic dopaminergic tone (6) and/or inhibition by endogenous and/or exogenous opioids (55, 56, 57). However, no patients were receiving dopamine infusions in the present study, and statistical analysis revealed no evident effect of analgesia with opiates in this population of sick men. Cytokines also are known to suppress the GnRH-LH-testosterone axis, directly or indirectly through increased CRH and/or cortisol (58, 59). The latter seems an unlikely mechanism in the critically ill men studied here, as morning cortisol concentrations were no longer elevated in this chronic phase of illness and did not correlate to LH concentrations. Relative adrenal insufficiency has been described in prolonged critically ill patients (5), which would explain the absent correlation between circulating cytokines and the hypothalamic-pituitary-adrenal axis. Thus, we postulate that increased aromatase activity may have induced bioactive E2 concentrations at the hypothalamus/pituitary level, which may have contributed to reduced basal and GnRH-stimulated LH release in this milieu. Such an E2 effect also occurs in feminizing adrenal and Leydig cell tumors (52, 60, 61), in the aromatase excess syndrome (62), and during iv infusion of E2 (51, 63).
The clear-cut stimulation of pulsatile LH secretion in response to GnRH administration for 24 h evoked a 174% rise in serum T after 48 h without inducing a detectable change in E2, SHBG, or DHEAS concentrations, indicating at least partial Leydig cell responsiveness. The effects of the experimentally controlled GnRH regimen on LH secretion and serum T faded after 5 days of treatment. The latter waning of responsiveness coincided with recovery of a transient drop in the E2/T molar ratio. As the exogenous GnRH drive was constant, this attenuation conceivably has a feedback-dependent pituitary origin. One explanation could be desensitization of the gonadotropes to recurrent GnRH pulses. However, as the same experimental regimen does not desensitize LH secretion when administered for 1–14 days to healthy starved young men (53) and healthy young or older men (49), this consideration appears unlikely. A second pathway could be feedback inhibition of gonadotropes exerted by the rise in T. As T-induced feedback inhibition at the pituitary level is mediated in part by aromatization to E2 (46), the rise in circulating T may have increased gonadotrope E2 levels via in situ aromatization. Alternatively, the set-point for feedback inhibition could be altered in critical illness, as has been suggested in healthy aging men (44, 64). Whereas in principle T precursors, such as progesterone and 17-hydroxyprogesterone, or a metabolite, e.g. dihydrotestosterone, also might have exerted undue negative feedback, dihydrotestosterone does not seem to play a major role in LH feedback inhibition except at supraphysiological concentrations (45) (65). Finally, SHBG levels were low, and sex steroid binding to SHBG could be hampered in this metabolic milieu (e.g. by high circulating levels of nonesterified fatty acids), potentially relatively elevating free T or E2 levels not otherwise evident by measuring total levels of these sex steroids.
The observed rise in total T was moderate and was not into the normal range. This indicates that a primary Leydig cell failure could also be present, as previously suggested (8, 9), or that LH bioactivity is reduced in this context, as inferred in end-stage renal failure (66). It remains unclear how much of this inferred Leydig cell defect is due to cytokines (67) or to concomitant endocrine failure (4), such as hyposomatotropism and/or deficiency of IGF-I and its binding proteins (68, 69).
Despite the fact that the changes in LH secretion and sex steroids during the 5 days of GnRH administration were transient, biological effects emerged at the peripheral tissue level. Firstly, excess urea generation was substantially reduced with GnRH administration, which denotes decreased protein breakdown mainly in skeletal muscle (28). As GH, IGF-I and its binding proteins, cortisol, and thyroid hormones were unaltered by GnRH, and GnRH is not an anabolic agent by itself, this is conceivably a direct T-mediated effect (70, 71). The anabolic response could also indicate that the relatively small rise in total T obscured a more pronounced rise in free or bioavailable T. Alternatively, up-regulation of androgen receptors could have occurred.
Secondly, bone turnover measured at baseline was massively increased,
as both markers of bone formation (PICP) and resorption (DPD) were
5-fold elevated. This increase is more pronounced than in Paget’s
disease or primary hyperparathyroidism, as reported previously
(4, 12). The discrepancy between increased PICP and ALP
and decreased OC synthesis suggests a maturation defect that could
conceivably hinder new bone acquisition (72). The
stimulation of only immature osteoblasts at baseline, which tended to
be accentuated by GnRH treatment, may reflect a cytokine and/or gonadal
steroid effect (23), enhanced by IGF-I deficiency
(4). Here we observed that this maturation defect is
unaltered by GnRH treatment. Increased osteoclast function is probably
the result of an increased production of osteoclast-stimulating
cytokines (TNF and IL-6) (73) and loss of androgen
action. As these parameters were not (cytokines) or only transiently
and incompletely (androgens) corrected during GnRH administration, no
treatment effects were observed on bone resorption parameters
(DPD).
Finally, GnRH treatment activated markers of the immune/inflammatory cascade, as indicated by a rise in white blood cell count, C-reactive protein, and serum lactate. As immunostimulatory effects of sex steroids are largely attributable to E2 (74), the inferentially accelerated aromatization of T to E2 could play a role. In relation to the increase in serum lactate concentration, we postulate increased glycolysis in phagocytic cells (75), greater relative tissue hypoperfusion/cellular hypoxia, or reduced lactate metabolism. Whether these measures are potentially harmful or contribute to the so-called systemic inflammatory response syndrome and, hence, multiple organ failure (76) is not known.
In conclusion, hypogonadotropic hypogonadism in prolonged critically ill men is partially (but incompletely) overcome with exogenous iv GnRH pulses, thus pointing to combined hypothalamic-pituitary-gonadal origins of the profound hypoandrogenism evident in this context. Undue pituitary desensitization to GnRH or increased feedback inhibition could be involved. Peripheral tissues were sensitive to transient changes in sex steroids, as reflected by anabolic and inflammatory responses. In view of concomitant neuroendocrine disturbances in the somatotropic and thyrotropic axes during prolonged critical illness, evaluating the effectiveness of pulsatile GnRH intervention together with GH and TSH secretagogues will be important.
Acknowledgments
We thank the medical and nursing staff of the intensive care unit of Leuven University Hospital for their cooperation, in particular Drs. Lauwers, Ferdinande, Schetz, Verwaest, Vlasselaers, Ingels, Desmet, and Muller for the clinical patient management. Dr. J. Billen is acknowledged for the TSH and thyroid hormone determinations. We thank Paula Veldhuis, Sri Meka, Kevin Hardman, Viviane Celis, Tina Schreurs, Christiane Eyletten, Marianne Aerts, Ivo Jans, Willy Coopmans, Ilse Milants, and Drs. Ramael, Schoonheydt, and Fransen for expert technical assistance. Dr. Erik Mehuys (Ferring Pharmaceuticals Ltd., Aalst, Belgium) is acknowledged for generously providing Lutrelef, and Mr. Jean Hellers (Baxter, Brussels, Belgium) is thanked for the Vamp-devices.
Footnotes
1 This work was supported by NIH Grant ROI-AG-14799 and the
General Clinical Research Center of the University of Virginia (to
J.D.V.), the Fund for Scientific Research Flanders Belgium Grants
G.0242.01 (to R.B.) and G.0144.00 and G.3C05.95N (to G.V.d.B.), the
Research Council of the University of Leuven Grants OT 95/24 and OT
99/32 (to G.V.d.B.), and a Voorzorgskas Voor Geneesheren (VKG)
1998–1999 clinical investigator award (to F.W.). Presented in part at
the 82nd Annual Meeting of The Endocrine Society, Toronto, Canada,
2000. 
Received January 9, 2001.
Revised March 21, 2001.
Accepted March 30, 2001.
References
2-adrenergic agonist. J Clin Endocrinol
Metab. 64:755–762.This article has been cited by other articles:
 |
|
 |
|
  D. I. Spratt, R. S. Kramer, J. R. Morton, F. L. Lucas, K. Becker, and C. Longcope Characterization of a prospective human model for study of the reproductive hormone responses to major illness Am J Physiol Endocrinol Metab, July 1, 2008; 295(1): E63 - E69. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. I. Spratt, J. R. Morton, R. S. Kramer, S. W. Mayo, C. Longcope, and C. P. H. Vary Increases in serum estrogen levels during major illness are caused by increased peripheral aromatization Am J Physiol Endocrinol Metab, September 1, 2006; 291(3): E631 - E638. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Van Vliet, M. A. Spruit, G. Verleden, A. Kasran, E. Van Herck, F. Pitta, R. Bouillon, and M. Decramer Hypogonadism, Quadriceps Weakness, and Exercise Intolerance in Chronic Obstructive Pulmonary Disease Am. J. Respir. Crit. Care Med., November 1, 2005; 172(9): 1105 - 1111. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Kaufman and A. Vermeulen The Decline of Androgen Levels in Elderly Men and Its Clinical and Therapeutic Implications Endocr. Rev., October 1, 2005; 26(6): 833 - 876. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Johnsen, S. E. Dolan, K. V. Fitch, K. M. Killilea, J. L. Shifren, and S. K. Grinspoon Absence of Polycystic Ovary Syndrome Features in Human Immunodeficiency Virus-Infected Women Despite Significant Hyperinsulinemia and Truncal Adiposity J. Clin. Endocrinol. Metab., October 1, 2005; 90(10): 5596 - 5604. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Van den Berghe, D. Van Roosbroeck, P. Vanhove, P. J. Wouters, L. De Pourcq, and R. Bouillon Bone Turnover in Prolonged Critical Illness: Effect of Vitamin D J. Clin. Endocrinol. Metab., October 1, 2003; 88(10): 4623 - 4632. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Mesotten, P. J. D. Delhanty, F. Vanderhoydonc, K. V. Hardman, F. Weekers, R. C. Baxter, and G. van den Berghe Regulation of Insulin-Like Growth Factor Binding Protein-1 during Protracted Critical Illness J. Clin. Endocrinol. Metab., December 1, 2002; 87(12): 5516 - 5523. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Lee, R. Miselis, and C. Rivier Anatomical and Functional Evidence for a Neural Hypothalamic-Testicular Pathway that Is Independent of the Pituitary Endocrinology, November 1, 2002; 143(11): 4447 - 4454. [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 |
) or pulsatile iv GnRH
administration (0.1 µg/kg every 90 min; 
). The infusions were
started after the 0600 h sample of the baseline night profile and
continued until 0600 h of night 5. Pulsatile LH secretion rose
substantially after 24 h of GnRH administration compared with
placebo. This effect waned, but remained detectable after 5 days.
