Abstract. In 10 hyperthyroid women studied in the follicular phase of the menstrual cycle, basal plasma PRL was normal, but PRL release after TRH was significantly suppressed compared with that in 11 control women. The suppressed PRL response to TRH was not explained by changes in serum estradiol or sex hormone-binding globulin. It recovered after treatment of hyperthyroidism. When normal women were treated with T4 (0.5 mg daily for 6 to 10 days), their mean serum free T4 level increased to about 70% of that in the hyperthyroid patients, whereas their serum free T3 levels increased to a lesser degree. During T4 administration, these women had PRL changes similar to those of the hyperthyroid patients. When the normal women took T3 (60–120) μg for 6 to 8 days), their serum free T3 increased to almost the level of the hyperthyroid patients, but the TRH stimulated PRL release remained close to the control level. The PRL increase after dopaminergic blockade with metoclopramide was significantly suppressed in hyperthyroid patients, and they had no PRL response to TRH after pretreatment with metoclopramide. In conclusion, the PRL changes in hyperthyroidism were reproduced by administration of T4, but not by administration of T3 to healthy women. The site of action is suggested to be pituitary, but additional hypothalamic effects cannot be excluded.
Eva Marie T Erfurth and Lars E Hagmar
Erfurth EMT, Hagmar LE. Decreased serum testosterone and free triidothyronine levels in healthy middle-aged men indicate an age effect at the pituitary level. Eur J Endocrinol 1995;132:663–7. ISSN 0804–4643
In an attempt to study further the age-specific influence on the hypothalamo-pituitary–gonadal axis as well as the hypothalamo–pituitary–thyroid axis, we have now investigated young and middle-aged men, considering possible confounding factors. Both serum total testosterone, free testosterone and the total ratio of testosterone to sex-hormone binding globulin were significantly lower among middle-aged men as compared with young men (p = 0.02, p = 0.002 and p = 0.0003, respectively). In accordance with these findings there was also a decrease in the luteinizing hormone response to gonadotrophin-releasing hormone in the middle-aged men (p= 0.02). Free testosterone was correlated significantly with the luteinizing hormone response (r = 0.32, p = 0.02). Serum free triiodothyronine was significantly higher among young men as compared with middle-aged men (p = 0.002) and the thyrotrophin-releasing hormone-stimulated thyrotrophin response was also higher in the young group compared with the middle-aged group. The present results may indicate that the age effect on serum levels of testosterone and free triidothyronine is mediated at the pituitary level.
Eva Marie Erfurth, Dept. of Internal Medicine, University of Lund, S-221 85 Lund, Sweden
Ellen Vinge, Eva Marie T Erfurth, and Stefan Lundin
In order to study the influence of the hypothalamic-pituitary-adrenal axis on the levels of endogenous digitalis-like substances (EDLS) in plasma and urine, eight healthy subjects (25–40 years old) were given dexamethasone 1 mg orally and tetracosactide (an ACTH analog) 0.25 mg iv, on separate occasions. The circulating levels of EDLS, TSH, PRL and AVP following administration of either test drug, and under control conditions, were measured by a RIA for digoxin and specific RIAs for each hormone. Plasma cortisol was measured by liquid chromatography. The area under the curve (AUC) of hormone levels between 08.00 and 09.30 was used for data comparisons. Urine was collected before and after each test dose, and analysed for cortisol levels by gas chromatography/mass spectrometry, and for digitalis-like activity both by RIA and by a bioassay measuring 86Rb-uptake into red blood cells. Dexamethasone suppressed the AUC of plasma and urine levels of cortisol (p=0.0001 and p<0.01, respectively) and immunoreactive EDLS (p=0.0007 and p<0.01), as well as serum levels of TSH (p=0.0002) and PRL (p=0.001), but did not alter AVP levels. The biological digitalis-like activity in the urine measured by the 86Rb-uptake assay was decreased, but not to a statistically significant degree. ACTH increased the levels of cortisol in plasma (p=0.0001) and urine (p<0.01) and the immunoreactive EDLS in plasma (p = 0.03), but not in urine. There were no effects of ACTH on TSH, PRL or AVP. There are alternative explanations for the discrepancy between the effects on EDLS levels in plasma and urine: methodological difficulties in quantitating EDLS, the doses of dexamethasone and ACTH used for the adrenal function tests, and that other hypothalamic or pituitary factors than ACTH may contribute to a significant degree in the regulation of EDLS levels. Taken together, the results of the present study support the hypothesis that EDLS is of adrenal origin, rather than hypothalamic or pituitary.
Eva Marie T. Erfurth, Robyn Attewell, and L. Pavo Hedner
Fifteen premenopausal women were investigated in the follicular phase of the menstrual cycle with two TRH tests within an interval of 48 to 96 h. Ninety min before each TRH test either 10 mg metoclopramide or saline was injected iv in randomized order. The same procedure was repeated in the following menstrual cycle after pretreatment with T4 (0.5 mg daily for 6-14 days). At least 2 months later the same procedure was repeated with T3 pretreatment (60-120 μg for 6-8 days) in 9 of them. A multiple regression analysis was used in modelling the relationships between TSH release and serum free T3, T4 and estradiol levels, adjusting for the presence of metoclopramide and the order of the test. No correlation was found between the TSH response to TRH and the serum estradiol level. The TSH response to the second TRH test was approximately half the first one, both in the control situation and after treatment with T4 or T3. The blunting of the TSH response to the second TRH administration was significantly reduced by metoclopramide, both at normal and elevated thyroid hormone levels, suggesting that a dopaminergic mechanism takes part in the blunting.
Helene Holmer, Vera Popovic, Bertil Ekman, Cecilia Follin, Ann Britt Siversson, and Eva Marie Erfurth
Data on bone mineral density (BMD) are lacking in adults with childhood onset (CO)–craniopharyngioma (CP) with hypothalamic damage from the tumor. In patients with CO GH deficiency, BMD increases during GH treatment.
The aims were to evaluate BMD in adults with CO–CPs on complete hormone replacement, including long-term GH and to evaluate the impact of hypothalamic damage on these measures.
Design and participants
BMD (dual-energy X-ray absorptiometry), markers of bone turn over, physical activity and calcium intake were assessed in 39 CO–CP adults (20 women), with a median age of 28 (17–57) years, in comparison with matched population controls.
Late puberty induction was recorded in both genders, but reduced androgen levels in females only. Only CP women had lower BMD (P=0.03) at L2–L4, and reduced Z-scores at femoral neck (P=0.004) and L2–L4 (P=0.004). Both genders had increased serum leptin levels (P=0.001), which significantly correlated negatively with BMD at L2–L4 (P=0.003; r=−0.5) and 45% of CP women had Z-score levels ≤−2.0 s.d. Furthermore, 75% of those with a Z-score ≤−2.0 s.d. had hypothalamic involvement by the tumor. Calcium intake (P=0.008) and physical activity (P=0.007) levels were reduced in CP men only. Levels of ostecalcin and crossLaps were increased in CP men only.
Despite continuous GH therapy, low BMD was recorded in CO–CP females. Insufficient estrogen and androgen supplementation during adolescence was the main cause, but hypothalamic involvement with consequent leptin resistance was also strongly associated with low BMD in both genders.
Helene Holmer, Bertil Ekman, Jonas Björk, Carl-Henrik Nordstöm, Vera Popovic, AnnBritt Siversson, and Eva-Marie Erfurth
Craniopharyngioma patients without GH therapy are at an increased cardiovascular disease (CVD) risk and particularly concerning women. No previous study on long-term GH therapy in adults with childhood onset (CO) craniopharyngioma was identified.
To investigate CVD risk in adults with CO craniopharyngioma on complete hormone replacement, including long-term GH therapy, and to investigate the impact of disease-related factors on CVD risk.
Design and participants
In a cross-sectional study of operated CO craniopharyngiomas (1958–2000) from a defined area of Sweden (2.5 million), we enrolled 42 patients (20 women) with a median age of 28 years (range 17–57) and assessed CVD risk of 20 (4–40) years after first operation. Comparisons were made with matched controls and between patients with tumor growth into the third ventricle (TGTV) versus non-TGTV. GH therapy was 10–12 years in women and men.
In comparison with controls, both male and female patients had increased body mass index, fat mass, insulin, and leptin levels. Overall, while not significantly increased in male patients, 55–60% of female patients had a medium–high CVD risk, compared with 10–20% in controls. An increased CVD risk (all P<0.05) and higher levels of fat mass and insulin were recorded in the TGTV group versus the non-TGTV group. Late puberty induction and lack of androgens were shown in female patients.
Adult patients with CO craniopharyngioma, especially those with TGTV, have persistently increased CVD risk. Conventional hormone substitution, including GH, is insufficient to normalize CVD risk, suggesting an important role for irreversible hypothalamic dysfunction.
Axel Netterlid, Helena Mörse, Aleksander Giwercman, Emir Henic, Kristina E Akesson, Eva Marie Erfurth, and Maria Elfving
Objective: Female childhood cancer survivors (CCS) are at risk of several late effects, among them metabolic syndrome (MetS) and premature ovarian insufficiency (POI). The objective is to study if POI is associated with risk of MetS and increased cardiovascular risk in CSS.
Design: A cross-sectional study with a median time since cancer diagnosis of 25 (12 – 41) years. Patients and controls were recruited from the South Medical Region of Sweden.
Methods: The study included 167 female CCS, median age 34 (19 – 57) years, diagnosed with childhood cancer at median age 8.4 (0.1 – 17.9) years together with 164 controls, matched for age, sex, ethnicity, residence, and smoking habits. All subjects were examined with fasting glucose, insulin, HbA1c, and lipid profile. Fat mass was calculated with dual-energy X-ray absorptiometry (DXA), questionnaires for medication were obtained. Detailed information of cancer treatment was available.
Results: POI was present in 13% (22/167) among CCS (hypothalamic/pituitary cause excluded) and in none among controls. MetS was present in 14% (24/167) among all CCS (p = 0.001), in 23% (5/22) of those with POI (p < 0.001), compared with 4% (6/164) among controls. OR for MetS in all CCS compared with controls was 4.4 (95% CI 1.8, 11.1) (p = 0.002) and among CCS with POI the OR was 7.7 (CI 2.1, 28.1) (p = 0.002).
Conclusion: The prevalence of MetS was higher in females treated for childhood cancer compared with controls and the presence of POI significantly increased the risk of developing MetS.
Christopher J Child, Daniel Conroy, Alan G Zimmermann, Whitney W Woodmansee, Eva Marie Erfurth, and Leslie L Robison
Speculation remains that GH treatment is associated with increased neoplasia risk. Studies in GH-treated childhood cancer survivors suggested higher rates of second neoplasms, while cancer risk data for GH-treated and untreated hypopituitary adults have been variable. We present primary cancer risk data from the Hypopituitary Control and Complications Study (HypoCCS) with a focus on specific cancers, and assessment of recurrence rates for pituitary adenomas (PA) and craniopharyngiomas (CP).
Incident neoplasms during HypoCCS were evaluated in 8418 GH-treated vs 1268 untreated patients for primary malignancies, 3668 GH-treated vs 720 untreated patients with PA history, and 956 GH-treated vs 102 untreated patients with CP history.
Using population cancer rates, standardised incidence ratios (SIRs) were calculated for all primary cancers, breast, prostate, and colorectal cancers. Neoplasm rates in GH-treated vs untreated patients were analysed after propensity score adjustment of baseline treatment group imbalances.
During mean follow-up of 4.8 years, 225 primary cancers were identified in GH-treated patients, with SIR of 0.82 (95% CI 0.71–0.93). SIRs (95% CI) for GH-treated patients were 0.59 (0.36–0.90) for breast, 0.80 (0.57–1.10) for prostate, and 0.62 (0.38–0.96) for colorectal cancers. Cancer risk was not statistically different between GH-treated and untreated patients (relative risk (RR)=1.00 (95% CI 0.70–1.41), P=0.98). Adjusted RR for recurrence was 0.91 (0.68–1.22), P=0.53 for PA and 1.32 (0.53–3.31), P=0.55 for CP.
There was no increased risk for all-site cancers: breast, prostate or colorectal primary cancers in GH-treated patients during HypoCCS. GH treatment did not increase the risk of PA and CP recurrences.
Pat Kendall-Taylor, Peter J Jönsson, Roger Abs, Eva Marie Erfurth, Maria Koltowska-Häggström, David Anthony Price, and Johan Verhelst
Background: Craniopharyngioma is a parasellar tumour that, although benign, tends to behave aggressively. It can occur at any age but most commonly presents in childhood or adolescence.
Objectives: To investigate the frequency and severity of problems associated with craniopharyngioma, using the large international database (KIMS) for adult patients with GH deficiency (GHD), and to assess the differences between the adult onset (AO, aged 18 or above) disease and adults with childhood onset (CO) craniopharyngioma.
Design: Inclusion criteria were: an established diagnosis of craniopharyngioma, severe GHD and no recent GH treatment. These criteria were fulfilled by 393 (184 female, 209 male) patients; 241 had AO (mean age 28.7±8.7 years) and 152 had CO disease (age 42.0±12.3 years). Disease history, clinical features and anthropometric data were recorded at the time of enrolment in the database, and body composition, serum IGF-I, serum lipids and quality of life (QoL) were assessed.
Results: Peak age at onset of craniopharyngioma was 15–20 years. Ninety percent of patients had been treated surgically. CO patients were shorter than AO patients and had much lower IGF-I standard deviation scores (SDS). The majority had hypopituitarism and over 60% had diabetes insipidus. Body mass index (BMI) was higher in AO males (30.2±5.5) than in CO males (28.5±7.5); waist circumference was also greater. Obesity was more common in AO patients (51.8% vs 39.1%). Body composition did not differ between groups. Cholesterol and triglycerides were higher in AO than in CO patients, but high density lipoprotein (HDL)- and low density lipoprotein (LDL)-cholesterol did not differ. Quality of life, assessed by Quality of Life-Assessment of Growth Hormone Deficiency in Adults (QoL-AGHDA) and the Nottingham Health Profile, was markedly reduced in all groups with no significant differences between them; the QoL-AGHDA score correlated with age at onset of both craniopharyngioma and GHD, and also with BMI in AO patients.
Conclusions: These data emphasise the generally poor state of health of patients treated for craniopharyngioma, with respect to endocrine and metabolic function, and also the markedly reduced quality of life. In addition to GHD, most patients have evidence of hypothalamic damage with associated obesity, diabetes insipidus and hypopituitarism. Adults with CO craniopharyngioma were shorter, had lower IGF-I, lower BMI, less obesity and slightly lower blood lipid levels than patients with AO craniopharyngioma.
Christopher J Child, Alan G Zimmermann, Whitney W Woodmansee, Daniel M Green, Jian J Li, Heike Jung, Eva Marie Erfurth, and Leslie L Robison
GH and IGFs have mitogenic properties, causing speculation that GH treatment could increase risk of malignancy. While studies in GH-treated childhood cancer survivors have suggested a slight increase in second neoplasms, studies in GH-treated adults have been equivocal.
Incidence of de novo and second cancers was evaluated in 6840 GH-treated and 940 non GH-treated adult patients in the Hypopituitary Control and Complications Study pharmacoepidemiological database.
Evident cancer cases were evaluated in the main analysis, with sensitivity analyses including probable and possible cancers. Standardized incidence ratios (SIRs) for cancers were calculated using Surveillance, Epidemiology and End Results for the USA and GLOBOCAN for all other countries.
During the mean follow-up of 3.7 years/GH-treated patient, 142 evident cancer cases were identified, giving an overall SIR of 0.88 (95% confidence interval (CI) 0.74–1.04); 95% CIs included the value of 1.0 for each country examined. The SIR for GH-treated patients from the USA (71 cases) was 0.94 (95% CI 0.73–1.18), and for non GH-treated patients from the USA (27 cases) was 1.16 (95% CI 0.76–1.69). For GH-treated patients from the USA aged <35 years, the SIR (six cases) was 3.79 (1.39–8.26), with SIR not elevated for all other age categories; SIR for patients from the USA with childhood onset (CO) GH deficiency (GHD) was 2.74 (95% CI 1.18–5.41). The SIR for colorectal cancer in GH-treated patients (11 cases) was 0.60 (95% CI 0.30–1.08).
With relatively short follow-up, the overall primary cancer risk in 6840 patients receiving GH as adults was not increased. Elevated SIRs were found for subgroups in the USA cohort defined by age <35 years or CO GHD.