Hypercholesterolemia and Obesity in Adult Patients with Hypopituitarism: a Report of a Nation-Wide Survey in Japan

Endocrine Journal 2003, 50 (6), 759–765

Hypercholesterolemia and Obesity in Adult Patients with Hypopituitarism: A Report of a Nation-wide Survey in Japan

YOSHIO MURAKAMI AND YUZURU KATO

Department of Endocrinology, Matabolism and Hematological Oncology, School of Medicine, Shimane University, 89-1 Enya-cho, Izumo 693-8501, Japan

Abstract. Dyslipidemia and obesity are common in adult patients with hypopituitarism. Possible contributions of age, sex and hormone deficiencies to hypercholesterolemia and obesity in adult hypopituitary patients were analyzed in 1, 272 Japanese cases based on a database of a national survey on adult hypopituitarism. In patients on routine hormone replacement therapy, 30.5% of male and 40.7% of female subjects were considered hypercholesterolemic. In univariate analysis, hypercholesterolemia was more prevalent in female, aged, untreated Gn-deficient and TSH-deficient groups. In multivariate analysis, sex of female, age older than 40 yr and TSH deficiency were the independent contributing factors to hypercholesterolemia. Obesity (body mass index (BMI) O25 kg/m2) was more prevalent in male, TSH-deficient and ADH-deficient groups. Severe obesity (BMI O30) was observed in high prevalence in the youngest group. These findings suggest that hypercholesterolemia and obesity were prevalent in different age and gender groups in Japanese adult patients with hypopituitarism. Insufficient replacement of thyroid hormone and possibly gonadotropin deficiency might contribute to hypercholesterolemia. In contrast, hypothalamic dysfunction as well as hormone deficiencies might play roles in obesity in these patients.

Key words: Hypopituitarism, Hypercholesterolemia, Obesity (Endocrine Journal 50: 759–765, 2003)

IT has been demonstrated that mortality in hypo- was attributed to GH deficiency. Rosén et al. [9] spec- pituitary patients is increased compared to general ulated increased LDL and triglyceride concentrations population [1–4] in Western countries. In some stud- due to GH deficiency might explain the increased vas- ies, the increased mortality was attributed to cardio- cular death. However, disturbances and inappropriate vascular [1, 4] and cerebrovascular deaths [2]. In replacement of other hormones may contribute to the Japan, we failed to see the marked number of vascular dyslipidemia and consequent vascular death. Obesity, deaths or onset of vascular diseases during hormone another independent risk factor for vascular disorders, replacement in a national survey of adult patients with is also common in hypopituitary adults [10, 11]. hypopituitarism [5]. However, we found that a large In the present study, possible contributions of age, number of patients were treated with lipid-lowering sex and hormone deficiencies to hypercholesterolemia drugs [5]. and obesity in adult hypopituitary patients were ana- Dyslipidemia is common in adult patients with hypo- lyzed. pituitarism [6–8]. The increased vascular mortality

Subjects and Methods Received: December 24, 2002 Accepted: September 5, 2003 Patients Correspondence to: Dr. Yoshio MURAKAMI, M.D., D.M.S., Department of Endocrinology, Matabolism and Hematological Oncology, School of Medicine, Shimane University, 89-1 Enya- This study is based on a database of 1,464 case cho, Izumo 693-8501, Japan reports of adult patients with hypopituitarism obtained 760 MURAKAMI and KATO at a national survey in 2001. Inclusion criteria were present study. Thus 1,272 patients were eligible for hypofunction of at least one axis of the anterior pitui- the present analysis. Underlying diagnosis is shown in tary gland and age not less than 18 yr. Table 1. Three hundred eighty-two patients received Patients with GH- (n = 64) and ACTH-producing surgery (transsphenoidal: 300, cranial: 57, both trans- adenomas (n = 14) were excluded because of esta- sphenoidal and cranial: 14, others: 11). Conventional blished effects of these diseases on lipid profile and and stereotaxic irradiation was performed in 207 and body composition. Patients with Kallmann syndrome 48 patients, respectively. (n = 24) and DAX-1 mutation (n = 3) were also ex- ACTH, TSH, gonadotropin (Gn), PRL and ADH de- cluded because these disorders are known to affect ficiencies were evaluated by reporting doctors based specific axes of the pituitary gland. Eighty-seven case on combination of clinical findings and laboratory records did not contain adequate information for the tests according to the guidelines [12]. Available as-

Table 1. Etiologies of hypopituitarism in 1272 adult patients

Etiology male female total Pituitary adenoma non-functioning 120 94 214 Prolactinoma 29 36 65 TSH-producing 2 0 2 Gn-producing 5 1 6 not identified 19 16 35 Craniopharyngioma 102 86 188 Germinoma 65 39 104 Meningioma 6 14 20 Other hypothalamo-pituitary tumors 40 53 93 Sheehan syndrome 0 93 93 Hypophysitis 17 24 41 Trauma 17 2 19 Idiopathic 167 104 271 Others 603999 Unknown 12 10 22 Total 661 611 1272

Table 2. Hormone deficiency and replacement therapy in 1272 adult patients with hypopituitarism

deficient treated ACTH 931 898 TSH 850 787 Gonadotropins 737 459 Prolactin 181 0 AVP 409 371

Table 3. GH status ( g/L) in 1272 adult patients with hypopituitarism

Stimulation Number of patients peak GH?3 3cholesterol values were available in 903 arginine, glucagon or L-dopa stimulation was less than patients. Distribution of serum total cholesterol levels 5 g/L. When serum GH value was more than 5 g/L in stratified age groups are shown in Table 4. In men, after insulin, arginine, glucagon, L-dopa or GHRH 65 (13.7%) of 476 patients were on lipid-lowering stimulation, the patient was classified as a “GH re- drugs. In 80 of 411 men (19.5%) who were not taking sponder”. The other patients were designated as “not lipid-lowering drugs, serum total cholesterol levels determined”. were greater than 240 mg/dL. In women, 80 (18.7%) of 427 patients were on lipid-lowering drugs. In 94 Statistical analyses of 347 women (27.1%) who were not taking lipid- lowering drugs, serum total cholesterol levels were Univariate analysis was performed by 2-test. Be- greater than 240 mg/mL. cause age of patients did not show normal distribution, Subjects whose serum total cholesterol levels were patients were stratified into 3 age groups (<40 yr, less than 240 mg/mL without taking lipid-lowering 40–59 yr and O60 yr) for the analysis. Logistic regres- drugs were considered normocholesterolemic. The sion technique was used for multivariate analysis to other subjects were designated as hypercholesterol- determine factors contributing to hypercholesterolemia emic. Accordingly, 30.5% of male and 40.7% of fe- and obesity. Factors considered include sex, age, Gn male subjects were hypercholesterolemic. In univariate status, GH status, glucocorticoid replacement, TSH analysis, hypercholesterolemia was more prevalent deficiency and ADH deficiency. A P value less than in female, aged, untreated Gn-deficient and TSH- 0.05 was considered significant. deficient groups (Fig. 1). In multivariate analysis, sex of female, age older than 40 yr and TSH deficiency were the independent contributing factors (Table 5).

Table 4. Serum total cholesterol levels in 903 adult patients with hypopituitarism

lipid-lowering serum total cholesterol (mg/dL) sex age (yr) total drugs ?219 220–239 240–259 260–279 280–299 300? male18–39(+)7001019 (–)1042413663156 40–59(+)183432131 (–)771410543113 60–91(+)153421025 (–)872516770142 female18–39(+)3201208 (–)782013811121 40–59(+)84620121 (–)5481488395 60–91(+)358611051 (–) 66 27 14 14 8 2 131 762 MURAKAMI and KATO

Fig. 1. Prevalence of hypercholesterolemia in 903 adult patients with hypopituitarism. Statistical difference was analyzed by 2-test.

Table 5. Multivariate analysis of hypercholesterolemia in adult patients with hypopituitarism

Variable Category Hazard ratio (95% CI) Sex Male 1.00 Female 1.49 (1.12–1.99) p = 0.0063 Age (yr) <40 1.00 40–59 2.50 (1.67–3.72) >60 2.43 (1.60–3.68) p<0.0001 Gonadotropins No deficiency 1.00 Treated deficiency 1.32 (0.90–1.92) Untreated deficiency 1.45 (1.01–2.10) p = 0.1073 GH Responder 1.00 Deficiency 0.90 (0.55–1.46) “Not determined” 0.79 (0.53–1.20) p = 0.4858 Glucocorticoid replacement No 1.00 Yes 0.99 (0.68–1.43) p = 0.9486 TSH No deficiency 1.00 Deficiency 1.51 (1.04–2.18) p = 0.0279 ADH No deficiency 1.00 Deficiency 1.04 (0.76–1.43) p = 0.8041

When ACTH deficiency was substituted for gluco- percentage was lower in GH responder than groups of corticoid replacement, results of analysis were not af- GH-deficiency and “not determined” GH status, but fected (data not shown). the difference did not reach the statistical significance. Body mass index (BMI) was assessed in 654 pa- When independent contributing factors were deter- tients. In univariate analysis, obesity (BMI O25 kg/m2) mined by multivariate analysis, the above 3 factors was more prevalent in male, TSH-deficient and ADH- remained significant (Table 6). Severe obesity (BMI deficient groups (Fig. 2). There was a trend that the O30 kg/m2) was observed in high prevalence in the LIPID AND OBESITY IN HYPOPITUITARISM 763

Fig. 2. Prevalence of obesity in 654 adult patients with hypopituitarism. Prevalence of BMI O30 kg/m2 (closed box) and 30 kg/m2>BMI O25 kg/m2 (shaded box) are shown. Statistical difference was analyzed by 2-test.

Table 6. Multivariate analysis of obesity in adult patients with hypopituitarism

Variable Category Hazard ratio (95% CI) Sex male 1.00 female 0.62 (0.44–0.88) p = 0.0071 age (yr) <40 1.00 40–59 1.12 (0.71–1.77) >60 1.11 (0.67–1.84) p = 0.8712 Gonadotropins No deficiency 1.00 Treated deficiency 1.16 (0.72–1.85) Untreated deficiency 1.36 (0.85–2.19) p = 0.4449 GH Responder 1.00 Deficiency 1.17 (0.64–2.15) “Not determined” 1.54 (0.93–2.56) p = 0.1494 Glucocorticoid replacement No 1.00 Yes 0.87 (0.55–1.38) p = 0.5569 TSH No deficiency 1.00 Deficiency 1.97 (1.24–3.14) p = 0.0034 ADH No deficiency 1.00 Deficiency 1.48 (1.01–2.16) p = 0.0439 youngest group although prevalence of BMI O25 kg/m2 Discussion was not different among age groups (Fig. 2). Sub- stitution of glucocorticoid replacement by ACTH Hyperlipidemia is common in untreated hypopitu- deficiency did not affect the results. itarism and persists even after routine replacement therapy [6, 7]. In the present study, 30.5% of male and 40.7% of female subjects were hypercholesterolemic. These values are similar to those in foreign reports [13, 764 MURAKAMI and KATO

14]. We also found that age older than 40 years (yr), deficiency in whom glucocorticoid, thyroid and go- gender of female and TSH deficiency were the contrib- nadal hormone deficiencies were adequately treated. uting factors for hypercholesterolemia. Preference of The hypopituitary patients below the age of 50 yr had female gender to hypercholesterolemia is on the same proper sex hormone substitution. In their study, serum line with the reports that increased prevalence of total cholesterol concentrations of the patients were hyperlipidemia in hypopituitary patients over general not different from those in age- and sex-matched population was most striking in female patients [14, controls. In addition, 83 of 785 patients with adult- 15]. Thyroid hormone deficiency causes hyper- onset hypopituitarism received lipid-lowering drugs cholesterolemia and results in excess vascular risk. In during GH replacement therapy in a surveillance study patients with primary hypothyroidism, even sub- [25]. clinical hypothyroidism, defined as a normal circulat- Obesity is common in hypopituitary adults [10, 11]. ing thyroid hormone concentration with an elevated In the present study, gender of male and ADH defi- plasma TSH level, is associated with hyperlipidemia ciency were the contributing factors for obesity. Al- [16]. It is suggested, therefore, that many hypo- though prevalence of obesity was similar in stratified pituitary patients may tend to be under-replaced for age groups, BMI more than 30 kg/m2 was the most fre- thyroid hormone in Japan. quent in the youngest group. These associations may Although hypercholesterolemia was prevalent in be partly attributed to the etiology of hypopituitarism. patients with untreated Gn deficiency, Gn status was Generally, tumors of hypothalamic origin such as not a statistically significant factor in multivariate craniopharyngioma and germinoma are prevalent in analysis. Serum total cholesterol levels increase after young males. Taken together with the association with menopause in women [17] and estrogen replacement ADH deficiency, it is suggested that hypothalamic therapy elicits beneficial alterations of lipid profile dysfunction might be involved in obesity. by increasing HDL-cholesterol and decreasing LDL- TSH deficiency was also a statistically significant cholesterol [18]. In premenopausal women, however, determinant of obesity. Hypothyroidism per se associ- use of contraceptives results in an increase in total and ates with overweight which is attributed to water and LDL-cholesterol levels [19], suggesting complicated ef- sodium retention. Alternatively, GH deficiency might fects of estrogens on lipid profiles. Co-administration be another candidate involved in obesity because of progesterone may modulate the effects of estrogens. severity of GH deficiency is related to the severity of Progestagens with androgenic properties raise total deficiency of other pituitary hormones [26]. GH- cholesterol level [20] while medroxyprogesterone deficient adults had an increase in BMI [24], which acetate possessing minimal androgenic properties do has been ascribed to increased body fat [10] and not affect lipid profiles [20]. Thus effects of Gn defi- especially visceral fat [27]. In the univariate analysis, ciency and its treatments on lipid metabolism might be there was a trend that percentage of obesity was lower modulated by multiple factors. in GH responders than groups of chemically defined In this study, prevalence of hypercholesterolemia GH-deficiency and “not determined” GH status. It was not different among groups classified by GH sta- should be noted that the statistical power might be tus. It is widely accepted that GH treatment results in reduced because GH status was evaluated by provoca- a decrease in lipid levels in GH deficiency. However, tion tests only in small numbers of patients. this has not been consistently observed. It was report- In summary, hypercholesterolemia and obesity, two ed that GH replacement therapy decreased total cho- independent risk factors of cardiovascular diseases, lesterol level only in men [13, 21] and only transiently were prevalent in different age and gender groups in [22] in GH-deficient hypopituitary patients. In a Japanese adult patients with hypopituitarism. Insuffi- report, no effect of GH treatment was evident [23]. cient replacement of thyroid hormone and possibly Gn There have been several reports suggesting the major deficiency might contribute to hypercholesterolemia. role of other factors than GH deficiency in hyper- In contrast, hypothalamic dysfunction as well as hor- cholesterolemia. Rosén et al. [24] examined cardio- mone deficiencies might play roles in obesity in these vascular risk factors in adult patients with GH patients. LIPID AND OBESITY IN HYPOPITUITARISM 765

References

1. Rosén T, Bengtsson B-Å (1990) Premature mortality malities in growth hormone deficient adults. II. Carbo- due to cardiovascular disease in hypopituitarism. hydrate tolerance and lipid metabolism. Endocrinol Lancet 336: 285–288. Metab 1: 173–180. 2. Bülow B, Hagmar L, Mikoczy Z, Nordström C-H, Erfurth 16. Bastenie PA, Vanhaelst L, Bonnyns M, Neve P, EM (1997) Increased cerebrovascular mortality in pa- Staquet E (1971) Pre-clinical hypothyroidism: a risk tients with hypopituitarism. Clin Endocrinol 46: 75–81. factor for coronary heart disease. Lancet 203–204. 3. Bates AS, Hoff WV, Jones PJ, Clayton RN (1996) The 17. Gambrell RD, Teran A-Z (1991) Changes in lipids and effect of hypopituitarism on life expectancy. J Clin lipoproteins with long-term estrogen deficiency and Endocrinol Metab 81: 1169–1172. hormone replacement therapy. Am J Obstet Gynecol 4. Tomlinson JW, Holden N, Hills RK, Wheatley K, 165: 307–317. Clayton RN, Bates AS, Sheppard MC (2001) Associa- 18. Bush TL, Barrett-Connor E, Cowan LD, Criqui MH, tion between premature mortality and hypopituitarism. Wallance RB, Suchindran CM, Tyroler HA, Rifkind Lancet 357: 425–431. BM (1987) Cardiovascular mortality and noncontra- 5. Murakami Y, Yokoyama T, Oiso Y, Kato Y (2002) ceptive use of estrogen in women: results from the lipid Clinical features of adult hypopituitary patients in a research clinics program follow-up study. Circulation national survey. In: Kato Y (ed) Annual Report of Japan 75: 1102–1109. Study Group on Hypothalamo-Pituitary Disorders. 19. Varizi SM, Evans JC, Larson MG, Wilson PWF (1993) 170–176 (In Japanese). The impact of female hormone usage on the lipid 6. Cuneo RC, Salomon F, Watts GF, Hesp R, Sönksen PH profile. The Framingham offspring study. Arch Intern (1993) Growth hormone treatment improves serum Med 153: 2200–2206. lipids and lipoproteins in adults with growth hormone 20. Hirvonen E, Mälkönen M, Manninen V (1981) Effects deficiency. Metabolism 42: 1519–1523. of different progestogens on lipoproteins during post- 7. De Boer H, Blok GJ, Voerman HJ, Phillips M, menopausal replacement therapy. N Eng J Med 304: Schouten JA (1994) Serum lipid levels in growth 560–563. hormone-deficient men. Metabolism 43: 199–203. 21. Burman P, Johansson AG, Siegbahn A, Vessby B, 8. Ishibashi S, Murase T, Yamada N, Tanaka K, Takaku Karlsson FA (1997) Growth hormone (GH)-deficient F, Sato K (1985) Hyperlipidaemia in patients with men are more responsive to GH replacement therapy hypopituitarism. Acta Endocrinol 110: 456–460. than women. J Clin Endocrinol Metab 82: 550–555. 9. Rosén T, Wilhelmsen L, Bengtsson B-Å (1998) Al- 22. Edén S, Wiklund O, Oscarsson J, Rosén T, Bengtsson tered lipid pattern explains increased cardiovascular B-Å (1993) Growth hormone treatment of growth mortality in hypopituitary patients with growth hor- hormone deficient adults results in a marked increase mone deficiency. Clin Endocrinol 48: 525–526. in Lp(a) and HDL cholesterol concentrations. Arterio- 10. Rosén T, Bosaeus I, Tölli J, Lindstedt G, Bengtsson scler Thromb 13: 296–301. B-Å (1993) Increased body fat mass and decreased ex- 23. Whitehead HM, Boreham C, McIlrath EM, Sheridan B, tracellular fluid volume in adults with growth hormone Kennedy L, Atkinson AB, Hadden DR (1992) Growth deficiency. Clin Endocrinol 38: 63–71. hormone treatment of adults with growth hormone 11. Salomon F, Cuneo RC, Hesp R, Sonkesen PH (1989) deficiency: results of a 13-month placebo controlled The effects of treatment with recombinant human cross-over study. Clin Endocrinol 36: 45–52. growth hormone on body composition and metabolism 24. Rosén T, Edén S, Larsson G, Wilheimsen L, Bengtsson in adults with growth hormone deficiency. N Engl J B-Å (1993) Cardiovascular risk factors in adult pa- Med 321: 1797–1803. tients with growth hormone deficiency. Acta Endo- 12. Anonymous (2002) Guidelines for diagnosis and crinol 129: 195–200. treatment of hypothalamo-pituitary disorders. In: Kato 25. Attanasio AF, Bates PC, Ho KKY, Webb SM, Ross RJ, Y (ed) Annual Report of Japan Study Group on Strasburger CJ, Bouillon R, Crowe B, Selander K, Hypothalamo-Pituitary Disorders. 29–57 (In Japanese). Valle D, Lamberts SWJ (2002) Human growth hormone 13. Bengtsson B-Å, Abs R, Bennmarker H, Monson JP, replacement in adult hypopituitary patients: long-term Feldt-Rasmussen U, Hernberg-Ståhl E, Westberg B, effects on body composition and lipid status-3-year Wilton P, Wüster C (1999) The effects of treatment and results from the HypoCCS database. J Clin Endocrinol the individual responsiveness to growth hormone (GH) Metab 87: 1600–1606. replacement therapy in 665 GH-deficient adults. J Clin 26. Toogood AA, Beardwell CG, Shalet SM (1994) The Endocrinol Metab 84: 3929–3935. severity of growth hormone deficiency in adults with 14. Abs R, Bengtsson B-Å, Hernberg-Ståhl E, Monson JP, pituitary disease is related to the degree of hypo- Tauber J-P, Wilton P, Wüster C (1999) GH replace- pituitarism. Clin Endocrinol 41: 511–516. ment in 1034 growth hormone deficient hypopituitary 27. Bengtsson B-Å, Edén S, Lönn L, Kvist H, Stokland A, adults: demographic and clinical characteristics, dosing Lindstedt G, Bosaeus I, Tölli J, Sjöström L, Isaksson and safety. Clin Endocrinol 50: 703–713. OGP (1993) Treatment of adults with growth hormone 15. Beshyah SA, Herdersson A, Niththyananthan R, Sharp deficiency with recombinant human growth hormone. P, Richmond W, Johnston DG (1994) Metabolic abnor- J Clin Endocrinol Metab 76: 309–317.