cost of illness in adult patients with hypopituitarism

Cost of Illness in Adult Patients with HypopituitarismChrister Ehrnborg,1 Leona Hakkaart-Van Roijen,2 Björn Jonsson,3 Frans F.H. Rutten,2Bengt-Åke Bengtsson1 and Thord Rosén1

1 Research Centre for Endocrinology and Metabolism, Sahlgrenska University Hospital,Gothenburg, Sweden

2 Institute for Medical Technology Assessment, Erasmus University, Rotterdam, The Netherlands3 Department of Women and Child Health, Karolinska Institute, Stockholm, Sweden

Abstract Objective: To compare the healthcare costs of patients with hypopituitarismwiththose of individuals from the general population.Design:Aretrospective study of costs over 1 year. Estimates of direct and indirecthealth-related costs were calculated for patients from the general population usingexisting databases, and for patients with hypopituitarism using records of allpatients eligible to participate on 31 December 1989 who could be traced, werewilling to participate and had not been treated for acromegaly or Cushing’s dis-ease.Setting: The catchment area of the Endocrine Unit, Sahlgrenska Hospital,Gothenburg. The study was conducted from the societal perspective. Referencedata were collected from official regional and national registries.Patients: 199 patients with adult-onset hypopituitarism in whom replacementtherapy was given to maintain the adrenal, thyroid and gonadal (but not thesomatotropic) axes.Main outcome measures and results: Direct and indirect costs incurred bypatients with hypopituitarismwere higher than those incurred by individuals fromthe general population. The total direct costs per patient were Swedish Crowns(SEK)22 920 vs SEK12 080 (p < 0.003) in the general population, and the highestcosts were related to inpatient care. Of the patients aged 16 to 64 years, 22% haddrawn a disability pension versus the expected 11.3% (p < 0.003) in the generalpopulation, and the patients had amean sick leave of 38.4 days vs 23.5 (p < 0.001).Total excess costs for all patients with hypopituitarism were SEK 35 768 perpatient (p < 0.007).Conclusions: Patients with hypopituitarism incur more health-related costs thanindividuals from the general population. They also take more sick leave days andare more likely to claim a disability pension than members of the general popu-lation. Further cost analyses are needed to determine whether improvements indiagnostic and surgical procedures, and hormone replacement therapy, can reducethe healthcare costs of patients with hypopituitarism.

ORIGINAL RESEARCH ARTICLE Pharmacoeconomics 2000 Jun; 17 (6): 621-6281170-7690/00/0006-0621/$20.00/0

© Adis International Limited. All rights reserved.

Hypopituitarism acquired in adult life is gener-ally caused by pituitary or peripituitary tumoursand their treatment.[1] Acquired hypopituitarism isusually severe and patients require hormone sub-stitution to maintain function of the adrenal, thy-roid, gonadal and somatotropic axes. Cost of ill-ness is an aspect of pituitary disease that has beenlittle studied,[2,3] although the cost to society for spec-ific diseases, such as diabetes mellitus, has beenattracting increasing attention in recent years.[4-7]In this article, we review the cost to society of

illness and disability in patients with hypopituit-arism of adult onset, in whom anterior pituitaryhormone replacement (other than growth hormone)was given. We estimated the direct and indirecthealthcare costs in a group of patients with hypo-pituitarism over 1 year and compared the resultswith equivalent costs in the general population.The year in which the costs associated with hypo-pituitarism were assessed was 1989, a year thatpredated the establishment of growth hormone de-ficiency in adults as an indication for growth hor-mone replacement therapy.

Methods

Study Design

This was a cross-sectional register study con-ducted from the societal perspective, using datafrom official national and regional registries andpublications, together with data manually gatheredfrom patient case sheets.The study was approved by the Ethics Commit-

tee of the Medical Faculty of the University ofGothenburg. The use of databases was approved bythe National Data Authorities.

Patient Characteristics

Eligible patients were adults with hypopituita-rism diagnosed between 1956 and 1988 who hadbeen treated at the Endocrine Unit of SahlgrenskaHospital, Sweden on at least one occasion. Patients(or their relatives if the patient had died beforefollow-up in August 1995) received written and,

when necessary, verbal information about thestudy, and informed consent was obtained.A total of 311 adults with known hypopituita-

rism were alive and living in the catchment area ofSahlgrenska University Hospital on 31 December1989. Patients were excluded if they had beentreated for acromegaly or Cushing’s disease (n =70) or could not be traced or refused to participate(n = 42). Thus, the study cohort consisted of 199patients, 25 of whom had died before August 1995.Of the 199 patients [125 men, mean age 60.6

years (range 26 to 87); and 74 women, mean age53.1 years (range 21 to 83)], 107 had completeanterior pituitary deficiency and 92 had partialdeficiencies. The aetiology of hypopituitarism inthe participating patients is presented in table I, andthe routine substitution therapies received by allpatients, when appropriate, are detailed in table II.Growth hormone stimulation tests or serum lev-

els of insulin-like growth factor I were assessed in129 patients, 125 of whom were considered to begrowth hormone–deficient.

Reference Population and Related Cost Data

The direct costs for inpatient care in 1995 wereobtained from a Gothenburg database [DebiteringOch Registrering I Slutenvården (charging and reg-istration in inpatient care); DORIS] in which allinpatient stays and costs for procedures and hospi-tal days were registered for individuals living in theGothenburg catchment area. The database con-tained information on gender, age and hospital carefor 44 494 individuals. Data on gender and age dis-tribution of the Gothenburg catchment area popu-

Table I. Aetiology of hypopituitarism in the 199 patients included inthe study

Diagnosis No. of patients (%)Nonsecreting adenoma 110 (55.3)Prolactinoma 31 (15.6)Idiopathic pituitary disease 28 (14.1)Craniopharyngioma 21 (10.5)Meningioma 4 (2.0)Other tumours 5 (2.5)

622 Ehrnborg et al.

© Adis International Limited. All rights reserved. Pharmacoeconomics 2000 Jun; 17 (6)

lation in 1995 were supplied by the NationalBureau of Statistics. Using these data, age- andgender-standardised inpatient direct costs (‘ex-pected values’) were obtained for comparisons withthe study population. An official study of health-care costs in southern Sweden by the SwedishMin-istry of Health and Social Affairs, HSU 2000,[8]provided information on estimates of total directin- and outpatient costs per individual for differentage groups in 1994, which were assumed to be ap-plicable to the general Swedish population. Costdata from HSU 2000 were adjusted to 1995 levelsusing the official consumer price index. Unfortu-nately, these costs could be standardised accordingto age only.Controls selected from the national registry of

sick leave, held by the National Social InsuranceBoard (NSIB), were those individuals whose per-sonal identity numbers, which include date ofbirth, immediately followed the identity numbersof the patients.

Costs

The costs in the study were assessed in 1989 andthen compared with the actual costs for the 1995reference population after the patients’ costs weretransformed to the 1995 level. The transformationwas performed because no reliable information ontotal healthcare costs in 1989 was available. Costsare stated in Swedish Crowns (SEK). In 1995,$US1 was equivalent to SEK8.

Direct CostsAcoding sheet was developed to enable compa-

rable data on in- and outpatient care to be gatheredand evaluated. Costs incurred as a direct result ofpituitary surgery in 1989 (4 patients) were exclu-ded from the analysis.

Inpatient CareData for days in general care in hospital, days

in intensive care, laboratory tests, consultations,medical procedures and medication were recordedfor 1989. Information on inpatient stays was ob-tained from the Patient Registry, a centralised data-base that registers all inpatient stays in Sweden,

and the local database, DORIS. Additional infor-mation about inpatient stays was obtained fromlocal hospitals and, in some cases, the patients.Manual checks at the hospitals ensured that all in-patient stays had been registered.Medical charts for each inpatient stay were ex-

amined by one of our group (C.E.) and the costs ofinpatient care were calculated by multiplying thecost for each item in 1995 by the number of itemsused, according to manuals developed at Sahlgren-ska University Hospital.

Outpatient CareThe costs for outpatient care include the costs

of visiting family practitioners and hospital clinicsin 1989. Family practitioners’ records, the localhospital or the patients provided information aboutvisits. Components of outpatient care included thenumber of visits, laboratory tests, radiography andmedication. Costs for visits to family practitionerswere based on the 1995 hospital tariffs.

Indirect CostsIndirect costs included absence fromwork (sick

days) and awarded disability pensions. In Sweden,disability pensions and reimbursement for sickdays are given to individuals aged 16 to 64 years.The NSIB supplied information on reimbursedworking days due to sickness for 105 patients (outof 127 eligible patients aged 16 to 64 years) and5341 control individuals in 1989. The number ofsick days for a particular individual in 1 year canexceed 365 in NSIB registries because the number

Table II. Routine adrenal, thyroid and gonadal substitution therapyin the 199 study patients with hypopituitarism

Treatment Mean dosage Range (mg)

AdrenalCortisone 25 mg/day (po) 12.5-50

ThyroidThyroxine 0.125 mg/day (po) 0.05-0.20

GonadalTestosterone 250 mg/4 weeks (IM)Estradiol 1.8 mg/day (po)IM = intramuscular; po = oral.

Cost of Illness in Adult Hypopituitarism 623


of days is notified when the sick leave ends. In suchcases, we truncated the number of sick days to 365.

The NSIB also supplied information on patientsreceiving disability pensions in 1989. Using thesedata, together with age- and gender-specific datafrom the general Swedish population in December1989, supplied by the NSIB,[9] the number of pa-tients who would have been expected to draw adisability pension was calculated.

Excess Total CostsThe cost to society of sick leave and disability

pensions was assumed to be comparable to themean income per fully employed Swedish workerin 1995, which means that 1 sick day would corre-spond to SEK984 per working day.[10] The numberof working days per year in Sweden is 253, givingan annual cost of SEK248 952 for 1 disability pen-sioner.

The expected costs for disability pensions werecalculated using this estimate and national statis-tics showing the expected frequency of disabilitypensioners, according to age and gender. Data onthe mean number of sick days for controls, accord-ing to gender and age, gave an estimate of the ex-pected cost of sick days in the patient sample.

The total observed cost for each patient was cal-culated as the sum of the estimated total directcosts, cost for disability pensioners and sick leavecosts. Excess cost was defined as the differencebetween observed costs for patients with hypopitu-itarism and costs expected in the general popula-tion.

Statistical Methods

As the main focus of the study was on meancosts, descriptive statistics include means, standarddeviations and 95% confidence intervals. The dis-tributions of the variables of interest were skewed.Despite this, and supported by the central limit the-orem, Student’s one-sample t-test was used to ana-lyse the difference between actual cost data andexpected cost data obtained from official registries.Student’s independent-sample t-test was used foranalysing differences between groups. The stand-ardised incidence ratio was defined as the numberof observed patients divided by the number of ex-pected patients drawing a disability pension. Sta-tistical tests were performed and confidence inter-vals calculated for standardised incidence ratiosbased on the assumption that the observed numbers

Table III. Total annual direct and inpatient costs ($US1 = SEK8) per patient for the 199 patients with hypopituitarism included in the study,and corresponding total direct costs for the Swedish population (HSU 2000) and total inpatient costs for residents of Gothenburg (DORIS).Values are means ± standard error of the mean, with median values in parentheses

Parameter Costs for all patients (SEK) Costs for men (SEK) Costs for women (SEK)

Reference dataHSU 2000–standardised costs 12 080 (9328) 13 224 (9328) 10 144 (9328)

DORIS-standardised total inpatient costs 5416 (4232) 6592 (6368) 3448 (2552)

Patient dataTotal direct cost 22 920 ± 2984† (7880) 25 712 ± 4168** (9576) 18 208 ± 3808* (6360)

laboratory cost 984 ± 120 (424) 1008 ± 160 (496) 936 ± 184 (384)

nursing cost 12 240 ± 2528 (0) 14 088 ± 3488 (0) 9120 ± 3384 (0)

specialist cost 3120 ± 192 (2664) 3104 ± 208 (2664) 3152 ± 384 (2664)

medicine cost 3400 ± 256 (2088) 3648 ± 344 (2392) 2984 ± 360 (1672)

procedures cost 2112 ± 344 (224) 2584 ± 520 (344) 1328 ± 272 (32)

surgery cost 1064 ± 304 (0) 1288 ± 448 (0) 696 ± 336 (0)

Total inpatient costa 14 584 ± 2744‡(0) 16 856 ± 3824 (920) 10 744 ± 3552 (0)

a Total inpatient cost is a subset of total direct cost.

DORIS = Debitering Och Registrering I Slutenvården (charging and registration in inpatient care); HSU 2000 = Investigation from the SwedishMinistry of Health and Social Affairs (needs and resources in healthcare – an analysis);[8] *p < 0.05 vs DORIS; **p < 0.008 vs DORIS; †p <0.003 vs HSU 2000; ‡p < 0.001 vs DORIS.

624 Ehrnborg et al.


came from a Poisson distribution and the expectednumber was constant.

Results

Direct Costs

The total direct cost per patient with hypopitu-itarism was significantly greater (p < 0.003) thanthe estimated total direct cost per individual in theSwedish population (i.e. SEK12 080 in HSU 2000)[table III]; the mean (95% confidence interval) to-tal direct cost per patient was SEK22 920 (17 000to 28 800). The highest cost components were fornursing,medicines and specialists. Costswere higherfor men than for women with hypopituitarism, butthis difference was not statistically significant.The mean cost for the inpatient care of a patient

with hypopituitarism was SEK14 584 (9200 to24 400). This was significantly higher (p < 0.001)than the corresponding cost (estimated to beSEK5416) for inpatient care of individuals belong-ing to the Gothenburg population, which wasstandardised for age, gender and habitation (i.e.DORIS-standardised inpatient costs) [table III].Inpatient costs for patients with hypopituitarismwere also significantly elevated compared with thepatients in DORIS when men (p < 0.008) andwomen (p < 0.05) were considered separately. Themost frequent causes of hospitalisation in patientswith hypopituitarismwere pituitary-related (33%),cardiovascular (18%) and musculoskeletal (8%)disorders. The number of hypopituitary hormonedeficiencies had no significant effect on the directcosts for the patients, and nor did a specific hor-mone deficiency such as thyroid or adrenal defi-ciency.

Indirect Costs

Disability PensionOf the 127 patients who were eligible for dis-

ability pensions in the study year, 28 (22%) haddrawn a pension, which was significantly (p <0.003) higher than expected [14.4 (11.3%)]. Thestandard incidence ratio (observed/expected num-ber) was 1.94 (1.22 to 2.66). There was no statisti-

cally significant differencebetweenmenandwomen;the standard incidence ratio was 1.53 (0.63 to 2.43)for men and 2.35 (1.23 to 3.47) for women.

Sick LeavePatients with hypopituitarism took a signifi-

cantly greater number of days sick leave than con-trols (p < 0.001) [table IV]. This difference wassignificant for both men (p < 0.001) and women(p < 0.002). The number of hypopituitary hormonedeficiencies had no significant effect on the indi-rect costs of the patients, and nor did a specifichormone deficiency such as thyroid or adrenal de-ficiency.

Excess Total Costs of Patients with HypopituitarismThe excess direct annual costs for a patient with

hypopituitarism was estimated to be SEK10 840.The annual indirect excess costs for a patient withhypopituitarism were estimated to be SEK24 928,giving total excess costs of SEK35 768 per patientwith hypopituitarism (p < 0.007). When restrictingthe calculations to patients aged 16 to 64 years(n = 127), the total excess annual costs were esti-mated to be SEK50 424 per patient (p < 0.001)[table V]. There was no significant difference be-tween men and women in the total excess costs perpatient.

Discussion

Direct and indirect health-related costs incurredby patients with hypopituitarism were found to be

Table IV. Numbers of reimbursed sick days in 1989 for patientswith hypopituitarism and for age- and gender-matched controls.Values aremeans ± standard deviation, withmedians in parentheses

Participants No. ofobservations

No. of reimbursedsick days

All patients 105 38.4 ± 79.6**(7.0)All controls 5341 23.5 ± 65.1 (2.0)Male patients 55 32.4 ± 69.1†(7.0)Male controls 2653 21.6 ± 62.9 (1.0)Female patients 50 45.1 ± 89.9*(7.0)Female controls 2688 25.3 ± 67.1 (3.0)**p < 0.001 vs controls; †p < 0.001 vs male controls; *p < 0.002 vsfemale controls. Mann-Whitney U-test was used in all compari-sons.



higher than those incurred by the general popula-tion.In recent years, the cost to society of specific

diseases, such as diabetes mellitus, has been at-tracting increasing interest.[4-7] The present studyis, to our knowledge, one of the first attempts todetermine the economic costs to society of adultswith hypopituitarism. A Belgian study[2] has re-ported, similarly, that the costs to society, in termsof lost production and medical resource use, arehigher in patients with hypopituitarism than in thegeneral population.The overall purpose of our studywas to examine

the costs to society of patients with hypopituitarismreceiving routine replacement therapy (i.e. corti-sone, L-thyroxin and gonadal steroids) before theintroduction of growth hormone therapy. This waswhy the year 1989 was chosen, as this was the lastyear (at least in Sweden) when growth hormonetherapy was not routinely offered to patients withhypopituitarism. After this initial study, which hasshown increased costs to these patients with hypo-pituitarism, we plan a forthcoming study in whichwe intend to study the effects of the costs to thesepatients when growth hormone therapy is added.The methods used in our study ensured that

practically all available information on the costs tosociety of illness in patients with hypopituitarismwas available for analysis. Some records of outpa-tient visits were not always available; therefore,costs may have been underestimated. The totalcosts for the patients might also have been under-estimated because nonparticipating patients (n =42) are, as a rule, less healthy than those participat-

ing. The official statistics on sick leave and disabil-ity pensions for 1989 were considered accurate andreliable; information was confirmed by manualchecks at social insurance offices.The direct costs for patients in 1989 were com-

pared in the present study with actual costs for ref-erence populations in the mid-1990s, after havingtransformed patient costs to 1995 levels. Transfor-mation of patient costs was performed because noreliable information on total healthcare costs forthe Swedish population, or the population of thecatchment area, in 1989 was available. Therefore,we used the information from a 1994 study ofhealthcare costs in southern Sweden (HSU 2000)and from the 1995 Gothenburg database of charg-ing and registration in inpatient care (DORIS). Ref-erence material on indirect costs used in this studywas essentially based on data obtained from offi-cial statistics; thus, there is no reason to questiontheir accuracy.We realise that there may have been changes in

the general treatment of patients during the period1989 to 1995, caused by improved medical thera-pies and secular trends, and resulting in, for exam-ple, a change in the length of hospital stays. 1989patient costs may, therefore, have been overesti-mated. On the other hand, the use of new and moreexpensive diagnostic procedures such as magneticresonance imaging instead of computed tomogra-phy may have rendered relatively lower 1989 pa-tient costs. Overall, we are convinced that thesefactors had no major effect on the main results ofthe study.

Table V.Observed and excess costs ($US1 = SEK8) for patients with hypopituitarism, and estimated expected costs for the general population

Costs All patients (n = 199) Patients aged 16 to 64 years (n = 127)observed costs(SEK)

expected costs(SEK)

excess costs(SEK)

observed costs(SEK)

expected costs(SEK)

excess costs(SEK)

Direct 22 920 12 080 10 840 19 992 8624 11 368Indirectdisability pensions 35 032 18 056 16 976 54 888 28 288 26 600sick leave 22 592 14 640 7952 35 400 22 944 12 456

Total 35 768* 50 424**

SEK = Swedish Crowns. * p < 0.007 versus estimated total annual costs for individuals from the general population; ** p < 0.001 versus theestimated total annual costs for individuals from the general population.

626 Ehrnborg et al.


Indirect cost of disease is defined as the valueto society of production lost due to absence fromwork and disability and death. It should beemphasised that we do not value human life itself,but we analyse only the economic effects of dis-ease. Other important aspects of illness, such aspain and suffering, are not dealt with here andshould preferably be measured in terms of qualityof life. We have estimated indirect costs due to ab-sence from work and disability. We estimated theindirect costs according to the human capitalmethod. This method calculates the potential indi-rect costs of diseases. If, for example, someone be-comes disabled at the age of 35 years, it is assumedthat the total earnings from that age until the ageof retirement represent the value of productionlost. So, the human capital method assumes that allpotential years of life lost, short term absence fromwork and long term disability would have beenspent in full productivity. This concept of indirectcosts in the economic analysis of healthcare is notuncontroversial.Many authors have suggested that the produc-

tion losses for society are overestimated becauseof the current methodology used due to economiccircumstances.[7,11,12] For short term absences, workmay be taken over or postponed. For long termabsences, work can be taken over by an unem-ployed person or by the reallocation of jobs to otheremployees. An alternative approach, called thefriction cost method, takes into account these eco-nomic circumstances that limit production lossescaused by disease. According to this approach, pro-duction losses are assumed to be confined to theperiod needed to replace a sick worker – the fric-tion period. The length of this period and the re-sulting indirect costs depend on the situation in thelabour market.[6] Estimates of the indirect costs ac-cording to the friction cost method are only avail-able for the patient group. Because of a lack ofdetailed data, we could not calculate the indirectcosts according to the friction cost method for thegeneral population. For reasons of comparisonwith the general population, we present only theestimates based on the human capital approach. So,

this estimation calculates the loss in 1 year, basedon the prevalence of disability during 1989.The patients included in this study are repre-

sentative of patients with hypopituitarism in gen-eral, as almost all of the eligible patients weretraced. About two-thirds of the patients were testedfor growth hormone deficiency, and practically allof these were considered to be growth hormone–deficient. It is reasonable to suppose that a similarproportion of the remainder were also growthhormone–deficient as growth hormone is generallythe first hormone to be lost in patients with hypo-pituitarism.[13]The major costs for the patients were related to

hospital admissions; the main causes of these ad-missions were ‘pituitary disease’, cardiovasculardisorders and disorders of the musculoskeletal sys-tem. Increased atherosclerosis and premature deathfrom cardiovascular disease are seen among pa-tients with hypopituitarism.[14-16] Other factorsthat may contribute to the number of hospital ad-missions among patients with hypopituitarism in-clude: (i) fractures and myopathy due to low mus-cle mass and reduced bone mineral density;[1,17,18](ii) low physical activity due to a sedentary life-style in general; (iii) balance problems due to dis-turbed visual capacity; and (iv) decreased percep-tion secondary to the underlying pituitary tumour.We believe that the known morbidity of patientswith hypopituitarism was reflected well in the re-ported causes of hospital admission.The effect of the underlying hormone substitu-

tion therapy must also be considered. Adequatesubstitution therapy with corticosteroids and thy-roxine was provided, with no clinical or laboratorysigns of overdosage; however, a previous minoroverdosage, possibly aggravating bone loss and re-sulting in myopathy, cannot be excluded. No clin-ical signs of continuing overdosage of corticoste-roids were noted. On the contrary, many of thehospital admissions designated as resulting from‘pituitary disease’ were patients with infection orstress, and parenteral administration of glucocor-ticoids was indicated. The long term consequencesof untreated pituitary deficiencies, which may



have been present for some years prior to diagno-sis, could also have contributed to the morbidityseen among patients with hypopituitarism.

Conclusion

Results from the present study show that pa-tients with hypopituitarism use more healthcare re-sources, are more likely to draw a disability pen-sion and have more absenteeism from work thanthe general population. Further cost analyses areneeded to determine whether improvements in, forexample, diagnostic and surgical procedures andsubstitution therapy can reduce the healthcare costsof patients with hypopituitarism.

Acknowledgements

Paul Verboom, MSc, is gratefully acknowledged for hissupport in completing the analyses. The authors are indebtedto Pharmacia & Upjohn and the Swedish Medical ResearchCouncil (Grant 244115707) for economic support. This arti-cle was presented in part at the Annual Meeting of theEndocrine Society, Minneapolis, Minnesota, USA, in June1997.

References1. Sönksen PH. Replacement therapy in hypothalamo–pituitary

insufficiency after childhood: management in the adult. HormRes 1990; 33 Suppl. 4: 45-51

2. Hakkaart-van Roijen L, Beckers A, Stevenaert A, et al. Theburden of illness of hypopituitary adults with growth hormonedeficiency. Pharmacoeconomics 1998; 14: 395-403

3. Sanmarti A, Lucas A, Hawkins F, et al., Collaborative ODA(Observational GH Deficiency in Adults) Group. Observa-tional study in adult hypopituitary patients with untreatedgrowth hormone deficiency (ODA study): socio-economicimpact and health status. Eur J Endocrinol 1999; 141 (5):481-9

4. Jönsson B. Diabetes: the cost of illness and the cost of control.An estimate for Sweden 1978. Acta Med Scand Suppl 1983;671: 19-27

5. Finkler SA. The distinction between cost and charges. Ann In-tern Med 1982; 96 (1): 102-9

6. KoopmanschapMA, van Ineveld BM. Towards a new approachfor estimating indirect costs of disease. Soc Sci Med 1992; 34(9): 1005-10

7. Gerard K, Donaldson C, Maynard AK. The cost of diabetes.Diabetes Med 1989; 6 (2): 164-70

8. An official investigation: needs and resources in healthcare – ananalysis [in Swedish]. Stockholm: Ministry of Health and So-cial Affairs, 1996

9. The National Social Insurance Board. Disability pensions: costand number of reimbursed pensioners in December 1989 [inSwedish]. Stockholm: Riksförsäkringsverket

10. National Bureau of Statistics. Statistical yearbook of Sweden1998. Stockholm: Official Statistics of Sweden, 1998

11. Lindgren B. The economic impact of illness. In: Abshagen U,Munnich FE, editors. Cost of illness and benefits of drug treat-ment. Munich: W Zuckschwerdt Verlag, 1990: 12-20

12. Drummond M. Cost-of-illness studies: a major headache. Phar-macoeconomics 1992; 2 (1): 1-4

13. ThornerMO, VanceML, Horvath E, et al. The anterior pituitary.In: Wilson JD, Foster DW, editors. Williams textbook of en-docrinology. 8th ed. Philadelphia (PA): WB Saunders Com-pany, 1992: 221-310

14. Rosén T, Bengtsson B-Å. Premature mortality due to cardiovas-cular disease in hypopituitarism. Lancet 1990; 336: 285-8

15. Markussis V, Beshyah SA, Fisher C, et al. Detection of prema-ture atherosclerosis by high-resolution ultrasonography insymptom-free hypopituitary adults. Lancet 1992; 340: 1188-92

16. Bülow B, Hagmar L, Mikoczy Z, et al. Increased cerebrovascu-larmortality in patients with hypopituitarism. Clin Endocrinol1997; 46: 75-81

17. Rosén T, Hansson T, Granhed H, et al. Reduced bone mineralcontent in adult patients with growth hormone deficiency.Acta Endocrinol 1993; 129: 201-8

18. Rosén T, Wilhelmsen L, Landin-Wilhelmsen K, et al. Increasedfracture frequency in adult patients with hypopituitarism. EurJ Endocrinol 1997; 137: 240-5

Correspondence and offprints: Dr Christer Ehrnborg, Endo-crine Division, Department of Medicine, Sahlgrenska Uni-versity Hospital, S-413 45 Gothenburg, Sweden.

628 Ehrnborg et al.


cost of illness in adult patients with hypopituitarism

Documents