73 l)erSimonian R, Laird N. NMeta-analysis in clinical trials. Corntrolled Clin Trials1986;7: 177-88.

74 P'etrie A. Lecture notes on nzedical statistics. Oxford: Blackwell ScientificPublications, 1978.

75 Van Brummelen P, Koolen Ml. Differences in sodium sensitivits in humanhypertensives. Clin Invest Med 1987;10:581-5.

76 Anonymous. Salt and blood pressure: the next chapter [Editorial]. Lancet1989;i: 1301-3.

77 MacMahon S, Peto R, Cutler J, et al. Blood pressure, stroke, and coronaryheart disease. Part 1, prolonged differences in blood pressure: prospectiveobsersational studies corrected for the regression dilution bias. Lancet1990;335:765-74.

78 Collins R, Peto R, NIacMahon S, et al. Blood pressure, stroke and coronaryheart disease. Part 2, short-term reductions in blood pressure: overview ofrandomised drug trials in their epidemiological context. Lancet 1990;335:827-38.

79 Rutan (GH, iNMcDoniald RH, Kullcr LH. A historical perspcctive of elevatedsystolic vs. diastolic blood prcssure from an epidemiological and clinicaltrials viewpoint. 7 Clin Epidemiol 1989;42:663-73.

80 British Hypertension Society Working Party. Trcating mild hypertension.BMJ 1989;298:694-8.

81 Bucknall CA, Morris GK, Mlitchell JRA. Physicians' attittides to fourcommon problems: hypertension, atrial fibrillation, transient ischaemicattacks and angina pectoris. BAt7 1986;293:739-42.

82 James WPT, Ralph A, Sanchez-Castillo CP. The dominance of salt inmanufactured food in the sodium intake of affluent societies. Latncei1987;i:426-9.

(Accepted 29_7anuarv 1991)

Mortality, neoplasia, and Creutzfeldt-Jakob disease in patientstreated with human pituitary growth hormone in the UnitedKingdom

C R Buchanan, M A Preece, R D G Milner

Department of Growth andDevelopment, Institute ofChild Health, LondonWC1N 1EHC R Buchanan, MRCP,clinical lecturerM A Preece, FRCP, professorofchild health and growth

Department of Paediatrics,Sheffield University,Children's Hospital,Sheffield S1O 2THR D G Milner, FRCP,professor ofpaediatnics

Correspondence to:Professor Preece.

BMJ 1991;302:824-8

AbstractObjective-To determine the cause of death

and incidence of neoplasia in patients treated withhuman pituitary growth hormone.Design-A long term cohort study established to

receive details of death certification and tumourregistrations through the Office of PopulationCensuses and Surveys and NHS central register.Patients-All patients (1246 male, 662 female)

treated for short stature with pituitary growthhormone under the Medical Research Council work-ing party and health services human growth hormonecommittee.Main outcome measures-Death or development

of neoplasia.Results-110 patients died (68 male, 42 female;

aged 0 9-57 years) from 1972 to 1990. Fifty threedeaths were from neoplasia responsible for growthhormone deficiency (27 craniopharyngioma, 24other intracranial tumour, two leukaemia); two fromhistiocytosis X; and 13 from pituitary insufficiency.Six patients died of Creutzfeldt-Jakob disease, six ofother neurological disorders, and eight of acuteinfection. Other deaths were apparently unrelated togrowth hormone deficiency or its treatment. Seven-teen tumours (in 16 patients) were identified duringor after growth hormone treatment. Four were inpatients with previous intracranial neoplasia and twowere after cranial irradiation. Thirteen were intra-cranial, the others being Hodgkin's lymphoma,osteosarcoma, carcinoma of colon, and basal cellcarcinoma.

Conclusions - Recuffence or progression of intra-cranial tumours and potentially avoidable metabolicconsequences of hypopituitarism were the maincauses of death. Growth hormone treatment prob-ably did not contribute to new tumour development.Creutzfeldt-Jakob disease after pituitary growthhormone treatment continues to occur in the UnitedKingdom. This cohort must remain under long termreview.

IntroductionTreatment of short stature with human pituitary

growth hormone was first described in 1958.' A similarpreparation was introduced in the United Kingdomunder a Medical Research Council working party in1959, and from 1977 pituitary growth hormone wasavailable through the Health Services Human GrowthHormone Committee.2 By 1985 over 850 patients in the

United Kingdom and Eire were receiving growthhormone, and its use for causes of short statureother than classical growth hormone deficiency wasunder reassessment.3 The only adverse effect ofthis treatment then known was the occurrence ofgrowth hormone antibodies, which rarely inhibited theresponse.24

In 1985 pituitary growth hormone treatment becameassociated with fatal Creutzfeldt-Jakob disease, includ-ing one case in the United Kingdom,`8 and suchpreparations were withdrawn almost world wide. Ourstudy was then instigated to review mortality andmonitor prospectively the long term health of patientsin the United Kingdom treated with growth hormone.We report the mortality and cancer registration datafrom 1959 to December 1990.

Patients and methodsPatients treated with growth hormone from 1959 to

1985 in the United Kingdom and Eire were identifiedfrom Health Services Human Growth Hormone Com-mittee records. These were collated with the NHScentral register and registrar's offices in Scotlandand Northern Ireland to ascertain deaths and cancerregistrations and "flag" survivors so that future deathsand cancer registrations could be notified. In addition,deaths from Creutzfeldt-Jakob disease and similarneurological disorders from 1971 to 1984 werereviewed; none of these patients had been treated withgrowth hormone. About 40 patients from the UnitedKingdom received commercial growth hormone pre-parations and were excluded from the study, aswere patients who received only recombinant growthhormone. 15

Cause of death was determined from death certifi-cates, case notes, and discussion with doctors caringfor the patients. Information from necropsy was avail-able in 44 cases. Comparison between groups was by X2or unpaired Student's t test, with significance at the 5%level.

ResultsOverall, 1908 patients were registered as having

received pituitary growth hormone (fig 1). Treatmentstarted at ages from 1 month (congenital hypo-pituitarism) to 55 years (idiopathic hypopituitarism).Figure 2 shows their distribution by year of birth, andtable I shows diagnoses when treatment started.Patients deficient in two or more anterior pituitary

824 BMJ VOLUME 302 6 APRIL 1991

on 10 June 2021 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.302.6780.824 on 6 A

pril 1991. Dow

nloaded from

http://www.bmj.com/

hormones were designated "panhypopituitary. "Follow up data were insufficient to reclassify patientsaccording to evolution of disease-for example, fromisolated growth hormone deficiency to panhypo-pituitarism. Such revision would have been limited

FIG 1-Cumulativeyearly totalsof the number ofpatients treatedwith human pituitary growthhormone (line) and the number ofdeaths (bars)

-

a)

-o0)0z

Year

TABLE II-Causes of death after treatment with human pituitarygrowth hormone

No of deaths

Cause of death M F Total

Craniopharyngioma 15* 12 27Other intracranial tumour 17 8 25

Medulloblastoma 6 2 8Glioma 2 1 3Astrocytoma 4** 3 7Germinoma 1 1Ependymoma 1 1Other 4 1 5

Other neoplasia:Acute leukaemia 2 2Adenocarcinoma of the colon 1 I

Histiocytosis X 2 2Hypopituitarism 8 5 13Accident or trauma 6 2 8Infection 3 5 8Neuromuscular disorders 3 3 6Creutzfeldt-Jakob disease 4 2 6Other 7 5 12

Total 68 42 110

120-

100-cn4-I

a) 80-

m 60-0

z 40

FIG 2-Number ofpatientstreated with human pituitarygrowth hormone byyear of birth

FIG 3 -Number ofdeaths aftertreatment with human pituitarygrowth hormone by age at death

I

20-

19C204640Year of birth

14

12-

c, 1 0 M

-0

0

z4

Osteosarcoma of the fibula (case 13) developed distantfrom the irradiation field (anterior and lateral beams tothe pituitary area). One patient developed Hodgkin'slymphoma five years after hormone treatment, andone developed the myelodysplastic syndrome aftereight years' treatment for isolated growth hormonedeficiency; Fanconi's anaemia was then diagnosed.'7

Follow up from diagnosis and initial managementof patients with craniopharyngioma, including thoseidentified after starting hormone treatment, rangedfrom 6 to 37 years (mean 17 3, SD 6-7, n=227;missing= 5). When the eight patients dying from othercauses and the five (one death) with missing data wereexcluded 76 patients received radiotherapy duringinitial tumour management; eight of these died oftumour recurrence or complications. This number wasnot significantly different from the 18 deaths among143 patients who did not receive early irradiation.Furthermore, there were proportionately no moredeaths from craniopharyngioma (27 of 232) than otherknown intracranial primary tumours (24 of 202); themean follow up from diagnosis and initial treatmentof these tumours w%s 13 4 years (SD 4 9, n=200;missing=2). Deaths from craniopharyngiomaoccurred at a mean of 10-1 years (SD 5 9; range 1-14)after tumour diagnosis, which did not differ from the7-3 years (51; 0 5-20; 23;1) for other intracranialtumours. Only 25% of patients with intracranialtumours began growth hormone within two years ofdiagnosis, whether this was a craniopharyngioma(mean interval 4-1 years, SD 3-8, n=226; missing=6)or other tumour (4-0, 3-2, 200;2). These patientssurvived early complications and were likely to have abetter prognosis.

Both patients who died of leukaemia had hada previous relapse. One relapsed within a year ofstarting growth hormone (previous central nervoussystem relapse twice and testicular once). The otherbegan treatment four years after testicular relapse butrelapsed again seven months later only to die inremission after bone marrow transplantation. Meta-static colonic carcinoma occurred in a boy withhypopituitarism who had had previous surgery forimperforate anus; his sister died of acute childhoodleukaemia.

HYPOPITUITARISM

Thirteen deaths were attributed to hypopituitarism(table IV), including six of 11 deaths in children undersix years (the others were from medulloblastoma,bronchiolitis, pulmonary fibrosis after nitrogenmustard chemotherapy for medulloblastoma, subduralhaemorrhage, and prolonged convulsion in a child withsepto-optic dysplasia). The three youngest patientswith panhypopituitarism were identified by hypo-

glycaemia in the neonatal period. Only the 15 year oldwas not established on glucocorticoid replacementtreatment before starting growth hormone treatmentaged 5 years. Gonadotrophin deficiency was suggestedby hypospadias, and thyrotrophin and adrenocortico-trophin deficiency (with hypoglycaemic convulsions)developed within two years.

CREUTZFELDT-JAKOB DISEASE

The first of the deaths from Creutzfeldt-Jakobdisease (table V) antedated this study and was exten-sively reported.56 The other cases were ascertainedbefore death, and all but the sixth have been confirmedby neuropathological findings of spongiform en-cephalopathy. All presented with clinical featurestypical of spongiform encephalopathy after pituitaryhormone treatment18: rapidly progressive cerebellardysfunction, ataxia as an early feature (with or withoutmyoclonus), followed by dementia and death within 12months of onset.Three patients (cases 2, 5, and 6) had neither had

neurosurgery nor received donor blood products ortissues. No single batch of growth hormone wascommon to the six patients, but all received growthhormone prepared before column chromatographywas introduced into the purification procedure. Onepatient (case 3) received only four batches of hormoneand another (case 4) never received hormone preparedby the Raben method.2 Three (cases 2, 4, and 6)completed their treatment with recombinant growthhormone.

Other deaths from neuromuscular disordersoccurred in two patients with primary neuromusculardisease and four with epilepsy who died duringseizures without evidence of hypoglycaemia.

OTHER CAUSES OF DEATH

Among six patients who died from respiratoryinfection two had cardiorespiratory complications postpartum (isolated growth hormone deficiency) or post-operatively (jejunal bypass for obesity associated withhypothalamic dysfunction). A girl with hydrocephalusdied of recurrent bacterial meningitis. Other deathsincluded respiratory failure in Prader-Willi syndrome,aortic dissection in Turner's syndrome, asthma in ahypopituitary boy, and a 14 year old boy found hanged("open" verdict at inquest). One death, 17 years aftertreatment for subfrontal neuroblastoma, remainedunexplained.

DiscussionThis is the largest reported study of mortality in

growth hormone deficient patients receiving growthhormone treatment, although subgroups of this cohort

TABLE III- Tumours diagnosed after treatment with human pituitary growth hormone

Time afterAge when start oftumour growth

diagnosed hormoneCase Sex Tumour (years) Indication for growth hormone treatment (years) Outcome

1 F Pituitary* 19 Panhypopituitarism 2 Alive2 F Hypothalamic* 14 Growth hormone deficiency, diabetes insipidus 2 Alive3 F Third ventricle* 12 Growth hormone deficiency, diabetes insipidus 2 Alive4 F Germinoma of pituitary 27 Growth hormone deficiency, diabetes insipidus 11 Alive5 F Germinoma of fossa hypophysialis 13 Isolated growth hormone deficiency 1 Alive6 F Germinoma 21 Isolated growth hormone deficiency, delayed puberty 6 Alive7 F Pinealoma 19 Short stature, delayed puberty 2 Alive8 F Astrocytoma 18 Isolated growth hormone deficiency, temporal lobe epilepsy, tuberous sclerosis 3 Dead9 M Anaplastic astrocytoma 19 Isolated growth hormone deficiency 3 Dead10 M Craniopharyngioma 11 Isolated growth hormone deficiency 2 Alive11 M Anaplastic astrocytoma (1988) (medially right temporal lobe) 21 Craniopharyngioma (1982), 54 Gy cranial irradiation 6 Dead12 M Craniopharyngiomat 18 Panhypopituitarism 1 Alive

Basal cell carcinoma (nose) 25 813 F Osteosarcoma of fibula (1979) 14 Chromophobe adenoma (1974), 45 Gy cranial irradiation 2 Deadt14 M Hodgkin's lymphoma (lymphocyte depleted) 25 Isolated growth hormone deficiency 10 Alive15 M Adenocarcinoma of transverse colon 22 Isolated growth hormone deficiency 14 Dead16 M Astrocytoma (1988) (left frontal region) 17 Acute lymphoblastic leukaemia, 35 Gy craniospinal irradiation 3 Dead

*History unavailable. tNo cranial irradiation. tAsphyxia in house fire.

BMJ VOLUME 302 6 APRIL 1991826

on 10 June 2021 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.302.6780.824 on 6 A

pril 1991. Dow

nloaded from

http://www.bmj.com/

TABLE IV-Deaths from hypopituitarism in patients who received huiman pituitary growth hormone

Age Year ofCause of death Diagnosis (years) death

Hypoglycaemia Panhypopituitarism 3 1985Hypoglycaemia, cerebral infarction, infection Septo-optic dysplasia 3 1987Electrolyte disturbance Panhypopituitarism 4 1972Electrolyte disturbance, diabetes insipidus Septo-optic dysplasia 5 1979Gastroenteritis, diabetes insipidus Craniopharvngioma 5 1974Hypoglyvcaemia Panhypopituitarism 5 1986Epiglottiditis Craniopharyngioma 13 1985Bronchopneumonia, diabetes insipidus Germinoma 14 1985Adult respiratory distress syndrome, viral pneumonia, hypopituitarism Panhypopituitarism 15 1989Acute laryngotracheobronchitis, pneumonia Glioma 17 1987Respiratory infection, asthma, intracranial haemorrhage Craniopharyngioma 18 1973Bronchopneumonia, diabetes insipidus Histiocytosis X 18 1976Hypopituitarism Panhypopituitarism 57 1977

TABLE V-Clinical details of six patients dyingfrom Creutzfeldt-Jakob disease after treatment with humanpituitary growth hormone

Period and duration(years) of pituitary growth Date and age (years)

Case Sex Diagnosis at start of treatment hormone treatment at death

1 M Craniopharyngioma (surgery 1964) 1972-77 (4-5) February 1985 (22)2 M Isolated growth hormone deficiency 1974-85 (9-6) April 1988 (20)3 M Craniopharyngioma (surgery 1968) 1970-74 (3 0) April 1990 (34)4 M Craniopharyngioma (surgery 1972) 1976-85 (9 0) April 1990 (20)5 F Isolated growth hormone deficiency 1974-85 (10-9) July 1990 (20)6 M Isolated growth hormone deficiency 1975-85 (10-3) December 1990 (21)

have been reported. 19-21 Half the deaths resulted fromrecognised brain tumours that were related to thegrowth hormone deficiency. Because of the patients'ages it was predictable that accidents and infectionwould be other main causes of death.22 This mortalitypattern was found in growth hormone treated patientsin France, although the mortality rate was one third ofthat in our study.2'The mortality findings for patients with cranio-

pharyngioma were no better than for those with malig-nant intracranial tumours, representing late tumourrecurrence for both groups. Deaths among patientswith craniopharyngioma were apparently not reducedby early radiotherapy yet cranial irradiation may havecontributed to two fatal second intracranial tumours.Children with tumours of the central nervous systemseemed to have an increased risk of developing asecond tumour in the central nervous system,24 anddevelopment of second tumours within radiotherapyfields after latencies such as the 6-10 years found by ushas been reported.29Our study could not assess not-fatal tumour recur-

rences, and a control group not treated with growthhormone was not available. We cannot thereforeconclude whether growth hormone contributed totumour recurrence and death. Studies of subgroupsof these patients showed no excess recurrence ofcommoner intracranial tumours compared withpatients who had not received growth hormone'920;small studies suggest no increased risk of tumourrecurrence.26 27

First intracranial tumours were identified in 10patients with "idiopathic" growth hormone deficiency.These were probably causally related to the hormonedeficiency but undetectable at the time. Improvedimaging now permits earlier localisation of tumoursalthough clinical features, such as epilepsy, delayedpuberty, and diabetes insipidus, present in some ofthese cases, should alert doctors to the need for regularreassessment and long term follow Up.- 29

Deaths from hypopituitarism should be prevent-able. The risk of hypoglycaemia in young children,aggravated by acute infection, and the added compli-cation of diabetes insipidus cannot be overemphasised.Education should minimise the risks of hypopituitarycrises. The importance of periodic evaluation of pitui-tary function in growth hormone deficient patients

is shown by the death of one boy whose panhypo-pituitarism evolved over several years. Morbidityassociated with non-fatal metabolic disturbances dueto hypopituitarism needs to be reviewed.There have been recent reports of leukaemia

developing after growth hormone treatment9'4 but weidentified no new cases of leukaemia; both instances ofrelapse were at high risk of further relapse. Of fourextracranial malignancies two were second primarytumours (osteosarcoma and basal cell carcinoma).The first (reported previously as osteosarcoma of thefemur25) had only cranial irradiation and neitherhad chemotherapy. Without a control group it isimpossible to assess the relevance of growth hormonedeficiency or treatment to increased risk of tumourformation. Short stature is associated with conditionswith increased risk of malignancy and such patientshave received growth hormone. These conditionsinclude Fanconi's anaemia (two patients died fromnarrow aplasia and infection), neurofibromatosis,Bloom syndrome, Kearns-Sayre syndrome, andbilateral retinoblastoma. These patients should bemonitored with particular care for tumour develop-ment.World wide there have now been 16 cases of

Creutzfeldt-Jakob disease after pituitary growthhormone treatment, the incidence among pituitarygrowth hormone recipients in the United States beingsimilar to that in the United Kingdom (P Brown andJ-C Job, personal communications). Additionally, twocases of the disease have occurred after treatment inAustralia with extracted pituitary gonadotrophins (AHarding, personal communication30). Although thefeared "epidemic" of this disease8 has not yet occurred,the rising number of cases suggests that peopleshould remain alert to public health implications;transmission of Creutzfeldt-Jakob disease may occurthrough donor tissue, infected blood, and urine.7 Fourpatients with the disease in the United Kingdomreceived hormone concurrently in 1974, all six havingtreatment between 1974 and 1976, when about 400patients were undergoing treatment in the UnitedKingdom. As no single hormone batch was common tothese six patients, multiple contaminations with theinfective agent must have occurred. One case of thedisease in the United States was subclinical, diagnosedon review of necropsy tissue."' Only 40% of patientswho died in our study underwent necropsy, andevidence of the disease may have been missed. Allpatients treated with pituitary growth hormone have asmall but presently undefined risk of developingCreutzfeldt-Jakob disease, and general practitionersand neurologists need to be aware of this possibility.Pathologists should heed precautions advised in thesecases,32 and expert neuropathological studies of allfuture deaths among these patients are necessary.The oldest surviving patient treated with pituitary

growth hormone is only 47 years old. Lifelong surveil-lance of these patients is essential to assess fully boththe benefits and possible adverse consequences of theirtreatment.

We thank the physicians who provided clinical details oftheir patients, particularly Dr C M C Allen, Dr H Besterman,Professor H S Jacobs, Dr P Kennedy, and Dr A B Kurtz; DrJanice Anderson, Professor L W Duchen, and ProfessorR 0 Weller for the neuropathological findings in cases ofCreutzfeldt-Jakob disease; the Health Services HumanGrowth Hormone Committee; Miss Joanne Nash and MissCarole Lawrence for their help in collating the data. Thisstudy was funded by the Department of Health.

I Raben N. Treatment of a pituitary dwarf with human growth hormone. J ClinEndocrinolMetab 1958;18:901-3.

2 Milner RDG, Russell-Fraser T, Brook CGD, et al. Experience with humangrowth hormone in Great Britain: the report of the MRC working party.Clin Endocrinol((Oxf) 1979;11:15-38.

BMJ VOLUME 302 6 APRIL 1991 827

on 10 June 2021 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.302.6780.824 on 6 A

pril 1991. Dow

nloaded from

http://www.bmj.com/

3 Buchanan CR, Law CM, Milner RDG. Growth hormone in short, slowlygrowing children and those with Turner's syndrome. Arch Dis Child1987;62:912-6.

4 Preece MA. Experience of treatment with pituitary derived growth hormonewith special reference to immunological aspects. In: Milner RDG, Flodh H,eds. Immunological aspects of human growth hormone. Oxford: MedicalEducation Services, 1986:9-16.

5 Powell-Jackson J, Weller RO, Kennedy P, et al. Creutzfeldt-Jakob diseaseafter administration of human growth hormone. Lancet 1985;ii:244-6.

6 Weller RO, Steart PV, Powell-Jackson JD. Pathology of Creutzfeldt-Jakobdisease associated with pituitary-derived human growth hormone adminis-tration. Neuropathol Appl Neurobiol 1986;12:117-29.

7 Preece MA. Creutzfeldt-Jakob disease: implications for growth hormonedeficient children. Neuropathol Appl Neurobiol 1986;12:509-15.

8 Brown PB, Gaidusek C, Gibbs CJ, Asher DM. Potential epidemic ofCreutzfeldt-Jakob disease from human growth hormone therapy. N EnglJMed 1985;313:728-30.

9 Watanabe S, Tsunematsu Y, Fujimoto J, Komiyama A. Leukaemia in patientstreated with growth hormone. Lancet 1988;i: 1159.

10 Delemarre-Van de Waal HA, Odink JH, de Grauw TJ, de Waal FC.Leukaemia in patients treated with growth hormone. Lancet 1988;i: 1159.

11 Fisher DA, Job J-C, Preece MA, Underwood LE. Leukaemia in patientstreated with growth hormone. Lancet 1988;i: 1159-60.

12 Frisch H, Thun-Hoenstein L, Balzar E. Leukaemia and growth hormone.Lancet 1988;i:1335.

13 Redman GP, Shu S, Norris D. Leukaemia and growth hormone. Lancet1988;i: 1335.

14 Stahnke N, Zeisel HJ. Growth hormone therapy and leukaemia. EurJf Pediatr1989;148:591-6.

15 Milner RDG, Barnes ND, Buckler JMH, et al. United Kingdom multicentreclinical trial of somatrem. Arch Dis Child 1987;62:776-9.

16 Tanner JM, Whitehouse RH, Hughes PCR, Vince FP. Effect of humangrowth hormone treatment for 1 to 7 years on growth of 100 children withgrowth hormone deficiency, low birth weight, inherited smallness, Turner'ssyndrome and other complaints. Arch Dis Child 1971;46:745-82.

17 Standen GR, Hughes IA, Geddes D, Jones BM, Wardrop CAJ. Myelodys-plastic syndrome with trisomy 8 in an adolescent with Fanconi anaemia andselective IgA deficiency. Amj Hematol 1989;31:280-3.

18 Brown PB. The decline and fall of Creutzfeldt-Jakob disease associated withhuman growth hormone therapy. Neurology 1988;38:1135-7.

19 Clayton PE, Shalet SM, Gattamaneni HR, Price DA. Does growth hormonecause relapse of brain tumours. Lancet 1987;i:711-3.

20 Clayton PE, Price DA, Shalet SM, Gattamaneni HR. Craniopharyngiomarecurrence and growth hormone therapy. Lancet 1988;i:642.

21 Whitehead HM, Hadden DR, Carson DJ. The Northern Ireland experience ofgrowth hormone therapy for short stature. Ulster MedJ 1989;58:153-60.

22 Office of Population Censuses and Surveys. Mortality statistics. Review of theregistrar general on deaths by cause, sex and age in England and Wales. London:HMSO, 1988. (Series DH2 1986.)

23 Gouiard J, Entat M, Maillard F, et al. Human pituitary growth hormone(hGH) and Creutzfeldt-Jakob disease: results of an epidemiological surveyin France, 1986. Int7Epidemiol 1988;17:423-7.

24 Farwell J, Flannery JT. Second primaries in children with central nervoussystem tumours. J Neurooncol 1984;2:371-5.

25 Kingston JE, Hawkins MM, Draper GJ, Marsden HB, Wilson LMK. Patternsofmultiple primary tumours in patients treated for cancer during childhood.Br7 Cancer 1987;56:331-8.

26 Arslanian SA, Becker DJ, Lee PA, Drash LA, Foley TP. Growth hormone andtumor recurrence. Findings in children with brain neoplasms and hypo-pituitarism. AmJ Dis Child 1985;139:347-50.

27 Rodens KP, Kaplan SL, Grumbach MM, Teller WM. Does growth hormonetherapy increase the frequency of tumor recurrence in children with braintumors? Acta Endocrinol (Copenh) 1987;283(suppl): 188-9.

28 Sherwood MC, Stanhope R, Preece MA, Grant DB. Diabetes insipidus andoccult intracranial tumours. Arch Dis Child 1986;61:1222-5.

29 Stanhope R, Preece MA, Brook CGD, Grant DB. Is diabetes insipidus duringchildhood ever idiopathic? BrJ Hosp Med 1989;41:490-1.

30 Cochius JI, Mack K, Burns RJ, Alderman CP, Blumbergs PC. Creutzfeldt-Jakob disease in a recipient ofhuman pituitary-derived gonadotrophin. AustNZJMed 1990;20:592-3.

31 New MI, Brown P, Temeck JW, et al. Preclinical Creutzfeldt-Jakob diseasediscovered at autopsy in a human growth hormone recipient. Neurology1988;38: 1133-4.

32 Committee on Health Care Issues, American Neurological Association. Pre-cautions in handling tissues, fluids, and other contaminated materials frompatients with documented or suspected Creutzfeldt-Jakob disease. AnnNeurol 1986;19:75-7.

(Accepted 107anuary 1991)

Coppetts Wood Unit,Royal Free HospitalDepartment of Infectiousand Tropical Diseases,London N10M G Jacobs, MB, senior houseofficerM G Brook, MD, seniorregistrarW R C Weir, MRCP,consultant physicianB A Bannister, FRCP,consultant physician

Correspondence andrequests for reprints to:Dr Brook.

BMJ 1991;302:828-9

Dengue haemorrhagic fever: arisk of returning home

M G Jacobs, M G Brook, W R C Weir,B A Bannister

The number of imported cases of infection withdengue virus each year is rising as a consequence ofincreased travel and an increase in the worldwideincidence of dengue.' Almost all travellers who becomeinfected with the virus experience a benign, ifuncomfortable, febrile illness. We report on twoadult residents of the United Kingdom who visitedtheir country of origin and developed a severe, lifethreatening complication of this infection known asdengue haemorrhagic fever.

Case reportsCase I-A 73 year old Malaysian woman had been

living in the United Kingdom for many years. Shespent two months visiting Malaysia and on her returncomplained ofunremitting fever, anorexia, and nausea.By the fourth day she had developed nose bleeds,haemoptysis, haematuria, and a generalised petechialrash. Tests showed severe thrombocytopenia (plateletcount 3 x 1I /1) with a low white cell count (totalwhite cells 1 6x 0l/l; neutrophils 0 48x 10/Il) andhaemoglobin concentration (86 g/l). Virus relatedhaemophagocytic syndrome was diagnosed on bonemarrow aspiration. Management consisted of treat-ment with broad spectrum antibiotics appropriate forneutropenia and platelet transfusions. She made anuneventful recovery. Infection with dengue virus wasdiagnosed by an eightfold rise in antibody titres.

Case 2-A 41 year old Pakistani man returned to theUnited Kingdom, where he had been living for manyyears, after a two week visit to Pakistan. He developeda confluent petechial rash after six days of fever,nausea, diarrhoea, and headache. Tests showedthrombocytopenia (platelet count 18 x 109/1) and a

severe coagulopathy (partial thromboplastin time 66 s;thrombin time >120 s). He was treated with bloodproducts and made an uneventful recovery. Infectionwith dengue virus was diagnosed by an eightfold rise inantibody titres.

CommentDengue viruses, comprising four distinct serotypes

of flavivirus, are transmitted from infected to sus-ceptible humans by Aedes mosquitoes. This vectoris ubiquitous in the tropics, where dengue is a majorpublic health problem. The first infection withdengue virus, usually in early childhood, may passunrecognised or cause a self limiting febrile illness.This is followed by lifelong homotypic immunity, butafter a few months of cross protection the subjectis susceptible to infection with other serotypes.1Considerable evidence suggests that haemorrhagiccomplications are associated with subsequent infectionwith a second serotype. This explains an earlyobservation that foreign visitors who have notbeen exposed to dengue virus previously developuncomplicated dengue during outbreaks of denguehaemorrhagic fever among indigenous people.4 Itis believed that "immune enhancement" of viralreplication underlies the development of severe,haemorrhagic disease: non-neutralising, cross reactingantibodies opsonise virus of a second serotype andenhance uptake into mononuclear phagocytes, inwhich the virus replicates.'

Large epidemics of dengue haemorrhagic feveramong children are frequently reported in the tropics,particularly in South East Asia and the Caribbean, withan associated mortality of 2-10%. This syndrome israre in adults and has not, we believe, been reportedbefore in travellers returning to the United Kingdom.We presume that our patients acquired denguehaemorrhagic fever as a result of infection with asecond serotype of dengue virus after a delay of manyyears since their first exposure to the virus. Thispotential problem should be remembered wheneverpeople return to the tropics, and they should be

828 BMJ VOLUME 302 6 APRIL 1991

on 10 June 2021 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.302.6780.824 on 6 A

pril 1991. Dow

nloaded from

http://www.bmj.com/