318 HAEMODILUTION IN ACUTE STROKE: RESULTS OF THE ITALIAN HAEMODILUTION TRIAL* ITALIAN ACUTE STROKE STUDY GROUP Summary In a multicentre clinical trial 1267 patients with hemispheric stroke of duration 12 h or less and haematocrit of 35% or more were prospectively randomised to either haemodilution (by venesection and replacement of the same volume of dextran 40 in saline solution) or control treatment groups. In the haemodiluted group mean haematocrit declined from 43% to 37% at 48 h and this fall was maintained for seven days. A plain computed tomographic scan was obtained in all but 37 patients. 87% of the strokes were infarcts and 13% were haemorrhages. After six months the numbers of dead or severely disabled patients were equally distributed in the two treatment groups, and this was true also within the ischaemic and haemorrhagic subgroups. Furthermore, haemodilution did not improve outcome either in the group with very recent ischaemic stroke (< 6 h) or in the subgroup with highest haematocrit (>45%). Thus, moderate haemodilution does not improve the outcome in acute stroke patients. Introduction EXPERIMENTAL work has shown that a functional neurological deficit can be reversed and the size of a necrotic lesion reduced if reperfusion or improvement of the collateral flow is obtained in the first hours after an ischaemic brain insult.1 Therefore treatments have been proposed to improve cerebral blood flow (CBF) in the early phase of stroke, including haemodilution therapy, which increases cerebral perfusion by lowering the haematocrit, the major determinant of whole-blood viscosity.z°3 When haematocrit is kept higher than 33 %, haemodilution increases CBF without reducing oxygen-carrying capacity,’ 4 and this effect is not confined to patients with polycythaemia. Despite the interest in this form of treatment, no clinical trials have been conducted in very acute phases of the disease. We report the results of a multicentre controlled prospective trial conducted on 1267 stroke patients randomised to haemodilution or control treatment within 12 h after their first symptoms. The outcome was evaluated at the time of discharge from acute care and again after six months. Patients and Methods All patients admitted to hospital for their first hemispheric stroke were enrolled. We then excluded comatose patients in whom there was no evidence of unilateral motor deficit and, at the other extreme of clinical severity, patients whose symptoms had resolved by the time of randomisation. (Advanced age was not an exclusion criterion.) After the initial clinical evaluation patients with haematocrit of < 35 % or with medical features that might interfere with the haemodilution procedure were also excluded. The principal features of the protocol have been described elsewhere.5 In the period November, 1984, to December, 1986, thirty-one Italian centres randomised 1267 cases within 12 h of onset of their first symptoms. The randomisation lists were centrally located and stratified for centre and severity of the neurological deficit. Informed consent from the patient or a relative was obtained in all *Clinical Coordinating Centres.—Istituto di Clinica Neurologica, Università di Milano, and III Cattedra di Clinica Neurologica, Università di Roma, Italy. Names of all participants are listed at the end of the article. TABLE I-CHARACTERISTICS AT RANDOMISATION cases. After preliminary blood sampling the haemodilution procedure was as follows: 20 ml haptenic dextran (’Promiten’, Pharmacia) was given iv, followed by 350 ml dextran 40 in 0-9% saline (’Rheomacrodex’, Glaxo) infused in 60 min. 350 ml of blood was removed by venesection in 15-30 min, starting 10 min after the dextran 40 infusion to avoid the risk of hypovolaemia. Haematocrit was determined again 6-12 h and 12-24 h after the first procedure. The dextran 40 (350 ml) and venesection (350 ml) were repeated in patients who still had a haematocrit of over 35 % on either occasion - ie, no patient had more than three haemodilutions. Since it was not possible to use a sham procedure the patients in the control group received only the most appropriate general treatment, the same as that provided for the patients in the haemodilution group. Haematocrit values were determined daily for the next seven days. A computed tomographic (CT) scan without contrast medium was obtained in 1230 patients (97%), within 48 h after stroke in 908 (74%). The patients were evaluated at the time of discharge from the acute care hospital (neurological status being recorded as coma, stupor, severe motor deficit, slight motor deficit, deficit other than motor) and six months after randomisation, by telephone interview, to assess their disability status. We chose to evaluate disability status by telephone so as to maintain blindness with respect to treatment allocation and to avoid interobserver variability. In fact the same member of the clinical coordinating centre (GP) conducted all the interviews with patients, relatives, and family doctors. The follow-up was completed in all but 1 patient. The interviews were conducted directly with the patients in 70% of cases. The family doctor gave most of the information on the causes of death and on the treatments during follow-up. The disability status due to stroke was assessed according to a modified version of the Rankin classification:6 1. No significant disability: able to carry out all previous usual activities. 2. Slight disability: unable to carry out all previous activities but able to attend to bodily needs without assistance. 3. Moderate disability: requiring some help for bodily needs and with walking. 4. Moderately severe disability: requiring some help for bodily needs and unable to walk without assistance or physical device.
Transcript
318
HAEMODILUTION IN ACUTE STROKE:RESULTS OF THE ITALIAN HAEMODILUTION
TRIAL*
ITALIAN ACUTE STROKE STUDY GROUP
Summary In a multicentre clinical trial 1267 patientswith hemispheric stroke of duration 12 h or
less and haematocrit of 35% or more were prospectivelyrandomised to either haemodilution (by venesection andreplacement of the same volume of dextran 40 in salinesolution) or control treatment groups. In the haemodilutedgroup mean haematocrit declined from 43% to 37% at 48 hand this fall was maintained for seven days. A plaincomputed tomographic scan was obtained in all but 37patients. 87% of the strokes were infarcts and 13% werehaemorrhages. After six months the numbers of dead orseverely disabled patients were equally distributed in thetwo treatment groups, and this was true also within theischaemic and haemorrhagic subgroups. Furthermore,haemodilution did not improve outcome either in the groupwith very recent ischaemic stroke (< 6 h) or in the subgroupwith highest haematocrit (>45%). Thus, moderatehaemodilution does not improve the outcome in acute strokepatients.
Introduction
EXPERIMENTAL work has shown that a functional
neurological deficit can be reversed and the size of a necroticlesion reduced if reperfusion or improvement of thecollateral flow is obtained in the first hours after anischaemic brain insult.1 Therefore treatments have been
proposed to improve cerebral blood flow (CBF) in the earlyphase of stroke, including haemodilution therapy, whichincreases cerebral perfusion by lowering the haematocrit,the major determinant of whole-blood viscosity.z°3 Whenhaematocrit is kept higher than 33 %, haemodilutionincreases CBF without reducing oxygen-carrying capacity,’ 4and this effect is not confined to patients with
polycythaemia. Despite the interest in this form of
treatment, no clinical trials have been conducted in veryacute phases of the disease. We report the results of amulticentre controlled prospective trial conducted on 1267stroke patients randomised to haemodilution or controltreatment within 12 h after their first symptoms. Theoutcome was evaluated at the time of discharge from acutecare and again after six months.
Patients and Methods
All patients admitted to hospital for their first hemispheric strokewere enrolled. We then excluded comatose patients in whom therewas no evidence of unilateral motor deficit and, at the other extremeof clinical severity, patients whose symptoms had resolved by thetime of randomisation. (Advanced age was not an exclusioncriterion.) After the initial clinical evaluation patients withhaematocrit of < 35 % or with medical features that might interferewith the haemodilution procedure were also excluded. The
principal features of the protocol have been described elsewhere.5In the period November, 1984, to December, 1986, thirty-one
Italian centres randomised 1267 cases within 12 h of onset of theirfirst symptoms. The randomisation lists were centrally located andstratified for centre and severity of the neurological deficit.Informed consent from the patient or a relative was obtained in all
*Clinical Coordinating Centres.—Istituto di Clinica Neurologica, Universitàdi Milano, and III Cattedra di Clinica Neurologica, Università di Roma,Italy.Names of all participants are listed at the end of the article.
TABLE I-CHARACTERISTICS AT RANDOMISATION
cases. After preliminary blood sampling the haemodilution
procedure was as follows: 20 ml haptenic dextran (’Promiten’,Pharmacia) was given iv, followed by 350 ml dextran 40 in 0-9%saline (’Rheomacrodex’, Glaxo) infused in 60 min. 350 ml of bloodwas removed by venesection in 15-30 min, starting 10 min after thedextran 40 infusion to avoid the risk of hypovolaemia. Haematocritwas determined again 6-12 h and 12-24 h after the first procedure.The dextran 40 (350 ml) and venesection (350 ml) were repeated inpatients who still had a haematocrit of over 35 % on either occasion- ie, no patient had more than three haemodilutions. Since it wasnot possible to use a sham procedure the patients in the controlgroup received only the most appropriate general treatment, thesame as that provided for the patients in the haemodilution group.Haematocrit values were determined daily for the next seven days.A computed tomographic (CT) scan without contrast medium wasobtained in 1230 patients (97%), within 48 h after stroke in 908(74%). The patients were evaluated at the time of discharge fromthe acute care hospital (neurological status being recorded as coma,stupor, severe motor deficit, slight motor deficit, deficit other thanmotor) and six months after randomisation, by telephone interview,to assess their disability status. We chose to evaluate disability statusby telephone so as to maintain blindness with respect to treatmentallocation and to avoid interobserver variability. In fact the samemember of the clinical coordinating centre (GP) conducted all theinterviews with patients, relatives, and family doctors. The
follow-up was completed in all but 1 patient. The interviews wereconducted directly with the patients in 70% of cases. The familydoctor gave most of the information on the causes of death and onthe treatments during follow-up. The disability status due to strokewas assessed according to a modified version of the Rankinclassification:6
1. No significant disability: able to carry out all previous usualactivities.2. Slight disability: unable to carry out all previous activities butable to attend to bodily needs without assistance.3. Moderate disability: requiring some help for bodily needs andwith walking.4. Moderately severe disability: requiring some help for bodilyneeds and unable to walk without assistance or physical device.
319
Haematocrit in haemodiluted patients (open symbols) and controlpatients (closed symbols).
Means and SD.
5. Severe disability: unable to attend to bodily needs withoutassistance and/or to walk.6. Very severe disability: bedridden, incontinent, and requiringconstant nursing care.Statistical evaulation was by the log-rank test’ (two-tailed
assessment). We used a dichotomous subdivision to comparepatients with favourable outcome (surviving cases ranking 1 or 2 onthe disability scale) with patients with unfavourable outcome (deador if surviving ranking from 3 to 6 on the disability scale). For allcases and for each subgroup the number of subjects in whom anunfavourable outcome was observed (0) was compared with thenumber expected (E) in the absence of any real effect of treatment.The variance of O-E was used to estimate the odds ratio (OR) andits confidence intervals.8 If treatment was beneficial the OR wouldtend to be < 1 ’0.
Results
702 (55%) of the 1267 patients were randomised within 6h after their first symptom and 565 (45%) within 6-12 h.There were 633 in the haemodilution group and 634 in thecontrol group. The two groups were comparable for all thebaseline characteristics, the differences for age andhaematocrit not being statistically significant (table I).The haemodilution procedure reduced the mean
haematocrit from 42-7% (SD 3-9) at baseline to 39-6% (4-4)at 6-12 h, 37.9% (3-9) at 12-24 h, and 37-2% (3-8) at 48 h.The 13% reduction of haematocrit was maintained for atleast seven days (figure). In the control group the meanhaematocrit did not vary significantly from the first to theseventh day. The haemodilution procedure required threebloodlettings (1050 ml in 24 h) in 359 (57%) of the 633haemodiluted patients, two (700 ml) in 115 (18%), and only
TABLE II-CASE FATALITIES AND SEVERITY OF NEUROLOGICAL
IMPAIRMENT AT DISCHARGE FROM ACUTE CARE HOSPITAL
*xz=114; df= 5, p (two tailed) = 0.95.
TABLE III-OUTCOME SIX MONTHS AFTER RANDOMISATION
*1 case was lost to follow-up; tincluding haemorrhagic infarcts.
one (350 ml) in 147 (23%). Venesection was not performedin 12 patients (2%), who were kept in the haemodilutiongroup in accordance with the intention-to-treat principle.On analysis of the results in the total population and in the
subgroups the outcome of patients allocated to thehaemodilution group did not differ significantly from that ofpatients allocated to the control group. By the time ofdischarge from acute care 103 patients (16%) in the treatedgroup and 93 (14%) in the control group had died; thesurviving patients were equally distributed for severity ofneurological impairment (table II). After six months the casefatalities and disability status of the two study groups againdid not differ, and this was true within the ischaemic andhaemorrhagic subgroups as well (table ni).
In 14 patients the final diagnosis of stroke was notconfirmed (8 brain tumour or metastasis, 3 subdural
haematoma, 2 postepileptic focal motor deficit, and 1
subarachnoid haemorrhage) and in 37 cases CT scan was notdone.Of the 520 acute ischaemic stroke patients who received
haemodilution treatment 263 (51%) had an unfavourableoutcome (dead or severely disabled) at six months,
TABLE IV-UNFAVOURABLE OUTCOME (DEAD OR DEPENDENT) SIX MONTHS AFTER RANDOMISATION
320
compared with 260 (49%) of 531 in the control group. Theodds ratio between the haemodilution and control groupswas 1-07 and though the analysis tended to favour thecontrols there was no statistically significant difference(2p=060, table iv).
. In the patients with very recent ischaemic stroke ( < 6 h)no advantage was found for haemodilution (OR 1 01,2p=0-93), and in the subgroup with highest haematocrit( > 45%) again there was no significant difference betweenthe two groups (OR =1 40, 2p = 0’19). 18 patients hadside-effects possibly related to haemodilution: cutaneousanaphylactic reaction in 2, readily reversed on termination ofdextran 40 infusion; acute heart failure in 7, soon aftervenesection; transient hypotension in 5; rupture of aorticaneurysm in 1; neurological deterioration in 1; uppergastrointestinal haemorrhage in 1; and jaundice in 1. 2
patients died during or immediately after the haemodilutionprocedure.
Discussion
This study has yielded strong evidence that reduction ofthe haematocrit from 43% to 37% does not improve short orlong term outcome after an acute hemispheric stroke. Thelack of effect of haemodilution is evident in both ischaemicand haemorrhagic strokes as well as in patients with veryrecent lesions and in patients with high haematocrit at’entry. The sample size was adequate. Previous controlledstudies9-1z with haemodilution and plasma volume
expanders have been performed with fewer cases and,because of the clinical variability of acute stroke patients, theresults probably have low statistical value. Moreover themode of haemodilution adopted in our study differssomewhat from that used in previous studies;hypovolaemic13 and hypervolaemic14 haemodilution, bothpotentially harmful, were avoided by replacing the bloodremoved with the same volume of dextran infusion. Twoother factors that could explain our negative results must beemphasised. The first is the interval between symptoms andtreatment. Positron emission tomography has shown thatregions of low flow and preserved oxygen metabolismpersist for no longer than 24-48 h after stroke. 15 All ourpatients were treated within 12 h, and 55 % within 6 h. Thisis the main difference between the present trial and theScandinavian Haemodilution Study16 in which only 28% ofcases were entered within 12 h. Nevertheless the results ofthe two studies are comparable. Moreover, in the presentstudy, analysis of the subgroup of patients in whomtreatment was started within 6 h and completed within amaximum of 30 h after stroke does not show any benefit ofhaemodilution in terms of fatalities and disability at sixmonths.
Secondly, the value of haematocrit reduction must also beconsidered. Wood and Kee4 suggested that a haematocrit of33% is ideal for improvement in blood flow and oxygendelivery. However, since an increase in flow is to a largeextent achieved by an increase in cardiac output, whichtends to decline in elderly subjects,17 35% was fixed as theoptimum in the present study, and even this strategy provednot to be free of adverse effects. The mean haematocritreduction of only 13% may be regarded as a limitation,14 butthe reduction of 16% obtained in the patients withhaematocrit > 45% did not secure any further reduction ofthe unfavourable events.
In agreement with Grotta18 we conclude that therapeutictrials aimed at improving CBF in the acute stage of cerebral
infarction have not shown any benefit, and this applies inparticular to haemodilution.This work was supported in part by the National Research Council of Italy,
Special Project of Preventive Medicine (subproject atherosclerosis). TheItalian Acute Stroke Study Group consists of the following investigators:
Clinical Coordinatzng Centre.-L. Candelise (principal investigator); R.Colombo, M. Rango, G. Pinardi (Universita’ degli Studi, Milano, Istituto diClinica Neurologica); C. Fieschi, C. Argentino, F. Giubilei, M. Rasura(Universita’ di Roma, III Cattedra di Clinica Neurologica); organisationalstaff: S. Cappellari, A. Casazza, L. Grilli (Istituto di Clinica Neurologica,Milano).
Monitoring Committee.-C. P. Warlow, Department of Clinical
Neurosciences, University of Edinburgh; R. Peto, Clinical Trials ServiceUnit, Radcliffe Infirmary, Oxford.
Participating Centres (in alphabetical order of place).
Universita’ degli Studi, Ancona, Istituto delle Malattie del SistemaNervoso (L. Provinciali, M. Del Pesce, P. Di Bella, F. Logullo); PresidioOspedaliero, Aosta, Servizio di Neurologia (A. Mamoli, G. D’Alessandro,M. Camerlingo, E. Bottacchi, P. Gambaro); Presidio Ospedaliero, Arezzo,Unita’ Operativa Neurologia (M. Ghezzi, A. Bianchi, G. M. Malentacchl)and Servizio TC-Radiologia (C. Refi); Universita’ degli Studi, Ban, I ClinicaNeurologica (F. Federico, G. Deluca, V. Lucivero, A. Fiore, P. Lamberti, E.Ferrari); Universita’ degli Studi, Brescia, Clinica Neurologica (L. A. Vignolo,V. Di Monda, D. Vangi, G. Dalla Volta, C. Lazzari, M. Magoni); OspedaleRegionale Busto Arsizio- VA, Divisione Neurologica (P. Perrone, D. Porazzi,G. Bonzi, P. Secchi); Presidio Ospedaliero, Chiari-BS, Divisione
Neurologica (L. Faggi, E. Donati, L. Callea); Universita’ degli Studi, Chiett,Clinica Neurologica (D. Gambi, L. Marchionno, L. Aquilone, R. Cutarella,S. D’Annunzio); Istituti Ospedalieri, Cremona, Divisione Neurologica (L.De Lorenzi, G. Chiodello, G. Galli, M. Guarneri); Universita’ degli Studi,Firenze, Istituto di Clinica Neurologica I (D. Inzitari, P. Marini, A.Ginanneschi, P. Nencini, D. Sita’) and Ospedale di Careggi, Medicinad’Urgenza (P. Cioni); Ospedale Civile S. Martino, Genova, DivisioneNeurologica (L. Garello, P. Piola, G. A. Ottonello, M. Fonzari); PresidioOspedaliero, Mantooa, Divisione Neurologica (F. Mironi, P. Previdi);Ospedale Generale Regionale, Niguarda-Ca’ Granda, Milano, DivisioneNeurologica (F. Erminio, S. Cappa, R. Sterzi, N. Galbiati); Universita’ degliStudi, Milano, Istituto di Clinica Neurologica (G. Scarlato, L. Candelise, R.Colombo, M. Rango, G. Pinardi) and Ospedale Maggiore, Medicinad’Urgenza (A. Randazzo, F. Porro); Ospedali Riuniti, Parma, DivisioneMedicina Generale IV (0. Ponari, M. Squeri, A. Mombelloni); Universita’degli Studi, Pavia, Clinica Neurochirurgica (G. L . Brambilla, G.
Sangiovanni, V. Silvani) and Clinica Medica II (P. Fratino, D. Rogledi, M.G. Ranieri); Universita’ degli Studi, Perugza, Istituto di Semeiotica Medica(G. Nenci, E. Boschetti, A. Scortecci, M. Taramelli) and Neuroradiologia (C.Chiurulla); Univesita’ degli Studi, Perugia, Clinica Neurologica (L. Agostini,S. Ricci, P. Brustenghi, M. G. Scarcella) and Neuroradiologia (E. Signorini);Universita’ degli Studi, Perugia, II Clinica Medica Generale e TerapiaMedica (U. Senin, G. Aisa, G. Valigi) I Clinica Medica Generale e TerapiaMedica (A. Del Favero), and Clinica Neurologica (V. Gallai); OspedaleRegionale, Sede "R. Silvestrini", Perugia, Divisione Medicina (F. Santucci,P. Caselli, G. L. Piccinin); Universita’ degli Studi, Pisa, Clinica Neurologica(A. Muratorio, C. Giraldi, A. Martini, L. Valpentesta, M. C. Masi, D.Bizzarri); Presidio Ospedaliero, Poggibonsi-SI, Divisione Medica (M.D’Ettorre, M. B. Biotti, V. Catenni) and Universita’ degli Studi di Siena,Clinica Neurologica (G. M. Fabrizi) and Neuroradiologia (C. Veatun);Ospedale S. Maria delle Croci, Ravenna, Divisione Neurologica (G. G.Rebucci, G. Bissi, F. Benazzi); Universita’ degli Studi, Roma, III Cattedra diClinica Neurologica (C. Argentino, M. Rasura, F. Giubilei, M. Silvestrini,M. L. Sacchetti, D. Toni); Ospedale Regionale, Saronno- VA, DivisioneNeurologica (D, Zerbi, A. Gomitoni); Ospedale Regionale Bolognini)Seriate-BG, Divisione Neurologica (C. A. De Fanti, A. Brambilla, L. C.Erli); Ospedale Regionale, Sondrio, Divisione Neurologica (M. Lozza, STomaiuolo, I. Votta, D. Baldini); Ospedale Civile, Taranto, DivisioneNeurologica (L. Sallusto, N. Gigante, A. Calogero, F. Pignatelli, S. Intemo’,B. Rotunno, F. Lincesso) and Divisione Neuroradiologica (C. Scapati),Ospedale SS. Trinita’, Varallo Sesia-VC, Divisione Neurologica (E.Nardozza, M. Dellasette); Ospedale Civile Maggiore, Verona, DivisioneNeurologica (M. Montanari, P. Bovi, G. Gambina, L. Deotto) andNeuroradiologia (A. Beltramello); Stabilimento Ospedaliero, IceM.Divisione Neurologica (F. Ferro Milone, G. D’Andrea, P. Parisen-Totdin,F. Bortolon, M. Morra, A. Cananzi).
Correspondence should be addressed to Dr Livia Candelise, Istituto dtClinica Neurologica, Universita degli Studi, Via F. Sforza 35, 20122 Milano,Italy.
321
1. Heiss WD. Flow thresholds of functional and morphological damage of brain tissue.Stroke 1983; 14: 329-31.
2. Thomas DJ. Haemodilution in acute stroke. Stroke 1985; 16: 763-64.3. Grotta J, Ackerman R, Correia J, Fallick G, Chang J. Whole blood viscosity
parameters and cerebral blood flow. Stroke 1982; 13: 296-301.4 Wood JH, Kee DB. Hemorheology of the cerebral circulation in stroke. Stroke 1985;
16: 765-725. Italian Acute Stroke Study Group. The Italian hemodilution trial in acute stroke.
Stroke 1987; 18: 670-76.6. Rankin J. Cerebrovascular accidents in patients over the age of 60 II, prognosis Scott
Med J 1957; 2: 200-15.7. Peto R, Pike MC, Armitage P, et al. Design and analysis of randomized clinical trials
requiring prolonged observation of each patient. I. Introducton and design. Br JCancer 1977; 35: 1-39.
8. Yusuf S, Peto R, Lewis J, Collins R, Sleight P. Beta blockade during and aftermyocardial infarction: an overview of the randomised trials. Prog Cardiovasc Dis1985; 27: 335-71.
9 Gilroy J, Barnhart M, Meyer J S. Treatment of acute stroke with dextran 40. JAMA1969; 210: 293-98
10. Spudis EV, De La Torre E, Pikula L. Management of completed strokes with dextran40. A community hospital failure. Stroke 1973, 4: 895-98.
11. Matthews WB, Oxbury JM, Grainger KMR, Greenhall RCD. A blind controlled trialof dextran 40 in the treatment of ischaemic stroke. Brain 1976; 99: 193-206.
12. Strand T, Asplund K, Eriksson S, Hagg E, Lithner F, Wester PO. A randomisedcontrolled trial of hemodilution therapy m acute ischemic stroke. Stroke 1984; 15:980-89.
13. Kiraly JF, Feldmann JE, Wheby MS. Hazards of phlebotomy in polycythemicpatients with cardiovascular disease. JAMA 1976; 236: 2080-81.
14. Wood JH, Simeone FA, Kron RE, Snyder LL. Experimental hypervolemichemodilution physiological correlations of cortical blood flow, cardiac output, andintracranial pressure with fresh blood viscosity and plasma volume. Neurosurgery1984; 14: 709-23.
15. Wise RJS, Bernardi S, Frackowiack RSJ, Legg NJ, Jones T. Serial observations on thepathophysiology of acute stroke. Brain 1983; 106: 197-222
16. Scandinavian Stroke Study Group. Multicenter trial of hemodilution in acuteischemic stroke. I. Results in the total patient population. Stroke 1987; 18: 691-99.
17 Kusunoki M, Kimura K, Nakamura M, Isaka Y, Yoneda S, Abe H. Effects ofhematocrit variations on cerebral blood flow and oxygen transport in ischaemiccerebrovascular disease. J Cereb Blood Flow Metab 1981; 1: 413-17.
18. Grotta JC: Can raising cerebral blood flow improve outcome after acute cerebralinfarctions? Stroke 1987; 18: 264-67.
GROWTH WITHOUT GROWTH HORMONE:THE "INVISIBLE" GH SYNDROME
T. BISTRITZER
J. C. LOVCHIKS. A. CHALEW
A. A. KOWARSKI
Department of Pediatrics, University of Maryland School ofMedicine, Baltimore, Maryland 21201, USA
Summary Growth hormone (GH) deficiency, diag-nosed by radioimmunoassay (RIA)
measurements of GH in blood after provocation or incontinuous 24 h samples of venous blood, is usuallyassociated with growth failure. 4 non-obese boys have beenidentified who had normal linear growth despite apparentGH deficiency by standard RIA. All 4 patients had normalGH concentrations as measured with an IM-9 cell radio-
receptor assay (RRA). The RRA/RIA ratio of the 4 patientssignificantly exceeded that of controls. Thus, these patientssecrete a molecule with normal GH receptor binding andbioactivity which is "invisible" to the standard GH RIA.This variant GH is possibly expressed from the humanGH-V gene or a mutant allele.
Introduction
GROWTH hormone (GH) has a crucial role in promotingnormal linear growth during childhood. Deficient growthhormone secretion leads to growth failure. The GH radio-immunoassay (RIA) is the standard method for determin-ation of growth hormone levels in blood samples. Sub-normal concentrations of GH by RIA in samples obtainedafter provocation or in 24 h blood collections are theestablished diagnostic tests for GH deficiency. 1-5 Allchildren who appear to have GH deficiency are expected toshow growth impairment.Growth without GH is difficult to reconcile with our
present understanding of the role of GH in promotinggrowth. Several children with craniopharyngioma havebeen reported to show normal or even excessive growth aftersurgical removal of the tumour and subsequent documen-tation of GH deficiency.6-9 The growth of these patients hasbeen attributed to hyperinsulinism secondary to obesityinduced by hypothalamic damage.8 In another child lackingGH normal growth was attributed to a putative growthfactor that was not detected with a GH radioreceptor assay. 10We describe here 4 non-obese healthy children with
normal growth despite GH deficiency by standard diag-nostic tests in which the GH levels were measured withRIA. In all 4 patients, however, GH levels were normal
Fig 1-Standard curve of the IM-9 cell hGH radioreceptor assay.
when measured with a radioreceptor assay (RRA). Wepostulate that the 4 children secrete a variant GH that bindsto the GH receptors, mimics the biological activity of GH,but is invisible to the standard RIA.
Patients and Methods
Subjects4 patients were referred to the endocrine diagnostic unit at the
University of Maryland for diagnostic GH testing. The referringdiagnosis in case 1 was acromegaly; the other 3 children werethought to have growth hormone deficiency. The growth velocity ofall 4 patients was found to have been consistently normal for severalyears before the study. Patient 1 was growing slightly above the 95thpercentile, patient 2 was growing along the 5th percentile, patient 3along the 25th percentile, and patient 4 along the 10th percentile. A24 h integrated concentration test (IC-GH) was done in all 4children. In addition, an insulin-arginine tolerance test (IATT) wascarried out in patients 3 and 4. Blood samples for GH RRA weretaken from venous blood obtained during the GH provocative testsof patients 3 and 4, and during the IC-GH tests of patients 1 and 2.For each patient the samples selected for RRA were those with thehighest GH levels by RIA. In the light of the apparent GHdeficiency of these children, patients 2-4 were started on a trial ofGH therapy, 01 1 mg/kg three times a week.
