J. clin. Path., 27, Suppl. (Roy. Coll. Path.), 8, 12-18

HaemoglobinopathiesThe pathophysiology of beta-thalassaemia majorC. B. MODELL

From the Department of Paediatrics, University College Hospital, London

The diagnosis of ,B-thalassaemia major is by thetypical blood film and the presence of more than20% of fetal haemoglobin. The defect in 3-thalas-saemia is in the switch from y to f-chain synthesisand it is clear that the failure to switch on the :chains is in some way related to a failure to switchoff the y chains, perhaps because the same mutatedfactor is responsible, in opposite senses, for the con-trol of both. One of the most remarkable featuresof the disease is the heterogeneous distribution ofHbF among the peripheral red cells, which indicatesimmense variation, from cell to cell, in the degreeto which HbA synthesis is depressed and HbF syn-thesis permitted. However, unlike the situation inhereditary persistence of fetal haemoglobin, y-chainsynthesis is always repressed to a substantial extent,so that the total non-a chain, i e, combined /, 8,and y-chain synthesis is reduced to one quarteror less of normal (Weatherall and Clegg, 1972).Alpha-chain synthesis appears to proceed normally,so that unpaired oc chains accumulate rapidly in thethalassaemic cells. These are unstable (Huehns andShooter, 1966) and precipitate to form inclusionbodies (fig 1); they may also interfere with the func-tioning of the nuclear and the cell membranes.The fundamental lesion in f-thalassaemics may

then be described as a defect in f-chain productionwith consequent precipitation of unpaired cx chains.Figure 2, on the other hand, shows two brothers fromQatar with thalassaemia major. Both brothers areseverely stunted, have typical facial and skull changes,gross hepatosplenomegaly and stick-like arms andlegs. Both pass dark urine. In common with otherthalassaemic patients they show a degree of dis-ability which seems out of proportion to the degreeof the anaemia when compared with, e g, iron defi-ciency or HbS disease. The question that should beasked is, how does a mutation affecting the controlof Hb f-chain synthesis produce these final conse-quences for the patients?The current idea is that most of the gross pathology

can be related not only to a deficiency of adulthaemoglobin, but also to the existence of precipi-tates of unpaired occhains in the developing red

12

p%....

*t ..

* .o*:s.

Fig I Electron micrograph taken by Dr Siro Koyamaof red cells in the peripheral blood in thalassaemiamajor. Alpha-chain precipitates dominate the picture.

cells. These appear to be responsible for prematurecell death in the bone marrow and hence for in-effective erythropoiesis and dark urine; and alsofor the abnormal red cells in the peripheral bloodthat are the cause of hypersplenism. The pathologyof fl-thalassaemia major is summarized in figure 3.Though this paper is not to be concerned with

treatment it should be mentioned here that there havebeen considerable recent advances in management.Modern treatment is directed first towards correct-ing the anaemia by maintaining the haemoglobinwithin the normal range, as this suppresses thehyperactivity of the bone marrow and eliminates most

.I!w...E ...R

.41.40,

copyright. on O

ctober 17, 2020 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.s3-8.1.12 on 1 January 1974. D

ownloaded from

The pathophysiology of beta-thalassaemia major

ht 131cmn

cmn

Vi

Fig 2 Side andfront viewsof two brothersfrom Qatarwith minimally treatedthalassaemia major. Theirgrowth is severely stunted,there is gross hepatospleno-megaly, and the arms andlegs are like sticks. Photographkindly provided by Dr HughJolly.

16YR 5YR 16YR

HbP-chein | hffctive erYthropoiesis ANAEMIA

0X-chain excee Short red cell life ~ with derk urin

faciesBone changes fractures

Plasma vol.1 cirrhosisOveractive marrow /endocrine damage

\ IGI. Fe absorption (d turbed puberty)

cardiac failureEffective protein malnutrition stunting

utohoousAbnormal red cells - HYPERSPLENISM ._-Ied cell life-:.donor

with inclusions \Zitmarrow activity.stuntineucopenia

thrombocytopeniaTRANSFUSIONAL SIDEROSIS

Fig 3 Outline of the pathophysiology of thalassaemiamajor.

of the pathology that springs from it; secondly,towards removing the spleen when it can be shownto be increasing the transfusion requirement,and this is very common; and thirdly towardscontrolling transfusional siderosis by the intensiveuse of iron-chelating agents. Evidence that this treat-ment is effective is gradually accumulating (Con-

stantoulakis, Economidou, Karagiorga, Katsantoni,and Gyftaki, 1974; Modell and Beck, 1974; O'Brien,1974), and there is every reason to hope that theprognosis in thalassaemia, both in the short and thelong term, is improving steadily. To return to thepathology, I would like to trace in some detail theconsequences of the ineffective erythropoiesis, andto follow this with a shorter discussion of hyper-splenism.

Ineffective Erythropoiesis

Ineffective erythropoiesis is the outstanding featureof thalassaemia major, and is unparalleled in anyother condition. The bone marrow expands to occu-py all the space available, yet the reticulocyte countis rarely much above normal. In the bone marrowthe myeloid: erythroid ratio is reversed, and dynamicstudies show an arrest of maturation in the earlypolychromatic erythroblasts. Figure 4, based on thework of Wickramsinghe, McElwain, Cooper, andHardisty (1970), shows the numbers of basophilicand early polychromatic erythroblasts in the dif-ferent stages of the cell cycle compared with thenormal. There is substantial failure of the thalas-

13

copyright. on O

ctober 17, 2020 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.s3-8.1.12 on 1 January 1974. D

ownloaded from

C. B. Modell

100_UooC) N 0 R M A L

Ln60T

,i 40 -

lo _ 01 GlS02 01 S G2

T HA ELRAYSHSRAOEMEIHAMRAOJOLARO 80_

> 60 -

cf) 40 -

_LE20 _

Gl S G2 x,T.osis Gl S G2BASOP H LIC EARLY POLYCHROMATIC

E RY THROBLAS TS ER YTHROB LA ST S

Fig 4 The percentage of basophil erythroblasts andearly polychromatic erythroblasts in different stages ofthe cell cycle, in normal (above) and thalassaemic(below) bone marrow. Two complete cell cycles arerepresented. G1 is the growth phase fbllowing mitosiswhen the cells are still diploid, S is the phase ofDNAsynthesis leading up to the next cell division, and G2is the growth phase preceding division. Then mitosisoccurs and the whole cycle is repeated. There is astriking arrest, in thalassaemic bone marrow, at the GIphase amwng early polychromatic erythroblasts.(Based on the work of Wickramsinghe and associates,1970.)

saemic polychromatic erythroblasts to begin thecycle of DNA synthesis that should lead up tothe final mitosis of the erythroid pathway. Thisarrest occurs at a stage when Hb synthesis has reallyonly just begun (Thorell, 1948), yet the evidence isthat the intracellular accumulations of oc chainsalready occur at this stage (Yataganas and Fessas,1969), and though they are not conspicuously at-tached to the cell membrane, they may be associatedwith the nuclear membrane, indent it, and alsoappearas intranuclear inclusions(Polliack,Yataganas,and Rachmilewitz, 1974). Thus there is good morpho-logical evidence of early intracellular and particu-larly, nuclear, damage by abnormal protein pre-cipitates.The arrested cells may die, mature more slowly,

or pass into the peripheral blood as they are. Thecells that do emerge into the circulation are stronglyheterogeneous in size, shape, and content of HbFand HbA. They are defective in several ways, eg,

the membrane is visibly abnormal (Hoffman,

Wolman, Hillier, and Parpart, 1956), excessivelysensitive to oxidation (Stocks, Offerman, Modell,and Dormandy, 1972), and leaky to Na+ (Cividalli,Locker, and Russell, 1971). 51Cr studies with autolo-gous cells show at least two populations, onevery short lived (Bailey and Prankerd, 1958) butthe others with a relatively long life span (51Cr T* =6 5 - 25d: mean about 17 days) (Smith, Erlandson,Stem, and Schulman, 1960; Malamos, Belcher,Gyftaki, and Binopoulos, 1961; Nightingale, Prank-erd, Richards, and Thompson, 1972; Blendis,Modell, Bowdler, and Williams, 1974). This and thedifferentially long survival of HbF-containing cells(Loukopoulos and Fessas, 1965) support the ideathat red cells appear in the circulation as a resultof a process of natural selection in the bone marrow.In any case, the production of red cells is neveradequate, and the consequent anaemia stimulatesthe bone marrow to ever-increasing activity.The total extent of the ineffective erythropoiesis

has not yet been adequately measured in thalas-saemia. Sturgeon and Finch (1957) and Malamoset al (1961) found marrow activity increased tosix times normal by ferrokinetic methods. However,this must represent a gross underestimate as thismethod, in common with all other methods formeasuring ineffective erythropoiesis, depends onhaemoglobin turnover, whereas in thalassaemia thereis defective Hb synthesis, and indeed most of thecell damage and death appear to occur before thecells are well haemoglobinized, ie, erythropoiesismust be at least 90% ineffective in severe thalas-saemia major.

Clinically, the 24-hour urinary uric acid shouldreflect marrow cell death, as the uric acid is derivedalmost entirely from DNA and RNA breakdown.Figure 5 shows that, though the plasma uric acid isusually in the normal range, the 24-hour urinaryuric acid calculated on a weight basis is many timesabove the adult normal. High transfusion, whichreduces bone marrow activity, reduces the urinaryuric acid towards, but not into, the normal range.It is not surprising that of the patients shown, onehas had a uric acid stone, and one has big kidneyswith a concentration defect.The urine also contains dark pigments which,

like the uric acid, reflect the degree of ineffectiveerythropoiesis because they are a direct result ofoc-chain precipitation. They are not bilirubin orporphyrins, but abnormal haem breakdown pro-ducts, conjugated dipyrryl methenes (Kreimer-Birnbaum, Pinkerton, Bannerman, and Hutchin-son, 1966). These are also found in the urine ofpatients with Heinz-body anaemia, and are probablyproduced in both conditions by intralysosomaldigestion in reticuloendothelial cells of the haemo-

14

copyright. on O

ctober 17, 2020 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.s3-8.1.12 on 1 January 1974. D

ownloaded from

The pathophysiology of beta-thalassaemia major

40

321-1-

203C42

03" 2

E

0C])41

C-)C 1l

4zFe A

is

16

14- 0

8 NORMAL RANGE -ADULT

4

I a I. I a I I a a. . _

1 3 5 7 9 11 13

AGE, YEARS

Fig 5 Twenty-four-hour urinary uric aciipatients with thalassaemia major. One ofmarked with arrows has had a uric acid stother has big kidneys with a concentrationl

globin precipitates. They are of clinical value, as a* *: clinical diagnosis of thalassaemia major can be

made when they are present in a child of suitableparentage with anaemia and hepatosplenomegaly.The gross bone marrow expansion consequent

on severe ineffective erythropoiesis has secondaryeffects on the bones and the plasma volume, ongrowth, and on gastrointestinal iron absorptionwhich may lead to the full-blown picture of haemo-chromatosis. These effects will be examined in turn.

Figure 6 illustrates the typical bony lesions, and* it is worth noticing the nutrient foramina in the

phalanges (marked with arrows). These are notnormally visible, and suggest the large proportion of

MALES the circulation that must be diverted through thegrossly expanded marrow. This probably accountsfor the plasma volume expansion that has long been

,,,,, , known in thalassaemia major (Erlandson, Schul-15 17 + man, and Smith, 1960), but has been attributed to

splenic enlargement, as plasma volume expansiond excretion in is common in association with splenomegaly.the patients However, direct measurements in untransfusedtone, and the thalassaemic patients before and after splenectomytdefect. (Blendis et al, 1974) show little change in the ex-

Fig 6 The typicalbone changes ofthalassaemia. Onthe left is aradiograph of thehand ofa 12-year-old untransfusedthalassaemic patientcompared with anormal 12-year-oldhand on the right.Note the nutrientforamina (arrowed)visible in the firstphalanges.

15

a

0*

.

copyright. on O

ctober 17, 2020 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.s3-8.1.12 on 1 January 1974. D

ownloaded from

C. B. Modell

Fig 7 Growth failure in thalassaemia major. Eachgrowth curve is based on an actual growth curve. Growthfailure at presentation (top L) is related to progressivebone marrow expansion. In early childhood it is relatedto excessive marrow expansion, either due to a trans-fusion scheme that is too low (top R), or in an effortto compensate for hypersplenism (bottom L). Growthfailure at puberty (bottom R) has an endocrine basisand does not respond to changes in the transfusionscheme.

panded plasma volume, which remains about 1-7times normal, so this must be attributable to some-thing else: it is probably the result of blood volumeexpansion in response to the shunting of a largefraction of the cardiac output through the bonemarrow. Since further expansion of the red cellmass is not possible, the blood volume expansionoccurs in the plasma compartment alone. Thisexpansion is reduced to normal by high transfusionwhich suppresses the bone marrow (Modell, inpreparation).The effect of ineffective erythropoiesis on growth

is generally seen in early childhood, and is quitedifferent from the endocrine growth failure thatoccurs in older, iron-loaded patients. Both typesare represented in figure 7. Early failure of growthoccurs only in association with a grossly overactivemarrow, eg, at presentation, or when the transfusionscheme is too low, or when there is compensatingbone marrow expansion in the presence of hyper-splenism. It can always be corrected by raising the

transfusion scheme or by splenectomy, as appro-priate. Clinically, such stunting is associated withanorexia, which resembles the anorexia of patientswith malignancies, in which there is also consider-able cell death. This suggests that stunting is dueto the release of some appetite suppressant bycells dying in the bone marrow, rather than tothe sequestration by the thalassaemic cells of somefactor necessary for growth. The stunting does notrespond to folic acid therapy.

Clinically, the most striking aspect of the dis-turbance in growth is the lack of muscular tissue,which gives the limbs their characteristic stick-likeappearance, and causes the patient to complain oflimited exercise tolerance due to weakness ratherthan to shortness of breath. It is reflected in the24-hr urinary creatinine. Since 2% of muscle creatineis broken down to creatinine per day, and this is theonly source of urinary creatinine, the 24-hr urinarycreatinine is constant and is a rough estimate ofmuscle mass. Figure 8 shows that low-transfused

DE

zZV

4:

zcc

* high transfusion* low transfusion

AGE, YEARS

Fig 8 Urinary creatinine excretion in thalassaemia major,expressed as mg/kg/24 hr. The normal curve is taken,from Stearns et al, 1958: Untransfused patients (0)and those on a low transfusion scheme (0) have a lowercreatinine excretion than those on a high transfusionscheme (-). The three sets ofpoints connected by arrowsshow the effect oJ shifting patients from a low to ahigh transfusion scheme.

or untransfused patients have a much lower creatin-ine excretion than high-transfused patients, whosemuscle mass appears to be in the normal range, inaccordance with their normal physical appearanceand exercise tolerance.

Finally, the gross bone-marrow overactivity leadsto excessive gastrointestinal iron absorption: of thevarious factors controlling this, bone marrow activ-ity is the most influential (Erlandson, Walden,

16

copyright. on O

ctober 17, 2020 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.s3-8.1.12 on 1 January 1974. D

ownloaded from

17The pathophysiology of beta-thalassaemia major

Stem, Hilgartner, Wehman, and Smith, 1962) prob-ably because the marrow cells selectively removeiron from the avid end of transferrin (Fletcher andHuehns, 1967). Absorption can reach 80% of dietaryintake (Erlandson et al, 1962), and in small childrenthe rate of iron loading from absorption may rivalthe rate of iron loading from transfusion. The table,

Mean Mean IronHaemoglobin Absorption(gldl) (% of oral dose)

Normal child 12-6-14 3-27Thalassaemia major

Before transfusion 6-7 45After transfusion 11-2 16-3

Table Oral iron absorption (5 mg FeSO4 taken fasting) in16 patients with thalassaemia major before and aftertransfusion''From Erlandson et al (1962)

based on the work of Erlandson and her associates(1962), shows that transfusion suppresses the in-creased iron absorption, and this is one of the com-pensations of a high transfusion scheme. Even in theabsence of transfusion, thalassaemia major is aniron-loading disease, but patients on maintenancetransfusion become iron loaded faster, and it isthis group that we have been able to study by meansof liver biopsies and endocrine studies. Thoughin them haemochromatosis is the result of treat-ment, this is a convenient point to discuss it. It isnow clear that unless the iron overload is treatedby long-term chelation therapy, these patientsdevelop frank cirrhosis by 12 to 16 years of age.Though no pituitary abnormality has been demon-strated, they commonly develop relative adrenalinsufficiency from as early as 8 years ofage (McIntosh,1974), and this may contribute to the sudden deaththat seems characteristic of the disease. Diabetesoccurs sporadically, possibly only in patients with afamily history of the disease, and hypoparathyroid-ism is common in heavily iron-loaded patients. Thetypical growth failure startiiig around the time ofpuberty, and the delayed puberty itself, have not beenadequately explained. Death from iron overloadis usually due to cardiac failure and may be sudden.It is to be hoped that long-term therapy with theiron-chelating agent desferrioxamine will preventthe development of this distressing pathology.

Hyperplasia of the Reticuloendothelial System

The remaining striking aspect of thalassaemia isthe reticuloendothelial hyperplasia represented byhepatosplenomegaly. Splenic enlargement is not

Fig 9 Electron micrograph taken by Dr Siro Koyama,of cells in the red pulp of the spleen from a thalassaemicpatient. A red cell containing inclusions is passingbetween two reticuloendothelial cells (R.E.) from thesplenic cord (bottom R) towards the splenic sinus (top L).The inclusions have been stroked down into the tail ofthe red cell, which will be removed by thereticuloendothelial cells. The rest of the red cell willreturn, 'pitted', to the circuilation.

primarily due to extramedullary erythropoiesis:it happens because the spleen is presented with ab-normal red cells. The red cells are expected totraverse the walls of the splenic sinuses between thereticuloendothelial cells, which stroke down thepassing erythrocytes so that any inclusions (a chains,nuclei, Howell-Jolly bodies) are squeezed into theirtails. Then the reticuloendothelial cells selectivelyingest this part of the red cell and the remainderpass back into the circulation (Koyama, Aoki, andDeguchi, 1964; Slater, Muir, and Weed, 1968).This process is illustrated in figure 9, an electronmicrograph taken by Dr Koyama of material fromthe spleen of one of our patients. This processnecessarily involves some delay and the spleenbecomes engorged, enlarges, and finally in most casescauses some accelerated destruction of autologousor donor red cells or both. It then becomes appro-

copyright. on O

ctober 17, 2020 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.s3-8.1.12 on 1 January 1974. D

ownloaded from

C. B. Modell

priate to remove the spleen. Before splenectomy theperipheral blood film shows relatively few nucleatedred cells but after splenectomy nucleated cells, whichwould normally be 'pitted', appear in the circula-tion in large numbers. From the fact that thesepersist indefinitely, it is clear that the pitting functionis never taken over by any other organ.

Hypersplenism sometimes causes further gross

plasma volume expansion, and can lead to deathfrom thrombocytopenia. In patients on a lowtransfusion scheme it may exacerbate the disease tothe point where they die in heart failure; and inpatients on a high transfusion scheme, by increasingthe blood requirement, it may accelerate iron over-

load. For these reasons opinion is now swingingin favour of splenectomy in thalassaemia whenindicated after a long period of being against it.

Conclusion

In conclusion, in /-thalassaemia major the cellularevent which produces much of the pathology isprecipitation of unpaired haemoglobin chains.This causes ineffective erythropoiesis, bone-marrowexpansion, and hypersplenism from which the gross

physical pathology follows. In a fundamentalapproach to treatment it would be reasonable toattempt to reduceac-chain synthesis as well as to seekto stimulate /- or y-chain synthesis.

I am grateful to Professors L. B. Strang and E. R.Huehns for help and advice, to Dr S. Koyama forthe electron micrographs, to Dr Hugh Jolly for thephotograph of two patients with severe thalassaemiamajor, and to the Sir Halley Stewart Trust for a

research grant.

References

Bailey,I. S., and Prankerd, T. A. J. (1958). Studies in thalassaemia.Brit. J.Haemat.,4, 150-155.

Blendis, L., Modell, C. B.,Bowdler, A. J, andWilliams, R. (1974).Some effects of splenectomy in thalassaemia major. Brit. J.Haemat., in press.

Cividalli, G., Locker, H., and Russell, A. (1971). Increased perme-

ability of erythrocyte membrane in thalassemia. Blood, 37, 716-724.

Constantoulakis, M., Economidou, J., Karagiorga, M., KatsantoniA., and Gyftaki, E. (1974). Combined long-term treatment ofhemosiderosis with desferrioxamine and DTPA in homo-zygous lI-thalassemia. Ann. N. Y. Acad. Sci., 232, 193-200.

Erlandson, M. E., Schulman, I., and Smith, C. H. (1960). Studies on

congenital hemolytic syndromes.III. Rates of destruction and

production of erythrocytes in sickle cell anemia. Pediatrics,25, 629-644.

Erlandson, M. E., Walden, B., Stern, G., Hilgartner, M. W., Wehman,J., and Smith, C. H. (1962). Studies on congenital hemolyticsyndromes. IV. Gastrointestinal absorption of iron. Blood,19, 359-378.

Fletcher, J., and Huehns, E. R. (1967). The uptake of plasma iron byimmature red cells. J. Physiol. (Lond.), 192, 38-39P.

Hoffman, J. F., Wolman, I. J., Hillier, J., and Parpart, A. K. (1956).Ultrastructure of erythrocyte membranes in thalassemia majorand minor. Blood, 11, 946-956.

Huehns, E. R., and Shooter, E. M. (1966). Further studies on theisolation and properties of oc-chain sub-units of haemoglobin.Biochem. J., 101, 843-851.

Koyama, S., Aoki, S., and Deguchi, K. (1964). Electron microscopicobservations of the splenic red pulp with special referenceto the pitting function. Mie. med., J. 14, 143-188.

Kreimer-Birnbaum, M., Pinkerton, P. H., Bannerman, R. M., andHutchison, H. E., (1966). Dipyrrolic urinary pigments in con-genital Heinz-body anaemia due to Hb Koln, and in thalas-saemia. Brit. med. J., 2, 396.

Loukopoulos, D., and Fessas, P. (1965). The distribution of hemo-globin types in thalassemic erythrocytes.J.,clin. Invest., 44, 231-240.

McIntosh, N., (1974). Endocrinopathy in Thalassaemia Major, in press.Malamos, B., Belcher, E. H., Gyftaki, E., and Binopoulos, D. (1961).

Simultaneous radioactive tracer studies of erythropoiesis andred-cell destruction in thalassaemia. Brit. J. Haemat., 7,41 1-429.

Modell, C. B., and Beck, J. (1974). Long-term desferrioxamine inthalassemia. Ann. N. Y. Acad. Sci., 232, 201-210.

Nightingale, D., Prankerd, T. A. J., Richards, J. D. M., and Thompson,D. (1972). Splenectomy in anaemia. Quart. J. Med., 41,261-267.

O'Brien, R. T. (1974). Ascorbic acid enhancement of desferrioxamine-induced urinary iron excretion in thalassemia major. Ann.N. Y. A cad. Sci. , 232, 221-225

Polliack, A., Yataganas, X., and Rachmilewitz, E. A. (1974). Ultra-structure of the inclusion bodies and nuclear abnormalities in0-thalassemic erythroblasts. Ann. N. Y. Acad. Sci., 232, 261-282.

Slater, L. M., Muir,W. A., andWeed, R.I. (1968). Influence ofsplenectomy on insoluble hemoglobin inclusion bodies inthalassemic erythrocytes. Blood, 31, 766-777.

Smith, C. H., Erlandson, M. E., Stern, G., and Schulman,I. (1960).The role of splenectomy in the management of thalassemia.Blood, 15, 197-211.

Stearns,G., Newman, K. J., McKinley, J. B., and Jeans, P. C. (1958).The protein requirements of children from 1-10 years of age.Ann. N. Y. Acad. Sci., 69, 857-868.

Stocks, J., Offerman, E. L., Modell, C. B., and Dormandy, T. L.(1972). The susceptibility to autoxidation of human red celllipids in health and disease. Brit. J. Haemat., 23, 713-724.

Sturgeon, P., and Finch, C. A. (1957). Erythrokinetics in Cooley'sanemia. Blood, 12, 64-73.

Thorell, B. (1948). The relation of the synthesis of hemoglobin to thecellular growth during normal and certain pathologicalconditions. Acta. path. microb. scand, 25, 54-65.

Weatherall, D. J., and Clegg, J. B. (1969). Disordered globin synthesisin thalassemia. Ann, N. Y. Acad. Sci., 165, 242-252.,

Weatherall, D. J., and Clegg, J. B. (1972). The Thalassaemia Syn-dromes, 2nd ed. Blackwell, Oxford.

Wickramasinghe,S. N., McElwain, T. J., Cooper, E. H., and Hardisty,R. M.(1970). Proliferation oferythroblasts in beta-thalassaemia.Brit. J. Haemat., 19, 719-727.

Yataganas, X., and Fessas, P.(1969). The pattern of hemoglobinprecipitation in thalassaemia and its significance. Ann. N. Y,Acad. Sci., 165, 270-287.

18

copyright. on O

ctober 17, 2020 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.s3-8.1.12 on 1 January 1974. D

ownloaded from