Lack of influence of fetal hemoglobin levels orerythrocyte indices on the severity of sickle cellanemia.

D R Powars, … , L S Chan, S P Azen

J Clin Invest. 1980;65(3):732-740. https://doi.org/10.1172/JCI109720.

Persons with sickle cell anemia who have elevated fetal hemoglobin or lowered erythrocytemean corpuscular volume are reputed to have less severe clinical manifestations and agreater probability of survival. This study examines the relationship between seven clinicalindicators of morbidity in sickle cell anemia and seven hematological parameters that werecollected from 214 patients. Risks of sickle cell crisis, acute chest syndrome, hospitaladmissions, cerebrovascular accident, aseptic necrosis, meningitis/septicemia, and deathwere used as indicators of morbidity. The hematological parameters included percent fetalhemoglobin, absolute fetal hemoglobin, percent hemoglobin A2, hemoglobin concentration,packed cell volume, mean corpuscular volume, and mean corpuscular hemoglobinconcentration. Statistical analyses of the data showed no relationship between thehematological parameters and six of the seven clinical indicators of the severity of sicklecell anemia. The only significant finding was an increased risk of stroke in those patientswith lower levels of fetal hemoglobin. Therefore, with this exception, there is no predictablerelationship between morbidity and mortality in sickle cell anemia and levels of fetalhemoglobin or erythrocyte indices. Thus, the general belief that there is an associationbetween severity of sickle cell anemia and the levels of fetal hemoglobin has not beenestablished.

Research Article

Find the latest version:

http://jci.me/109720/pdf

Pdf

Lack of Influence of Fetal Hemoglobin Levels or

Erythrocyte Indices on the Severity of Sickle Cell Anemia

DARLEENR. POWARS,Departmtlenit of Pediatrics, Untiversity of Souitherni (CaliforniaSchool of Medicine, Los Angeles, Californiia 90033

WALTERA. SCHROEDER,Divisiorn of ChemiistrmJ anid Chemttical Eniginieerinig,California Institute of Technology, Pasaderna, Californial 91125

JOYCE N. WEISs, LINDA S. CHAN, an(d STANLEYP. AZEN, Divisioni of Biomnettril,University of Southern Californtia School of Medicine,Los Angeles, California 90033

A B S T RA C T Persoins with sickle cell anemiiia whohave elevated fetal hemoglobin or lowered erythrocytemean corpuscular volume are reputed to have lesssevere clinical manifestations anid a greater prob-ability of survival. This study examiines the relation-ship between seven clinical indicators of morbidityin sickle cell anemia and seven hematologicalparameters that were collected fromn 214 patients.Risks of sickle cell crisis, acute chest syndromiie,hospital admissions, cerebrovascular accident, asepticnecrosis, meningitis/septicemia, and death were usedas indicators of morbidity. The hematological param-eters included percent fetal hemoglobin, absolutefetal hemoglobin, percent hemoglobin A2, hemoglobinconcentration, packed cell volume, mean corpuscuilarvolume, and mean corpuscular hemoglobin concen-tration. Statistical analyses of the data showed norelationship between the hematological paramiietersand six of the seven clinical indicators of theseverity of sickle cell anemia. The only significantfinding was an increased risk of stroke in thosepatients with lower levels of fetal hemnoglobin.Therefore, with this exception, there is no pre-dictable relationship between morbidity and mortalityin sickle cell anemia and levels of fetal heimoglobinor erythrocyte indices. Thus, the general belief thatthere is an association between severity of sickle cellanemia and the levels of fetal hemoglobin has notbeen established.

INTRODUCTION

The variable clinical expression of sickle cellanemia has initiated a world-wide search for bio-

Received for publication 10 September 1979 and(I in revisedforn 7 November 1979.

732

chemical, hemiiatological, or enIvir-oinmental facwtorsthat mayi influenice the severity of the illniess. Thereis a wvidely held opinion that elevation of fretalhemoglobin (Hb-F)m or lowered erythrocyte imeaiiicorpuscular volume (MCV) is associated wvith a cle-creased severity of cliniical miianjifestationis and(I ainincrease(l survival. This paper reports the restilts of astatistical assessmiient of the relationship between thehematological factors and(l the imortality anid morbidityof 214 patienits with sickle cell aniemiiia.

Backgroundil (of stU(lld. The concept that ele\vationof Hb-F favorably influenices morbidity has supportfroml both biochemlical ancd clinical stuidies (1). Thuts,increased proportionis of Hl-F increase the minimniumgelling concenitration of mixttures of hemnoglobinis Sandl F (2, 3), modify the polyvnerization kinetics(4, 5), and increase the delay timiie to polymiierizationiin S-F mixtures (3, 6). Hb-F is (lecrease(l in ir-reversibly sickle(d SS cells (7) and greater numilbers ofirreversibly sickle(d SS cells produice increased viscosityof oxygenatedl whole SS bloo(d (8). However, thekinetics of Hb-S polymerizationi suggest that thevariable clinical expressioni inay ressult from severalkinetic paths to polymerization anid to imiarginalchanges in free energy (9). \Miale et al. (10) haveshown irregular shifts of Hb-F contenit amiiong theindlividual erythrocytes an(l have attempted to dem-onstrate a correlationi of this shift with growth and(Idevelopmnent of the SS patient. Moreover, accordIinigto Dover et al. (11), wide variationi exists in thepro(luctioni anid survival of F cells in persoius wVithsickle cell aniemlia.

'Abbreviatiotns uised int this paper: Hb), hemoglobini;MCHC, imeant corpuisecular hemoglobin concenitrationl; \I CV,mean corpuscular volumiie; PCV, packe(d cell volllme; SS,sickle cell anemia.

J. Clint. Invest. (C The Atmierican Society for Clinical Investigationt, inc. * 0021-97.38/80103/0'l7.32109 $1.00Voltumiie 65 Alarch 1980 732-740

Those persoins with the determinianit for the hereditarypersistence of Hb-F in the tratns position to S haveessentially, anl asymlptomiiatic condition (12, 13). Furthersupport for the concept that Hh-F is protective hascome from observations of sickle cell anemia inseveral populationis (14-20). In these populations themiiean Hb-F was high (10-25%), although within eachpopulation the range was wide (a few to 40%). Theuniversal suggestion has been that the relatively mildexpression of sickle cell anemia in these populationisis caused by the high Hb-F.

Following Lehmann's observations (21) that sicklecell anemiiia appeared to be a less severe disease in

vestern Africa where a-thalassemia is frequent,it has beein postulated that a-thalasseiiiia anid( sicklecell anemiiia in the same individual will protect the SSpatient because of lowered MCVor mean corpuscularhenmoglobin concenitration (MICHC) (22, 23). Lowerinigof the Hb-S coineeintratioin does lengtlhen the time forHb-S poly merization and (lecreases the numbers ofirreversibly sickled SS cells (24, 25), and Clark et al.(26) report that MCfIC is a major deterninlanit ofthe meimibranie clhanges that are associated wvithsickling.

METHODSStudy population. The study population consists of 214

"S-only" patients w,hose hemnatological and cliniical datawere available for analysis. An S-only patienit is definie(d asone whose hemiioglobinl, by electrophoresis and columnchrotnatography, has only Hb-S as the major hemoglobini(absenice of Hb-A) and whose cells have a positive sicklingtest. Consequently, not only may this category includepatienits with homozvgous sickle cell anemnia, but alsothose patients with S- &-thalasseniia, S-8,8-thalassemia,S-hereditary persistence of fetal hemiioglobin, and anyvariety of a-thalassemiiia. Excltude(d fromn the study wereinfants who had not vet attained a reasonably stable level ofHb-F, regularly transfused patients, and( SS patients withconcomitant a-chain substitutions. Of the 214 patients, 206have honmozygous sickle cell anemnia, 6 have S-#°-thalassenmia,and 2 have S-6B,-thalassenmia.

Clinic and medical care for patients wvith sickle cellanenmia has beei pirovi(led as part of a continuing programsince 1964. Although many SS patients have been knownfrom birth or shortly thereafter, others have been diagnosedat a later age. Consequenitly, an extensive medical historyis available for some huimdre(ls of SS patients. Since 1972,extensive biochemnical and genetic characterizations of manyof the patients have been inade.

Laboratort parameters. The hematological parametersused in the statistical analysis were percenit Hb-F, absoluiteHb-F, Hb-A2, hemoglobin concentrationi, packed cell vol-umiie (PCV), MICV, and ICHIC. All blood samples weretestedl with so(lium metabisulfite for morphologic evi-deniee of sicklinlg. Erythrocyte indices and hemoglobinlevels were determined by standard methodology. Micro-hematocrit determinations wvere obtained on all samples.

The hemoglobin type of all patients has beeni examinedby one or more of three electrophoretic procedures (cellutloseacetate or starch gel at alkaline pH or agar at aci(d pH) anidby chromatography. Hb-F, Hb-A2, and other comnponients were

determined quantitatively by microchromatographic pro-cedlures (27-29). These procedures for Hb-F have theadvantage of accuracy at all Hb-F levels and are applicableto large numbers of determinations with small samples ofblood. The quantity of lib-F has been calculated both asthe percenitage of total hemoglobin and as the absolutevalue in imilligramns per decaliter of whole blood. Stainingfor Hb-F by the Kleihauer-Betke method (30) has beenapplied to the cells of 144 patients.

In no inistance has a hemoglobin of the study groupbeent examinied by aminio acid sequencing.

Clitnical parameters. Collection of morbidity data ac-companiies an ongoinig study of the demographic char-acteristics of the population. Data from medical historywere transferred to computer cards and current data arecollected regularly from each patient. Regardless of thereason for which the patient may be seen, blood is collectedso that appropriate hematological and biochemical data forthe visit may be incorporate(d into the data base. At the sametime, the diagnosis of any clinical complications and anassessmiient of the state of well-beinig are recorded.

The clinical indicators of nmorbiditv used in this studv weretotal hospital admissionis for any cause, total hospitalizedvaso-occlusive sickle cell crisis, acute chest syndrome (whichincludes proven bacterial pnieumnonia, strongly suspectbut nonibacteriologically proven pneumonia, viral pneumonia,proven pulmonary infarcts, and multiple microthromboticchest syndrome), bacterial meninigitis and septicemia, asepticnecrosis in anv major bony site, stroke (cerebral infarct andsubarachnioi(d hemorrhage), and( death.

Statistical mtietho(ds. The seven clinical parameters werestudied in two categories. The first categorv provided meas-ures of imorbiditv through inmultiple occurrences of sickle cellcrisis, acute chest syndrome, and hospital admission. Eachparamneter in this group was analyzed using both multipleregressioni and discriininant analysis techniques (31). Theseconid categorv of clinical parameters included cerebro-vascular accidenit, aseptic necrosis, mneningitis/septicemia,andl death. The first occurrence of these clinical param-eters, which often significantly changes the subsequentcourse of the patient, was analyzed by survival analysistechmiiquies (32, 33). In the mutltiple regression analysis, thedependenit variable is the risk of the recurrent clinicalevent, defined as the ntumber of episodes per person-yearof observation or rate of occurrence. The independentvariables are age at entry and the laboratory parameter understudv.

In the discrimiiniant analysis, strata of each laboratoryparamiieter were used. Cut points of the strata were choseneither to separate normal from abnormal levels, or to dividethe paraimieter into equal groups with high, medium, andlov values. The dependent variable in the analysis was thelaboratory paramiieter and the indepenident variables were ageat entry anod the risk of the clinical parameter as de-fined above. In view of the marked effect of age onclinical imianifestation (34), age at entry was "forced" into the(liscrilninant eqiuation to adjust for differences in averageage at entry amnonig the strata defined by each laboratoryparameter.

The survival analysis also used strata of the laboratoryparameters as in the discrimiiinant analysis. The analysiscompare(d the risk of the clinical parameter, defined as theprobability of "terminiation" by the first occurrence of thecliniical event, among patients in different strata. Here, "ter-minination" refers to the end point of the patient's ob-servation for that particular event under analysis. The logrank test wvas used in the survival analysis to determinestatistical significance among the strata (35).

Fetal Hemoglobitn anid Sickle Cell Antemia 733

TABLE ISummary of the Hemnatologic Laboratory Parameters

NumberLaboratory parameters of persons Mean±SD Range

%Hb-F 214 12+7 3-36Absolute Hb-F, gmldl 214 1,080±717 240-3,888MCV,fl* 179 88±10 55-117Hb-A2, %* 212 3.2±0.6 1.5-5.9Hb, gmkll 214 8.5±1.4 5.4-13.7PCV, % 214 25.4+4.3 16-41MCHC,gmldl 203 33.9±1.5 28-38

* These data include results from individuals with S-,3°- andS-8,8-thalassemia.

RESULTS

Characteristics of the study population. The studypopulation of 214 persons had an average age of 19.2yr. The sex distribution was 114 males (53%) and 100females (47%). 19% of the patients are currently<10 yr of age, 37% are between 10 and 19, 31% arein the third decade, 9% are in the fourth decade, and4% are >40 yr of age. 94% (202) are currentlyalive. The cumulative number of years of observationin the 214 patients is 2,126 person yr. A total of 34%(72 patients) entered the study before age 3 and have anaverage (±SD) of 12.6±7.7 yr of observation; 32% (70patients) entered between ages 3 and 10 and have anaverage of 10.2 ±6.6 yr of observation; 16%(34 patients)entered between ages 11 and 20 and have an averageof 7.3+4.6 yr of observation and 18% (38 patients)entered between ages 21 and 56 and have an averageof 6.7+4.5 yr of observation.

Description of the laboratory parameters. Thecells of all patients in this study had the expectedmorphologic changes with sodium metabisulfite, andtheir hemoglobins showed the anticipated electro-phoretic and chromatographic behaviors. A representa-

tive laboratory value was chosen for each patient in anattempt to select values that reflect his "steady-state" level. Thus, some patients have had as many as10 determinations of Hb-F over a period of severalyears. In the 144 patients to whose cells the test wasapplied, the Kleihauer-Betke method (30) detected aheterogeneous distribution of Hb-F with markedlyvariable staining in individual erythrocytes.

Of the various parameters, hemoglobin concentra-tion perhaps tends to vary the most with time in theindividual patient. The most recent steady-statehemoglobin level and hematocrit have been used forthis analysis. Table I summarizes various aspects ofthe laboratory parameters that were used in the study.Figs. 1 and 2 depict the frequency distributions ofthe percentage of Hb-F and of the MCVof the cases.

Description of the clinical parameters. The per-centage of the 214 patients with at least one episodeof each of the seven selected clinical indicators ofseverity of illness is depicted in Fig. 3. Table IIsummarizes the risk of sickle cell crisis, chest syn-drome, and hospital admission in terms of frequencyof episodes per person-year observed, and the risk ofstroke, aseptic necrosis, meningitis/septicemia, anddeath in terms of the cumulative probability oftermination by the first occurrence of these episodesat various ages.

Statistical analysis. No significant differenceswere observed in the multiple correlations of the riskof the clinical parameters with age at entry and thelaboratory parameters. Neither the partial correlation ofthe laboratory parameter with risk, given age at entry,nor of age at entry with risk, given the laboratoryparameter, was significant. The maximum value of thesquare of the multiple correlation coefficient, whichmeasures the goodness-of-fit of the analysis, was foundto be <3%. Thus, no laboratory parameter was found

FIGURE 1 The frequency distribution of the percentage of FIGURE 2 The frequency distribution of MCVvalues in 179Hb-F in 214 persons with sickle cell disease. persons with sickle cell disease.

734 D. R. Powars, W. A. Schroeder,J. N. Weiss, L. S. Chan, and S. P. Azen

to predict the risk of sickle cell crisis, cheor hospitalization.

In the discrimiinianit analyses, no si~ferences were detected in the averageor in the risk of the clinical paramiieteigroups defined by the laboratorx paranlIII presents the average age at entry aneage of occurrence of sickle cell crisis, cheand hospitalization for the selected grouilaboratory parameter.

Table IV summilarizes the results ofanalysis of stroke, aseptic necrosis, mileicenmia, and death for the samle groupinigs otory parameter. No signiificanit dlifferencesfor the risk of aseptic necrosis, meninlgitior death amonig groups of patienits defineclthe laboratory parameters. No significacnwere found for the risk of stroke amiiong Utients define(d by levels of MCV, Hb-A2, FMCHC. Fig. 4 presents the survival cursnecrosis for the three grotips of patienits vlevels of percenit Hb-F. Fig. 5 presenitscurves of stroke for the same groups. The Istroke was found to be significantly cliffgroups of patients defined by percenit Hb-FHb-F; a decreased probability of stroke wvwith increased levels of Hb-F.

DISCUSSION

There are major difficulties in developinclinical severity for sickle cell anemiiia (various indices of severity have been usetThese indices, in our opinion, have usumapplicable. Age separates the populaticell anemia patients by the nature of tIconcditions and the outcomiie (39). In geim

100

90

so80

70

. 60

I.C 40lUo 30

20

10

93.0%

FT; R80.4%

F.n1 72

56.5%

F.n1211

17.3%3.1S

_n - 37 ;|

HOSPITAL VASO- CHEST BACTERIAL ASEPTICADMISSIONS OCCLUSIVE SYNDROMEMENINGITIS NECROSIS

CRISES SEPTICEMIA

FIGURE 3 Percentage of patients with one orof the seven cliniical indicators of morbiditv. Si.chest syndrome, and hospital admission arecurrent clinical events, whereas death, asestroke, and septicemia are analyzed as one-timevents.

st syndrome,

gnificaint dif-age at entry

rs across theweters. Table1 the average-st synlldromiie,pings of each

the survivaliniigitis/septi-

f each labora-were eviident5s/septiceniia,by any onie ofit clifferencesgroups of pa-11, PCV, andlkes of asepticvith differenit

TABLE I I

Siooooiarill of tie Clinical Parameters

Multiple episo(le A.erage rate of' octcrrenceclinical paranmeters per person-\ ear SI)

Sickle cell crisis 0.71 1.23Chest sxi dronile 0. 17 0.30Hospitalization 1.17 1.68

Cimiuilative p)rohahilit\ of teri iniationi at

Sinlgle episode Agez Age Age Age Age Agecliniical parameters 5 1t) 15 2(0 3f) 4ff

Stroke 0.06 0.11 0.16 0.17 0.19 0.32Aseptic

inecrosis 0.06 0.10 0.15 0.19 0.27 0.57Mleniiligitis/

septicemicla 0.43 0.46 0.49 0.51 0.55 0.63Death 0.01 0.01 0.02 0.04 0.12 0.37

<7 yr of age often have acute infectiouis illnesses,the survival wvhereas, in adults, overwvhelminig problems of'specific

probability of organi damage becomiie the pre(lominiant featuire of theFerent amonig manifestationis of sickle cell aniemiiia. Once permiianienit

aan( absolute (lamiiage to a vital organ occutrs, the problemn of theas associate(d malfuncetioninlg organ overshadows those that canl be

relate(d simnply to sickle cell anemiiia itself: Con-se(iuelntlv, wve have choseni the seven major cliniicalcoomplicationis alreacdy listedl as measures of clinicalseverity.

Ig aln inex of Fetal hiemtoglobini and(I ,orbi(litij. The resuilts of the

(l5) althou3gh present study do not agree with the generally heldC (19,36-3 ) opinlioIn that high percentages of Hb-F ancd lowv valuesillX nlOt beenl of M\CV have an amelioratinig effect upon the severityonof soikle of sickle cell anemiiia. Thus, the statistical analyses

et associated have failed, with one exceptioni, to uncover anveral, childreni relationship between the severity of sickle cell aniemiiia

(as mieasured by the stated indicators) and the levelsof Hb-F, Hb-A2, or other hematological parameters.The single exception, to be discussed below, isthe association between Hb-F and stroke. Contraryto the general views expressed in the literature, con-clusions similar to ours were drawn by Ozsovlu andAltinoz (40) fromii a study of 95 Turkish cases. Theyfelt that these cases, with less than one crisis peryear, were clinically milder than "African cases."However, in our group of Black Americans, the risk of

8.4s 5^ crisis was 0.71±1.23/yr of observation. Turkish andn; r18 nl21 American SS patients seem to be similar clinically,CVA DEATH hematologically, and in range of Hb-F. Serjeant (41)

reported on 60 SS patients over 30 yr of age who weremore episodes alive, reasonably well, and did not have elevatedckle cell crisis, Hb-F levels (41). Recently he and his collaborators (42)frequenitly re- report mean percent Hb-F of 5.18+3.59 among 123 SS~ptic necrosis,

e (termination) cases and 6.50±+3.73 among 41 probable S-,80-thalassemia cases. Although these values are some-

Fetal Hemtoglobitn atid Sickle Cell Anteem ia 735

TABLE IIIDiscriminant Analysis of the Risk of Sickle Cell Crisis, Chest Syndrome, and Hospitalization

Sickle cell crisis Chest syndrome Hospitalization

Mean Mean rate Mean rate Mean rateNumber age at per person- P per person- P per person- P

Laboratory parameters of patients entry year value year value year value

%Hb-F c 8 78 11.4 0.73 0.20 1.138 < %Hb-F c 13 65 9.0 0.74 0.97 0.16 0.55 1.13 0.85%Hb-F > 13 71 8.8 0.68 0.16 1.24

Total 214 9.8 0.71 0.17 1.17

Abs Hb-F c 650 70 10.8 0.73 0.19 1.15650 < Abs Hb-F c 1150 73 8.9 0.79 0.64 0.19 0.44 1.25 0.82Abs Hb-F > 1150 71 9.8 0.61 0.14 1.10

Total 214 9.8 0.71 0.17 1.17

MCVc 75 20 9.2 1.25 0.11 1.6575 < MCVc 95 125 9.1 0.62 0.11 0.20 0.48 1.07 0.28MCV> 95 34 14.0 0.87 0.18 1.50

Total 179 10.0 0.74 0.18 1.22

HbA2 < 2.5 25 8.8 0.78 0.22 1.602.5 c HbA2 ' 4.0 171 9.8 0.68 0.70 0.17 0.64 1.08 0.28HbA2 > 4.0 16 12.0 0.95 0.13 1.43

Total 212 9.9 0.71 0.17 1.17

Hb < 8 71 8.3 0.73 0.19 1.138 c Hb < 9 73 10.1 0.73 0.91 0.18 0.85 1.30 0.67Hb 2 9 70 11.1 0.68 0.16 1.07

Total 214 9.9 0.71 0.17 1.17

PCVc 23 71 8.8 0.71 0.19 1.2023 < PCVc 26 68 8.4 0.73 0.91 0.17 0.74 1.24 0.66PCV> 26 75 12.1 0.69 0.16 1.07

Tbtal 214 9.8 0.71 0.17 1.16

MCHCc 32 29 9.4 0.53 0.09 0.90MCHC> 32 174 9.3 0.69 0.52 0.18 0.09 1.15 0.46

Total 203 9.3 0.66 0.16 1.11

what lower than our mean percent Hb-F of 12.3±7.0,the methodologies for Hb-F determination were dif-ferent. In general, alkaline denaturation methodsas used in Jamaica and Turkey give lower readingsparticularly at higher levels (above 20%) of Hb-Fso that these apparent differences may not be sig-nificant. Wrightstone and Huisman (43) reported amean Hb-F (alkaline denaturation) of 6.6% in 117Black SS individuals.

The strongest support for the concept that a highlevel of Hb-F ameliorates the severity of sickle celldisease comes from the ongoing reports of Arabiansickle cell anemia patients (14-19). Their mean Hb-Famong 183 children was 25%and among 54 adults was24%. Nevertheless, 95% of their SS cases covered arange of 12 to 40% Hb-F whereas in our group 40%of the cases fall within this range. As reported byPerrine et al. (18), the Arabian SS cases show con-siderable variability of clinical expression. Informa-

tion is not currently available regarding the correla-tion of risk of sickle crisis, stroke, chest syndrome, etc.,and Hb-F levels in the Arabian SS population. Note that75% of the Arabian SS patients were less than 20 yr ofage when reported (18), including many adolescentpatients, an age range in which sickle cell anemiatends to be mnost quiescent in its manifestations. The15 mildly affected Indian SS patients had characteristicsmuch like those of the Saudi Arabians; the age ranigewas from 8 to 26 yr and only two were >20 (20).

Some of our S-only cases as well as those observedby others (18, 19) have remarkably high Hb-F levels(above 30%). Nevertheless, such individuals dem-onstrate the hematologic values and other char-acteristics of classic sickle cell anemia. Several suchpatients in the present study have had serious ill-ness. Contrariwise, some cases with Hb-F of the orderof 5% have had minimal clinical problems. Somequestions are unanswered: Are the Saudi Arabian or

736 D. R. Potvars, W. A. Schroeder,J. N. Weiss, L. S. Chart, ani(d S. P. Azeni

TABLE IVAnalysis of Termination Probabilities of Stroke, Aseptic Necrosis, Meningitis/Septicemia, and Death

Stroke Aseptic necrosis Meningitis/septicemia Death

Number Number Number Numberof with 1+ P with 1+ P with 1+ P P

Laboratory parameters patients eipsodes* valuet episodes* value t episodes* valuet valuet

%Hb-F c 8 78 14 (788) 15 (788) 15 (t44) 7 (893)8 < %Hb-F c 13 65 3 (670) 0.003 9 (651) 0.59 13 (63) 0.68 1 (709) 0.21%Hb-F > 13 71 1 (518) 4 (485) 9 (502) 4 (524)

Total 214 18 (1,976) 28 (1,924) 37 (1,849) 12 (2,126)Abs Hb-F c 650 70 15 (698) 13 (712) 13 (688) 6 (813)650 < Abs Hb-F < 1150 73 3 (739) 0.0001 11 (713) 0.46 15 (644) 0.85 3 (773) 0.85Abs Hb-F > 1150 71 0(540) 4 (500) 9(518) 3 (540)

Total 214 18 (1,977) 28 (1,925) 37 (1,850) 12 (2,126)

MCVs 75 20 1 (203) 2 (191) 5 (177) 2 (205)75 < MCVc 95 125 11 (1,065) 0.47 17 (1,068) 0.80 24 (998) 0.52 7 (1,173) 0.80MCV> 95 34 2 (360) 5 (327) 3 (332) 2 (376)

Total 179 14 (1,628) 24 (1,586) 32 (1,507) 11 (1,754)

HbA2< 2.5 25 1 (179) 2 (162) 4 (182) 0 (185)2.5 c HbA2 ' 4.0 171 16 (1,576) 0.83 21 (1,579) 0.98 29(1,484) 0.91 10(1,718) 0.71HbA2> 4.0 16 1 (201) 5(163) 3 (173) 2 (203)

Total 212 18 (1,956) 28 (1,904) 36 (1,839) 12 (2,106)

Hb < 8 71 9(598) 7(633) 10(599) 5(676)8 c Hb < 9 73 7 (706) 0.11 11 (653) 0.64 13 (683) 0.29 3 (764) 0.50Hb 2 9 70 2 (672) 10 (639) 14 (568) 4 (686)

Total 214 18 (1,976) 28 (1,925) 37 (1,850) 12 (2,126)

PCVc 23 71 8 (592) 7 (610) 11 (570) 3 (656)23 < PCVc 26 68 7 (651) 0.77 11 (613) 0.55 11 (645) 0.51 4 (720) 0.98PCV> 26 75 3 (734) 10 (702) 15 (633) 5 (750)

Total 214 18 (1,977) 28 (1,925) 37 (1,848) 12 (2,126)

MCHCc 32 29 1 (303) 5 (279) 6 (265) 2 (305)MCHC> 32 174 17 (1,576) 0.22 21 (1,559) 0.61 30 (1,493) 0.86 10 (1,724) 0.98

Total 203 18 (1,879) 26 (1,838) 36 (1,758) 12 (2,029)

* Entry in parentheses is the total number of patient-years of observation.t The log rank test was applied to the survival analysis using life table technique.

Indian SS patients different from the American,Jamaican, or Turkish cases? If so, is the Hb-Fresponsible? What is the genetic explanation?

Erythrocyte indices and morbidity. Our data havebeen unable to show any association of Hb-A2, MCV,or MCHCvalues with the expression of clinical dis-ease. Although the number of patients with S-,l3-thalassemia is small, the severity of expression in themseems equal to that in sickle cell anemia. AlthoughHb-A2 influences the polymerization kinetics of Hb-Sin the same way as Hb-F (6), the concentrationrequired is much above the 5-6% in the patient withS-f30-thalassemia (42). Consequently, as do otherinvestigators, we conclude that interacting a- or,8-thalassemia in an S-only patient does not sufficientlychange the rheology of the erythrocyte to affect theclinical course in any demonstrable way (42, 44-48).

Wewould not be inclined to support the concept thatlowered MCV, which is induced by iron deficiency,changes the course of the illness (49).

Fetal hemoglobin and stroke. From the statisticalanalysis of 18 stroke patients (Table IV), it follows thatthe risk of stroke increases in those patients withlowered Hb-F (P 0.003 and 0.0001). A previous studyof 12 stroke patients, some of whom are included inthis study, failed to show a significant difference inHb-F between the group of stroke patients and thegroup of nonstroke patients when age-matched cohortanalysis was used (50). It should be emphasized thatthis opposite conclusion was derived from a smallersample. In contrast, the other indicators of morbidity,which are based on large numbers of data, show norelationship between Hb-F or the other laboratoryparameters.

Fetal Hemoglobin and Sickle Cell Anemia 737

If it is correct that a decreased level of Hb-F in-creases the risk of stroke, then this briefly de-scribeci hypothesis may explain why Hb-F levels,but none of the other indicators of morbiditv, doinfluence the risk of stroke. In sickle cell anemia thedistorted damaged erythrocvte must be a factor instroke. As the cell becomiies less deformable, increasedhemolytic breakdown occurs, particularly in the smallarterial capillary bed, with the resultant local increasein ionized and nonionized calciumii conicentration.Those patient's with less Hb-F would have a greaterpopulation of erythrocytes with reduced flexibility.Oxygenated sickle cells preferentially agglutinate tocapillary endothelial cells in tissue cultures in asimiilar mannier to barilacles on1 a ship hull (51). Con-se(quently, one could envision a specific phenomenonionly in the brain that would be related to the com-bine(d effects of less flexible, but still oxygenated ervth-rocytes, and increased localizedl calciumii ionization withresultant neurovascular spasm anid increased plateletfunction. The brain capillary system, which is ex-quisitely sensitive to calciumn ionl, would interact withthe calciumn overloade(d ervthrocytes, anid barnacle likeplaques of erythrocytes would accumllulate and producean obstructed circulationi. Thus, it may be suggestedthat there is sufficient variability of erythrocytevascular interrelationships in different organs to ac-count for a variability of predictive factors. There-fore, a given f;actor might be influential in producingdamage to one organ but not to another. The statisticalanalyses implicate Hb-F concentrationi in producingthe risk of stroke but not of aseptic necrosis. Theanatomy and pathophysiology of brain and bone aresufficiently different that different predictive factorsare not surprising. In sickle cell anemia, where the

1.000

0.875

-i< 0.750

r co

3 o 0.6254 O)

OUl, Z 0.500

X L 0.375

£ 0.250

0.125

-% Hb-F >13

8<%Hb-F<13%Hb-FI8

_ ~ *_. ___

- * I . . .

0 5 10 15 20 25 30 35 40 45 50AGE IN YEARS

FIGURE 4 Survival curves for aseptic necrosis in three groupsof patients as defined by percent Hb-F levels (lemonstrating nosignificant association.

1.000

0.875 [

< UJ 0.7505 0

-Q

' 0.3755 0.500

>.00m0375

mmcoM 0.250

aQ125

FF

_-

_-_ - %Hb-F>13

-- 8C%Hb-FI13--

%Hb-FS8

35 40 45 500 5 10 15 20 25 30AGE IN YEARS

FIGuRE 5 Survival curves for cerebral vascular accidentin three groups of patients as defined by percentage of Hb-Flevels (lemonstratinig a significant association between the riskof cerebral vascular accident and levels of fetal hemoglobinbelow 13%.

pathophysiology is related to many interdependentphysicochemiiical and functional factors, it may besimplistic to expect to find a sinigle factor that wouldreproducibly produce the same risk of dcamage in allsituationis.

The extraordinarily small values of the squiare ofmultiple correlations of the risk of the clinical variableswith age at entry and the laboratory variable indicatethat other factors not considered in this study may havean influence on the risk. The inability of the analyses toreveal significant relationships may be due to sm-nall sam-ple sizes, especially when subgroups of patients werestudied. Large collaborative studies are called for. Thecumulative effects of environmental, sociological,physiological, anid physicochemiiical factors, coupledwith the remarkable age-related clinical variabilitythat is observed in sickle cell anemia, overshadow theinfluence of identifiable hematological or biochemicalparameters on the overall disease process and theclinical severitv.

ACKNOWLEDGMENTS

We acknowledge the cooperation of Dr. Cage Johnison,Svlvia Lee, R.N., P.H.N., M.P.H., anid Pat Corley, R.N., ofthe clinical staff, and of Jean Hilburn, Laurence Pace, LoisKay, Stephen Baker, and Colleen Warford for laboratory de-terminations. The support of Dr. Bernard Portnoy, anid thesecretarial assistance of Joani Dreniikowski and Gail Atkinsonare appreciated.

This investigation was supported in part by grant HL-15162 from the National Institutes of Health, U. S. PublicHealth Service, and contribution 6035 from the Divisionof Chemistry and Chemical Engineering, CaliforniaInistittute of Technology, Pasadena, Calif.

738 D. R. Powcars, WV. A. Schlroe(ler, J. N. Weiss, L. S. Chani, anid S. P. Azen

% \-% N.1% "-

v

l1

REFERENCES

1. Bertles, J. F. 1974. Human fetal hemoglobin: sig-nificance in disease. Antn. N.Y. Acad. Sci. 241: 638.

2. Bertles, J. F., R. Rabinowitz, and J. Dobler. 1970. Hemo-globin interaction: modification of solid phase com-position in the sickling phenomenon. Science (Wash.D. C.). 169: 375.

3. Bookchin, R. NI., and R. L. Nagel. 1973. Conformationalrequiremeints for the polymllerization of hemoglobin S:stu(lies of mixed liganded hybrids.J. Mol. Biol. 76: 233.

4. Hofrichter, J., P. D. Ross, and W. A. Eaton. 1976. Aphysical description of hemoglobin S gelation. In Pro-ceedinigs of the Sysmposiumll on Molecular and CellularAspects of Sickle Cell Disease. J. I. Hercules, G. L.Cottam, NI. R. Waterman, and A. N. Schechter, editors.Department of Health, Education, and Welfare publica-tion number 76-1007, Bethesda, MId. 185.

5. Cottam, L., K. Shibata, and NI. R. Waterman. 1978. Ef-fectors of the rate of deoxyhemoglobin S polymeriza-tion. in Biochemical and Clinical Aspects of HemoglobinAbnormalities. W. S. Caughey, editor. Academic Press,Inc., New York. 695.

6. Nagel, R. L., R. NI. Bookchin, J. Johnson, D. Labie, H.Wajcman, WV. A. Isaac-Sodeve, G. R. Honig, G. Schiliro,J. H. Crookston, and K. Mlatsutomo. 1979. Structuralbases of the inhibitory effects of hemoglobin F andhemoglobin A2 on the polymerization of hemoglobin S.Proc. Natl. Acad. Sci. U. S. A. 76: 670-672.

7. Bertles, J. F., and P. F. A. Milner. 1968. Irreversiblysickled ervthrocvtes: a consequence of the heterogeneous(listribution of hemoglobin types in sickle cell anemia.

J. CliGu. Incvest. 47: 1731.8. Chien, S., S. Usami, and J. F. Bertles. 1970. Abnormal

rheology of oxygenated blood in sickle cell anemia. J.CliGn. Incvest. 49: 623.

9. Briehl, R. 1978. Sickle cell: a metastable disease.Natn re (Lon1doni) 276: 666.

10. MIiale, T. D., D. L. Lawsoni, L. S. Rotolo, and F. S.Rotolo. 1978. Semiquantitative comiiputerized imageanaly sis of fetal hemoglobin distribution patterns insickle cell anemiiia and its variants. Hemtioglobin. 2: 291.

11. Dover, G. J., S. H. Bover, S. Charache, and K. Heintzel-mani. 1978. Individual variation in the production andsurvival of F cells in sickle cell disease. N. Enigl. J.AMe(1. 299: 1428.

12. Conlev, C. L., D. J. Weatherall, S. N. Richardson, NI. K.Shepard, and S. Charache. 1963. Hereditary persistenceof fetal hemoglobin: a study of 79 affected persons in 15Negro families in Baltimore. Blood 21: 261.

13. Fessas, P., and G. Stamotovannopoulos. 1964. Hereditarypersistence of fetal hemoglobin in Greece. A study and acomparison. Blood 24: 223.

14. Ali, S. A. 1970. Mlilder variant of sickle-cell disease inArabs in Kuwait associated with unusually high level offoetal haemoglobin. Br. J. Haemnatol. 19: 613.

15. Haghshenass, NI., F. Ismail-Beigi, J. B. Clegg, and D. J.WVeatherall. 1977. MIild sickle-cell anaemia in Iranassociated with high levels of fetal haemoglobin. J.Mled. Genet. 14: 168.

16. Gelpi, A. P. 1970. Sickle cell disease in Saudi Arabs.Acta Haemnatol. (Basel). 43: 89.

17. Perrine, R. P., NI. J. Browni, D. J. XVeatherall, J. B.Clegg, anid A. May. 1972. Benign sickle cell anaemiia.Lanicet. II: 1,163.

18. Perrine, R. P., MI. E. Pembrey, P. John, S. Perrine, andF. Shoup. 1978. Natural history of sickle cell anemia in

Saudi Arabs: a study of 270 subjects. Ann. Intern. Med.88: 1.

19. Pembrey, M. E., W. G. Wood, D. J. Weatherall, and R. P.Perrine. 1978. Fetal haemoglobin production and thesickle gene in the oases of eastern Saudi Arabia.Br. J. Haematol. 40: 415.

20. Brittenham, G., B. Lozoff, J. W. Harris, S. M. Mayson,A. Miller, and T. H. J. Huisman. 1979. Sickle cellanemia and trait in Southern India: further studies.Anii. J. Hematol. 6: 107-123.

21. Lehmann, H.. and R. G. Huntsman. 1974. Man'sHaemoglobins. J. B. Lippincott Co., Philadelphia,Pa. 2nd edition. 268.

22. Weatherall, D. J., J. B. Clegg, J. Blankson, and J. R.McNeil. 1969. A new sickling disorder resulting frominteraction of the genes for haemoglobin S anda-thalassemia. Br. J. Haematol. 17: 517.

23. Van Enk, A., A. Lang, J. M. White, and H. Lehmann. 1972.Benign obstetric history in women with sickle-cellanaemia associated with a-thalassemia. Br. Med. J. 4:524.

24. Seakins, NI., W. N. Gibbs, P. F. Milner, and J. F. Bertles.1973. Erythrocyte Hb-S concentration. An importantfactor in the low oxygen affinity of blood in sicklecell anemia. J. Clini. Intvest. 52: 422.

25. Serjeant, G. R., B. E. Seijeant, P. Desai, K. P. Mason,A. Sewill, and J. M. England. 1978. The determinantsof irreversibly sickled cells in homozygous sicklecell disease. Br. J. Haemiatol. 40: 431.

26. Clark, M. R., N. Mohandas, and S. B. Shohet. 1978. De-formnability of irreversibly sickled cells, influence ofMCHC.Blood. 52(Suppl. 1.): 96. (Abstr.)

27. Schroeder, W. A., L. Evans, L. Grussing, E. C. Abraham,T. H. J. Huisman, H. Lam, and J. B. Shelton. 1976.Quantitative microchromatographic determination ofhemoglobin F in patients with hemoglobins S and/or C.Am. J. Hemiiatol. 1: 331-342.

28. Schroeder, W. A., L. A. Pace, and T. H. J. Huisman.1978. Microchromatography of hemoglobins. VIII. Ageneral qualitative and quantitative method in plasticdrinking straws and the quantitative analysis of Hb-F. J.Chrotncatogr. 145: 203.

29. Huisman, T. H. J., W. A. Schroeder, A. N. Brodie, S. M.Mayson, and J. Jakway. 1975. 'Microchromatography ofhemoglobins. III. A simplified procedure for the de-termination of hemoglobin A2.4. Lab. Clitn. AMed. 86: 700.

30. Kleihauer, E., H. Braun, and K. Betke. 1957. Demon-stration von fetalem hamoglobin in den erythrocyteneines blut ausstriches. Klimi. XVochenischr. 35: 637.

31. Afifi, A. A., and S. P. Azen. 1979. Statistical Analysis:A Computer Oriented Approach. Academic Press, Inc.,New York. 2nd edition.

32. Peto, R., MI. C. Pike, P. Armitage, N. E. Breslow, D. R.Cox, S. V. Howard, N. NIantel, K. McPherson, J. Peto,and P. G. Smith. 1976. Design and analysis of randomizedclinical trials requiring prolonged observation of eachpatient. I. Introduction and design. Br. J. Cancer. 34:585.

33. Peto, R., NI. C. Pike, P. Armitage, N. E. Breslow,D. R. Cox, S. V. Howard, N. Mantel, K. McPherson, J.Peto, and P. G. Smnith. 1977. Design and analysis ofrandomized clinical trials requiring prolonged observa-tion of each patient. II. Analysis and examples. Br.

J. Catncer. 35: 1.34. Powars, D. R., anid WV. A. Schroeder. 1978. Progress in

the natural historv studies of the clinical severity ofsickle cell disease: epidemiologic aspects. In Bio-

Fetal Hemoglobitn atnd Sickle Cell Atnemia 739

chemical and Clinical Aspects of Hemrolgobin Ab-normalities. W. S. Caughey, e(litor. Academiiic Press, Inc.,New York. 151.

35. Maantel, N. 1966. Evaluationi of survival (lata an(l twonew rank order statistics arisinig in its consideration.Cantcer Chemt other. Rep). 50: 163.

36. Steinberg, M. H., B. J. Drieliing, F. S. Morrison, andlT. F. Necheles. 1973. Mild sickle cell (lisease. Clinicaland laboratory studlies. JAMA (J. Atmi. ,Mled. Assoc.).224: 317.

37. Karayalcin, G., F. Rosnier, K. Y. Kimn, P. Chan(dra, anidA. J. Aballi. 1975. Sickle cell anemiiia. Clinical mani-festations in 100 patients an(d review of the literatuire.Amii. J. Med. Sci. 269: 51.

38. Konotey-Ahulu, F. I. D. 1971. Compiputer assisted anialysisof data on 1697 patients atteniding the sickle cell/haemoglobinopathy clinic of Korle Bu Teaching HIospital,Accra, Ghana. Clinical features: I. Sex, genotype, agerheumatism and daetvlites fre(Iuenicies. Ghania Me!d. J.10: 241.

39. Powars, D. R. 1975. Natural history of sickle celldlisease: the first ten years. Semini. Hematol. 12: 267.

40. Ozsoylu, S., and N. Altinioz. 1977. Sickle-cell anaemiiain Turkey. Evaluation of 97 (ases (with p)arents' finid-ings). Scatn(l. J. Haematol. 19: 85.

41. Seijeant, G. R., R. Richards, P. R. H. Barbor, anid P. F.Milner. 1968. Relatively benigni sickle cell ainaemllia in 60patients aged over 30 in the West Indlies. Br. MIed. 1.3: 86.

42. Sejeant, G. R., A. Xi. Sonimiiereux, NI. Stevenson, K.Mason, and B. E. Serjeant. 1979. Comnparison of sicklecell-,p-thalassemia with homozygous sickle cell (lisease.Br. J. Haematol. 41: 83.

43. Wrightstonie, R. N., and( T. H. J. Iluissman. 1974. On thelevels of hemoglobins F an(d A2 in sickle-cell aniemiiiaau(I somle related (lisor(lers. Am.jI. Clin. Pathol. 61: 375.

44. Natta, C. 1978. Failture of the a-thalassemnia genie todecrease the severity of sickle cell anemia. Bloodl. 51:1163.

45. joishy, S. K., P. F. Griner, and( P. T. Rowvlev. 1976.Sickle f8-thalassemia: identical twins differing in severityiniplicate nonigenetic flactors influencing couirse. Amrl.Hema( tol. 1: 23.

46. Honig, G. R., M. Koshy, R. G. Mason, and(i L. N. Vida.1978. Sickle cell synv(roiiies. 11. The sickle cell anemiiia-a-thailassemilia synCdrome. J. Pediatr. 92: 556.

47. Steinberg, NI. H., and( B. J. D)reiling. 1976. Clinical,hematologic an(l biosynthetic stuidies in sickle cell-,l-thalassemiiia: a comnparisoni with sickle cell anemia.Am.J. lIemtiatol. 1: 35.

48. Felice, A. E., B. Webber, A. MIiller, S. MI. MIavsoni, H. F.Hairris, J. B. Ilenison, N1. E. Gravely, ani(l 1T. H. J.Huiismani. 1979. The associationi of sickle cell anlemliawith heterozygotis au(I homozvgous a-thalasseminia-2:in vitro HB chain synthesis. Am. J. H'ei1atol. 6: 91-106.

49. Lincoln, T. L., J. Aroesty, andl P. Miorrison. 1973. Ironi-deficienicy aniemlia andl sickle cell disease: a hypothesis.Lantcet. II: 260.

50. Povars, D., B. Wilson, C. Imbus, C. Pegelow, ancI J.Allen. 1978. The niatuiral historv of stroke in sicklecell (lisease. Amw. J. MXIcd. 65: 461.

51. flebbel, R. P., 0. Yaimtada, C. F. Mloldow, J. G.White, and J. WV. Eaton. 1978. Clinical adherentce ofsickle erythrocytes to vascular endothelitiin. Blood.52(Snppl. 1.): 98. (Abstr.)

740 D. R. Potars, WV. A. Schroe(ler, J. N. WVeiss, L. S. Chant, anid S. P. Azeni