Influence of lipids and obesity on haemorheological parameters in patients with deep vein thrombosis
It is not well established whether haemorheological alterations constitute independent risk factors for deep vein thrombosis (DVT). We have determined in 149 DVT patients and in 185 control subjects the body mass index (BMI), the haemorheological profile: blood viscosity (BV), plasma viscosity (PV), fibrinogen (Fg), erythrocyte aggregation (EA), erythrocyte deformability (ED) and plasma lipids. In the crude analysis BMI, Fg, PV, EA, triglycerides (TG) and ApoB were statistically higher and HDL cholesterol (HDL-Chol) statistically lower in DVT patients than in controls. No differences in BV and ED were observed. After BMI adjustment, Fg, PV and EA remained statistically higher in DVT cases than in controls (P = 0.013; P = 0.012; P = 0.013; P = 0.028, respectively). When the risk of DVT associated with these variables (using cut-offs that corresponded to the mean plus one SD of the control group) was estimated, EA > 8.2 and PV > 1.28 mPa . s were significantly associated with DVT even further adjustment for lipids and obesity (OR = 2.78, P = 0.004; OR = 1.91, P = 0.024, respectively). However, PV did not remain statistically significant after additional adjustment for Fg. When we consider together all the analyzed variables in order to control every variable for each other, TG > 175 mg/dl (OR = 3,2, P = 0.004) and BMI > 30 kg/m(2) (OR = 3.5, P = 0.003), were also independently associated with a greater risk of DVT. Our results suggest that increased EA constitute an independent risk factor for DVT. However, when associated to hyperlipidaemia and obesity it further increases thrombotic risk.
©2007 Schattauer GmbH,Stuttgart
Influence of lipidsand obesityonhaemorheological parameters in
patients withdeep vein thrombosis
Haemorheologyand ThrombosisUnit, Department of Clinical Pathology, La Fe University Hospital,Valencia, Spain;
Service,LaFeUniversity Hospital,Valencia, Spain;
Research Centre, La Fe University Hospital,Valencia, Spain;
Epidemiology Unit, School of Medicine,University of Valencia, Spain;
CIBER Fisiopatología de la obesidadynutrición, Instituto de Salud
Carlos III, Madrid, Spain
It is not well establishedwhether haemorheological alterations
constitute independent riskfactors fordeep vein thrombosis
(DVT).We have determinedin149 DVTpatientsand in 185 con-
trol subjects the bodymass index(BMI), the haemorheological
profile: blood viscosity (BV), plasma viscosity (PV), fibrinogen
(Fg),erythrocyte aggregation (EA), erythrocyte deformability
(ED)and plasma lipids.Inthe crudeanalysis BMI, Fg,PV, EA, trig-
lycerides (TG)andApoB were statisticallyhigher and HDLcho-
lesterol (HDL-Chol) statisticallylower in DVTpatientsthan in
controls.No differencesinBVand ED were observed.After BMI
adjustment, Fg, PV and EA remainedstatisticallyhigher in DVT
cases than in controls (P=0.013; P=0.012; P=0.013; P=0.028, re-
spectively).Whenthe risk of DVTassociated with these vari-
Deep vein thrombosis,lipids,obesity,haemorheology
ables (using cut-offs thatcorrespondedtothe mean plus one SD
of the control group) wasestimated,EA>8.2 and PV>1.28 mPa
were significantlyassociated with DVTevenfurther adjustment
forlipids and obesity (OR=2.78,P=0.004;OR=1.91,P=0.024,re-
spectively).However,PVdid not remainstatisticallysignificant
after additional adjustmentfor Fg.When we consider together
all theanalyzedvariablesinorder to control everyvariable for
each other,TG>175mg/dl (OR=3,2,P=0.004) and BMI>30 kg/m
(OR=3.5, P =0.003),werealsoindependentlyassociated with a
greaterriskofDVT.Our results suggestthat increasedEAcon-
stitute an independent risk factor forDVT.However,whenas-
sociated to hyperlipidaemiaand obesity it further increases
ThrombHaemost 2007; 98: 621–626
Blood Coagulation, Fibrinolysis and CellularHaemostasis
Hemorheologyand Thrombosis Unit
Department of Clinical Pathology
La Fe University Hospital
Avda.Campanar 21,Valencia 46009, Spain
Te l.: +34963862714, Fax: +34 96 1973109
Thisstudy waspartially supportedbygrants from Instituto de SaludCarlos III(CIBER
CB06/03/0035 and RedRECAVARD06/0014/0004), Spain.
Accepted after resubmission May29, 2007
Prepublished onlineAugust18, 2007
Deep vein thrombosis (DVT) is amultifactorialdiseaseinwhich
inherited and acquiredrisk factorsplayanimportant pathogenic
role(1). Since Virchow, it is assumedthat blood flowalterations
(haemorheological alterations) mayfavour the development of
venous thrombus (2).The rheologicalhypothesis for venous
thrombogenesis is supported by the association of manyrisk fac-
torsfor DVTwith systemic rheological alterations(3, 4) and the
association of some rheologicalvariableswith DVTincase-con-
trol studies(5–10). However, there is not enoughevidenceto
prove thathaemorheological alterationsconstitute independent
risk factorsfor DVTwhen patients areevaluatedout of the acute
phase,when the temporary fibrinogen increase (the protein
whichmodulates most of the haemorheologicalparameters)is
stronglymarked. Moreover, studiesconductedsofar consist
mainlyofasmall sample size (6,7,9,11, 12), do not determine
allthe rheologicalparameters (5–8,10), do not excludepatients
with malignant, inflammatory, autoimmune or infectious dis-
eases (11, 12)that entail markedfibrinogenincreases and,inad-
dition, theydonot always considerthe presenceofconcomitant
classical cardiovascularrisk factors, i.e.obesity,hypertension,
diabetes, dyslipidaemia and tobacco (6–9, 11, 12), whichcan not
be overlookedastheyaffect rheological blood behaviour,acting
as confounders. In particular, both obesity (13–18) and hyper-
lipidaemia (16,19–26) appear to be associated with atwo-to
four-foldincreased risk of DVT. The prothrombotic pathogenic
mechanisms of obesity, as well as hyperlipidemia, arenot clearly
Vayá et al. Lipids, obesity and haemorheologyinDVT
established. One possible mechanismcould be by modifying the
blood flowcharacteristics, giventhat lipids and fibrinogen,
whichare increased in both conditions,may modulate most of
the haemorheological parameters.
The aim of the present study wastoassess the overallhae-
morheological profile six to 12 monthsafter the acuteepisode in
patients with DVT, i.e.blood viscosity (BV), plasmaviscosity
(PV) and erythrocyte aggregation (EA) and deformability (ED),
to find out whether some haemorheological parameters consti-
tute independent risk factorsfor DVTand to determine whether
the lipid alterationsthemselves or those due to obesity could
modify rheologicalparameters and thus promote DVT.
One hundred and forty-nine patients withpreviouslydocu-
mented DVT(86 male, 63 female) aged42±12 years were re-
ferred to our Haemorheology and Thrombosis Unit between Ja-
nuary2003 and October 2006. Theyall hadexperienced afirst
DVTepisode six to 12 months before sampling (mean, 9±3
months).Exclusion criteria were: organic, malignant, infectious,
autoimmune or inflammatorydiseases,i.e. those pathologies
that could influencerheologicalprofile,aswellastreatment with
rheological drugs. Those patients with inherited(antithrombin,
protein Corprotein Sdeficiency, factor VLeiden,prothrombin
G20210A mutation) or acquired (lupus anticoagulant or anticar-
diolipin antibodies) thrombophilic risk factorswere also ex-
cluded as their thromboembolic episodescould be explained by
The control group wasmade up of 185subjects (aged 42 ±13
years,89male, 96 female) undergoing aroutine check up at our
hospital, without aprevioushistoryofDVT confirmed witha
validatedquestionnaire (27).Cardiovascularrisk factorswere
recorded and consideredfor both groups.These included obesity
(body mass index[BMI]>30 kg/m
), current tobacco use(>1
cigarette/day), hypertension (diastolic blood pressure >90
mmHg), hyperlipidaemia (total cholesterol >220 mg/dl and/or
triglycerides>175 mg/dl), fasting glucose >126 mg/dl,orwere
receiving anypharmacologicaltreatment for hypertension, hy-
perlipidaemia, or diabetes. Subjects gave their informedconsent
to participate in the study,which wasapprovedbythe Hospital
Circumstantial patient thrombotic risk factors, i.e.medical
(oral contraceptives,varicoseveins,pregnancy, bed rest >1
week, heartfailure, chronic obstructive pulmonarydisease,
obesity,etc), surgery, immobilization, and traumawere recorded
(secondaryDVT), and thosethrombotic eventsthat occurred
withoutany circumstantialrisk factor were considered sponta-
neous DVT. From the 149 DVTpatients, 111 showedsome cir-
cumstantial risk factors(75%) and 38 did not (25%). In the 111
DVTpatients, the percentages of circumstantialrisk factorswere
as follows:medical 41.7%, surgical 17.5%, immobilisation
20.5% and trauma20.3%. In addition, 16 of the 149 DVTpa-
tients hadapulmonaryembolism(PE). DVTwas documented
with ultrasonographyorvenography, and PE withventilation
perfusion scanning, pulmonary angiography, or helicoidalcom-
Bloodwas collected at leastsix monthsafter the acuteevent
(range 6–12 months;mean: 9±3months). After a12-hour over-
night fast, blood wasdrawn between 8and 10 a.m. by venipunc-
ture into standard vacuumtubes containing EDTAK
logical and haematological measurements, 0.1 volof0.129 M
trisodium citrateasananticoagulant for fibrinogen measure-
ment, or into plain tubesfor glucose and lipids determination.
Rheological parameters were examinedwithin 2hours of blood
collection, to avoid deterioration of the rheologicalred blood cell
Bloodviscosity (BV) wasdeterminedinaBrookfield DVIII vis-
cosimeter(Engineering, Stougthon MA,USA)atnativeand
45% correctedhematocrit with autologous plasma, at twoshear
rates of 230s
and 23 s
,at37ºC. Plasma viscosity (PV) was
measured in acapillary plasma viscosimeter (Fresenius GmbH,
Germany) at 37ºC (29).Erythrocyte aggregation (EA) wasdeter-
Roetgen, Germany) (30) after adjustingthe haematocrit to 45%
with autologous plasma, during completedstasis (EA
while the sample wassubjected to alow shear rateof3s
Thegreater the tendencyofred blood cell to aggregate,the
higher the aggregability index. Erythrocyte deformability(ED)
renne GmbH)at12, 30 and 60 Pascals(31). The higher the ery-
throcyte elongation index(EEI), the more deformable the red
blood cell is. Fibrinogen (Fg) wasmeasured using coagulometric
oratory, Milan, Italy). Haematocritwas measured by microcen-
trifugation at 15,000 xgfor 10 minutes.
Totalcholesterol (T-Chol), low-density lipoprotein-choleste-
rol (LDL-Chol), high-densitylipoprotein-cholesterol (HDL-
Chol),triglycerides,and glucose were measured by enzymatic
techniques, usingaDAX 72 autoanalyser (BayerDiagnostics
Division, Tarrytown, NY,USA). ApolipoproteinsAIand Bwere
Basic haematological parameters and red cellindices, in-
cluding meancellvolume (MCV), meanhaemoglobin concen-
tration (MHC),and meancorpuscularhaemoglobin concen-
tration (MCHC), were measured in aSysmexME8.000 autoana-
lyser(TOAMedical Electronics,Kobe, Japan).
BV wasdeterminedinarandom sample of 102 cases and 105
controls and ED in 40 cases and 59 controls. Casesand controls
were well matchedfor age and sex. Patients and controls samples
were processed simultaneously.
All continuousvariableswere checked fornormal distribution.
TG, glucose and Fg valueswere log-transformed forstatistical
testing.The data areexpressed as mean±one standard deviation
(SD). Student'st-test wasusedtoassess the meandifferences in
continuous variablesbetween patients and control subjects, and
-testswere used to comparedifferences in percentages be-
tween patients and controls. Pearson correlation coefficients
were calculatedtodescribe the bivariate correlation amongvari-
Vayá et al. Lipids, obesity and haemorheologyinDVT
ables. Multiple linealregression analysis (covariance method)
wasusedtoadjust for the influenceofBMI or the otherparam-
eters on the unadjusteddifferences of means in lipids,fibri-
nogen, plasma viscosity and erythrocyteaggregation between
DVTpatients and controls. In these analyses,weuseddifferent
models dependingonthe adjusted variable.First, we adjusted
each variable thatwas statistically significant in the crudemodel
for all the other variablesthat were statistically significant in the
unadjustedanalysis. Taking into account that after these adjust-
ment BMI wasthe onlyvariable thatremained highly signifi-
cant,wefittedadditional multivariate models in whicheach lipid
and haemorheological parameters were only adjustedfor BMI.
Logistic regression models (crude and adjusted for potential
confounders) were fittedtoestimate the odds ratio (OR) and
95%confidenceinterval(CI)ofDVT associated withthe pres-
enceofobesity and/or the other haemorheologicalparameters.In
these analyses,dichotomisedvariablesaccording to cut-off
points that corresponded to the meanplus one SD of the control
group were used. Standard regression diagnostic procedures
were used to ensurethe appropriateness of the models.
Forstatistical inference, abilateral p-value<0.05 wascon-
sideredstatistically significant. All analyses were calculated
using the StatisticalPackagefor Social Sciences (SPSS,version
14) for Windows.
Table1shows the percentage of DVTpatients and controls with
cardiovascularrisk factors. DVTpatients showedahigherper-
centage of hyperlipidaemia (46% vs. 35%, P=0.040) and obesity
(23% vs. 6%, P=0.001) than controls. Whenhypercholesterole-
mia (T-Chol>220 mg/dl)and hypertriglyceridemia (TG
>175 mg/dl)were consideredseparately, DVTpatients showeda
borderline increased percentage of hypercholesterolemia
(P=0.040) and ahigherpercentage of hypertriglyceridemia
(P=0.001). The percentage of hypertensives, diabetics and
smokers wassimilarinDVT patients and control subjects.
Table2shows BMI, glucose,lipids and haemorheological
parameters in the 149 DVTpatients and 185control subjects.
OnlyBMI,TG,Apo B, B/AIratio, Fg, PV,EA
nificantly higher and HDL-Chol lowerincases thanincontrols.
After multivariate adjustment for the potentialconfounding fac-
tors, onlyBMI remainedsignificantly different (P<0.001). After
adjusting for BMI, onlyFg, PV,EA
cantly different betweenpatients and controls.
No statistical differences in rheological parameters were ob-
served between patients with spontaneous and secondaryDVT.
No differences in erythrocyteindiceswere observed between
patients and controls (MCV: 91 ±10vs. 90 ±5fl,P=0.782;
MCH: 31 ±2vs. 30 ±2pg, P= 0247; MCHC:33±1vs. 33 ±1%,
P=0.875).The Pearson bivariate correlations between haemor-
heologicaland lipidparameters were statistically significant
(P<0.01) between EA
and BMI (r=0.334), T-Chol(r=0.268),
LDL-Chol(r=0.269), Apo B(r=0.344), TG (r=0.356), Fg
(r=0.355), PV (r=0.412) and there wasanegative correlation be-
and HDL-Chol(r=-0.152) ( P <0.05 in all cases).In
addition, PV correlated with BMI (r=0.196), Apo B(r=0.186),
TG (r=0.153) ,and Fg (r=0.508) (P<0.01), and T- Chol(r=0.136)
Giventhe multiple correlations betweenthe rheologicalpa-
rameters,lipids and BMI,toestimate the risk of obesity associ-
ated with these parameters, alogistic regression analysis was
carried out. Instead of continuous variables, dichotomized vari-
ables(according to cut-off points that corresponded to the mean
plus one SD of the control group) were used for the analysed pa-
rameters to betterestimate the risk associatedwith high values.
Crude and multivariate models were estimatedfor each variable.
Table3shows OR of DVTassociatedwith haemorheological pa-
>8.2, PV>1.28 mPa
s, Fg>320 mg/dl) and obesity
). In the crudeModel Iall parameters were sig-
nificantly associatedwith ahigherrisk of DVT. This association
remained statisticallysignificant even further adjustment for li-
pids (ModelII) and BMI (Model III). When PV wasadditionally
adjustedfor Fg (ModelIV) this parameter didnot remained stat-
,BMI and Fg remained statis-
tically significant in ModelIV.
Moreover, when we consideredtogether allthe analysedvari-
ablesinorder to control everyvariable for each other and to know
whether theyconstitute independent risk factorsfor DVT, we ob-
tained that TG >175 mg/dl (OR=3.2, 95% CI=1.4–7.3,
P = 0.004), Fg>320 mg/dl (OR=2.7, 95% CI=1.3–5.6, P=0.008),
>8.2 (OR=2.8, 95% CI=1.3–5.9, P=0.007) and BMI >30
(OR=3.5, 95%CI=1.5–8.1, P=0.003),but not PV >1.28
snor T- Chol>220 mg/dl remained significantly associated
withDVT.Therefore TG, Fg, EA
and BMI constitute indepen-
dent risk factorsfor DVTwhen theyreach high values(higher
than meanplus one SD).
The results obtained in the present study indicate that DVTpa-
tients do not showahigherBVcompared to the control group,
Ta ble1:Age,gender, BMIand incidence (%) of cardiovascular
risk factor in DVTpatients and controls.
Age (mean ±SD) 42 ±1242±130.982
Male/female 86/63 89/96 0.082
)27.6 ±4.23 24.5 ±3.5 0.001
*Hyperlipidaemia (%) 46 35 0.040
***Hypertriglyceridemia(%) 18.7 0 5.7 0.001
Obesity (%) 23 0 60.001
Hypertension(%) 13.5 0 8.2 0.114
Diabetes (%) 0 4.2 0 2.2 0.326
To bacco (%) 31.1 38 0.183
*T-Chol>220mg/dland/or TG>175 ml/dl or were receiving anypharmacologicaltreatment.
** T- Chol>220 mg/dl or were receiving anypharmacologicaltreatment.
*** TG>175 ml/dl or were receiving anypharmacologicaltreatment.
Vayá et al. Lipids, obesity and haemorheologyinDVT
both at native and 45%correctedhaematocrit. Little information
is availableonDVT case-control studies regarding BV evalu-
ation. Ourresults agree with thosepreviouslyreported (5,6,9,
11, 12) in not having found differencesbetween both groups. It is
well known thatBV, when determinedathigh shear rate, is aglo-
balrheologicalparameter,depending on haematocrit, Fg con-
centration and ED.Given thatDVT patients in thepresentstudy
did not showalower ED either, it is reasonable that, at the shear
rates used (230 s
and 23 s
)whereBVbasically depends on ED,
the formerdid not showany significant differencewith respect
to the control group. The significant increase in Fg observedin
cases compared with controls doesnot seemtobehigh enoughto
producesignificant increases in BV.Our resultsagree with sev-
eral studieswhereDVT patients show significant Fg increases
even three monthsormore after the acute event (5,6,8–12). In
line with these results, in the population-basedcasecontrol study
Ta ble2:BMI, glucose,li-
Haematocrit (%) 0 43 ±4 0 42 ±4
s) 001.25±0.07 001.22 ±0.06
003.85 ±1.28 003.46 ±1.02
007.58 ±1.36 006.99 ±1.17
Nativeblood Viscosity 230s
s) 004.34±0.58 004.29 ±0.52
s) 006.61±1.0 006.43 ±1
s) 004.63±0.39 004.61 ±0.31
s) 007.09±0.69 007.06 ±0.61
Erythrocyte elongation index 12 Pa (%) 0 47 ±3 0 47 ±4
Erythrocyte elongation index 30 Pa (%) 0 53 ±3 0 54 ±4
Erythrocyte elongation index 60 Pa (%) 0 56 ±3 0 57 ±3
Valuesexpressed as mean±SD.*The corresponding P-value of each variable was adjusted forthe othervariablesthat were statisticallysignificant in the un-
adjustedanalysis. **Thecorresponding P-value was adjusted onlyfor BMI.
) 0 27.6 ±4.23 0 24.5 ±3.5
Glucose(mg/dl) 0 91 ±14 0 90 ±14.34
To talcholesterol (mg/dl) 211 ±47203 ±37
HDL-Cholesterol (mg/dl) 0 50 ±15 0 55 ±13
LDL-Cholesterol(mg/dl) 133 ±38130 ±34
Triglycerides (mg/dl) 127 ±91 0 94 ±48
Apo AI (mg/dl) 144 ±32145 ±25
Apo B(mg/dl) 103 ±30 0 95 ±23
B/AI 000.74 ±0.26 000.67 ±0.21
Fibrinogen (mg/dl)282 ±72252 ±46
Ta ble3:Risk of DVTassociated withhaemorheological parametersand BMI. Crudeand adjusted models.
POR(95% CI) P
>8.20.004 2.85 1.35–6.010.006
PV>1.28 0.024 1.47 0.81–2.71 0.205
Fg>320 0.001 +2.91 1.42–5.94 0.004
BMI>30 ---- ++4.19 1.84–9.52 0.001
s; Fg:mg/dl; BMI:kg/m
.*ModelII. Adjusted forlipids: T- Chol>220 mg/dl and TG >175 mg/dl. ** ModelIII. Adjusted forlipids and BMI >30kg/m
.*** ModelIV. Adjusted forlipids, BMI >30kg/m
and Fg >320 mg/dl. +Adjustedfor lipidsand BMI >30kg/m
.++Adjustedfor lipidsand Fg >320 mg/dl.
Model II adjusted*
Model III adjusted**
Vayá et al. Lipids, obesity and haemorheologyinDVT
conducted by Koster (8), in which199 patients with afirst DVT
episode were compared with 199 healthysubjects, those with
plasma Fg higher than 500 mg/dl had an almost four-fold in-
creased risk of DVT.
With respect to ED,asmentioned, we did not observe differ-
ences between cases and controls at anyofthe shearstresses
tested. There have beennostudiespublished,sofar,wherethis
rheologicalparameter has been evaluatedusing ektacytometric
techniques, as in the present one. Previous studies(32, 33) that
found controversialresults regarding ED in DVTpatients, used
older filtration methodsthat maybeinfluencedbyextra-erythro-
cytic factorssuch as remnant leucocytes, redblood cellaggre-
gatesorplasma factors(34, 35), whereas ektacytometry does
As regards the otherrheologicalparameters analysed in the
present study,itisimportant to highlight thatDVT patients have
shown increased PV and EA when compared with controls. It is
known thatboth rheologicalparameters are influenced by plas-
ma lipids (10, 36–39) and Fg levels (40–42), whichhavealso
shown to be increased in our study. Additionally,obesity is also
associated with increased Fg and TG levels (14,16, 21, 43–46).
In the present study,the risk associated with obesity (BMI
≥ 30 kg/m
)was 3.5(95% CI 1.5–8.1), whichisinline with pre-
viousstudiescarried outbyour group where the risk of DVTas-
sociatedwith obesity was2.5 (95% CI 1.2–5.1) (16).For this rea-
son, the first question we should consideris: Is the increasein
these rheological parameters,PVand EA, due to an independent
increase in Fg and lipids or is their increase associated to obes-
ity? In other words,ifpatients with DVTwere notobese:Would
theystill have increased PV and EA?Our logistic regression
models (Table 3) indicatethat EA
>8.2 and PV >1.28 mPa
were significantly associatedwith DVTevenfurther adjustment
for lipids (ModelII) and obesity (Model III). These results sug-
gest that PV and EA areincreased in DVTpatients independently
of lipidlevelsand the obesity status. However, PV didnot remain
statisticallysignificant after additionaladjustment for Fg
(ModelIV), giventhe high influenceofFgonthis rheological
parameter.Therefore,inDVT patients EA
>8.2 constitutes arisk
factor forDVT,increasing the risk even though patients would
not be obese.However,the fact of being obesewould further in-
crease the risk.
Several authors have also found EA to be increased in DVT
patients 12 months after the acute event (6,10–12). Somere-
searchers have found higher EA only in thosepatients with per-
sistent thrombotic risk factorsbut not in those with transient
thromboticrisk factors(11). In this sense,wehavenot observed
anydifferences in this rheological parameter in connection with
the natureofthe thromboticepisode, i.e. spontaneous or second-
aryDVT.Itmust be emphasised that some studies include pa-
tients with malignancy(11, 12) where increased Fg could have
beenresponsible for erythrocyte hyperaggregability.Otherau-
thors (7) performed the study six weeks after the acuteepisode,
where the acutephase could have influenced results and,inaddi-
tion, most studies did not measure plasma lipids whichmay also
account forerythrocyte hyperaggregability.Inour study,the in-
fluence of acute phase or malignancyhas been excluded,and the
role playedbylipids on EA hasalso been evaluated, allowingus
to concludethat, according to our logistic regression analysis,
>8.2 is an independent thromboticrisk factor,increasing the
risk almost threetimes.
As regards PV in DVTpatients, onlysome studies did deter-
mine this rheological parameter,findingittobenormal (9, 10) or
increased (5, 6) in association with higher cholesterol and TG
levels (5). Moreover, Balendra et al. (5) reported PV to be an in-
dependent risk factor forDVT.Although the present study has
found PV to be increased in DVTpatients and associatedwith li-
pids and mostlywith Fg levels, it doesnot constituteanindepen-
dent risk factor forDVT (Model IV). Thisisconsistent with the
pathophysiologyofthe different thromboticlocations,asin-
creases in PV mostly promote thrombus development in the
microcirculatory areas and small arteries, whereas increases in
EA promote thrombus formation in low-shear areas suchas
pocket valves in the lowerextremities.
In conclusion, the resultsobtained in the present study,per-
formedinalarge group of patients with aDVT episode in the
previous 6–12 months,allowustoconclude that rheologicalpa-
rameters play arole in the pathogenesis of the thromboticevent.
IncreasedEAconstitute arisk factor forDVT independently of
lipidlevelsand obesity status. However, when associatedwith
hyperlipidaemia and obesity it further increases thromboticrisk.
It is reasonable to advise losingweightand to decrease plasma
lipidlevelsinorder to reduce DVTrisk.
tional anticoagulant treatment. Clin Hemorheol 1992;
12: 495 Abstract.
7. Chabanel A, HorellouH,Conard J, et al. Red blood
cell aggregability in patients withahistoryoflegvein
thrombosis: influenceofpost-thrombotic treatment.
Brit JHaematol 1994; 88: 174–179.
8. Koster T, Rosendaal FR,Reitsma PH, et al. Factor
VIIand fibrinogen levels as risk factors forvenous
thrombosis. Acase-control study of plasma levels and
DNApolymorphisms--the Leiden Thrombophilia
Study (LETS).Thromb Haemost1994;71: 719–722.
9. MiraY,VayáA,Martínez M,etal. Hemorheologi-
cal alterations and hypercoagulable stateindeep vein
thrombosis. Clin Hemorheol Microcirc 1998;19:
10. Vaya A, Mira Y, Martinez M,etal. Biological risk
factors fordeep vein thrombosis. Clin Hemorheol
Microcirc 2002;26: 41–53.
11. Alt E,Banyai S, Banyai M,etal. Blood rheology in
deepvenous trombosis--relationtopersistent and tran-
sient risk factors.Thromb Res 2002; 107:101–107.
12. Krieger E, vanDer Loo B, Amann-Vesti BR,etal.
C-reactiveproteinand redcell aggregationcorrelate
with late venousfunctionafter acute deepvenous
thrombosis. JVascSurg2004;40: 644–649.
13. Hansson PO,Eriksson H, Welin L,etal. Smoking
and abdominal obesity: risk factors forvenous throm-
boembolismamong middle-aged men: "the study of
menbornin1913". Arch Intern Med 1999;159:
SamamaMM. An epidemiologic study of risk fac-
torsfor deep vein thrombosis in medicaloutpatients: the
Sirius study.ArchInternMed 2000; 160:3415–3420.
15. Tsai AW,Cushman M, Rosamond WD,etal. Car-
diovascular risk factors and venous thromboembolism
incidence: thelongitudinal investigationofthrom-
1. Rosendaal FR. Venous thrombosis: amulticausal
disease. The Lancet 1999; 353: 1167–1173.
2. Virchow R. Thrombose undEmbolie (1846–1856).
Verlag-von JohannAmbrosius Barth, Leipzig, Ger-
3. Lowe GDO.Thrombosisand hemorheology.In:
Clinical Hemorheology. Dordrecht/Boston,Martinus
NijhoffPublishers 1987: 195–226.
4. Lowe GDO.Rheological influences on thrombosis.
Baillieres ClinHaematol 1994; 7: 573–589.
5. Balendra R, RumlevA,Orr M, et al.Blood lipids,
coagulation, fibrinolysis and rheology in spontaneous
provendeep vein thrombosis. Brit JHaematol 1991; 77
(Suppl 1): 83 Abstract.
6. Koppensteiner R, HofmannS,Atteneder M, et al.
Blood rheologyinacutedeep venous thrombosis. Fol-
lowuptoone year after theacuteevent under conven-
boembolismetiology.Arch InternMed 2002; 162:
16. Vaya A, Mira Y, Ferrando F, et al.Hyperlipidaemia
and venous thromboembolisminpatientslacking
thrombophilic risk factors.BrJHaematol 2002; 118:
17. Abdollahi M, Cushman M, Rosendaal FR. Obesity:
risk of venous thrombosisand theinteraction with co-
agulationfactor levels and oral contraceptive use.
ThrombHaemost 2003; 89: 493–498.
18. Fuchs S, Heyse T, RudofskyG,etal. Continuous
sis: arandomised comparison in traumapatients. J
BoneJoint SurgBr2005; 87: 1117–1122.
19. CogoA, Bernardi E, Prandoni P, et al.Acquired risk
factors for deep vein thrombosisinsymptomatic out-
patients. Arch InternMed 1994; 154: 164–168.
20. KawasakiT,Kambayashi J, Ariyoshi H, et al.Hy-
percholesterolemia as ariskfactor for deep vein throm-
bosis. ThrombRes 1997; 88: 67–73.
21. McCollMD, SattarN,Ellison J, et al.Lipoprotein
(a), cholesteroland triglycerides in womenwith venous
thromboembolism. Blood Coagul Fibrinolysis 2000;
22. GriffinJH, Fernández JA,Deguchi H. Plasma lipo-
proteins, hemostasisand thrombosis. Thromb Haemost
2001; 86: 386–394.
23. Gónzalez-Ordóñez AJ,Fernández-Carreira JM,
Fernández-Alvarez CR, et al. The concentrationsof
soluble vascularcelladhesion molecule-1 and lipids
are independently associated with venous thromboem-
bolism. Haematologica2003; 88: 1035–1043.
24. Doggen CJM, Smith NL, LemaitreRZ, et al. Serum
lipid levels and theriskofvenous thrombosis. Arte-
riosclerThromb Vasc Biol 2004;24: 1970–1975.
25. Deguchi H, Pecheniuk NM,EliasDJ, et al.High-
associated with venous thrombosisinmen. Circulation
2005; 112: 893–899.
26. Ageno W, Prandoni P, Romualdi E, et al. The meta-
bolicsyndrome and theriskofvenous thrombosis: a
case-control study.JThromb Haemost2006; 4:
27. FrezzatoM,TosettoA,Rodeghiero F. Va lidated
questionnaire for theidentificationofpreviousperson-
al or familialvenous thromboembolism. Am JEpide-
miol1996; 143: 1257–1265.
28. Guidelines for measurement of blood viscosisty
and erythrocytedeformability.International Commit-
teefor standarizationinHaematology, Expertpanel on
blood rheology.Clin Hemorheol 1986;6:439–453.
29. Jung F, Roggenkamp HG,Schneider R, et al.Das
Kapillarschlauch-Plasmaviskosimeter. Ein neues
Messgerätzur Quantifizierung der Blutplasmaviskosi-
tät. BiomedTech 1983; 28: 249–252.
30. Schmid-SchönbeinH,Volger E, TeitelP,etal. New
hemorheological techniques for theroutinelaboratory.
ClinHemorheol 1983; 2: 93–105.
31. Ruef P, Pöschl JMB, Linderkamp.The rheodyn
SSD for measuringerythrocytedeformability.Bior-
heology1995; 32: 357–358.
32. MiraY, Vayá A, Martínez M, et al.Effect of unfrac-
tionated sodium heparin at therapeuticdoses on thehe-
morheological parameters of patientswith deep vein
thrombosis. ClinHemorheol 1996; 16: 277–284.
33. Da SilvaMA, Cardoso L, GeraldesJ,etal. Haemor-
heological and coagulationconditions in patientswith
acutedeep venous thrombosiscompared with subjects
without vasculardisease. Clin Hemorheol 1991;11:
34. KennyMW, MeakinM,StuartJ.Methods for re-
movalofleucocytes and plateletsprior to studyofery-
throcytedeformability.Clin Hemorheol 1983;3:
35. StuartJ,Stone PCW,Bareford D, et al.Evaluation
of leucocyteremoval methods for studies of erythro-
cyte deformability.Clin Hemorheol 1985;5:137–147.
36. Razavian SM, Atger V, Giral M, et al.Influence of
HDL subfractions on erythrocyteaggregationinhyper-
cholesterolemia men. Artherioscl Thromb1994; 14:
37. Ruhenstroth-Bauer G, Hossmer G, Ottl J, et al.
Highlysignificant negative correlations between ery-
throcyteaggregationvalue and serum valueconcen-
trationofhighdensitylipoproteincholesterol in a
samplefrom anormalpopulationand in patientswith
coronaryheartdisease. European Journal of Clinical
Investigation1987; 17: 275-279.
38. Vayá A, Martínez M, Carmena R, et al.Red blood
Thromb Res 1993;72: 119-126.
39. KoenigW,Sund M, Ernst E, et al.Associationbe-
tween rheology and componentsoflipoproteins in
human blood. Circulation1992; 85: 2197–2204.
40. FalcóC,VayáA,Simó M, et al.Influence of fibri-
nogen levels on erythrocyteaggregationdetermined
with theMyrenne aggregometer and theSefam ery-
thro-aggregometer. ClinHemorheol Microcirc 2005;
41. Weng X, Roederer GO,BeaulieuR,etal. Contribu-
factors to erythrocyteaggregationinnormolipidaemic
and hyperlipidaemic individuals. ThrombHaemost
1998; 80: 903–908.
42. Vayá A, FalcóC,Réganon E, et al.Influence of
plasma and erythrocytefactors on red blood cellaggre-
gationinsurvivors of acutemyocardial infarction.
ThrombHaemost 2004; 91: 354–359.
43. Juhan-Vague I, Alessi MC.PAI-1, obesity,insulin
resistance andriskofcardiovascularevents. Thromb
Haemost 1997; 78: 656–660.
44. Margaglione M, Cappucci G, d'Addedda M, et al.
PAI-1 plasmalevelsinageneral populationwithout
clinical evidence of atherosclerosis:relationtoenvi-
ronmental and genetic determinants. Arterioscler
Thromb Vasc Biol 1998;18: 562–567.
45. Woodward M, Lowe GDO,RumleyA,etal. Epi-
demiology of coagulationfactors,inhibitors andacti-
vationmarkers:The Third GlasgowMonica Survey II.
Relationships to cardiovascular risk factors and preva-
lent cardiovasculardisease. BritishJournal of Haema-
tology 1997; 97: 785-797.
46. Goodman E, DolanLM, Morrison JA,etal. Factor
analysis of clustered cardiovascularrisks in adoles-
cence: obesityisthe predominant correlateofrisk
among youth. Circulation2005; 19: 1970–1977.
Vayá et al. Lipids, obesity and haemorheologyinDVT