7/23/2019 Dia Care-1992-Ganda-1245-50

1/6

O R I G I N L R T I C L E

Hyperfibrinogenemiq

An important risk factor for vascular

complications in diabetes

OM P. GANDA, MD

CHARLES F. ARKIN, MD

OBJECTIVE To evaluate the determina nts of elevated fibrinogen levels and the

impact of hyperfibrinogenemia on vascular complications in diabetes.

RESEARCH DESIGN AND METHODS

Plasma fibrinogen, glucose, HbA

lt

and lipids were measured in 116 ambulatory type I and type II diabetic patients with

(n = 59) or without (n = 57) clinical evidence of micro- or macrovascular compli-

cations. In 56 of these patients, factor VII activity and CRP also were measured.

Univariate and multivariate data analyses were conducted.

RESULTS

Overall mean SE fibrinogen levels in patients (339 7.3 mg/dl)

were elevated markedly compared with control subjects (248 9.1 mg/dl). Fibrin-

ogen levels were elevated disproportionately in patients with type II diabetes

(P < 0.0001), hypertension (P = 0.0001), obesity (P < 0.0001), and vascular com-

plicatio ns (P < 0.00 01). Fibrino gen w as correlated significantly with age

(P < 0.001), cholesterol (P = 0.002), CRP (P < 0.001), and factor VII activity

(P = 0.032), but not with plasma glucose, triglycerides, HDL cholesterol, or disease

duration. Stepwise multiple regression analyses revealed that type II diabetes and

presence of vascular complications were major determinants of fibrinogen. For

vascular complications, fibrinogen emerged as one of only three independent pre-

dictors, the other two being diabetes duration and hypertension.

CONCLUSIONS Fibrinogen frequently is elevated in diabetes and is an inde-

penden t p redictor of vascular com plications.

P

rospective, epidemiological studies

from Goteborg, Sweden (1), Lon-

don, UK (2), and Framingham, MA

(3), have identified elevated fibrinogen as

a risk factor for cardiovascular disease.

This relationship persisted in multivari-

ate analyses, taking into account tradi-

tional risk factors, which included smok-

FROM THE JOSLIN DIABETES CENTER AND THE DEPARTMENTS OF M EDICINE AND PATHOLOGY, NE W

ENGLAND DEACONESS HOSPITAL, AND THE HARVARD MEDICAL SCHOOL, BOSTON, MASSACHUSETTS.

ADDRESS

CORRESPONDENCE AND REPRINT REQUESTS TO O M P. GAND A, MD, JOSLIN DIABETES CENTER,

ONE

JOSLIN PLACE, BOSTON, MA 0 2 2 1 5 .

RECEIVED

FOR PUBLICATION 19 AUG UST 19 91 AND ACCEPTED IN REVISED FORM 11 MAY 19 92 .

TYPE

1 DIABETES, INSULIN-DEPENDENT DIABETES MELLITUS; TYPE 11 DIABETES, NON-INSULIN-DEPENDENT

DIABETES

MELLITUS; C R P , C-REACTIVE PROTEIN; O H A , ORAL HYPOGLYCEMIC AGENT; H D L , HIGH-DENSITY

LIPOPROTEIN.

ing, cholesterol, and hypertension. In the

Northwick Park Heart Study (2), hemo-

static factors, i.e., fibrinogen and factor

VII coagulant activity, were stronger pre-

dictors for ischemic heart disease than

was cholesterol. In another long-term

prospective study (4), fibrinogen levels

were a strong and independent predictor

of acute heart attacks after adjusting for

systolic blood pressure and cholesterol.

Along with evidence for a role of fibrin

deposition in the development of athero-

sclerotic lesions (5,6), such observations

provide support for the theory of throm-

bogenesis in the evolution of atheroscle-

rosis (7).

Abundant evidence has accumu-

lated to suggest that atherosclerosis is

accelerated in both type I (8,9) and type

II (10,11) diabetes. Traditional risk fac-

tors (hyper l ipidemia , hyper tens ion,

smoking, age, obesity) do not account

fully for the increased prevalence and

severity of vascular disease in diabetes

(12). It has been proposed that a hyper-

coagulable state in diabetes may con-

tribute at least in part (13). Of the

various hematological factors, elevated

fibrinogen as a risk factor in diabetes

has received little attention. A few

cross-sectional studies have indicated a

state of hyperfibrinogenemia in diabetes

compared with nondiabetic control sub-

jects (14-18), particularly in those with

preexisting micro- or macrovascular

complications. However, these studies

were not controlled for confounding var-

iables.

In this study, we examined the

relationship of plasma fibrinogen and

other clinical variables to vascular com-

plications in 116 patients with a wide

range of diabetes duration, severity, and

glycemic control.

RESEARCH DESIGN AND

METHODS

The study population

included 116 diabetic outpatients pre-

senting for an office visit to O.P.G. over a

period of several weeks. Pertinent infor-

mation, including the presence or ab-

DIABETES CARE, VOLUME 15, NUMBER 10, OCTOBER 1992

24 5

7/23/2019 Dia Care-1992-Ganda-1245-50

2/6

Hyperfibrinogenemia and vascular complications

Table

Study

population

CONTROL SUBJECTS

PATIENTS (ALL)

TYPE

1

TYPE11

N ( M / W )

30 (18/12)

116(55/61)

36 (19/17)

80 (36/44)

AGE (YR)

(RANGE)

46.4 2.55

(23-71)

56.6 1.44

(22-82)

41.3 2.3

(22-75)

63.5 1.2

(36-82)

DURATION OF

DIABETES

(YR)

(RANGE)

12.2 0.90

(1-60)

15.2 2.1

(1-60)

10.8 0.9

(1-35)

VASCULAR COMPLICATIONS

NONE

57

22

35

MICRO

32

11

21

MACRO

41

6

35

DIET

4

0

4

TREATMENT

OHA

30

0

30

INSULIN

82

36

46

Values for age and duration of diabetes are means SE.

Three type

I and 11

type

II

patients

had

both

micro- and

macrovascular complications.

sence of clinical evidence of vascular

complications in each patient, was re-

corded by the same researcher. The

means (and ranges) of age and diabetes

duration, and the sex, treatment modal-

ities,and status of vascular complications

are presented in Table 1. Microvascular

complications included presence of

background or proliferative retinopathy

and/or nephropathy. Retinopathy was

classified based on the funduscopic and

fluorescein angiographic assessment at

the Beetham Eye Unit at our institu-

tion. Nephropathy was defined by the

presence of overt, dipstick-positive pro-

teinuria in the absence of infection or

other discernible cause; most patients

had also additional evidence of diabetic

nephropathy. Macrovascular disease also

was defined by standard clinical criteria,

including a detailed checklist of history

and physical examination, routine and

stress electrocardiography, and noninva-

sive and/or invasive peripheral vascular

studies in most patients. Of the 80 pa-

tients with type II diabetes, 35 had var-

ious combinations of coronary artery dis-

ease (n = 22 ), perip hera l vascula r

disease (n = 13), defined by history of

lower extremity vascular bypass proce-

dure (n = 6) or evidence of absent pul-

sations, and/or amputation (n = 7); and

cerebrovascular disease, as indicated by

history of stroke, transient ischemic at-

tacks, or carotid bruit (n = 8). Twenty-

one patients, including 11 with mac-

rovascular disease, had microvascular

disease. Of the 36 type I diabetes pa-

tients, 14 had evidence of vascular com-

p l ic a t io ns , 11 ha d m ic rova sc u la r

disease, and 6 had macrovascular dis-

ease. Altogether, of 116 patients, 57 pa-

tients had no evidence of vascular com-

plications, and 59 had micro- and/or

macrovascular complications. Only 12

patients were smokers. Obesity, defined

as >120% of ideal body weight, was

prevalent in type II patients (63 of 80

[79%]), whereas only 3 type I patients

(8%) were obese. For comparison, 30

healthy nondiabetic laboratory techn olo-

gists or bloo d-ban k donors of similar age

range (23-71 yr) served as control sub-

jects.

Procedures

Random (fasting or nonfasting) blood

samples were obtained for glucose,

HbA

l5

lipids (total cholesterol, triglycer-

ides, HDL cholesterol), and fibrinogen

determinations. Plasma glucose and lip-

ids were determined by routine autoan-

alyzer methodology with enzymatic tech-

niques (19). HbA

x

was determined by an

electrophoretic method (20). For plasma

fibrinogen assay, the Dade thrombin

clotting time methodology was used

(21). In 56 of the patients (25 with an d

31 without vascular complications),

plasma factor VII activity (22) and CRP

(23) also were determined.

Statistics

Unpaired Student's t tests were con-

ducted to determine the significance of

observed differences between the means

of continuous and discontinuous varia-

bles. Nonparametric analyses (Kruskall-

Wallis test and Wilcoxon's signed-rank

test) were applied for the nonnormally

distributed differences in patients and

control subjects. Results are presented as

means SE. Simple and multiple re-

gression analyses using stepwise regres-

sions were performed with Minitab Soft-

ware Release 7.2.

2 25 3 36 4 45 6 55

Fibrinogen mg/dL)

I Controls Patients

Figure Distributions of fibrinogen

concen-

trations incontrol subjects n = 30) and diabetes

patients

n = 116). P