Problems Associated with Long-Term
Anticoagulant Therapy
Observations in 139 Cases
By JOHN W. POLLARD, M.D., MICHAEL J. HAMILTON, M.D.,
NORMAN A. CHRISTENSEN, M.D., AND RICHARD W. P. ACHOR, M.D.
LONG-TERM anticoagulant therapy is be-
ing used frequently in the management
of certain vascular and thromboembolic dis-
eases. Results predominantly favorable in re-
spect to prolonged survival and decreased
morbidity have been reported from use of
coumarin anticoagulants in these diseases.1-4
However, there has not been adequate evalua-
tion of the ease or difficulty with which a
satisfactory and practical program of pro-
longed anticoagulant therapy can be carried
out. The purpose of this paper is to evaluate
this aspect of anticoagulant therapy in ambu-
latory patients with vascular and thrombo-
Downloaded from http://ahajournals.org by on June 26, 2020
embolic diseases who have been observed
during the past 101½2 years.
The following specific problems were con-
sidered: (1) the adequacy of control of pro-
thrombin activity, (2) the hemorrhagic com-
plicatious occurring during treatment, (3)
the vascular and thromboembolic complica-
tions occurring during treatment, (4) the
reasons for discontinuing treatment, (5) the
vascular and thromboembolic complications
that occurred immediately following cessation
of treatment, and (6) the problems occasioned
by surgery in patients on treatment.
Material and Methods
Selection of Patients
Accepted for study were the cases of patients
who had received anticoagulant therapy continu-
ously for 2 months or longer while residing in
Rochester, Minnesota, and vicinity. The period
of the study extended from July 1949 through
December 1959. All of the patients had some
From the Mayo Clinic and the Mayo Foundation,
Rochester, Minnesota.
Read at the meeting of the American Heart Asso-
ciation, Bal Harbour, Florida, October 20 to 22, 1961.
Circulation, Volume XXV, February 1962
form of occlusive vascular or thromboembolic
disease. Myocardial infarction (43 per cent of
cases) and cerebral vascular insufficiency (31 per
cent) with or without infaretion were the two
most common indications for long-term anti-
coagulant therapy (table 1).
All of these patients were under the continuous
care of physicians of a single section* of the
Mayo Clinic and were followed as outpatients
during the period of treatment with long-term
anticoagulants. The physicians were specialists
in internal medicine and were experienced in the
use of anticoagulant therapy. All except six
patients were started on anticoagulants during
hospitalization for acute complications of their
underlying vascular or thromboembolic disease.
Of necessity, all patients, or the relatives and
friends caring for them, were considered coopera-
tive and of sufficient intelligence to follow the
program outlined.
The 139 separate periods of long-term anti-
coagulant therapy of more than 2 months each
involved 126 patients. Nine men and two women
had two separate periods of treatment, and one
man had three separate periods of treatment.
For purposes of presentation, 139 will be used
as the number of cases in the studv.
Procedure
Bishydroxycoumarin (Dicumarol) was the anti-
coagulant used in all cases. A modified Ouiek
one-stage prothrombin test5 was used in our labora-
tories to measure the prothrombin activitv.
Determinations of prothrombin activity were made
every 3 to 5 days during 1949 and 1950; since
that time the interval has been extended to as
much as 1 month in some cases but has averaged
approximately 14 days. On the days that the
prothrombin determinations were made the patients
were advised of the result (usually by telephone),
and the dose schedule of bishydroxycoumarin was
prescribed for the subsequent interval. An effort
was made by the physician to keep the prothrorn-
*Acknowledgement is hereby accorded other mem-
bers of this section: Drs. L. A. Smith, N. 0. Hanson,
D. E. Ralston, K. G. Berge, and G. W. Morrow, Jr.
311
 312
Table 1
Indication for Long-Term Anticoagulawnt Therapy
in 139 Cases
Cases
Indication Number Per cent
Myocardial infarction 60 43
Cerebrovascular insufficieney with
or without infarction 43 31
Pulmonary emboli 9 7
Emboli of cardiac origin 7 5
Thrownbophlebitis 6 4
Coronary artery insufficiency 6 4
Peripheral arterial occlusive disease 5 4
Cerebrovascular lesion and
myocardial infarct 2 1
Thrombosis of central retinal vein 1 1
Total 139 100
bin activity within the therapeutic range-that
is, between 10 and 30 per cent of normal, the
range generally accepted as effective and safe.6
The required dosage of bishydroxycoumarin varied
from patient to patient and from time to time in
the same patient without apparent reason. For
the majority of patients the dosage ranged from
25 to 125 mg. per day.
Technics for Specific Problems
Downloaded from http://ahajournals.org by on June 26, 2020
In estimating adequacy of control of the anti-
coagulant treatment in each patient, the pereent-
age of prothromibin values within the therapeutic
range was used to calculate the per cent of the
time each patient's prothrombin activity was kept
in the therapeutic range. Thus, if three of four
prothrombin determinations were within this range,
the patient was described as being satisfactorilyv
controlled 75 per cent of the tinme. Also, in estimat-
ing the adequacy of control for this study, exelni-
sion was made of prothrombin determinations ob-
tained during the first 2 months of therapy and
those obtained within 2 weeks before and 2 weeks
after any surgical procedure. These exclusions
seemed justified because the periods were adjust-
ment periods and not representative of the long-
range aspects of the problem.
In considering the hemorrhagic complications
that occurred in patients receiving anticoagulant
therapy, the effect of hypertension was studied.
Hypertension was arbitrarily said to be present
if more than half the blood pressure readings
recorded during the period of anticoagulation
and the 6 months prior to the period of anticoag-
ulation exceeded 175 mm. of mercury systolic or
100 diastolic.
In order to assess whether vascular and thrombo-
embolie complications occurred when long-term
anticoagulant therapy was discontinued, the course
POLLARD ET AL.
of patients who stopped treatment was studied.
The 2-month period after discontinuation was
selected arbitrarily as the critical period of follow-
up study because it seemed that if complications
related to the discontinuance of anticoagulant
therapy were to occur, they should occur within
this period.
Findings
Among the 139 cases that comprised the
group studied, the patients were men in 91
and women in 48. The average age was 59.4
years, with a range of 22 to 84 years. Most
of the patients were in the fifth, sixth, and
seventh decades of life; and only four persons
were less than 40 years of age. The months
of treatment averaged 18.3 and ranged from
3 to 69 months. Sixty-one patients were still
receiving treatment at the conclusion of the
study. Prothrombin times were obtained an
average of 2.6 times per month per case.
Hypertension was present in 27 (20 per cent
of cases.
Effectiveness of Control
In 79 cases (57 per ceiit) the prothrombin'
activity was within the therapeutic range at
least 80 per cent of the time (table 2). Two
separate groupings have been tabulated be-
cause late in 1956, following a review of suc-
cess in maintaining patients within the thera-
peutic range, a more diligent effort was made
to improve the control. This diligence raised
from 49 to 66 the percentage of cases with
prothrombin activities in the therapeutic
range more than 80 per cent of the time; but
even with such increased effort, in one third
of the cases the desired range was not main-
tained as much as 80 per cent of the period
of treatment.
Hemorrhagic Complications
In 47 of the 139 cases (34 per cent) there
were 61 hemorrhagic complications (table 3).
Some of these episodes-diagnosed as epi-
staxis, ecchymosis, and postmenopausal bleed-
ing-might have occurred whether or not the
patient was receiving anticoagulants. How-
ever, in listing hemorrhagic complications it;
seemed necessary to include all bleeding that
the physician considered significant enough to
note in the record.
Sixty-nine per cent of these bleeding epi-
Circulation, Volume XXV, February 1962
 LONG-TERM ANTICOAGULANT THERAPY 313

Table 2
Maintenance of Prothrombin Time within Therapeutic Range
Cases
Per cent Period prior to Period after Total
of time in Jan. 1, 1957 Jan. 1, 1957 1949-1959
therapeutic range No.* % No.* % No.*
90-100 17 22 27 30 35 25
80-89 21 27 33 36 44 32
70-79 18 23 13 14 25 18
60-69 14 17 9 10 16 12
50-59 4 5 5 5 10 7
<50 5 6 5 5 9 6
*Thirty-two cases overlapping January 1, 1957, were counted in both subperiods but only
once for the total.

sodes occurred while the prothrombin activity Table 3

of the patient was within the therapeutic Hemorrhagic Complications during Long-Term
range; 20 per cent occurred while the pro- Anticoagulant Treatment
thrombin activity was depressed below the Complication Number
therapeutic range (less than 10 per cent of Major bleeding* 15
normal); and 11 per cent occurred while the Ecehymosis 12
prothrombin activity was between normal and Hematuria 12
the therapeutic range (between 30 and 100 Gastrointestinal bleeding 6
per cent of normal). The average age of the Epistaxis 5
patients who had hemorrhagic complications Oral bleeding 4
Conjunctival hemorrhage 3
was 62.4 years. Postmenopausal bleeding
Downloaded from http://ahajournals.org by on June 26, 2020

Included among the 61 hemorrhagic com-
plications were 15 bleeding episodes that
caused major incapacity or were dangerous
to life-and indeed resulted in three deaths
(table 4). These 15 major hemorrhagic com-
plications occurred in 14 (10 per cent) of the
139 cases as here defined and involved 14 pa-
tients. All but one of those patients with
major bleeding had prothrombin activity
greater than 10 per cent of normal at the time
of the complication. The average patient age
in the 15 cases with major bleeding episodes
was 66.7 years.
Hypertension was more common among the
patients who had these hemorrhagic compli-
cations. Fourteen (30 per cent) of the 47 pa-
tients who had hemorrhagic complications and
six (43 per cent) of the 14 patients who had
major bleeding episodes had hypertension. As
noted previously, significant elevation of
blood pressure was found in only 20 per cent
of the total 139 cases.
The cases with hemorrhagic complications
did not differ markedly from the whole group
Circulation, Volume XXV, February 1962
2
Middle-ear bleeding 1
Bloody ejaculum 1
Total 61
*See table 4.
in over-all control of therapy. Among all cases,
57 per cent of the patients were within the
therapeutic range more than 80 per cent of
the time, and that standard was achieved in
59 per cent of those having bleeding compli-
cations and 50 per cent of those having major
hemorrhagic complications.
Vascular and Thromboembolic Episodes during
Treatment
In 21 cases (15 per cent) there were 23
vascular or thromboembolic episodes during
anticoagulant treatment. Thirteen were diag-
nosed as acute myocardial infarction, nine as
cerebral vascular insufficiency with or without
cerebral infaretion, and one as thrombophle-
bitis. Six of the 21 patients died as a result
of the vascular or thromboembolic episode-
four from myocardial infarction and one of
 314 POLLARD ET AL.

Table 4
Major Hemorrhagic Complications Encountered
Prothrombin activity
Age, Treatment, Time, % of
Complication yr. months sec. normal
Blood clot, common duct 52 22 33 21
Intra-abdominal hemorrhage 53 36 44 14
(same patient listed above)
(70* 25 48 12
Gastrointestinal bleeding, indeterminate source t 68 18 43 14
(65* 4 39 16
Bleeding duodenal ulcer 64 3 19 Is0
Bleeding duodenal ulcer followed by 85* 3 322 23
myocardial infaretion
(63* 15 23 45
HIematuria g 79 29 30 25
L80 37 32 23
Hematoma, right knee 69* 6 38 17
Subarachnoid hemorrhage 63 16 37 18
(66 8 57 9
Intracerebral hemorrhaget ,58 35 25 37
i 66* 10 47 12
*Patients had hypertension.
tResulted in death.

Table 5 peutic range. The other five occurred when

Long-Term Anticoagulant Treatment: Reasons the prothrombin activity was between normal
for Discontinuation in 78 Cases and the therapeutic range, that is, 30 to 100
Cases per cent of normal.
Downloaded from http://ahajournals.org by on June 26, 2020

Reason Number Per cent

Reasons for Discontinuing Treatment
Medical judgment 27 34 In 78 (56 per cent) of the 139 cases, anti-
Social problems 22 28
Hemorrhagic complicationis 14 18 coagulant therapy was stopped during the
Death 10 13 period of study. Medical judgment-the deci-
Bleeding 3 sion of the physician that the period of treat-
Thrombosis 6 nient had been adequate-was the most fre-
Other 1
Potential hemorrhagic complication 3 quent reason for discontinuing anticoagulant
Unknown 3 4 treatment. Other reasons for stopping therapv
Total 78 100 are listed in table 5. Social problems such as
vacations, uncooperativeness, and expense.
were the reasons for discontinuing anticoagu-
cerebral infaretioni. In the sixth case sudden lants in 22 cases. Thus in 16 per cent of the
death was attributed to acute coronary insuffi- 139 instances the patients did not continue
ciency, but necropsy was not done. Eleven prolonged anticoagulant therapy although the
(53 per cent) of the 21 patients in this group indications for its use were still presenit.
were within the tlherapeutic range more than Recurrence of Vascular or Thromboembolic Episodes
80 per cent of the time. The average age of after Stopping Therapy
the group was 58.9 years. Of special interest Eleven (17 per cent) of the 64 patients who
was the fact that nine of the 21 had also ex- stopped anticoagulant therapy and who could
perienced bleeding complications during the be followed had thromboembolie or vascular
period of their treatment with anticoagulants. episodes reappearing within 2 months. Myo-
Eighteen of the 23 vascular or thrombo- cardial infaretion and cerebrovascular insuffi-
embolic episodes occurred when the patient's ciency with or without cerebral infaretion
prothrombin activity was within the thera- occurred most commonly (table 6). Three of
Circulation, Volume XXV, February 1962
 LONG-TERM ANTICOAGULANT THERAPY 315

Table 6
Recurrent Vascular Thromboembolic Episodes after Discontinuation of Anticoagulant
Therapy
Vascular or thrombo- Reason Months of
embolic episode Initial indication discontinued Age treatment
Myocardial infaretion Vacation 55 5
Cerebrovascular insufficiency Couldn't come in 78 15
for prothrombin
Myocardial infaretion checks
(5 cases with 2 deaths) Myocardial iCnfarction Medical judgment 43 5
Myocardial infarction Medical judgment 70 4
Myocardial infaretion Recurrent cerebral 54 6
infarction
Cerebrovascular insuf Cerebrovascular thrombosis Medical judgment 77 39
ficiency with or with- Cerebrovascular thrombosis Pericarditis 67 8
out thrombosis (4 cases Cerebrovascular insufficiency Hematuria 62 7
with 1 death) Basilar artery insufficiency Vacation 65 5
Thrombosis, renal artery Cerebrovascular insufficiency Confusion 74 11
Peripheral arterial Peripheral arterial disease Hematuria 78 42
occlusion (amputation,
right leg)

the 11 patients died. The average duration of
treatment for this group of 11 was 13.3 months
and the average age was 65.7 years.
Surgery in Patients Receiving Therapy
Twenty patients had 30 surgical procedures
while on long-term anticoagulant therapy
(table 7). Eleven patients had 14 dental ex-
Downloaded from http://ahajournals.org by on June 26, 2020
untreated control group was available, no ef-
fort has been made to assess the effect of
anticoagulant therapy on the course or prog-
nosis of the underlying vascular or thrombo-
embolic disease. Without an adequate control
group we think conclusions as to the effec-
tiveness of coumarin therapy cannot be made.

tractions, seven patients had eight minor sur-
gical procedures, and six patients had eight
major operations.
The minor surgical (including dental) pro-
cedures were done with the patients' pro-
thrombin activities ranging from 18 to 100
per cent of normal. Three of the 14 dental
extractions were performed when the patients'
prothrombin activities were in the therapeutic
range. During major procedures the pro-
thrombin activities in all cases were near
normal. Excessive bleeding did not occur in
any case.
The only complication of surgery in any of
these patients was a nonfatal cerebral infare-
tion following dental extraction in a 57-year-
old man whose prothrombin activity had been
brought to normal for this procedure.
Discussion
An attempt has been made to assess the
ease or difficulty with which a satisfactory
and practical program of prolonged anticoag-
ulant therapy can be carried out. Because no
Circulation, Volume XXV, February 1962
It was disquieting to find that with a con-
scientious effort, a familiar drug, and the
services of a reliable laboratory, the prothrom-
bin activity was maintained within the thera-
peutic range more than 80 per cent of the
time in only 57 per cent of the cases. After
1957, knowledge of the previously less thani
desirable results stimulated an effort that
brought this figure to 66 per cent. If it is
necessary for prothrombin activity to remain
within the range of 10 to 30 per cent of nor-
mal in order to influence favorably vascular
and thromboembolic disease, our data suggest
that an even greater effort must be made. This
raises the problem of feasibility and practi-
cality, for the regimen outlined in this paper
represents considerable investment of the pa-
tient's time and money and requires continual
supervision by the physician.
Another interesting observation was the
number of hemorrhagic complications that oc-
curred in this group of patients while receiv-
ing anticoagulant treatment. The data suggest
that older patients and those with hyperten-
 316 POLLARD ET AL.
Table 7
Surgery in Patients Receiving Long-Term Anticoagulant Therapy
Prothrombin activity Hemorrhagic or
Number of at time of surgery thromboembolic
Surgical procedures procedures Time, sec. % of normal complication
Minor
Dental extractions 14* 17-37 100-18 1*
Excision of skin lesion 4 23-29 45-27 0
Excision of lymph node 1 27 31 0
Fulguration of rectal polyp 1 26 34 0
Dilatation and curettage 1 33 21 0
Knee injection, compound F 1 29 27 0
Major
Closed reduction inter-
trochanteric fracture 1 23 45 0
Herniorrhaphy, right inguinal 1 25 37 0
Arthrodesis, right hip 1 23 45 0
Triple arthrodesis, foot 1 23 45 0
Mitral commissurotomy 1 19 80 0
Correction of incarcerated
inguinal hernia 1 24 40 0
Cholecystectomy 1 19 80 0
Abdominal exploration 1 24 40 0
*The single complication of a hemorrhagic or thromboembolic type was cerebral thrombosis
in a patient whose prothrombin activity had been brought to 100 per cent for the extraction.

sion were m-ore often affected by hemorrhagic however, which reasserts itself once the re-
complications. If this is true, then probably straining therapy is discontinued. At any
Downloaded from http://ahajournals.org by on June 26, 2020

younger patients and those without hyperten-
sion can undergo anticoagulant therapy with
more safety.
The majority of the hemorrhagic complica-
tions as well as the majority of vascular and
thromboembolic episodes occurred while the
patients' prothrombin activities were in the
desired therapeutic range. This experience
suggests that the prothrombin activity as de-
termined by our technic is not an entirely
adequate measurement of the patients' bleed-
ing or thrombosing propensities while receiv-
ing long-term coumarin anticoagulant therapy.
It indicates further that with the present
therapeutic methods a substantial number of
hemorrhagic complications and a significant
number of vascular and thromboembolic epi-
sodes are inevitable.
A rather high rate of vascular and throm-
boembolic episodes was noted after discontin-
uation of anticoagulant therapy. This may
represent a rebound or overshoot phenomenon
of the clotting mechanism. It may reflect only
the natural course of the underlying disease,
rate, the 17 per cent incidence of these com-
plications after discontinuation of anticoagu-
lant treatment is high enough to emphasize
the possible danger of stopping therapy once
it has been undertaken. The additional fact
that 16 per cent of the patients who began re-
ceiving therapy discontinued it while the in-
dication for treatment was still present adds
significance to this figure. Here, then, is
another consideration the physician must
weigh when contrasting the gravity of the
disease to be treated against the calculated
risk of long-term anticoagulant therapy.
Summary and Conclusions
In 139 cases, 126 patients having ocelusive
vascular disease were treated with bishydroxy-
coumarin (Dicumarol) for periods varying
from 3 to 69 months, with a mean duration
of 18.3 months. Of particular concern were
the problems and limitations that confront
the physician managing long-term anticoagu-
lant therapy.
Fifty-seven per cent of these patients had
Circulation, Volume XXV, February 1962
 LONG-TERM ANTICOAGULANT THERAPY
prothrombin times in the therapeutic range
during more than 80 per cent of the time they
were under treatment (adjustment periods
excluded). One third of the cases were
marked by hemorrhagic complications, which
included major bleeding complications in 10
per cent of the total. The incidence of bleed-
ing complications, and especially the major
bleeding complications, was associated with
either an increased incidence of arterial hy-
pertension or increased age. Three patients
died as a result of a hemorrhagic complica-
tion that occurred during long-term antico-
agulant therapy. Sixty-nine per cent of the
bleeding episodes occurred when the pro-
thrombin activity was within the therapeutic
range (10 to 30 per cent of normal).
Sixteen per cent of the patients discontin-
ued anticoagulant therapy although the in-
dications for its use were still present. Eleven
(17 per cent) of the 64 patients who stopped
anticoagulant therapy and could be traced
had recurrent vascular or thromboembolic
episodes within 2 months. Three of these 11
patients died.
Downloaded from http://ahajournals.org by on June 26, 2020
Surgery was performed in 20 cases during
long-term anticoagulant therapy. The pa-
tienlts' prothrombin activities were at various
levels, though for major operations all were
near normal. The only complication was a
minor cerebral infaretion in a patient whose
To have striven, to have made an effort, to have been true to certain ideals-this alone
is worth the struggle.-SIR WILLIAM OSLER. Aphorisms From His Bedside Teachings and
Writings. New York, Henry Schuman, Inc., 1950, p. 89.
Circulation, Volume XXV, February 1962
317
prothrombin activity had been brought to
normal for surgery.
In view of the limitations in the practical
management of patients on long-term anti-
coagulant therapy, plus the inconvenience,
expense, and risks involved, the physician is
compelled to appraise carefully the indica-
tions for such treatment in each patient be-
fore undertaking its use.
References
1. SUZMAN, M. M., RUSKIN, H. D., AND GOLDBERG,
B.: An evaluation of the effect of continuous
long-term anticoagulant therapy on the prog-
nosis of myocardial infarction: A report of
82 cases. Circulation 12: 338, 1955.
2. OLWIN, J. H., AND PAUL, O.: Long-term anti-
coagulant therapy. Surg., Gynec. & Obst. 105:
61, 1957.
3. RUSSEK, H. I., AND ZOHMAN, B. L.: Limited use
of anticoagulants in acute myocardial infarc-
tion: Analysis of one thousand "good risk"
cases. J.A.M.A. 163: 922, 1957.
4. MANCHESTER, B.: The value of continuous (1
to 10 years) long-term anticoagulant therapy.
Ann. Int. Med. 47: 1202, 1957.
5. HURN, M., BARKER, N. W., AND MAGATH, T. B.:
The determination of prothrombin time follow-
ing the administration of dicumarol, 3,3'-meth-
ylemebis (4-hydroxycoumarin), with special
reference to thromboplastin. J. Lab. & Clin.
Med. 30: 432, 1945.
6. ALLEN, E. V., BARKER, N. W., AND HINES,
E. A., JR.: Peripheral Vascular Diseases. Ed.
2. Philadelphia, W. B. Saunders Company, 1955.