close

Вход

Забыли?

вход по аккаунту

?

Complications of long-term anticoagulation.

код для вставкиСкачать
Complications of Long-Term
Anticoagulation
G. W. Petty, MD," L. Lennihan, MD," J. P. Mohr, MD," W. A. Hauser, MD,"? J. Weitz, MD,S
J. Owen, MD,t and C. Towey, BS$
We used life-table techniques to determine risks of morbidity and mortality associated with long-term warfarin
treatment in an anticoagulation clinic. Cumulative risks for life-threatening complications and warfarin-related death
among all patients were 1% at 6 months, 5% at 1 year, and 7% at 2 and 3 years. Cox regression analysis using age as a
continuous variable failed to show an effect of age on cumulative risks of complication. T h e occurrence of a minor
complication during the course of therapy did not place patients at higher risk for developing a major complication
that would prompt discontinuation of therapy or cause death. There was no statistically significant difference between
the cumulative risks of patients anticoagulated for cerebrovascular disease and the cumulative risks of patients anticoagulated for other indications.
Petty GW, Lennihan L, Mohr JP, Hauser WA, Weitz J, Owen J, Towey C. Complications of
long-term anticoagulation. Ann Neurol 1988;23:570-574
The anticoagulation clinic of The Presbyterian Hospital in
the City of New York serves a middle- and lower-income
population in an urban setting. The clinic is conducted by a
nurse and a physician specializing in hematology with special
interest in thrombosis and hemostasis. Patients are screened
by clinic staff before being accepted into the clinic, and the
referring physician must stabilize the dose of warfarin
sodium prior to the first visit. Prothrombin times are
checked every week for the first 6 weeks, and every 2 to 4
weeks thereafter depending on the stability of anticoagulation. The prothrombin time is measured with a rabbit brain
thromboplastin preparation (Ortho, Raritan, NJ). The dose
of warfarin sodium is adjusted to maintain the prothrombin
time at approximately twice the control value. The prothrombin time control range in our laboratory is 11 to 13
seconds.
The Stzrdy
We retrospectively reviewed hospital records and clinic log
records of every patient referred to the anticoagulation clinic
of The Presbyterian Hospital in the City of New York from
January 1, 1977, through December 31, 1980. Information
collected included age, sex, indication for anticoagulation,
complications and when they occurred, prothrombin times
when complications occurred, duration of therapy, and reason for termination of therapy. Efforts were made to reach
via telephone or mail those patients lost to follow-up. Names
of those patients who died at home were submitted to the
New York City Department of Health for death certificate
information.
Longitudinal analysis of data was performed using the lifetable technique [3}. The starting point was defined as the
date of enrollment in the anticoagulation clinic. Patients were
withdrawn from the life-table analysis at the time of completion of treatment, or at the time of chart review if they were
still receiving warfarin. For purpose of life-table analysis, patients who were transferred to other institutions or who died
of causes unrelated to anticoagulation were counted as having completed courses of anticoagulation at that point.
End points were complications or deaths related to anticoagdation. Complications were defined as events of
an untoward nature attributable to anticoagulation. Lifethreatening complications were defined as gastrointestinal
hemorrhage, intracerebral hemorrhage, hemorrhagic cerebral infarction, subdural hematoma, and death due to a complication. Minor complications were defined as all complications other than life-threatening complications (Table).
Because not all life-threatening complications prompted dis-
From the *Department of Neurology, The Neurological Institute of
New York, the tSergievskv Center. College of Physicians and Sur-
Received Jan 12, 1987, and in revised form Dec 1. Accepted for
publication Dec 1, 1987.
The use of warfarin in the treatment of stroke and
transient ischemic attack (TIA) is controversial because the frequency of complications attributable to
therapy may exceed the benefit [l, 21. We reviewed
the experience of the anticoagulation clinic of T h e
Presbyterian Hospital in the City of N e w York to
determine morbidity and mortality associated with
long-term anticoagulation. These data will be useful in
designing studies of warfarin therapy for the treatment
of stroke and TIA, as well as other conditions.
Methods
The Clinic
570 Copyright 0 1988 by the American Neurological Association
Complications of Warfarin Therapy in the Patients Studied”
Minor Complications
Life-Threatening Complications
Bleeding from hip
fracture
Bleeding from hemorrhoid
Epistaxis
Hematuria
Hemop tysis
intramuscular hematoma
Laceration bleeding
Occult blood in stool
Oral bleeding
Subconjunctival hemorrhage
Vaginal bleeding
Gastrointestinal hemorrhage
Hemorrhagic cerebral infarction
Subdural hematoma
Unspecified intracranial hemorrhage
“Patients enrolled from January 1, 1977, through December 31,
1980, in the anticoagulation clinic of The Presbyterian Hospital in
the City of New York.
continuation, and because the cause of death could not be
established in 3 patients, “best-case” and “worst-case” analyses were performed. In the best-case life-table analysis,
complications were defined as deaths attributable to anticoagulation and complications prompting discontinuation of
anticoagulation. In the worst-case life-table analysis, complications were defined as all complications, deaths due to anticoagulation, and deaths of unknown cause. Statistical differences between survival curves were determined using the
Generalized Wilcoxon (Breslow) and Generalized Savage
(Mantel-Cox) methods [31. Cox regression analysis was performed using age as a continuous variable (31.
Results
During the period studied, 310 patients were started
on 321 courses of anticoagulation. Eleven patients had
two courses of anticoagulation. Only first courses were
used for life-table analysis. Chronological data were
insufficient for life-table analysis for 11 patients because of unavailability of the chart or clinic log records.
Among the 299 (96.4%) patients for whom chronological data sufficed for life-table analysis, 9 died at
home and 3 were lost to follow-up. Death certificates
were located for 6 of the patients who died at home,
and in no case was death attributed to a complication
of anticoagulation by the medical examiner. These 6
patients were counted as having completed courses of
warfarin therapy for purpose of life-table analysis.
Death certificates could not be located for 3 patients,
and these deaths were counted as anticoagulant related
in the worst-case life-table analysis. Mail and telephone
inquiry determined that of the 3 patients who were
lost to follow-up, 1 moved, 1 was alive and no longer
taking anticoagulants at the time of the study, and 1
died after stopping therapy. These patients were
counted as having completed treatment at the time of
their last clinic follow-up.
The patients’ age range was 17 to 95 years (mean,
58.1 years; median, 60 years). Ninety (30.1%) were
70 years old or older. One hundred eighty-seven
(62.5%) were women. Indications for anticoagulation
included deep vein thrombosis (97, 32.4%), pulmonary embolus (85, 28.4%), prosthetic heart valve (46,
15.4%), atrial fibrillation (26, 8.7%), TIA or stroke
(20,6.6%), congestive cardiomyopathy (9, 3.0%), and
peripheral vascular disease (7, 2.3%). Duration of
treatment ranged from 1 month to 88 months (mean,
15.4; median, 6), for a total of 4,615 patient-months of
anticoagulation.
All complications are listed in the Table. Twelve
patients developed complications (4%, 3.1 per 100
patient-years of anticoagulation) that prompted the referring physician to discontinue therapy (gastrointestinal hemorrhage, 3; stool positive for occult blood, 2;
hematuria, vaginal bleeding, laceration, subdural hematoma, hemorrhagic cerebral infarction, hemoptysis,
and bleeding after hip fracture, 1 each). Prothrombin
times were greater than twice control for 3 of the 8
patients with complications causing discontinuation for
whom these measurements were available.
Life-threatening complications occurred in 13 patients (4.396, 3.4 per 100 patient-years of anticoagulation). Prothrombin times were greater than twice control in 5 of the 10 patients with life-threatening
complications for whom these measurements were
available. There were 3 anticoagulant-related deaths
(l%, 0.8 per 100 patient-years of anticoagulation).
Two of these patients died from acute subdural hematoma. The third patient was a 67-year-old woman with
metastatic breast cancer who was admitted for coma. A
radionuclide brain scan demonstrated an area of increased uptake in the left parietal region. The clinical
diagnosis was intracranial hemorrhage. She died before
a computed tomogram of the head could be obtained,
and no autopsy was performed. Prothrombin times
were greater than twice control at the time of the development of all 3 fatal complications.
Standard life-table techniques were used to generate
survival curves expressing percent survival without
complication due to anticoagulation as a function of
time (Figs 1-4). For all patients (Fig l), cumulative
risks for life-threatening complications and warfarinrelated death were 1% at 6 months, 5% at 1 year, and
7% at 2 and 3 years. Cumulative risks for all patients
using the best-case analysis (see Fig 1) were 2% at 6
months, 7% at 1 year, 12% at 2 years, and 14% at 3
years. The worst-case analysis (see Fig 1) resulted in
cumulative risks of 8% at 6 months, 21% at 1 year,
30% at 2 years, and 34% at 3 years. Cox regression
analysis using age as a continuous variable in the bestcase analysis failed to show an effect of age on cumulative risk.
For patients who experienced minor complications
during the course of treatment (Fig 2), best-case cumuPetty et al: Anticoagulation Complications
571
---.Best Case Patients Anticoagulated for Reasons other lhan Stroke or TIA
All Patients N=299
-Best Case Patients Anticoagulatedfor Stroke or TIA
....................................................................................................
I
.cj,,l~,llll,l,,l
I
Life Threatening
Best Case
- Worst Case
30
,
0
-
'
36 ' 42 ' 48 ' 54 '
$0'
66
?2
i8
6
12
18
24
30
Time (months)
~~
Fig 1 . Life-table estimates of percentage of patients without warfarin-related complicationfor all patients. Risk of warfarinrekzted complication at any point is 100% minus the percentage
of patients without complication at that point.
----Best
-6esI
N=279
N=20
(
(
,
42 48 54 60 66 72
Time (months)
36
78
~~~
Fig 3. Best-case life-table estimates o f percentage o f patients surviving without warfarin-related complicationfor patients treated
for stroke or transient ishemic attack (TIA)compared with patients treated for indications other than stroke or TIA.
._-_.
Worst Case Patients Anticoagulated lor Reasons other lhan Stroke or TIA
Case Patients wllhoul Prior Minor Complications N=268
Case Patients with Pr~orMinor Complications N=31
-Worst Case Patients Anticoagulated lor Stroke or TIA
N=279
N=20
N=10
........
90
L __._
~
90
-_- - - -_
L_
- - - - - - - - - - - __7
.l
.l
.
6
12
18
24
30
36
42
48
54
60 66
72
78
Time (months)
Fig 2. Best-case life-table estimates of percentage of patients surviving without warfarin-related complicationfor patients with
minor complications prior to discontinuation of therapy compared
with patients without prior minor complications.
Fig 4. Worst-caselife-table estimates of percentage of patients
surviving without warfarin-relatedcomplicationfor patients
treatedfor stroke or transient ischemic attack (TIA) compared
with patients treatedfbr indications other than stroke or TIA.
lative risks were 0% at 6 months, 7% at 1 year and 2
years, and 20% at 3 years. For patients who had no
minor complications (see Fig 2), best-case cumulative
risks were 2% at 6 months, 7% at 1 year, and 13% at
2 and 3 years. Survival curves for these groups were
not significantly different at the p = 0.05 level.
For the 20 patients who were anticoagulated for
TIA or stroke, best-case cumulative risks (Fig 3 ) were
6 months, and 17% at 1 year, 2 years, and 3
years. Worst-case cumulative risks for these patients
(Fig 4 ) were 6% at 6 months, 22% at 1 year, 43% at 2
years, and 54% at 3 years. For the 279 patients who
were anticoagulated for reasons other than TIA or
stroke, the best-case cumulative risks (see Fig 3) were
2% at 6 months, 5% at 1 year, 11% at 2 years, and
13% at 3 years. Worst-case cumulative risks for these
572 Annals of Neurology Vol 23 No 6 June 1988
0% at
patients (see Fig 4) were 8% at 6 months, 20% at 1
year, 28% at 2 years, and 31% at 3 years. Survival
curves for these two groups were not significantly different at the p = 0.05 level.
Discussion
In our anticoagulation clinic, the risks of life-threatening complications and anticoagulant-related deaths
were 3.4 and 0.8 per 100 patient-years, respectively.
These figures compare favorably with previously
reported rates of 0 to 22 major complications and 0 to
9 deaths per 100 patient-years of anticoagulation 14).
However, it is more useful to those designing clinical
trials to express complications in terms of cumulative
risks determined by life-table analysis. The advantage
of this type of analysis is that cumulative risks for complications are expressed as a function of duration of
therapy, not simply as the number of complications
per 100 patient-years of therapy. To our knowledge,
only one previous study has analyzed complications of
long-term outpatient anticoagulation using this technique. Petitti and associates examined the courses of
patients anticoagulated for venous thromboembolism
{5). In their study, the cumulative risks of major
hemorrhage were 10% at 3 months, 18% at 1 year,
26% at 2 years, and 41% at 5 years. Their figures are
higher than the cumulative risks of life-threatening
complications observed in our clinic, which are only
1% at 6 months, 5% at 1 year, and 7% at 2 and 3
years (see Fig 1). These differences may be related to
differences between our definition of life-threatening
complication and the definition of mjor hemorrhage used
by Petitti and associates. Our study is a historical cohort study. Since our patients did not return to clinic
at standard intervals of follow-up, and a standard inventory of questions was not used, it is possible that
some types of complications, particularly minor complications, may have been underestimated. Nevertheless, all of our patients were followed in a single
anticoagulation clinic by one nurse and one physician
who were responsible for patient education and ascertainment of complications. In addition, all patients
were followed by the referring primary physician. We
believe that these factors make it unlikely that significant numbers of complications were not ascertained.
To determine if advanced age increased the risk of
hemorrhagic complications of warfarin therapy, we
performed a Cox regression analysis using age as a
continuous variable. Age had no effect on cumulative
risk. We conclude that, in our anticoagulation clinic,
older patients are not at increased risk for developing
hemorrhagic complications. Other studies have failed
to demonstrate a correlation between increased age
and risk of hemorrhagic complications during longterm anticoagulation {6, 71, although some have demonstrated such a correlation [8, 9).
The occurrence of a minor complication in a patient
taking warfarin invariably arouses concern that such a
patient may be at greater risk for developing a lifethreatening complication. To address this issue, we
compared the survival curve of patients who had no
minor complications prior to the cessation of therapy
with the survival curve of those who did have minor
complications prior to cessation of therapy (see Fig 2).
These curves were not significantly different at the p
= 0.05 level. This finding cannot be attributed to earlier cessation of therapy in the group with minor complications, because the mean and median durations of
treatment were longer for these patients than for the
entire cohort. Although patients who develop minor
complications may be monitored more closely, our
data suggest that the development of a minor complication during long-term anticoagulation does not predict future life-threatening complication and is not itself an indication for shortening therapy.
Based on a review of previous studies involving
heterogeneous patient groups given various forms of
anticoagulation treatment, the suggestion has recently
been made that patients anticoagulated for cerebrovascular disease are at greater risk for developing serious
hemorrhagic complications compared with patients
anticoagulated for other indications 147. To address
this issue, we compared best-case and worst-case lifetable analyses of patients anticoagulated for cerebrovascular disease with those of patients anticoagdated for reasons other than cerebrovascular disease
in our clinic (see Figs 3 and 4). Differences in survival
curves for these two groups were not significant at the
p = 0.05 level. Our data do not demonstrate that
patients anticoagulated for cerebrovascular disease are
at greater risk for developing hemorrhagic complications compared with patients anticoagulated for other
reasons, but given our sample size in these two groups
(279 patients without stroke or TIA, 20 patients with
stroke or TIA) and using an a of 0.05 and a p of 0.2,
the risk ratio would have had to exceed 3.1 for us to
have demonstrated a statistically significant difference
{lo). To detect a statistically significant difference assuming complication rates of approximately 30% for
patients anticoagulated for cerebrovascular disease and
approximately 20% for patients with ischemic heart
disease and venous thromboembolism, as suggested by
the review by Levine and associates {4) and assuming
the same distribution of patients according to indications for anticoagulation as in this study, over 2,000
patients would be required [ 10).
The prothrombin time was greater than twice the
control value in all 3 patients with fatal hemorrhagic
complications, and in half of the patients with lifethreatening complications. Recent studies suggest that
anticoagulation complication rates may be reduced
without sacrificing therapeutic efficacy by employing
Petty et al: Anticoagulation Complications 573
more sensitive thromboplastin reagents, thereby maintaining lower levels of anticoagulation t l l , 123. Such
an approach may further lower the risk of morbidity
and mortality associated with long-term anticoagulation with warfarin.
Presented at the 11Ith Annual Meeting of the American Neurological Association, Boston, MA, Oct 7, 1986.
References
Levine M, Hirsh J. Hemorrhagic complications of long-term
anticoagulant therapy for ischemic cerebral vascular disease.
Stroke 1986;17:111-116
Sage JI. The use and overuse of heparin in therapeutic trials.
Arch Neurol 1985;42:315-317
BMDP Statistical Software Manual. Berkeley, CA: University
of California Press, 1985
Levine MN, Raskob G, Hirsh J. Risk of haemorrhage associated
with long term anticoagulant therapy. Drugs 1985;30:444-460
574 Annals of Neurology Vol 23 No 6 June 1988
5. Petitti DB, Strom BL, Melmon KL. Duration of warfarin anticoagulant therapy and the probabilities of recurrent thromboembolism and hemorrhage. Am J Med 1986;81:255-259
6. Errichetti AM, Holden A, Ansell J. Management of oral anticoagulant therapy. Experience with an anticoagulant clinic. Arch
Intern Med 1985;144:1966-1968
7. Forfar JC. A seven year analysis of hemorrhage in patients on
long term anticoagulant therapy. Br Heart J 1979;42:128132
8. Coon WW, Willis PW 111. Hemorrhagic complications of anticoagulant therapy. Arch Intern Med 1974;133:386-392
9. Wintzen AR, de Jonge H, Loehger EA, Bots GTAM. The risk
of intracerebral hemorrhage during oral anticoagulant treatment: a population study. Ann Neurol 1984;16:553-558
10. Schlesselman J, Sample size requirements in cohort and casecontrol studies of disease. Am J Epidemiol 1974;99:381-384
11. Hirsh J, Deykin D, Poller L. “Therapeutic range” for oral anticoagulant therapy. Chest 1986;89:1 IS-15s
12. Hull R,Hirsh J, Jay R, et al. Different intensities of oral anticoagulant therapy in the treatment of proximal-vein thrombosis. N
Engl J Med 1982;307:1676-1681
Документ
Категория
Без категории
Просмотров
0
Размер файла
463 Кб
Теги
terms, long, complications, anticoagulation
1/--страниц
Пожаловаться на содержимое документа