close

Вход

Забыли?

вход по аккаунту

?

Survival prognosis and causes of death in rheumatoid arthritis.

код для вставкиСкачать
706
SURVIVAL, PROGNOSIS, AND CAUSES OF DEATH IN
RHEUMATOID ARTHRITIS
DONALD M. MITCHELL, PATRICIA W. SPITZ, DONALD Y. YOUNG, DANIEL A. BLOCH,
DENNIS J. McSHANE, and JAMES F. FRIES
The factors associated with mortality were examined in a prospective longitudinal study, over an average
of 12 years, with 94% followup of patients diagnosed as
having rheumatoid arthritis. Of 805 patients, 233 died
during the period of the study. Survivorship of rheumatoid arthritis patients was approximately 50% less than
that of population controls. Survivorship was decreased
by the traditional demographic variables of greater age
and male sex; however, a significant independent effect
of variables reflecting disease severity (American Rheumatism Association functional class, rheumatoid factor
titer, number of involved joints) was identified by
multivariate analysis. Seventy-nine excess deaths (i.e.,
those that would not have been expected in a control
population) were due in part to disease-related causes,
to infections, and to gastrointestinal complications of
therapy. Treatment with gold or prednisone did not
seem to affect survivorship or cause of death, except for
the clustering of deaths of patients with vasculitis within
the prednisone group. Our findings indicate that rheumatoid arthritis, a chronic disabling disease, is also
associated with a major decrease in survivorship.
-__
From the Department of Medicine, University of Saskatchewan, Saskatoon, Saskatchewan, Canada, and the Department of
Medicine, Stanford University School of Medicine, Stanford, California
Supported by grant no. AM-21393 from the NIADDK, by
grant no. HS-03802 from the NCHSR, and by a grant from the
Arthriltis Society of Canada.
Donald M. Mitchell, MD; Patricia W. Spitz, RN; Donald Y.
Young, PhD; Daniel A. Bloch, PhD; Dennis J. McShane, MD;
James F. Fries, MD.
Address reprint requests to Dr. James F. Fries, Department
of Medicine, HRP Building, Room 109C, Stanford University
School of Medicine, Stanford, CA 94305.
Submitted for publication July 29, 1985; accepted in revised
form November 4, 1985.
Arthritis and Rheumatism, Vol. 29, No. 6 (June 1986)
Rheumatoid arthritis (RA) is a common chronic
illness, which causes great disability and chronic pain,
and has a major adverse economic impact upon patients. Considerable evidence also suggests that RA
can be fatal (1-18). This prospective longitudinal study
of a consecutive cohort of 1,196 patients with a clinical
diagnosis of rheumatoid arthritis, followed for an
average of 12 years, addressed 4 questions: 1) Do
patients with rheumatoid arthritis die prematurely? 2)
Which patient and disease characteristics influence
survival? 3) What are the causes of death in rheumatoid arthritis and how d o they differ from those of the
general population? 4) D o e s t r e a t m e n t affect
survivorship or cause of death?
PATIENTS AND METHODS
During the period from 1966 through 1974, 1,196
patients over age 16 with a clinical diagnosis of rheumatoid
arthritis were treated as either inpatients or outpatients at
the Rheumatic Disease Unit, University of Saskatchewan,
Saskatoon, Canada (Figure 1). Because of Saskatoon’s position as the sole referral center in northern Saskatchewan
Province, and because of the characteristics of the health
system in the province which encourage referral of significant illness, these patients comprise the large majority
(estimated 75%) of all patients with definite or classic rheumatoid arthritis in northern Saskatchewan.
Most subjects were seen as outpatients. Approximately 500 specific observations per patient were made by a
board-certified rheumatologist at study onset and were entered and stored on computer. Additional standard clinical
data (19) gathered throughout the study course were similarly stored on computer. After 1974 the data were transferred to and subsequently maintained at American Rheumatism Association Medical Information System (ARAMIS)
facilities at Stanford University (20).
Intensive followup studies of these patients were
begun in 1978 and completed in 1982; approximately half
MORTALITY IN RA
707
Diagnosis Maintained
Throughout Course
(1.018)
Diagnosis Changed
During Course
(1781
Tertiary Analysis
rpi,
/I\
Follow-Up
Lost to
Follow-up
Refused
Assessment
Auewnent
Fkgurre 1. Survivorship in 1,196 patients diagnosed with rheumatoid
arthritis and treated at the Rheumatic Disease Unit, University of
Saskatchewan.
were still being treated at the arthritis clinic at time of
followup. The Saskatchewan hospital services plan supplied
current addresses of patients who still lived in the province
of Saskatchewan. They were contacted by a standard protocol involving letters, questionnaires, and subsequent telephone calls (21,22). Further search included letters to family
members, including family members not residing in Saskatchewan. Referring physicians were also queried.
Names of patients not traced were submitted to the
Vital Statistics Branch of the Department of Health, which
then searched death records. Causes of death listed on death
certificates were checked against records of autopsies,
which had been performed in 28% of cases. Deaths were
classified by organ system, by a schema closely matched
with the standard Saskatchewan categories, with expansion
in particular areas of interest such as “rheumatic” causes of
death.
Of 1,196 patients with a clinical diagnosis of rheumatoid arthritis, the initial diagnosis was maintained throughout
the course in 1,018 (mean age 51.8 years; fema1e:male ratio
2.2: 1). Of the 178 patients in whom the diagnosis changed,
71 were considered at followup to have osteoarthritis and 40
to have no rheumatic disease at all. Final diagnoses for these
40 patients were arthralgia in 8, psoriatic arthritis in 8,
ankylosing spondylitis in 7, and a variety of other conditions
in others. Patients for whom the diagnosis was changed were
similar in age and sex distribution to patients with RA;
however, they had lower mortality rates. Ten-year survival
was 92% for women and 89% for men in the non-RA group
compared with 85% and 75%, respectively, for the RA
group. The non-RA group was not analyzed further.
For this study, patients with RA, when initially seen,
were divided into 2 groups: 805 with definite or classic
rheumatoid arthritis (satisfying 5 or more ARA criteria [23])
and 213 with “possible/probable” rheumatoid arthritis (satisfying 2 4 ARA criteria) (Figure 1). When patients were first
seen, duration of disease ranged from a few months to 40
years, with a mean of approximately 10 years. This is typical
of first referral time in other studies (1-18).
Patients who died were older (mean 59.6 years) than
those surviving (mean 48.9 years), and were more frequently
men (female : male ratio 1.6:2.6). Primary analyses in this
study are based on the 805 individuals with unequivocal
rheumatoid arthritis: of these, 538 were alive at followup,
233 had died, and 34 were lost to followup. Patients lost to
followup were significantly older (mean 58.5 years), with
longer disease duration (mean 23.1 years), than those followed ( P < 0.01).
Elimination from the analysis of the 178 patients
whose diagnosis changed and the 213 patients with possible
or probable rheumatoid arthritis could have introduced bias
into the study, since patients had to survive long enough to
manifest symptoms that enabled categorization as definite or
Table 1. Variables examined for predictive power from study entry*
Sext
Aget
Anorexia?
Paresthesias
Latex fixation titer
Log latex fixation titert
Rales
Systolic murmur
Ankle edema
Muscle weakness
Rose-Waaler titer
Joint pain
Joint swellingt
ARA functional classt
Symmetryt
Rheumatoid nodulest
Grip strengtht
Metatarsal synovitis
Tarsomandibular joint pain
Albumin
LDHt
SGOTt
Radiographic gradet
Number of active jointst
Year of study entryt
* ARA = American Rheumatism Association; LDH
glutamic oxaloacetic transaminase.
t Variables included in Cox analysis.
=
% bands t
Westergren erythrocyte sedimentation rate
Proteinuriat
Acetylsalicylic acid
Butazolidine
Indomethacin
Prednisone
Durationt
Morning stiffness
Number of ARA criteria?
Erosions on radiographt
Weighted joint count
lactate dehydrogenase; SGOT = serum
MITCHELL ET AL
708
100
G
80
z>
a 60
3
v)
I-
Expected from Age-Adjusted Population Data,
Female, Sstkatchemn, 1976
'
Observed in 554 Female Patients with Definite or
Classical Rheumatoid Arthritis
u
K
Lu
n
Standard Error
0
I
I
I
5
10
15
YEARS AFTER ENTRY
Figure 2. Survivorship in 554 women with rheumatoid arthritis (@)
in Saskatchewan compared with women in the general population
(---
classic RA. A secondary survival analysis that included all
1,018 patients was performed to compare results with those
obtained in the definitelclassic group.
Statistical analysis involved preliminary univariate
analyses followed by stepwise Cox survival analysis
(BMDP2L). Thirty-seven variables (Table 1) were identified
by univariate analysis.
Survival figures for Saskatchewan Province are published every 5 years. Using age-specific survival rates from
1976, which is the modal death year, a population control
survival curve was created to represent the same age and sex
distributions as those of the patients.
Symbolically, % and ti represent, respectively, age at
study entry and number of years of observation of patient i.
The percent survival of the control population for t years
after study entry was represented by s(t), and s(t) was
estimated by s(t) in the equation:
N(ti 2 t) is the number of patients observed for at
least t years after entry, b(tlap) is the estimated probability
(fromi the Saskatchewan Province survival figures) that a
patient of age %* survives at least t more years, and i* is the
set of i such that ti 2 t.
The survival curve for the patient group was estimated by the Kaplan-Meier procedure, which adjusts for
those patients who had censored observation periods or
were lost to followup (Figures 2 and 3).
RESULTS
Do patients with rheumatoid arthritis die earlier
than members of the general population? We found that
survival rates for patients with rheumatoid arthritis
were lower than those in the general population,
20
I
0
0
Standard Error
I
I
I
5
10
15
YEARSAFTER ENTRY
Figure 3. Survivorship in 251 men with rheumatoid arthritis (0)
in Saskatchewan compared with men in the general population
(-
1.
- - -Expected from Age-Adjusted Population Data,
Male, Saskatchewan, 1976
-Observed
in 251 Male Patients with Definite or
Classical Rheumatoid Arthritis
40
--).
matched for age and sex, and also lower than those in
patients whose diagnosis changed during the study
course. These results were similar for men and
women, whether the diagnosis was definite or classic
RA or probable RA, and for all RA patients combined
(Table 2 ) . The decrease in survivorship (Figures 2 and
3 ) , which began early in the disease course and increased throughout, represented an increase in standardized mortality ratio (SMR) of approximately 50%
(SMR = 151). Of 233 deaths, 79 were considered
excess, in that they would not have been expected in a
control population directly standardized for sex distribution. If these trends are projected to the median age
at death, the median age at death in male RA patients
would be 4 years younger than the median age at death
in the general male population, and the median age at
Table 2.
Survivorship of patients with rheumatoid arthritis
% surviving
Patients
Women (no.)
Saskatchewan Province (control subjects)
All initial subjects (830)
All rheumatoid arthritis (700)
Definite rheumatoid arthritis (554)
Diagnosis changed (130)
Men (no.)
Saskatchewan Province (control subjects)
All initial subjects (366)
All rheumatoid arthritis (318)
Definite rheumatoid arthritis (251)
Diagnosis changed (48)
10
15
5
years years years
96
93
92
93
95
90
85
84
85
92
82
75
73
73
92
93
89
88
89
96
82
75
73
73
89
68
61
60
61
59
MORTALITY IN RA
709
Table 3. Predictive effect of single variables on survivorship of
Datients with definite or classic rheumatoid arthritis
% surviving
Variable (no. patients
at study entry)
5
years
10
years
15
years
Age <50 (361)
Age >50 (444)
ARA functional class I (335)*
ARA functional class 11-IV (408)
No. active joints <14 (334)
No. active joints 214 (313)
No proteinuria (394)
Proteinuria (67)
Rales absent (316)
Rales present (26)
Rheumatoid factor 5512 (267)
Rheumatoid factor 2 1,024 (296)
Gnp strength >200 (183)
Gnip strength <200 (198)
Women (554)
Men (251)
99
85
96
88
94
90
91
81
94
65
93
88
96
89
93
89
95
71
89
76
88
54
81
62
77
62
65
51
78
27
71
61
79
55
73
61
85
77
80
64
89
42
85
73
86
76
85
73
* PiRA = American Rheumatism Association.
de<athin female RA patients would be 10 years younger
than in the general female population.
What factors are associated with early death?
Our results showed survivorship in rheumatoid arthritis patients to be associated with age and sex, the
traditional demographic predictors of mortality, as
well as with indices of disease severity and variables
associated with comorbidity (Table 3).
The 37 variables in Table 1, identified by
univariate analysis, were further analyzed by Cox
survival analysis, performed in a stepwise manner
(Table 4). This regression model yielded 7 variables for
which there was significant improvement in the model
at the 0.05 level. Age was the most important variable
and male sex was also important. These 2 traditional
deimographic markers are usually important regardless
of the disease or population being studied. Three of the
variables yielded-ARA functional class, the number
of active joints, and the latex fixation titer-are generally considered to be indicators of disease severity.
Grip strength, which narrowly missed being included
in the model, also denotes disease severity and decreased function. The sixth variable, proteinuria (generally mild), represents disease activity in a few patients, drug toxicity in others, and comorbid conditions in yet others. Not included in the Cox model, but
statistically significant, was the presence of pulmonary
rales, which is caused by rheumatoid lung disease in
approximately half of patients and by concurrent congestive heart failure in the remainder; this is also a
variable that represents both disease severity and
comorbidity. Stepwise logistic regression (results not
shown) demonstrated closely similar results.
Cox survival analysis uses “time to death” as a
dependent variable; thus, “year at entry” is a calendar
year covariate and represents the effect on the hazard
function for patients who entered the study at a given
calendar year. This was the seventh significant variable, which suggests that death rates from rheumatoid
arthritis are increasing over time, a somewhat puzzling
result. A possible explanation is that the 34
“unlocatable”patients, who were older and had longer
disease duration than those observed, were concentrated in the first years of study entry. If the unknown
death rates of these subjects exceeded those of the
others, then there was inadequate ascertainment of
actual death rates in the early years.
The significance of these results can be demonstrated using relative risk, shown in Table 4. A comparison of the predictive effect of single variables is
shown in Table 3. The Mantel-Haenszel test indicated
that differences were highly significant in each instance (P < 0.01).
Table 4. Seven variables examined by Cox sun4val analysis (BMDP2L) as predictors of death in
rheumatoid arthritis patients
~
~~
Step
Variable
1
Age
American Rheumatism Association
functional class
Year
Sex
Proteinuria
No. active joints
Log latex fixation titer
2
3
4
5
6
7
~
~~
~
Improvement,
chi-square
Improvement,
P
Relative
risk*
203
18
0
0
1.O7lyear
1.49Iunit
16
9
9
9
6
0
0.003
0.003
0.002
0.014
1.07lyear
1.61 MIF
1.35/+unit
1.OUjoint
1.06llog
* Relative risk = the relative multiplicative effect of the i-th covariate on the hazard function
corresponding to a 1-unit change in the i-th covariate only. See Patients and Methods for details.
710
MITCHELL ET AL
Table 5. Distribution of causes of death in study population of rheumatoid arthritis patients and in
the general population of Saskatchewan Province (1979)
~
Study population
Audited cause of
death
Death
certificate
Cause*
No.
%
No.
%
Cardiovascular
ASHD
Myocardial infarction
Arteriosclerosis
CVA
Aortic aneurysm
Rheumatic heart disease
Cardiac arrhythmia
Pericarditis
Hypertensive heart disease
Cor pulmonale
Congestive heart failure
Cardiomyopathy
Other
Malignancies
Ca colon
Ca lung
Leukemia
Ca prostate
Ca ovary
Ca breast
Ca stomach
Lymphoma
Ca kidney
Ca gall bladder
Ca metastatic
Ca tongue
Malignant melanoma
Ca brain
Other
Infectious/sepsis
Pneumonia
Septicemia
Other
Rheumatic diseases
Vasculitis
Rheumatoid lung
Amyloidosis
Rheumatoid arthritis
C 1 Z 2 subluxation
Septic arthritis
Adrenal insufficiency
Rheumatoid pericarditis
SLE
Drug toxicity
Other
Gastrointestinal
Respiratory
Renal
Trauma, suicide
CNSlpsychosis
Hematologic
Other
I07
56
17
11
10
5
3
2
1
I
1
0
0
0
33
7
5
4
3
3
2
2
1
1
1
1
1
1
1
0
34
22
11
1
25
5
5
4
4
2
1
I
108
32
28
11
11
5
3
5
1
1
1
8
2
0
30
7
4
4
3
2
2
2
1
1
1
0
1
43.0
29.6
25.9
10.2
10.2
4.6
2.8
4.6
0.9
0.9
0.9
7.4
1.9
0
2
3
2
42.6
52.3
15.9
10.3
9.4
4.7
2.8
1.9
0.9
0.9
0.9
0
0
0
13.2
21.2
15.2
12.1
9.1
9.1
6.1
6.1
3.0
3.0
3.0
3.0
3.0
3.0
3.0
0
13.6
64.7
32.4
2.9
10.0
20.0
20.0
16.0
16.0
8.0
4.0
4.0
4.0
4.0
4.0
0
6.0
4.8
3.2
4.0
0.8
1.2
0.8
2
23.3
13.3
13.3
10.0
6.7
6.7
6.7
3.3
3.3
3.3
0.0
3.3
3.0
3.0
0
13.0
60.0
36.7
3.3
6.9
12.5
12.5
12.5
31.3
0
6.3
6.3
12.5
6.3
0
0
4.3
6.1
3.0
4.3
1.3
0.4
0.9
Total
25 1
100.0
23 1
100.0
* ASHD
1
I
I
0
15
12
8
10
1
1
0
30
18
11
1
16
2
2
2
5
0
1
1
2
1
0
0
10
14
7
10
3
1
13.0
Saskatchewan
Province %t
48.0
22.5
35.5
5.8
17.5
2.2
2.1
1.8
0.1
1.6
0.3
7.9
1.6
1.2
22.8
11.8
16.8
3.8
4.9
3.7
13.7
6.4
0.8
2.1
1.7
3.9
0.4
2.5
0.1
27.3
5.7
95.2
1.5
3.3
0.6
8.9
0
12.6
12.6
0
0
0
0
10.3
26.0
29.6
3.7
4.0
1.7
4.0
3.1
0.5
5.9
100.0
arteriosclerotic heart disease; CVA = cerebrovascular accident; Ca = cancer; SLE =
systemic lupus erythematosus; CNS = central nervous system.
t Saskatchewan data taken from all deaths (6,944) over the age of 20 among persons in Saskatchewan
Province in 1979, age- and sex-matched to the study population.
=
MORTALITY IN RA
71 1
What are the causes of death in rheumatoid
arthritis? A comparison of the distribution of causes of
death between the rheumatoid arthritis population and
the general population of Saskatchewan Province (Table 5) showed a specific pattern of excess mortality
among the RA patients. The 2 major causes of death
(cardiovascular disease and malignancy) were similar
in the RA population and in the general population.
However, RA patients showed a marked increase in
deaths resulting from infection or sepsis and associated with the rheumatoid process itself. There was
also some increase in deaths from gastrointestinal
causes: principally, hemorrhage or perforation.
Ainong the arthritis-related causes of death were
vasculitis, rheumatoid lung disease, amyloidosis, nonspecific debilitation, and subluxation of the upper
cervical spine. There was 1 case of phenylbutazone
bone marrow suppression. Infectious complications
were usually located in the lungs, with a high incidence
of generalized sepsis.
Using death certificates, comparisons of causes
of death are always flawed because of the frequent
inaccuracies in data. Though the actual cause in the
study group is carefully determined, comparison with
the general population is difficult, since these data
depend upon death certificates. We carefully audited
all deaths, using physician records and hospital charts
as well as death certificates (available for 231 patients).
We found that death certificates underreported causes
of death due to rheumatic diseases, but all other
comparisons in Table 5 are accurate.
Table 6 compares causes of death by decade of
age, beginning with age 50. The same patterns of
mortality occurred in each of the various age groups.
The differences between RA patients and the general
population were particularly striking in individuals
who died between the ages of 50 and 70. Thus, deaths
tended to be more specifically related to the RA
process in the younger group and related to more usual
causes in older people. Within the category of malignancies, we searched carefully for particular malignancies that might be associated with RA or its treatment.
No particular tumor was significantly more common or
less common than expected on the basis of chance.
Leukemia and lymphoma (5 cases) were unusual
events. Data are not available to compare age-specific
rates for specific causes of death. Because of the
premature mortality in the RA group, one would
expect an apparent reduction in causes of death that
are strongly age-related (e.g., malignancy, atherosclerosis). Thus, one should not conclude that a decrease
in such conditions is associated with rheumatoid
arthritis.
What are the effects of treatment upon mortality
in rheumatoid arthritis? Since disease severity is related to premature mortality, it seemed possible that a
therapeutic agent which reduced disease severity
might decrease mortality. Conversely, a medication
such as prednisone, which is believed to lower resistance to infection, might be expected to increase
infectious complications. The original regression analyses included the medications most frequently used by
RA patients (prednisone, gold, phenylbutazone, and
indomethacin). Results using these variables failed to
add predictive strength to the Cox model.
Similarly, we were unable to find differences in
the causes of death between subjects who were receiving gold or prednisone and those who were not. The
exceptions were 6 deaths that occurred in vasculitis
patients treated with steroids; these patients had been
receiving steroids before they developed vasculitis.
Concern has been raised about the carcinogenicity of long-term gold salt administration. Cancer
data for patients treated with gold and those not
receiving gold therapy revealed that malignancies
were evenly distributed between the 2 groups. The
frequency of malignancy in the gold group actually
tended to be lower than those not so treated; thus, it
Table 6. Causes of death in rheumatoid arthritis (RA) patients and in the general population of
Saskatchewan (Sask) Province by decade of age (%)
Cause
Age 50-59
RNSask
(n = 371640)
Age 60-69
RNSask
(n = 8511,241)
Age 70-79
RA/Sask
(n = 9011,766)
Age 80+
RA/Sask
(n = 3412,636)
Cardiovascular
Malignancies
Infections
Rheumatic diseases
Gastrointestinal
Other
27/29
14141
2212
311
1113
23/24
44/39
14/38
1312
611
613
17/17
53/53
3/24
1014
311
313
28/15
41/58
29114
6/10
611
916
911 1
MITCHELL ET AL
712
seems unlikely that there is significant carcinogenesis
related to treatment with intramuscular gold compounds in rheumatoid arthritis patients. Our results
were similar to those of another recent careful study
(17), in which a suggested increase in leukemia or
lymphoma could not be confirmed.
DISCUSSION
Many studies of survivorship and prognosis for
rheumatoid arthritis patients have been published
since 1953 (1-18). In general, researchers report increased mortality rates among RA patients. Cobb and
coworkers (1) indicated shorter survival for men with
rheumatoid arthritis whose disease began before age
50. Uddin et a1 (2) observed the cumulative survival
rate to be lower than expected for RA patients of both
sexes and at each year of followup, when compared
with the general population of the Ontario Province.
Bywaters et a1 (3) reported that overall mortality in RA
is greater than expected, but no tests of statistical
significance of these differences were reported. Duthie
and coworkers (4) reported that death rates from all
causes are higher in rheumatoid arthritis patients than
those in a general population of comparable age and
sex, again with no report of statistical significance.
Rasker and Cosh (5) reported that the mean age at
death was 55.7 years for both sexes in 43 of 100 RA
patients who died during an 18-year followup period,
compared with a mean age at death of 71.9 years in the
general population of England and Wales.
Moesmann (6) reported on 2 series of patients
with rheumatoid arthritis. In the study of patients with
disease onset after age 60, the mortality rate for all
patients (men and women) was 2.6 times the expected
mortality rate. In a separate series of younger RA
patients, the mortality rate was 1.3 times that to be
expected in the population of the area. Kellgren and
O’Brien (7) reported increased mortality in patients
with seropositive as compared with seronegative disease. Reah (8) noted that 80 (42.5%) of the deaths in
185 rheumatoid patients followed for a mean of 12.75
years were related to “prolonged incapacity.” A study
by Vandenbroucke et a1 (17) of 165 RA patient deaths
produced estimates of increased mortality and specific
causes of death that closely parallel our findings. Cosh
(18) has recently reviewed and commented upon the
available studies, closing with a call for better determination of factors “predictive of future trouble.”
Only Linos et a1 (9) failed to find as great an increase
in mortality rate in rheumatoid arthritis patients (com-
pare also refs. 10-18), but that population-based study
may have included some patients who did not have
RA .
Previous studies have been criticized for hospital-based patient selection, incomplete followup, small
numbers, or uncertainty of diagnosis. In addition, no
previous study used prospective protocols that comprehensively detailed entry characteristics, and none
has systematically examined the predictive power of
specific entry variables.
It is generally agreed among clinicians that
certain features of rheumatoid disease may result in
death, including development of vasculitis, amyloidosis, rheumatoid lung disease, and cervical cord compression syndromes. Similarly, drug-related deaths
have been reported for patients receiving gold, penicillamine, phenylbutazone, immunosuppressive drugs,
or other agents.
The present study, which required 17 years for
its completion, meticulously recorded patient characteristics of a large, non-hospital-based regional population, by protocol at entry and throughout the course.
Followup information was available on 94% of patients
after an average of 12 years. The effects of rheumatoid
arthritis upon disability and economic factors in this
same patient population have been discussed in previous reports (24,25).
In this study, we found that patients with rheumatoid arthritis do die prematurely, both in comparison with population controls and with those whose
diagnosis changed after being enrolled in the study
with a diagnosis of RA. Moreover, the 34 patients lost
to followup had significantly greater age (58.5 years)
and disease duration (23 years) at study entry than
those observed, suggesting that the rate of death for
this group of patients was higher; thus, data presented
here may underestimate the true mortality rate.
Of approximately 79 excess deaths among the
RA patients studied, about 18 are accounted for by
excess infections: principally, pneumonia and sepsis.
About 20 deaths appeared to be disease-associated,
with causes ranging from vasculitis to cervical
subluxation and rheumatoid lung disease. Approximately 8 deaths appeared to result from drug side
effects; these were principally caused by upper
gastrointestinal bleeding or perforation, associated
with prednisone and nonsteroidal drugs. Of interest,
only 1 putative drug death was hematologic and none
were renal, while gastrointestinal complications of
drugs accounted for 7 excess deaths.
The remaining 33 excess deaths cannot be ex-
MORTALITY IN RA
plaimed by specific direct causes and do not fall into
specific categories, which suggests a nonspecific effect
of host debilitation similar to accelerated aging. The
debilitated host, not unexpectedly, is at increased risk
of death from all causes, since such hosts are less able
to withstand the effects of any kind of perturbation (26).
The factors present in RA patients when first
seen, that best predicted decreased survivorship, were
age and sex. As expected, the effects of these variables
are of major importance, and closely parallel those
seen in nonrheumatoid populations. However, factors
relating to the severity of the disease, such as ARA
functional class, latex fixation titer, and number of
involved joints, also contributed to the observed excess deaths among RA patients. Again, this observation suggests that compromise of the host, both physically and immunologically, increases death rates.
Physical disability (represented by the ARA
functional class) and disease severity are both important. In univariate analyses, all disease severity variablles tested, including erythrocyte sedimentation rate,
joint swelling, muscle weakness,-radiographic grade
and erosions, morning stiffness, symmetry of afihritis,
grip strength, rheumatoid nodules, and metatarsal
syrnovitis, were statistically associated with decreased
survivorship at the 0.001 level. The more severely
affected patient was at greater risk. No drug effect on
survival could be ascertained with the 2 drugs (prednisone and intramuscular gold) that were used with
frequency sufficient for analysis.
The decreased survivorship of rheumatoid arthriitis patients is impressive; it appears to be of the
same order of magnitude as that in patients with
Hodgkin’s disease (all stages), diabetes mellitus (type
I and type 11), stroke (age-adjusted), and postmyocardial infarction (age-adjusted) (27). The broad
distribution of modes of death in different categories
has prevented general recognition of the adverse effects of rheumatoid arthritis upon longevity. Rheumatoid arthritis is a major chronic disease which, in
addition to causing deformity and disability, is demonstrably associated with decreases in survivorship.
ACKNOWLEDGMENTS
The authors greatly appreciate statistical review and
criticism by Dr. Byron W. Brown and Dr. Lincoln Moses,
and thank the Vital Statistics Branch, Saskatchewan Health
713
Department, for search of death records and provision of life
table data.
REFERENCES
1. Cobb S, Anderson F, Bauer W: Length of life and cause
of death in rheumatoid arthritis. N Engl J Med 249:
553-556, 1953
2. Uddin J, Kraus AS, Kelly HG: Survivorship and death
in rheumatoid arthritis. Arthritis Rheum 13: 125-130,
1970
3. Bywaters E, Curwen M, Dresner E, Dixon A: Ten year
follow-up of rheumatoid arthritis. Ann Rheum Dis
20: 198-199, 1961
4. Duthie JJR, Brown PE, Knox JDE, Thompson M:
Course and prognosis of rheumatoid arthritis. Ann
Rheum Dis 16:411-422, 1957
5. Rasker JJ, Cosh JA: Cause and age at death in a
prospective study of 100 patients with rheumatoid arthritis. Ann Rheum Dis 40:115-120, 1981
6. Moesmann G: Malignancy and mortality in subacute
rheumatoid arthritis in old age. Acta Rheum Scand
15:193-199, 1969
7. Kellgren JH, O’Brien WM: On the natural history of
rheumatoid arthritis in relation to the sheep cell agglutination test (abstract). Arthritis Rheum 5: 115, 1962
8. Reah TG: The prognosis of rheumatoid arthritis. Proc R
SOCMed 565313-817, 1963
9. Linos A, Worthington JW, O’Fallon WM, Kurland LT:
The epidemiology of rheumatoid arthritis in Rochester,
Minnesota: a study of incidence, prevalence, and mortality. Am J Epidemiol 111:87-98, 1980
10. Jacoby RK, Jayson MIV, Cosh JA: Onset, early stages,
and prognosis of rheumatoid arthritis: a clinical study of
100 patients with 11 year follow-up. Br Med J i:96-100,
1973
11. Koota K, Isomaki H , Mutru 0: Death rate and cause of
death in RA patients during a period of five years. Scand
J Rheumatol6:241-244, 1977
12. Duthie JJR, Brown PE, Truelove LH, Baragar FD,
Lawrie AJ: Cause and prognosis in rheumatoid arthritis:
a further report. Ann Rheum Dis 23:193-204, 1964
13. Burch TA: Mortality from rheumatoid arthritis in the
United States, 1959-1961, Population Studies in the
Rheumatic Diseases. London, Excerpta Medica, 1968,
pp 39-42
14. Constable TJ, McConkey B, Paton A: The cause of
death in rheumatoid arthritis. Ann Rheum Dis
37~569-578,1978
15. Monson RR, Hall AP: Mortality among arthritics. J
Chronic Dis 29:459-467, 1976
16. Lewis P, Hazelman BL, Hanka R, Roberts S: Cause of
7 14
17.
18.
19.
20.
21.
death in patients with rheumatoid arthritis with particular reference to azathioprine. Ann Rheum Dis 39:
457-461, 1980
Vandenbroucke JP, Hazevoet HM, Cats A: Survival and
cause of death in rheumatoid arthritis: a 25-year prospective follow-up. J Rheumatol 11: 158-161, 1984
Cosh JA: Survival and death in rheumatoid arthritis. J
Rheumatol 43: 117-1 19, 1984
Fries JF, Hess EV, Klinenberg JR: A uniform database
for rheumatic diseases. Arthritis Rheum 22: 1029-1033,
1979
McShane DJ, Harlow A, Kraines RG, Fries JF: TOD: A
software system for the ARAMIS data bank. Computer
12:3440, 1979
Fries JF, Spitz PW, Kraines RG, Holman HR: Measurement of patient outcome in arthritis. Arthritis Rheum
23:137-145, 1980
MITCHELL ET AL
22. Fries JF: The assessment of disability: from first to
future principles. Br J Rheumatol (suppl) 22:48-58, 1983
23. Ropes MW, Bennett GA, Cobb S, Jacox R, Jessar RA:
1958 revision of diagnostic criteria for rheumatoid arthritis. Bull Rheum Dis 9:175-176, 1958
24. Sherrer YS, Bloch DA, Mitchell DM, Young DY, Fries
JF: The development of disability in rheumatoid arthritis. Arthritis Rheum 29:494500, 1986
25. Lubeck DP, Spitz PW, Fries JF, Wolfe F, Mitchell DM,
Roth SH: A multicenter study of annual health service
utilization and costs in rheumatoid arthritis. Arthritis
Rheum 29:488-493, 1986
26. Fries JF: The compression of morbidity. Milbank Mem
Fund Q 61:397-419, 1983
27. Fries JF, Ehrlich GE: Prognosis. Menlo Park, CA,
Addison-Wesley , 1981
Документ
Категория
Без категории
Просмотров
0
Размер файла
737 Кб
Теги
death, survival, causes, arthritis, prognosis, rheumatoid
1/--страниц
Пожаловаться на содержимое документа