Juvenile rheumatoid arthritis in rochester minnesota 19601993. Is the epidemiology changing

AKI'~IKl'I'1S& I<III~~UMA'IISM
Vol. 39, No. 8, August 1996, pp 1385 1390
61 1996. American Collcgc of Rhcurnatology
.
. ..
1385
- .... .
. _.
_.
.. . ...-
JUVENILE RHEUMATOID ARTHRITIS IN
ROCHESTER, MINNESOTA 1960-1 993
Is the Epidemiology Changing?
LYNNE S. PETERSON, TOM MASON, AUDREY M. NELSON,
W. MICHAEL O'FALLON, and SHERINE E. GABRIEL
Objective. To examine trends in the incidence and
prevalence of juvenile rheumatoid arthritis (JRA) in
Rochester, Minnesota, over 33 years.
Methods. The diagnostic retrieval system of the
Rochester Epidemiology Project was utilized to screen
medical records of all Rochester residents with any
potential diagnoses of JRA from 1978 to 1993 (based on
the American College of Rheumatology 1977 revised
criteria). In addition, all cases of J R A from our previously identified cohort from 1960-1979 were verified,
and the 2 data sets were combined, resulting in an
incidence cohort spanning 33 years (1960-1993).
Results. Of the 1,240 medical records screened,
we identified 65 cases of JRA diagnosed between 1960
and 1993 (48 females, 17 males). The average followup
for cases was 12.7 years (range 0-34 years) for a total of
833 person-years of observation. A bimodal distribution
of age at diagnosis was observed, with peaks between 0
and 4 years and 9 and 15 years. Seventy-two percent of
patients had pauciarticular-onset, 17%had polyarticularonset, and 11% had systemic-onset disease. Progression
of pauciarticular to polyarticular disease occurred in
11% of the cases. The overall age- and sex-adjusted
incidence rate was 11.7 per 100,000 population (95%
confidence intervals 8.7, 14.8). The incidence rate per
100,000 population was 15.0, 14.1, and 7.8 for the time
Supported in part hy research grants from the Arthritis
Foundation (MAP 118) and the USPiIS, NII-I (AR-30582).
Lynne S. Peterson, MD, Tom Mason, MD, Audrey M.
Nelson, MD, W. Michael O'Fallon. PhD, Shcrine E. Gabriel, MD.
MSc: Mayo Clinic, Rochcstcr, Minnesota.
Address reprint requests t o Sherine E. Gabriel, MD, MSc,
Department of I Iealth Sciences Rcscarch, Mayo Clinic Rochester, 200
First Street SW, Rochester, MN 55905.
Submitted for publication November 22. 1995; acceptcd in
revised form April 3, 1996.
periods 1960-1969, 1970-1979, and 1980-1993, respectively (P = 0.024). A 3-year, centered, moving average,
which was used to display time trends in incidence,
suggested a cyclical pattern, with incidence peaks in
1967, 1975, and 1987.
Conclusion. An overall decrease in the incidence
rate over the last decade was observed, most marked in
the pauciarticular- and systemic-onset subtypes. This
decrease, along with the observed cyclical pattern, suggest that environmental factors may influence disease
frequency.
Juvenile rheumatoid arthritis (JRA) is a wellrecognized chronic inflammatory condition in children.
The diagnosis of JRA is currently established by criteria
developed by the American College of Rheumatology
(ACR; formerly, the American Rheumatism Association) in 1977 (1). Three subtypes of JRA have been
defined based on clinical manifestations within the first 6
months of illness. These subtypes include pauciarticularonsct ( 5 4 joints involved), polyarticular-onset ( 2 5 joints
involved), and systemic-onset (predominance of extraarticular fcaturcs, such as fever and rash) disease. The
etiology of J R A is unknown; however, several studies
have reported variation in the incidence, suggesting an
infectious etiology (2,3).
Previous epidemiologic studies of JRA have been
limitcd by a referral-based sampling frame and short
followup periods (4-13). We conducted a populationbased study of the incidence of JRA between 1978 and
1993 and re-abstractcd and verified the data from our
previously published study (1960-1979) (14) in order to
describe trends in the incidence and long-term outcome
of this disorder in the community.
1386
PETERSON ET AL
Table 1. Characteristics of 65 Rochester, Minnesota residents with juvenile rheumatoid arthritis
bctwccn 1960 and 1993
Onset type
Number of paticnts
Fcnialcs
Males
Age at diagnosis, mean t- SD years
Females
Males
Months to diagnosis. mean z S1)
Females
Malcs
Rochcstcr followup. mcan I
SD years*
E’cmalcs
Males
Overall
Pauciarticular
Polyarticular
65
48
17
47
36
11
11
8.4 z 4.9
7.5 x 5.0
11.6 z 3.2
8.6 - 13.9
7.3 I
13.5
12.5 L 15.2
12.6 z 8.0
11.4 i 8.0
16.1 x 7.3
5.8 i 4.3
6.2 ? 4.2
5.0 % 5.3
7.6 -c 9.8
9.9 ? 10.8
1.7 I 0 . 1
12.0 t- 10.8
14.1 L 10.9
6.4 I
10.1
7.8
6.9
10.4
7.9
7.6
8.7
12.7
12.3
13.6
* Followup based o n incidence cases only (n
-
2 4.9
4.8
2 4.2
12.7
? 12.7
i 13.1
t- 8.6
I
8.6
t- 8.6
?
+
Systemic
8
3
6.4
3.0
11.0
4.1
5.7
2.0
13.7
15.1
11.8
7
4
3
t 5.3
-C 3.6
? 3.0
-C 7.5
? 10.1
L 1.5
+- 8.9
19.1
110.3
59).
PATIENTS AND METHODS
Thc Rochcstcr Epidemiology Projcct (REP) database
contains information on virtually all mcdical care provided to
rcsidcnts of Olmstcd County (15). Iliagnoscs made in the
clinics, hospitals, home visits, or at autopsy are entered into
a ccntralized index, which cnablcs thc identification and retrieval of pertinent medical records. This system ensures
virtually complete ascertainment of all clinically defined discase among residents in Olmsted County (including Rochester,
Minncsota).
In this study, we idcntificd all Rochcstcr rcsidcnts with
J R A that was first diagnosed between January 1, 1960 and
Dcccmbcr 31, 1993. Using thc R E P data rcsourccs, wc obtaincd thc complctc (inpaticnt and outpatient) medical records
of all Rochcstcr rcsidcnts agc 518 bctwccn January 1, 1978
and Dcccmbcr 31, 1993 with the following diagnoses: arthritis,
JRA, rheumatoid arthritis, Still’s disease, rheumatic fcvcr,
systemic lupus ciythematosus, infectious arthritis, inflammatory arthritis, palindromic rheumatism, degencrativc joint disease, ostcoarthritis, polyarthralgia, polyarthritis, acute arthritis, Felty’s syndrome, spondylitis, ankylosing spondylitis,
collagen vascular and ovcrlap discascs, iritis, iridocyclitis,
uveitis, I.ymc discasc, joint effusion, joint swelling, or traumatic arthritis. A prcvious study reviewed 150 mcdical rccords
with a diagnosis of fever of unknown origin and found no cases
that mct thc critcria for JIiA (14); therefore, this diagnosis was
not rcscrccncd.
A pretested data collection form was used to collect
information on dcmographics, clinical manifestations, laboratory findings, and disease course at diagnosis and at the last
followup. All cases of juvenile ankylosing spondylitis, psoriatic
arthritis, and arthritis secondary to inflammatory bowel disease
were excludcd. Using thc samc data collection instrument, we
rcvicwed all 49 patients with JRA idcntificd in our previously
publishcd study who had been diagnosed bctwcen January 1,
1960 and December 31, 1978 (14). In addition, all medical
records that wcrc scrccncd but excluded from thc previous
incidence study during thc ycars 1978 and 1980 wcrc rcreviewed to cnsurc complctc ascertainment of cases and
reliability of data collection. All data collection was conducted
by the principal investigator (LSP).
A n incidcncc casc was dcfincd as a rcsidcnt of Rochcstcr, Minncsota, whosc first diagnosis of JRA, bascd on the
ACR 1977 critcria ( l ) , was madc bctwccn January 1,1960 and
Ilcccmbcr 31, 1993 and who was <16 years of age at the time
of diagnosis. The definition of a prevalence case was any
Rochester resident who was <16 years of age on either
January 1, 1980 or January I , 1990 and who had had a
diagnosis of JIiA (based on the ACK 1977 criteria) prior to
that datc. ‘I’hc timc to diagnosis was calculated from the datc
of symptom onset until the date of diagnosis of JRA.
The incidcncc ratc was cstiniatcd by dividing the
numbcr of incidcncc cascs, as dcscribcd abovc, by the Iiochester population of children < I 6 years of age. Overall ratcs
were age- and sex-adjusted to the 1970 United States whitc
population, as obtained from the decennial censuses, with
linear interpolation for the intercensal years. Adjustmcnt to
thc year 1970 was used to enable direct comparison with the
previous study by Towncr ct al (14). Nincty-fivc perccnt
confidcncc intervals wcrc computcd for the incidence rates
assuming the observed number of incidence cases in any
age-sex specific category is the realization of a Poisson random
variable. The annual incidence rates were graphically illustratcd using a 3-ycar, ccntcrcd, moving avcragc to eliminate
some of the random fluctuation over time. Trends in the
incidcncc rate ovcr the 33-year period wcrc analyzcd using
Poisson rcgrcssion analysis (16). This modcl was utilizcd to
evaluatc the rclationship bctwccn incidcncc (thc dependent
variablc) and, agc, sex, and chronological timc (indcpcndcnt
variables). The prcvalcncc ratcs were estimated by dividing the
number of prevalcncc cases, as described abovc, on January I ,
1980 and January 1, 1990 by thc Rochcstcr population of
childrcn <16 ycars of age on those datcs. Thc survival ratc was
cstimatcd using the Kaplan-Meier method (17).
RESULTS
Patient characteristics. We screened 1,240 medical records and identified 65 patients (48 female and 17
1387
JRA IN ROCHESTER, MN, 1960-1993
0.1
4-5
2-3
67
0-9
10-11
12-13
14-15
Age at diagnosis
Figure 1. Age at diagnosis in Rochester, Minnesota residents with
juvenile rheumatoid arthritis, by sex.
male) who first fulfillcd the ACR criteria for JRA
between 1960 and 1993 (Table 1). Of the 1,240 medical
records, 1,098 patients were excluded because they did
not fulfill the ACR 1977 criteria for JRA (i.e., age 2 1 6
years at diagnosis, symptom duration <6 weeks, or no
evidence of joint effusion, tenderness, limited range of
motion, or warmth) or they were non-Rochester residents at the time of diagnosis. The remaining 77 patients
may have fulfilled the ACR criteria but were found to
have other rheumatic conditions and were therefore
excluded: systemic lupus erythematosus in 12, HenochSchonlein purpura in 9, Lyme disease in 4, septic
arthritis in 10, juvenile ankylosing spondylitis in 8,
psoriatic arthritis in 4, serum sickness in 5, viral arthritis
in 7, transient hip synovitis in 9, reactive arthritis in 6,
and Kawasaki disease in 3. Of the 65 cases included, 6
patients were excluded from the incidence study for the
following reasons: 1 patient developed the disease prior
to 1960 and 5 patients werc not Rochester residents at
the time of diagnosis, and therefore, could not be
classified as incidence cases. However, each of thesc 6
patients met critcria for a prevalence case on January 1,
1970, January 1, 1980, andlor January 1, 1990.
The mean % SD age at diagnosis was 7.8 % 4.9,
8.4 2 4.9, 5.8 -I 4.3, and 6.4 +- 5.3 years of age for all
patients, the pauciarticular group, the polyarticular
group, and the systemic group, respectively (Table 1).
The age at diagnosis was consistently higher in males
than in females (Figure 1). For the overall group, a
bimodal agc at diagnosis was apparent, with peaks
between 0 and 4 years and 9 and 15 years (Figure 1). The
mean 2 SD time from first symptoms to diagnosis was
7.9 t 12.7 months overall, 8.6 5 13.9 months in the
pauciarticular group, 7.6 t 9.8 months in the polyarticular group, and 4.1 % 7.5 months in the systemic group.
The mean length of followup while a Rochester resident
was 12.7 % 8.6 years (range 0-34 years).
All JRA cases werc subdivided according to their
onset presentation. Seventy-two percent had pauciarticular onset, 17% had polyarticular onset, and 11%
had systemic onset. Progression of pauciarticular to
polyarticular disease, after the first 6 months of illness,
occurred in 7 patients (11%). Six of the 7 patients with
systemic-onset disease presented with pauciarticular
joint involvement, and the seventh patient had a polyarticular disease onset.
Incidence, prevalence, and survival rates. The
overall age- and sex-adjusted annual incidence rate for
1960-1993 was 11.7 per 100,000 (95% confidence interval 8.7, 14.8) (Table 2). The incidence rates for 19601969,1970-1979, and 1980-1993 were 15.0,14.1, and 7.8
per 100,000 population, respectively. The annual incidence of JRA was significantly higher in female than in
male patients (P = 0.0005). However, the obscrvcd
trends in annual incidence rates, over time, were similar
in both sexes. A statistically significant decline in incidence over the study period was demonstrated using
Table 2. Annual incidence of juvenile rheumatoid arthritis in Rochester, Minnesota residents*
Onset type
Ycars
Overall
Pauciarticular
Polyarticular
Systemic
1960-1 993
Females
Malcs
1960-1969
1970-1 979
1980-1993
11.7 (8.7, 14.8)
16.6 (11.6, 21.70
7.0 (3.7, 10.4)
15.0 (8.6, 21.4):
14.1 (8.0, 20.l)t
7.8 (3.9, 11.7)t
8.4 (5.8, 11.0)
12.2 (7.8, 16.7)
4.7 (1.9,7.5)
11.3 (5.7, 16.9)
9.0 (4.1, 14.0)
5.8 (2.5, 9.1)
2.0 (0.8,3.2)
3.0 (0.9, S.2)
1.1 (0.0,2.3)
2.5 (0.005,4.9)
1.8 (0,3.7)
2.0 (0.02, 4.0)
1.3 (0.3, 2.3)
1.4 (0.02,2.8)
1.3 (0.0,2.7)
1.2 (0,2.3)
3.3 (0.4.6.2)
0.0 (0,O)
* Valucs are the annual incidence per IOO,(K)O population, age-adjusted to 1970 US white population, with
95% confidence intervals shown.
I I r 0.024 for the 3-way comparison.
1 . )
PETERSON ET AL
1388
wrist, ankle, metacarpophalangeal, metatarsophalangeal, and elbow joints were most frcqucntly involved in
polyarticular diseasc. The knee, ankle, and second and
third PIP joints were most frequently involved in systemic-onset disease.
"1
DISCUSSION
O !
1960
I
1965
1970
1975
1980
1985
1990
1995
Calendar year of diagnosis
Figure 2. Annual incidence of juvenile rheumatoid arthritis in Rochester, Minnesota residents, 1960-1993 (estimatcd using a 3-year
moving average).
Poisson regression analysis ( P = 0.024). There was a
decline in both pauciarticular-onset and systemic-onset
JRA. No significant change in the incidence rate was
sccn in the polyarticular group over the 33 ycars.
Figure 2 illustrates thc trcnd in incidence rate
over the 33 years of the study, using the 3-year moving
average. There is a suggestion of pcaks in incidence in
1967, 1975, and 1987. The prevalence rate for JRA on
January 1, 1980 was 94.3 per 100,000 (95% confidencc
interval 40.9, 147.7), and on January 1, 1990 it was 86.1
pcr 100,000 population (95% confidencc interval 36.9,
135.3). The survival rate was not significantly different
from that of the Minnesota white population ( P =
0.599). Our incidence and prevalencc ratcs were somewhat lower than thosc rcportcd by Towner et a1 (14)
during the same time period. Threc patients identified
from Towner's study wcrc >16 years of age at the time
of diagnosis and therefore did not mcct thc JRA criteria.
Another patient in thcir cohort was not a Rochester
resident either at the time of diagnosis of JRA or on
prevalencc datcs January 1, 1970 or January 1, 1980.
Two othcr patients from Towncr's January 1, 1980
prevalence cohort wcre subsequently found to be nonresidents. All 6 of thcsc patients were excluded from our
study.
Joint involvement. The onsct of pauciarticular
disease occurred in all 12 months of the year and the
onset of polyarticular disease occurred in all months
except January, May, July, and September. The onset of
systcmic disease occurrcd in February through May and
in July.
Joint involvcment in each of the JKA subtypes at
diagnosis is illustrated in Tablc 3. The kncc and ankle
were t h e joints most frequently involvcd in pauciarticular discasc. The knee, proximal intcrphalangeal (PIP),
The overall incidence rate observed in this study
was similar to that in previous population-based studies
(3,14,18) and higher than that in previous referral-based
studies (2,4,6,7,9,10). Because mild cases of JRA are less
likely to be referred, epidemiologic studics conducted
in referral centers are subject to selection bias, which
may account for the lower rcportcd incidence rates of
JRA (4,6,7,9,10,19). Population-based studies, such as
ours, eliminate this referral bias (3,9). In addition,
comparability with other studies is difficult, because
many lack standardized classification criteria for JRA
and/or include other causes of chronic arthritis in addition to JRA.
We observed a decrease in thc incidence of JRA
over thc last decade in the systemic- and pauciarticularonset disease subtypes. These trends may reflect a
change in clinical practice (i.c., diagnoses of other
discascs) and not necessarily a decreasing incidence. In
thc era of better serologic tests, other diseascs which
may mimic JRA can be more accuratcly diagnosed.
Lyme disease, for example, which may present as an
oligoarthritis, has been recently characterizcd by serologic studies. We excluded from this cohort all cases of
patients who were diagnosed with Lyme disease who
may have otherwise fit diagnostic criteria for JRA (4
Table 3. Joint involvement at diagnosis in the JRA subtypcs'
Onset type
Shouldc r
Elbow
Wrist
MCP
PIP
I lip
Kncc
Ankle
MTP
Pauciarticular
(n = 47)
Polyarticular
(n = 11)
Systemic
(n - 7)
0
2
13
2
0
4
0
18
36
27
45
0
64
36
27
14
14
14
14
43
0
51
57
14
64
23
2
* Joint involvement consistcd of joint swelling or effusion or >2 of thc
following: limited range of motion, tenderness or pain on motion, or
increased warmth of joint. Values are percentages of patients. MCP =
mctacarpophalangeal; PIP = proximal interphalangeal; MTP = mctatarsophalangcal.
JRA IN ROCHESTER, MN, 1960-1993
patients). In addition, none of the 65 patients with JRA
were ultimately diagnosed with Lymc disease. Secondly,
more sensitive radiologic tests (i.e., magnctic rcsonance
imaging) are now available for earlier, more accurate
diagnoses of intraarticular abnormalities which may
present with joint effusions (ix., avascular necrosis).
This reasoning may also apply to systemic-onset JRA,
which has unique features that are distinctive from those
of polyarticular- and pauciarticular-onset disease. It is
possible that, in more recent years, physicians are making diagnoses other than systemic JRA (e.g., transient
synovitis of the hip, Takayasu arteritis, Kawasaki disease,
or rheumatic fever) to account for the numcrous and
varied symptoms that are typically seen in this illness.
A seasonal variation of systemic-onset JRA has
been reportcd by one author (20), but this has not been
confirmed by others (21). Data from the 7 patients in our
systemic-onset cohort suggested a clear seasonal variation of disease onset, between February and July. In
contrast, the onset of pauciarticular and polyarticular
disease occurred in esscntially cvery month. The unique
features of systemic-onset JRA and its seasonal variation
are suggestive of an infectious etiology. Although multiple potential infectious causes have been reported in
JRA (2,22-28), evidence thus far of an infcctious etiology for JRA remains tenuous.
The observed decrease in the incidence rate ovcr
the last decade may simply represent a cyclical trough
during this time period. In fact, our results suggest a
cyclical pattern in the incidence of JRA over the last 3
decades (Figure 2). When analyzing both sexes, peaks in
the incidence are suggested in 1967, 1975, and 1987.
Gare and Fasth (3), during a 5-year population-based
study performed in Sweden, noted a lower overall
incidence of JRA in 1984 (8.9 per 100,000) and in 1988
(8.3 per 100,000) compared with 1985 through 1987
(11.6-12.4 per 100,000). A recent study by Oen et a1 (2)
documented variability in the incidence rates of JRA
over time. They reported a 3-4-year cyclical pattern in
the incidence of JRA, in which incidence peaks correlated with the incidence of Mycohactenum pneumoniae
infections, based on serologic screening in 1979, 1982,
1986, and 1990-1991. Although variations in the incidence rate of JRA have been found in several studies,
the peaks found in our cohort do not coincidc with those
found in Gare’s and Oens’ studies (23). Thcsc differences may reflect multiple infectious agents or thc
influence of geographic location or other environmental
factors on JRA prescntation, or simply the natural
variation of the disease. It is of interest that a rcccnt
study of giant cell artcritis rcportcd a regular cyclic
1389
pattern in incidence rates ovcr time (29), supporting the
hypothesis of an infectious cause in that disease.
As with any study, our results must be interpreted
in light of our study limitations. There is always a
possibility of underascertainment of cases in medical
record review studies. Several steps were taken to ensurc
complete case asccrtainment, however. First, a pretested
data abstraction form was used, and any questions
regarding a correct diagnosis were reviewed with the
coinvestigators and rcsolved by consensus. Second, cases
wcrc idcntificd using the medical records linkage system
of the Mayo Clinic, which cnsures almost complete
ascertainment of all cases of any disease diagnoscd in
Olmsted County dating back 75 ycars (15). Third, a
random samplc of 50 medical records with a “possible”
diagnosis of JRA were reviewed by one of the coinvestigators (TM), who was blinded to the results of thc first
review. No additional cascs were found after this confirmatory rcview. Fourth, we included more screening
codes than wcrc used in the previous study of the
epidemiolob?yr of JRA (14) (i.e., Lymc disease, joint
cffusions, joint swelling). The addition of these codes
resulted in the identification of only 1 additional case.
Fifth, to ensurc that indexing of the Rochester Epidemiology Project was complete and up to date, we
rescreened thc 1993-1994 list in May 1995 to identify
caws where a delay in diagnostic coding may have
occurred. No ncw cases of JRA were identified. Finally,
all screening lists for the 1978-1993 time pcriod wcrc
reassembled to ensurc accuracy of our screening lists.
No additional cases were identified. For these reasons,
we believe that underascertainmcnt of cascs is an unlikely explanation for our low incidence rates.
Other potential limitations arise when using databases linked to medical records such as the REP (30).
For a disorder to be recorded in the database, the
condition must be recognized by a clinician, recorded
accurately, and retrieved. Conditions that are unlikely to
come to mcdical attention may not be detectcd in such a
data set. It is possible, therefore, that some cases of JRA
go undetected by patients and their physicians. Thus, our
estimates of the incidence of JRA may bc conservative.
Finally, some racial and ethnic groups are underrcpresented in Olmsted County, Minnesota, whcrc the population was 96% white in 1990 (US Census data). Thus,
the rcsults of our population-based study are only generalizable to the US white population.
In summary, the incidence rate of JRA in Rochester, Minnesota has decreased in the last decade, most
markedly in the pauciarticular- and systemic-onset subtypes. Fluctuations have been observed in the incidence
PETERSON ET AL
1390
of JRA over the last 3 decades, suggesting that environmental factors may play a rolc in the pathogenesis and
ctiology ol’ JKA. The lower incidence rates reported in
this study, together with the milder cases of JRA observed and fewer extraarticular manifestations identified, suggest that important referral bias may cxist in
previous reports describing the epidemiology of JRA.
ACKNOWLEDGMENTS
The authors wish to thank Mr. Jeffrey Eickholt and Mr.
Jamcs Kuipcr for performing the statistical analyses.
REFERENCES
1. Brewer EJ Jr, Bass J, Baum J, ( idy JT, Fink C, Jacobs J,
Hanson V, Levinson JE, Schallcr J, Stillman JS: Current proposed
revision of JRA criteria. Arthritis Rheum 20:195-199, 1977
2. Oen K, Fast M, Post1 B: Epidemiology of juvenile rheumatoid
arthritis in Manitoba, Canada, 1975-92: cycles in incidence. J
Rheumatol 22:745-750, 1995
3. Gare BA, Fasth A: Epidcmiology of juvcnilc chronic arthritis in
southwestern Sweden: a 5-year prospective population study.
Pediatrics 90:950-958, 1992
4. Hochberg MC, Linet MS, Sills EM: The prevalence and incidence
of juvenile rheumatoid arthritis in an urban black population. Am
J Public Health 73:1202-1203, 1983
5. Gewanter HI,, Roghmann KJ, Baum J: The prevalence of juvenile
arthritis. Arthritis Rheum 26599-603, 1983
6. Prieur AM, Gall EL, Karman F, Edan C, Lasserre 0, Goujard J:
Epidemiologic survey of juvenile chronic arthritis in France:
comparison of data obtained from two different regions. Clin Exp
Rheumatol 5:217-223, 1987
7. Hill R: Juvenile arthritis in various racial groups in British
Columbia. Arthritis Rhcum 20:162, 1977
8. Hicks RM: Rheumatic diseases in Hawaii. Arthritis Rheum 20:
161, 1977
9. Sullivan DB, Cassidy JT, Petty RE: Pathogenic implications of age
of onset in juvcnilc rhcumatoid arthritis. Arthritis Rheum 18:251255, 1975
10. Laaksonen AL: A prognostic study of juvenile rheumatoid arthritis. Acta Pediatr Scand 166:SI-S163, 1966
11. Froom J: Prevalence and natural history of chronic discascs in
childhood. In, Chronic Childhood Illncss: An Assessment of
Outcome. Edited by GD Graune, IB Pless. USDHEW Publication
#(NIH)76-877, 1976
12. Wilder MII: Prevalence of Chronic Skin and Musculoskeletal
Conditions, Unitcd States, 1969. National Health Survey, Series
10, No. 92, USIXIEW, 1974
13. Honham GS: Prevalence of Chronic Skin and Musculoskelctal
Conditions, United States, 1976. National llealth Survey, Series
10, No. 124, USDHEW, 1978
14. Towner SR, Michet CJ Jr, O’Fallon WM, Nelson AM: The
epidemiology of juvcnilc arthritis in Rochester, Minnesota 19601979. Arthritis Rheum 26:1208-1213, 1983
15. Kurland LT, Molgaard CA: The patient record in epidemiology.
Sci Am 245:54-63, 1981
16. McCullagh P, Nelder JA: Generalized Linear Models. New York,
Chapman Hall, 1983
17. Kaplan EL, Meier P: Non parametric estimation from incomplete
observations. J Am Stat Assoc 53:457-481, 1958
18. Kunnamo I, Kallio P, Pelkonen P: Incidence of arthritis in urban
Finnish childrcn: a prospcctivc study. Arthritis Kheum 29:12321238, 1986
19. Bywaters E G L Diagnostic criteria for Still’s disease. In, Population Studies of the Rheumatic Discascs. Bditcd by PM Bennctt,
PM Wood. New York, Excerpta Medica Foundation, 1968
20. Lindsley CB: Seasonal variation in systemic onset juvenile rheumatoid arthritis (letter). Arthritis Rhcum 30:838-839, 1987
21. Fcldman B: Is there a seasonal onset to systemic onset juvenile
rheumatoid arthritis (abstract)? Arthritis Rheum 36 (suppl9):S60,
1993
22. Grahame R, Armstrong R, Simmons N, Wilton JM, Dyson M,
Laurent K,Millis R, Mims CA: Chronic arthritis associated with
the presence of intrasynovial rubella virus. Ann Rheum Dis
42~2-13, 1983
23. Ogra PL, Chiba Y, Ogra SS, Dzierba JL, Herd J K Rubella-virus
infection in juvenile rheumatoid arthritis. Lancet 1:1157-1161,
1975
24. Chantler JK, Tingle AJ,Petty
ersistent rubella virus infection associated with chronic a
in children. N Engl J Med
313:1117-1123, 1985
25. Hyer RH, Gottlieb N L Rheumatic disorders associated with viral
infections. Semin Arthritis Rheum 8:17-31, 1978
26. Pritehard MH, Matthews N, Munro J: Antibodies to influenza A in
a cluster of children with juvenile chronic arthritis. Br J Rheumatol
27~176-180, 1988
27. Phillips PE: Infectious agents in the pathogenesis of rheumatoid
arthritis. Semin Arthritis Rhcum 16:l-10, 1986
28. Rahal JJ, Millian SJ, Noriega ER: Coxsackievirus and adenovirus
infection: association with acute febrile and juvenile rheumatoid
arthritis. JAMA 235:2490-2501, 1976
29. Salvarani C, Gabriel SE, O’Fallon WM, IIunder GG: The incidence of giant cell artcritis in Olmsted County, Minnesota:
apparent fluctuations in a cyclic pattern. Ann Intern Med 123:192194, 1995
30. Leibson CL, Ballard IIJ, Whisnant JP, Melton I J : Thc comprcssion of morbidity hypothesis: promise and pitfalls of using recordlinked data bases to assess secular trends in morbidity and
mortality. Milbank Q 70:127-154, 1992