- Email: [email protected]
Influence of prognostic features on the final outcome in rheumatoid arthritis: A review of the literature
Influence of Prognostic Features on the Final Outcome in Rheumatoid Arthritis: A Review of the Literature By D&sir&e M.F.M.
van der Heijde, Piet L.C.M.
van Riel, Martin
R
HEUMATOID ARTHRITIS (RA) is a systemic disease with chronic, symmetrical polyarthritis as the most prominent clinical feature. The etiology of RA is still unknown. A genetic predisposition is strongly suggested, because the frequency of HLA-DR4 among seropositive RA patients is about twice that of the general population.‘-3 A recent hypothesis is that an (unknown) initiating agent causes immunological reactions and inflammation in the presence of a certain genetic make-up. Environmental, social, psychological, economic, and genetic factors may also play a role in the expression of the disease. Recent studies indicate that in nearly 90% of the patients with erosive disease, these erosions occurred within 2 years of the onset of joint symptoms, 4 indicating that it is important to select in a very early stage those patients who are at risk to develop erosions. These patients are possible candidates for a more aggressive therapy. Little agreement exists in the literature as to which factors in early-stage disease predict an unfavorable final outcome. For the sake of clarity, the prognosis of RA has to be divided into the prognosis in terms of life expectancy and the prognosis as to disability. Survival in RA has been reviewed by Pinals.s We limit our review to the literature concerning the prognosis of the final outcome as measured by clinical and laboratory features and radiological damage. Any study of predictive factors concerning the
From the Department of Rheumatology, University Hospital Nijmegen. The Netherlands. Dbsirbe M.F.M. van der Heijde, MD: Department of Rheumatology: University Hospital Nijmegen; Piet L.C.M. van Riel, MD: Department of Rheumatology; University Hospital Nijmegen; Martin H. van Rijswijk, MD: Professor of Rheumatology, Department of Rheumatology, University Hospital Groningen, The Netherlands; Levinus B.A. van de Putte, MD. Professor of Rheumatology, Department of Rheumatology, University Hospital Nijmergen. Address reprint requests to D.M.F.M. van der Heijde, MD. Department of Rheumatology, University Hospital Nijmegen, Geert Grooteplein Zuid 8. 6525 GA Nijmegen. The Netherlands. 0 1988 by Grune & Stratton. Inc. 0049-0172/88/l 704-0002S5.00/0
284
H. van Rijswijk,
and Levinus B.A. van de Putte
outcome of newly-diagnosed RA patients has to fulfill various criteria: it should be a prospective study, including RA patients early after onset of the disease; follow-up should be performed regularly for a minimum of several years, preferably by the same person; every check-up should include standardized clinical, laboratory, and radiological parameters. Radiological follow-up is essential because it gives a cumulative index of the damage already established, and is not influenced by temporary flare-ups or remissions. Reviewing the literature, we found that no study fulfilled all these criteria. Most studies are performed in retrospect or in a prospective manner with well-established RA. Only a few studies had a follow-up period long enough to warrant conclusions, but their results were based on only a few measurement points during the observation period. Other studies did not include radiological parameters and based their conclusions on clinical or laboratory features only. The heterogeneous study designs and the above-mentioned shortcomings may be the most important reasons for the conflicting results of different studies. Some characteristics, such as age at onset, type of onset, race, sex, and HLA status, are invariable, and therefore it should be easier to draw conclusions from these characteristics than from variable parameters such as rheumatoid factor (RF), erythrocyte sedimentation rate (ESR), or the joint score. A few useful studies of these characteristics have been found. Nevertheless, there is no consensus about the influence of age, type of onset, etc, on the course of the disease. The most important reasons are the variance in outcome parameters and the study duration, as well as the fact that many of these studies are performed in retrospect. Measurements at the real onset of the disease are hardly available; however, the data at the first visit within one year of onset may be regarded as the data of onset because of the chronicity of the disease. Little or nothing is known of the implications on the final outcome of the appearance of clinical and laboratory parameters at the beginning of RA, but the existence of such parameters throughout the course is an
Seminars in Arthritis and Rheumatism, Vol 17, No 4 (May), 1989: pp 294-292
285
PROGNOSTIC FEATURES IN RA: A REVIEW
indicator of an unfavorable outcome. An example is the ESR: an increased ESR at the beginning of RA has no prognostic importance, but patients with a persistently increased ESR have an unfavorable final outcome. Another example is a parameter such as RF, which might be negative at the start but become positive during the course: different measurements in the course of time are needed to draw conclusions on the patients serological status. A pitfall in studies of prognostic factors during the course of the disease on the final outcome is the heterogeneous patient group: patients with varying disease Table 1. Summary of Study Designs of Reports Investigating
Study Duration krl
Follow-Up Frequency (M)
No. of Patients
duration are studied as one group during a defined period. For example, patients with a disease duration of 2 years and 14 years are observed for 3 years and regarded as being in the same period of disease. However, this is incorrect in principle, because RA has no linear course. METHODS We reviewed articles in English since 1940 on the prognosis of the final outcome of RA with regard to disability. The studies were found through a review of the bibliographies of recent studies, review articles, and inquiries from specialists on this subject, and was completed with a Medline computer search in English, French, and German from 1975 to 1987.
Correlations
of Variables
DiseaseDuration (MI
on the Final Outcome
kw (VT)
OutcomeParameters
Reference
P/R
4
P
5
l-3
102
6
P
5
l-2
50
k32
6.0
Clin, rad
7
P
3
1-2
50
30, 4
6.0
Clin, rad
8
P
l-5
141
9
P
10
250
10
P
9
200
11
P
+7
12
Once
3
13
P
12
Once
Rad
2.0 1 Clin k57
2.7
Clin. lab
10 f 6
*50
3.0
Clin
50, 6
1.8
Clin, lab, rad
1.3
Clin
130
k12.5
77
681
14
P
11
100
15
P
102
tl
16
R
4, 5 5-10
17
R
18
P
19
R
20
R
31
P
32
P
33
R
35
P
37
P
1
F/M
in RA
Rad
500 l-2
Clin
154
l/2-25
45
6
54, 4
0.9
Rad
134 > 60
1, 1 + 1, 9
67 k 4
3.0
Clin
47 f
k2.5
+3.0
Clin, lab
7% < 60
1.9
Once
34134
214.5
71140
1.4
Clin
6-9
l-3-8
200
41 + 14
2.2
Clin, lab, rad
20
o-5-10-20
112
+5, 2
k45.2
2.7
Clin, lab, rad
Once
235
Recent
+41
2
Once
72
51
2.0
Clin
l-4
102
tl
49, 2
1.3
Clin
50, 6
1.6
Clin, lab, rad
2.9
5
38
P
4. 5 15
3-l 1-15
100
tl
39
P
5
l-3
102
65
P
1
l-4
54
55
66
P
3
1421142
4. 5
57
10
k2.5
Rad
Rad *37
Clin, lab Clin, rad
67
R
1
1-6
56
5, 7-10
68
P
1
l-2
57
32 < 1
69
P
2-4
1-4
33
cl
35 f 9
1.8
70
P
2
1-4
33
<1
35 + 9
1.8
71
P
Once
46
*8
+55
Once
30190
12
50. 7
3.0
Clin, lab, rad
132
+13
56
2.5
Clin, rad
Lab, rad Clin, rad
t2.5
Clin, lab, rad Clin, lab Lab, rad
73
P
74
R
75
P
2
Once
103
12
61
3.0
Rad
76
P
5
1-4
54
<5
38, 5
3.0
Lab, rad
15
>15
51, 2
3.0
77
P
1
l-2
64
6+6
47
2.2
Clin, lab, rad
10
O-5-10
112
5. 2
45, 2
2.7
Clin, lab, rad
1
Abbreviations: P/R.
prospective/retrostive;
F/M, female/male ratio: clin, clinical; lab, laboratory; rad, radiological.
266
VAN DER HEIJDE ET AL
Studies with a follow-up period of tl year or insufficient patient information were excluded. Table 1 presents a summary of the study designs of the reports on demographic, clinical, and laboratory features used in our review.
with RA onset at an older age than in younger RA controls matched for disease duration. A more favorable outcome of RA at an older age also has been reported in other studies.“23 Some investigators *.+**found little difference in course and prognosis in older patients, or even a more unfavorable outcome.9*13*15*29-32 In one study,i9 48% of the older-onset patients and 72% of the younger-onset patients were seropositive. This tendency of less seropositivity among RA patients with disease onset at an older age also has been established in other studies.*O**’The data on seropositivity in other studies are insufficient. The difference in outcome may be attributable to the fact that older-onset patients suffer from a distinct subset of RA. Masi et al’ noticed that especially patients with a combination of younger age and acute onset showed a more favorable outcome. Various studies failed to show any difference in the course and prognosis of RA between an acute, subacute, or gradual onset.14*20~33-35 Other studies reported a better outcome in patients with an acute onset.8,9,15*27~36 The studies of Feigenbaum et al6 and Fleming et al” showed the influence of malaise and weight loss at onset on the overall prognosis. Patients with these general complaints showed a less favorable outcome. Sherrer et all3 did not find any difference between patients with or without weight loss at onset in terms of overall prognosis. Some investigators6~9~15*31*36,37 have suggested that a particular site or pattern of early
DEMOGRAPHIC AND CLINICAL FEATURES
Table 2 summarizes reports on variables in descriptive factors and disease manifestations. Little is known about different expressions of RA in the various races. A prospective study by Feigenbaum et al6 concluded that white patients had a less favorable outcome. There are many contradictory reports about the relation of sex to the subsequent course. Most studies”‘3 claim that males show a more favorable outcome, although some studies”” found no correlation or even a more unfavorable outcome.‘* Obviously, patient selection seriously influences such studies, and one might expect less sex difference in outcome if one starts with patients who already have established disease. In fact, one report’ concluded that initial attacks of RA occur with practically the same frequency in both sexes, but that the chronic form is more often found in females. A positive family history does not seem to influence the course of RA, as shown in different studies.‘3-‘5 Characteristics at Onset Studies of the prognosis with regard to age at onset are inconclusive. Deal et ali9 found significantly greater clinical improvement in patients Table
Summary of Reports Inveetigating the InRuence of Clinical and Demogrephic Variables on the Final Outcome in RA
2.
Better
Variables
WOrSe
NOlIe 6 (White)
Race 6, 7. 6, 9, 10, 11, 12, 13
Male sex
14. 15, 16, 17
18
13, 14, 15
Family history Older age at onset
19, 20, 21, 22, 23, 24
Younger age at onset
7 (acute) 9
Acute onset
8. 9, 15. 27, 36
25, 26, 27, 28
13, 15, 29, 30, 31, 32
14. 20, 33. 34. 35 6
Malaise at onset Weight loss at onset Site of onset Nodules
13
15
14
6’ 9T$ 15T§ 31$3611
Raynaud-like symptoms
6
Early radiological features
13, 15,31,39
NOTE. Numbers in table refer to referenced studies. *More than two joints of the upper extremity affected. tNumber of joints affected. SSymmetrical involvement. §Metatarsophalangeal
)IHand
and feet.
37s
7. 10, 11, 13, 16, 17, 16,36,38
joints and wrist.
PROGNOSTIC FEATURES IN RA: A REVIEW
287
lation, but also in families, has a positive correlation between RA and DR4 been reported? In Indian immigrants to the United Kingdom, DRl was correlated with RA.42 Recently, an association between the HLA class 2 determinant MC1 on the one hand and DRl, DR4, and RA on the other has been described by Duquesnoy.47 A primary association of RA with this antigen would explain the increase in both DRl and DR4. Moreover, it has been suggested that DR4 is not primarily associated with susceptibility to Clinical and Radiological Features RA, but correlates more with the clinical The existence of subcutaneous nodules is concourse.48*49de Jongh et a148found no difference in incidence of DR4 between RA patients and sidered by many investigators7*10~11*13*16~18~36~38 to non-RA patients in the outpatient Dutch populapredict an unfavorable prognosis. Feigenbaum et tion: in most studies the higher incidence of DR4 al6 found Raynaud-like symptoms in 24% of their is established in hospital populations of RA patients, and this was associated with an unfapatients and therefore biased for more severe vorable outcome. According to some investigators 13*15*31,39 patients with early radiological feaRA. Perhaps, therefore, DR4 is correlated only tures show an unfavorable outcome. This is not with severe RA and not with milder subsets as present in the outpatient population. Young et surprising because erosive disease as such is a a149 found more erosions in patients with DR4 sign of a more aggressive subgroup of RA, and than in RA patients without DR4. Other investithese patients therefore have a less favorable prognosis. gators found only a relation between DR4 and seropositivity and not between DR4 and HLA AND PROGNOSIS seronegativity. S&5’Dequeker et a14’ found a high The human genes concerned with expression incidence of DR4 among RA patients, especially of the histocompatibility antigens are situated on in men under 40 years, but they found no correlathe sixth chromosome. The four principal loci are tion between DR4 and seropositivity, nodules, A, B, C, and D. The gene products, or histocomantinuclear antibody (ANA), or keratoconjuncpatibility antigens A, B, and C, are detected tivitis. The same observations were made by serologically and the D antigen either serologiMaeda et a1.3 Queiros et alSo established a positally (HLA-DR) or by mixed lymphocyte reactive correlation between DR4 and RF, but not tion (HLA-D). In the mouse, genes concerned with regard to radiological erosions, nodules, with the immune response are found in associasicca syndrome, or age at onset. Silman et a153did tion with the H-2 complex (analogue to the HLA not find DR4 a useful predictor of poor outcome complex in man), close to the D/DR locus. at 3 years. On the other hand, Walker et al& Perhaps this can be extrapolated to humans. An reported less favorable clinical outcome in famiassociation has been noted between particular lies with members with DR4 positivity. In this HLA groups and certain diseases. study, homozygotes for DR4 did not differ in outcome from heterozygotes. Brackertz and HLA-DR4 Wernet54 described a rapidly progressive, At the seventh International Histocompatibilextraordinarily severe course of RA in homozyity Workshop, Stastny et al4 reported for the gotes for DR4 in families. first time a relation between the HLA-DR4 HLA-DR2 and HLA-DR7 antigen and RA. This has been confirmed, especially in whites, by many investigators.3p41-““. The While the frequency of DR4 among RA relative risk of developing RA for DR4-positive patients is significantly high as compared individuals as compared with DR4-negative indiwith the population, DR2 is significantly viduals ranges from 2.5 to 6.0 according to low.3*43*“*46,55*56 There is no agreement on the role various investigators. 2*3*4’*45 Not only in the popuof DR2 either: does DR2 play a role in susceptijoint disease is of prognostic significance, but there is no agreement as to which site or pattern. On the other hand, Jacoby et alI4 could not find a relationship between the site of onset and the prognosis. Some studies9*15 have found a relationship between the number of affected joints and overall prognosis: the more joints affected, the worse the final outcome. A few studies9v31*37mention early symmetrical involvement of the joints as a predictor of poor prognosis.
288
VAN DER HEIJDE ET AL
bility to RA or does it protect against an unfavorable outcome? Maeda et al3 could not establish an association between DR2 and RF or clinical outcome. In a study by Panayi et a15’ DR2 was correlated with seronegative RA. Griffin et als7 found a correlation with a more favorable outcome: DR2-positive patients showed less nodules, Sjijgren syndrome, and erosions, and a better outcome in the Ritchie index. These patients also showed a better response to second-line drugs. Less erosions in patients with DR2 were also reported by Young et a1.4gBarger et a156noticed in a group of patients treated with gold a significantly lower frequency of DR2 than in RA patients never treated with gold. They presumed that DR2-positive patients have a better prognosis and are therefore less in need of treatment with gold (and perhaps other second-line drugs?). There are only a few reports about DR7 in relation to RA. Queiros et alSo found a lower frequency of DR7 in RA patients and Barger et als6 noted a lower frequency in RA patients treated with gold. This last observation indicates that DR7 possibly protects against a more aggressive course of RA. Many studies established a correlation between certain HLA determinants (mostly DR3) and the response to and toxicity of secondline drugs. This item is left undiscussed here, but will be considered in detail in another report. HLA-B27
In Northern Sweden and in Finland a significant association between B27 and RA has been Rantapaa et a16’ found signifiobserved.5g*60V61 cantly more radiological sacroiliitis in RA patients with HLA-B27 than in HLA-B27-negative RA patients. The patients with HLA-B27 also showed more subcutaneous nodules and a less favorable functional class. They concluded that the presence of B27 is an indicator of a less favorable final outcome. This has also been stated by other investigators, although there is a difference of opinion about the degree to which HLA-B27 influences the prognosis.62*63 LABORATORY VARIABLES
RF and Prognosis
In Table 3 the laboratory features correlated with the course and prognosis of RA are sum-
Table 3. Summary
of Reports Investigating Correlations of
Laboratory Variables With the Course of RA Variables RF
18, 67, 68
6, 7, 6, 9, 12, 13, 14, 15, 16, 31, 32, 35, 36, 38, 64,’ 65,’ 66, 69, 70 69,70
IgA-RF IgM-RF
69 70
IgG-RF ANA
31
73
APF
74
IgA-IC
66
Clq > 250 pg/L
76 75
Haptoglobulin Ceruloplasmin
75 36, 67, 75
CRP ESR
7. 14. 75
10, 11, 13, 15,31, 32, 67. 77
H8
15, 31, 36
LDH
18
AST
8. 18
Cholesterol
18
NOTE. Numbers in table refer to referenced studies. Abbreviations: ANA,
antinuclear antibody; APF, antiperinu-
clear factor; CRP. C-reactive protein; H8, hemoglobulin: LDH, lactic dehydrogenase. *Combined with DR4.
marized. Since the introduction of the tests for RF, most reports state that the presence of RF is (strongly) predictive of an unfavorable course of the disease. Most investigatorS6-9,12-16,31,32,35,36,38,64-66 agree about an unfavorable final outcome correlated to seropositivity, although different variables are used, eg, the number of American Rheumatism Association (ARA) criteria and Ritchie index,14 total joint count,” and erosive changes.6 The studies of Amos et a1,6’ Dawes et a1,18and Jacoby et all4 revealed no correlation between seropositivity and radiological changes. A few studies mention a correlation between the height of the RF titer and disease activity6* or final outcome.14*35V38 A longitudinal study performed by Westedt et al68 revealed that the RF correlated significantly with clinical indices for joint activity, but had no prognostic value for the final outcome. The authors postulated that the RF is an epiphenomenon and not a pathogenetic factor. The same investigators found IgA-containing immune complexes (IgA-IC) to be a good marker of the development of erosive disease. IgA-IC corre-
PROGNOSTIC FEATURES IN RA: A REVIEW
lated significantly with the appearance of extraarticular features, but not with the joint disease activity. A combined effect of RF and DR4 on the severity of the disease is also suggested by some investigators.64*65 Subtypes of RF
With the further development of the enzymelinked immunoabsorbent assay (ELISA) and the radioimmunoassay technique, it is now possible to measure IgA, IgG, and IgM isotypes of RF. The group of Teitsson and Withrington29*69*70 found that fluctuations in IgA-RF levels correlated significantly with corresponding fluctuations in grip strength, ESR, and a composite index of disease activity. The IgG-RF was associated with changes in ESR and grip strength, but IgM-RF showed only a weak association with fluctuations in ESR and none with any other clinical parameter. In the same prospective study they established that all seven patients who presented with increased IgA-RF developed erosions of the hands and wrists. By contrast, none of the five patients who presented with isolated increased IgM-RF developed erosive disease. These data suggest that detection of IgA-RF in early RA indicates a poor prognosis. Some patients who are conventionally seronegative have high serum IgA-RF values.” Recently, Gioud-Paquet et al’* published a study comparing the Rose-Waaler test with the ELISA test for RF: 70% of patients seronegative in the RoseWaaler test were seropositive in the ELISA test for one or more subtypes of the RF (IgM-, IgA - , IgG -, IgE- RF). Perhaps this is one of the reasons why different investigators disagree on the clinical and radiological pattern of “seronegative” RA: in fact they are comparing heterogeneous groups of patients. Other Laboratory Variables
Many laboratory features are considered to be of prognostic significance by one or a few research groups. ANA was considered to indicate a poor prognosis by Linn et a1,73but this autoantibody did not add any value to a positive RF: RF-ANA - patients showed significantly fewer nodules and less destructive disease than RF-ANA+, RF+ANA-, and RF+ANA+ patients (which groups were identical). Kaarela)’ could not establish a relationship between ANA
and final outcome. Westgeest et a174 found a strong relationship between RF and ANA and no relation between antiperinuclear factor (APF) and ANA. They divided seronegative patients into two different groups, dependent on the APF status: the RF-APF+ patients showed a more severe form of disease as measured by the ARA functional class, and more extra-articular features and radiological destruction than RF-APFpatients. In a report by Sjiiblom et al,” serum haptoglobulin was significantly related to the rate of destruction, whereas serum ceruloplasmin (and other conventional acute-phase proteins) did not show such a relation. In a Japanese study by Ochi et a1,76serum Clq levels turned out to be a prognostic guide to articular erosions. They found no correlation between the Clq level and ESR or C-reactive protein (CRP). However, patients with initial Clq levels >250 pg/mL showed a persistently increased CRP over a period of years and had more erosive changes than patients with Clq levels ~250 pg/mL. The authors suggested that active RA can be classified into two subsets by Clq levels: one with persistent inflammation and a large number of joints with erosions, and another without persistent inflammation and with a small number of joints with erosions. A study by Dawes et all8 on the prediction of progressive joint damage in patients with RA receiving gold or D-penicillamine therapy, showed that patients without radiological destruction had lower serum antistreptolysine titer (AST) levels and serum lactic dehydrogenase (LDH) levels but higher serum cholesterol levels. The study included only a 6-month period to establish the radiological progression. Otten et al* noticed a poor prognosis in seronegative patients with a positive serum AST. ESR, CRP, and Hemoglobin
In various studies, an increase in ESR’0,11*67,77 or CRP36*67*75 is correlated with a less favorable course of the disease. An increased ESR was one of the parameters predicting a poor prognosis according to Fleming et al,” but Masi et al’ and Jacoby et al’* established that the level at an early point in the course did not seem to have prognostic implications. Patients with persistently low hemoglobin (Hb) values had a less
290
VAN DER HEIJDE ET AL
favorable prognosis according to WawrzynskaPagowska et a136 and Sjijblom et a1.75 Most laboratory features such as CRP, ESR, or Hb are just epiphenomena of active disease and in most studies correlate with clinical disease activity: an association with radiological progression is less often established. Whether CRP, ESR, or Hb levels at the beginning of RA have any predictive value for the final outcome remains uncertain due to the lack of prospective studies of recently developed RA over a long follow-up period. DISCUSSION
RA is a heterogeneous disease in which subgroups show different courses. As described above, there have been many attempts to discover predictive features in order to select those subgroups with different prognoses. Most investigators studied patients with well-established RA in a retrospective way. Only a few based their results on a prospective study with newly diagnosed RA patients. This may be the main reason for the contradictory results of the different reports. Moreover, it is difficult to untangle independent correlations. Many studies, for instance, regard subcutaneous nodules as predictive of a poor prognosis; but subcutaneous nodules are strongly associated with a positive RF.7 Nearly all studies indicate a strong correlation between a positive RF and more aggressive disease. The question therefore arises whether there is a real correlation between subcutaneous nodules and prognosis or whether there is only a seeming correlation due to the association of
subcutaneous nodules and RF. On the other hand, DR4 is associated with seropositive RA and a less favorable clinical course: is DR4 the cause of a poor prognosis and RF just one of the measurable parameters? To summarize, only a few factors are known as predictors of an unfavorable clinical outcome in early RA: female sex and a positive RF, especially IgA - RF. During the course of the disease patients with the following characteristics show a more unfavorable clinical course: persistently increased ESR or CRP, decreased Hb, or the appearance of subcutaneous nodules. A prospective study of newly diagnosed RA patients over a long period of time with frequent measurements has to be undertaken to assess the importance of the other characteristics as predictive features in the early course of the disease. With this knowledge patients can be selected to ensure the best risk-benefit ratio of therapy with second-line drugs.
SUMMARY
The literature on prognostic features on the final outcome in RA is reviewed. It is generally agreed that female sex and a positive RF are variables indicating a poor prognosis. Longstanding increased ESR and CRP values, decreased Hb, or the appearance of subcutaneous nodules are indicators of a less favorable clinical course. No conclusions can be drawn regarding other factors due to the incomplete and heterogeneous study designs.
REFERENCES 1. Stockman A, Zilko PJ, Major GAC, et al: Genetic markers in rheumatoid arthritis relationship to toxicity from n-penicillamine. J Rheumatol 13:269-273, 1986 2. Perrier P, Raffoux C, Thomas Ph, et al: HLA antigens and toxic reactions to sodium aurothiopropanol sulphonate and rqe.nicillamine in patients with rheumatoid arthritis. Ann Rheum Dis 44:621-624,1985 3. Maeda H, Juji T, Mitsui H, et al: HLA DR4 and rheumatoid arthritis in Japanese people. Ann Rheum Dis 40:299-302,198l 4. Brook A, Corbett M: Radiographic changes in early rheumatoid disease. Ann Rheum Dis 36:71-73, 1977 5. Pinals RS: Survival in rheumatoid arthritis. Arthritis Rheum 30:473-475, 1987 6. Feigenbaum SL, Masi AT, Kaplan SB: Prognosis in rheumatoid arthritis. A longitudinal study of newly diagnosed younger adult patients. Am J Med 66:377-384, 1979 7. Masi AT, Maldonado-Cocco JA, Kaplan SB, et al:
Prospective study of the early course of rheumatoid arthritis in young adults: Comparison of patients with and without rheumatoid factor positivity at entry and indentification of variables correlating with outcome. Semin Arthritis Rheum 5~299-326, 1976 8. Otten HA, Westendrop Boerma F: Significance of the Waaler-Rose test, streptococcal agglutination, and antistreptolysin titer in the prognosis of rheumatoid arthritis. Ann Rheum Dis 18:24-28,1959 9. Bywaters EGL, Curwen M, Dresner E, et al: Ten-year follow-up of rheumatoid arthritis. Lancet 2: 138 1, 1960 10. Duthie JJR, Brown PE, Truelove LH, et al: Course and prognosis in rheumatoid arthritis. Ann Rheum Dis 23:193-204, 1964 11. Cats A, Hazevoet HM: Significance of positive tests for rheumatoid factor in the prognosis of rheumatoid arthritis. Ann Rheum Dis 29:254-259,1964 12. Hill A, Greenbury CL: Clinical interpretation of
PROGNOSTIC FEATURES IN RA: A REVIEW
serological tests in rheumatoid arthritis, in Dixon ADJ (ed): Progress in Clinical Rheumatism. Boston, Little Brown, 1965, pp 42-55 13. Sherrer YS, Bloch DA, Mitchell DM, et al: The development of disability in rheumatoid arthritis. Arthritis Rheum 28:494-500, 1986 14. Jacoby RK, Jayson MIV, Cosh JA: Onset, early stages, and prognosis of rheumatoid arthritis: A clinical study of 100 patients with II-year follow up. Br Med J 2:96-100, 1973 15. Fleming A, Crown JM, Corbett M: Prognostic value of early features in rheumatoid disease. Br Med J 1:12431245,1976 16. Ragan C, Farrington E: The clinical features of rheumatoid arthritis. JAMA 181:663-667, 1962 17. Sharp JT, Calkins E, Cohen AS, et al: Observations on the clinical, chemical, and serological manifestations of rheumatoid arthritis, based on the course of 154 cases. Medicine (Baltimore) 43:41-58, 1964 18. Dawes PT, Fowler PD, Jackson R, et al: Prediction of progressive joint damage in patients with rheumatoid arthritis receiving gold or D-penicillamine therapy. Ann Rheum Dis 451945949, 1986 19. Deal CL, Meenan RF, Goldenberg DL, et al: The clinical features of elderly-onset rheumatoid arthritis. Arthritis Rheum 28:987-994, 1985 20. Terkeltaub R, Esdaile J, D&cary F, et al: A clinical study of older age rheumatoid arthritis with comparison to a younger onset group. J Rheumatol 10:418-424, 1983 21. Ehrlich GE, Katz WA, Cohen SA: Rheumatoid arthritis in the aged. Geriatrics 25:103-l 13, 1970 22. Adler E: Rheumatoid arthritis in old age. Isr J Med 2:607-613, 1966 23. Isemein L, Redon M: La polyarthrite chronique evolutive apres Page de 55 ans. Rev Rhum Mal Osteoartic 20:877-881, 1953 24. Cecil RL, Kammerer WH: Rheumatoid arthritis in the aged. Am J Med 10:439-445,195l 25. Brown JW, Somes DA: The onset of rheumatoid arthritis in the aged. J Am Geriatr Sot 10:873-881, 1967 26. Dordick JR: Rheumatoid arthritis in the elderly. J Am Geriatr Sot 4:588-591, 1956 27. Moesmann G: Clinical features in subacute rheumatoid arthritis in old age. Acta Rheum Stand 14:285-297, 1968 28. Corrigan AB, Robinson RG, Terenty TR, et al: Benign rheumatoid arthritis of the aged. Br Med J 1:444-446, 1974 29. Schnell A: The clinical features of rheumatic infection in the old. Acta Med Stand 106:345-350, 1941 30. Evers A: Die im Alter auftretende primar chronische polyarthritis. Z Rheumaforsch 24:280-283, 1965 3 1. Kaarela K: Prognostic factors and diagnostic criteria in early rheumatoid arthritis. Stand J Rheumatol 57:1-54, 1985 (suppl) 32. Scott DL, Coulton BL, Symmons DPM, et al: Longterm outcome of treating rheumatoid arthritis: Results after 20yearsLancet 1:1108-1111, 1987 33. Luukkainen R, Isomaki H, Kajander A: Prognostic value of the type of onset of rheumatoid arthritis. Ann Rheum Dis 42:274-275, 1983
291
34. Hart FD: Presentation of rheumatoid arthritis and its relation to prognosis. Br Med J 2:621-624, 1977 35. Reeback J, Silman A: Predictors of outcome at two years in patients with rheumatoid arthritis. J R Sot Med 77:1002-1005, 1984 36. Wawrzynska-Pagowska J, Brzezinska B, Brzozowska M, et al: Observations on the symptoms and signs of “early” rheumatoid arthritis in a prospective study. Acta Rheum Stand 16:99-105, 1970 37. Fleming A, Crown JM, Corbett M: Early rheumatoid disease. I. Onset. Ann Rheum Dis 35:357-360, 1976 38. Rasker JJ, Cosh JA: The natural history of rheumatoid arthritis: A fifteen year follow-up study. The prognostic significance of features noted in the first year. Clin Rheumato1 3:l l-20, 1984 39. Brook A, Fleming A, Corbett M: Relationship of radiological change to clinical outcome in rheumatoid arthritis. Ann Rheum Dis 36:274-275, 1977 40. Stastny P, Batchelor J, Morris P: HLA and disease, in Bodmer WR, Batchelor JR, Bodmer JC, et al (eds): Histocompatibility Testing 1977. Report of the seventh Histocompatibility Workshop and Conference. Copenhagen, Munksgaard, 1978, p 205 4 1. Dequeker J, van Wanghe P, Verdickt W: A systematic survey of HLA-A, B, C, and D antigens and drugs toxicity in rheumatoid arthritis. J Rheumatol 11:282-286, 1984 42. Woodrow JC, Nichol FE, Zaphiropoulos G: DR antigens and rheumatoid arthritis: A study of two populations. Br Med J 283:1287-1288, 1981 43. Speerstra F, Reekers P, van de Putte LBA, et al: HLA-DR antigens and proteinuria induced by aurothioglucase and D-penicillamine in patients with rheumatoid arthritis. J Rheumatol 10:948-953, 1983 44. Panayi GS, Griffin AJ, Wooley PH, et al: Genetic predisposition to gold and o-penicillamine toxicity. Arthritis Rheum 22:645,1979 45. Roitt IM, Brostoff J, Male D: Immunology. London, Gower Medical, 1985, pp 23-24 46. Walker DJ, Griffiths M, Dewar P, et al: Association of MHC antigens with susceptibility to and severity of rheumatoid arthritis in multicase families. Ann Rheum Dis 44:519525,1985 47. Duquesnoy RJ, Marrari M, Hackbarlh S, et al: Serological and cellular definition of a new HLA-DR associated determinant, MCl, and its association with rheumatoid arthritis. Hum Immunol 10:165-176, 1984 48. de Jongh BM, van Romunde LKJ, Valkenburg HA, et al: Epidemiological study of HLA and GM in rheumatoid arthritis and related symptoms in an open Dutch population. Ann Rheum Dis 43:613-619, 1984 49. Young A, Jaraquemada D, Awad J, et al: Association of HLA-DR4/Dw4 and DR2/Dw2 with radiologic changes in a prospective study of patients with rheumatoid arthritis. Arthritis Rheum 27:20-25, 1984 50. Queiros MV, Sancho MRH, Caetano JM: HLA-DR4 antigen and IgM rheumatoid factors. J Rheumatol 9:370373,1982 51. Dobloug JH, Forre 0, Kass E, et al: HLA antigens and rheumatoid arthritis. Arthritis Rheum 23:309-313, 1980 52. Walton K, Dyer PA, Grennan DM, et al: Clinical
292
features, autoantibodies and HLA-DR antigens in rheumatoid arthritis. J Rheumatol 12:223-226, 1985 53. Silman AJ, Reeback J, Jaraquemada D: HLA-DR4 as a predictor of outcome three years after onset of rheumatoid arthritis. Rheumatol Int 6:223-235, 1986 54. Brackertz D, Wernet P: A hereditary basis for rheumatoid arthritis. HLA-D/DR-alloantigens and disease severity: Population and family studies, in Paulus HE, Erlich GE, Linenlaub E (eds): Controversies in the Clinical Evaluation of Analgesic-Anti-inflammatory-Antirheumatic Drugs. Stuttgart, Schattauer, 1981, pp 453-462 55. Panayi GS, Wooley P, Batchelor JR Genetic basis of rheumatoid disease: HLA antigens, disease manifestations, and toxic reactions to drugs. Br Med J 2:1326-1328, 1978 56. Barger BO, Acton RT, Koopman WJ, et al: DR antigens and gold toxicity in white rheumatoid arthritis patients. Arthritis Rheum 27:601-605, 1984 57. Griffin AJ, Wooley P, Panayi GS, et al: HLA DR antigens and disease expression in rheumatoid arthritis. Ann Rheum Dis 43:218-221,1984 58. Rantapaa Dahlqvist S, Strom H, Bjelle A, et al: HLA antigens in rheumatoid arthritis patients with and without a family history of polyarthritis. Stand J Rheumatol 14:375380,198s 59. Bjelle A, Cedergren B, Rantapaa Dahlqvist S: HLAB27 in the population of northern Sweden. Stand J Rheumato1 11:23, 1982 60. Isomaki H, Koota K, Nissila MJ, et al: HLA-B27 and arthritis. Ann Clin Res 7:138, 1975 61. Rantapaa Dahlqvist S, Nordmark LG, Bjelle A: HLA-B27 and involvement of sacroiliac joints in rheumatoid arthritis. J Rheumatol 11:27-32,1984 62. Elhabali M, Scherak 0, Seidl G, et al: Tomographic examinations of sacroiliac joints in adult patients with rheumatoid arthritis. J Rheumatol6:417-425, 1979 63. Martel W, Duff IF: Pelvo-spondylitis in rheumatoid arthritis. Radiology 77:744-756, 1961 64. Roitt IM, Corbett M, Festenstein, et al: HLA-DRW4 and prognosis in rheumatoid arthritis. Lancet 1:990, 1978 65. G’Duffy JD, O’Fallon WM, Hunder GG, et al: An attempt to predict the response to gold therapy in rheumatoid arthritis. Arthritis Rheum 27:1210-1217, 1984
VAN DER HEIJDE ET AL
66. Isomaki H, Martio J, Sarna S, et al: Predicting the outcome of rheumatoid arthritis. Stand J Rheumatol 13:3338,1984 67. Amos RS, Constable TJ, Crockson RA, et al: Rheumatoid arthritis: Relation of serum C-reactive protein and erythrocyte sedimentation rates to radiographic changes. Br Med J 1:195-197, 1977 68. Westedt ML, Daha MR, Baldwin WM III, et al: Serum immune complexes containing IgA appear to predict erosive arthritis in a longitudinal study in rheumatoid arthritis. Ann Rheum Dis 45:809-815, 1986 69. Teitsson I, Withrington RH, Seifert MH et al: Prospective study of early rheumatoid arthritis. I. Prognostic value of IgA rheumatoid factor. Ann Rheum Dis 43:673-678, 1984 70. Withrington RH, Teitsson I, Valdimarsson H, et al: Prospective study of early rheumatoid arthritis. II. Association of rheumatoid factor isotypes with fluctuations in disease activity. Ann Rheum Dis 43:679-685, 1984 71. Withrington RH, Seifert MH: Predominant wrist disease in rheumatoid arthritis associated with high concentration of IgA rheumatoid factor. Br Med J 291:1388, 1985 72. Gioud-Paquet M, Auvinet M, Raffin T, et al: IgM rheumatoid factor (RF), IgA RF, IgE RF, and IgG RF detected by ELISA in rheumatoid arthritis. Ann Rheum Dis 46:65-71, 1987 73. Linn JE, Hardin JG, Halla JT: A controlled study of ANA+RFarthritis. Arthritis Rheum 21:645-651, 1978 74. Westgeest AAA, Boerbooms AMTh, Jongmans M, et al: Antiperinuclear factor: Indicator of more severe disease in seronegative rheumatoid arthritis. J Rheumatol 14:893-897, 1987 75. Sjoblom KG, Saxne T, Pettersson H, et al: Factors related to the progression of joint destruction in rheumatoid arthritis. Stand J Rheumatol 13:21-27,1984 76. Ochi T, Yonemasu K, Iwase R, et al: Serum Clq levels as a prognostic guide to articular erosions in patients with rheumatoid arthritis. Arthritis Rheum 27:883-887, 1984 77. Scott DL, Grindulis KA, Struthers GR et al: Progression of radiological changes in rheumatoid arthritis. Ann Rheum Dis 43:8-17, 1984
van der Heijde, Piet L.C.M.
van Riel, Martin
R
HEUMATOID ARTHRITIS (RA) is a systemic disease with chronic, symmetrical polyarthritis as the most prominent clinical feature. The etiology of RA is still unknown. A genetic predisposition is strongly suggested, because the frequency of HLA-DR4 among seropositive RA patients is about twice that of the general population.‘-3 A recent hypothesis is that an (unknown) initiating agent causes immunological reactions and inflammation in the presence of a certain genetic make-up. Environmental, social, psychological, economic, and genetic factors may also play a role in the expression of the disease. Recent studies indicate that in nearly 90% of the patients with erosive disease, these erosions occurred within 2 years of the onset of joint symptoms, 4 indicating that it is important to select in a very early stage those patients who are at risk to develop erosions. These patients are possible candidates for a more aggressive therapy. Little agreement exists in the literature as to which factors in early-stage disease predict an unfavorable final outcome. For the sake of clarity, the prognosis of RA has to be divided into the prognosis in terms of life expectancy and the prognosis as to disability. Survival in RA has been reviewed by Pinals.s We limit our review to the literature concerning the prognosis of the final outcome as measured by clinical and laboratory features and radiological damage. Any study of predictive factors concerning the
From the Department of Rheumatology, University Hospital Nijmegen. The Netherlands. Dbsirbe M.F.M. van der Heijde, MD: Department of Rheumatology: University Hospital Nijmegen; Piet L.C.M. van Riel, MD: Department of Rheumatology; University Hospital Nijmegen; Martin H. van Rijswijk, MD: Professor of Rheumatology, Department of Rheumatology, University Hospital Groningen, The Netherlands; Levinus B.A. van de Putte, MD. Professor of Rheumatology, Department of Rheumatology, University Hospital Nijmergen. Address reprint requests to D.M.F.M. van der Heijde, MD. Department of Rheumatology, University Hospital Nijmegen, Geert Grooteplein Zuid 8. 6525 GA Nijmegen. The Netherlands. 0 1988 by Grune & Stratton. Inc. 0049-0172/88/l 704-0002S5.00/0
284
H. van Rijswijk,
and Levinus B.A. van de Putte
outcome of newly-diagnosed RA patients has to fulfill various criteria: it should be a prospective study, including RA patients early after onset of the disease; follow-up should be performed regularly for a minimum of several years, preferably by the same person; every check-up should include standardized clinical, laboratory, and radiological parameters. Radiological follow-up is essential because it gives a cumulative index of the damage already established, and is not influenced by temporary flare-ups or remissions. Reviewing the literature, we found that no study fulfilled all these criteria. Most studies are performed in retrospect or in a prospective manner with well-established RA. Only a few studies had a follow-up period long enough to warrant conclusions, but their results were based on only a few measurement points during the observation period. Other studies did not include radiological parameters and based their conclusions on clinical or laboratory features only. The heterogeneous study designs and the above-mentioned shortcomings may be the most important reasons for the conflicting results of different studies. Some characteristics, such as age at onset, type of onset, race, sex, and HLA status, are invariable, and therefore it should be easier to draw conclusions from these characteristics than from variable parameters such as rheumatoid factor (RF), erythrocyte sedimentation rate (ESR), or the joint score. A few useful studies of these characteristics have been found. Nevertheless, there is no consensus about the influence of age, type of onset, etc, on the course of the disease. The most important reasons are the variance in outcome parameters and the study duration, as well as the fact that many of these studies are performed in retrospect. Measurements at the real onset of the disease are hardly available; however, the data at the first visit within one year of onset may be regarded as the data of onset because of the chronicity of the disease. Little or nothing is known of the implications on the final outcome of the appearance of clinical and laboratory parameters at the beginning of RA, but the existence of such parameters throughout the course is an
Seminars in Arthritis and Rheumatism, Vol 17, No 4 (May), 1989: pp 294-292
285
PROGNOSTIC FEATURES IN RA: A REVIEW
indicator of an unfavorable outcome. An example is the ESR: an increased ESR at the beginning of RA has no prognostic importance, but patients with a persistently increased ESR have an unfavorable final outcome. Another example is a parameter such as RF, which might be negative at the start but become positive during the course: different measurements in the course of time are needed to draw conclusions on the patients serological status. A pitfall in studies of prognostic factors during the course of the disease on the final outcome is the heterogeneous patient group: patients with varying disease Table 1. Summary of Study Designs of Reports Investigating
Study Duration krl
Follow-Up Frequency (M)
No. of Patients
duration are studied as one group during a defined period. For example, patients with a disease duration of 2 years and 14 years are observed for 3 years and regarded as being in the same period of disease. However, this is incorrect in principle, because RA has no linear course. METHODS We reviewed articles in English since 1940 on the prognosis of the final outcome of RA with regard to disability. The studies were found through a review of the bibliographies of recent studies, review articles, and inquiries from specialists on this subject, and was completed with a Medline computer search in English, French, and German from 1975 to 1987.
Correlations
of Variables
DiseaseDuration (MI
on the Final Outcome
kw (VT)
OutcomeParameters
Reference
P/R
4
P
5
l-3
102
6
P
5
l-2
50
k32
6.0
Clin, rad
7
P
3
1-2
50
30, 4
6.0
Clin, rad
8
P
l-5
141
9
P
10
250
10
P
9
200
11
P
+7
12
Once
3
13
P
12
Once
Rad
2.0 1 Clin k57
2.7
Clin. lab
10 f 6
*50
3.0
Clin
50, 6
1.8
Clin, lab, rad
1.3
Clin
130
k12.5
77
681
14
P
11
100
15
P
102
tl
16
R
4, 5 5-10
17
R
18
P
19
R
20
R
31
P
32
P
33
R
35
P
37
P
1
F/M
in RA
Rad
500 l-2
Clin
154
l/2-25
45
6
54, 4
0.9
Rad
134 > 60
1, 1 + 1, 9
67 k 4
3.0
Clin
47 f
k2.5
+3.0
Clin, lab
7% < 60
1.9
Once
34134
214.5
71140
1.4
Clin
6-9
l-3-8
200
41 + 14
2.2
Clin, lab, rad
20
o-5-10-20
112
+5, 2
k45.2
2.7
Clin, lab, rad
Once
235
Recent
+41
2
Once
72
51
2.0
Clin
l-4
102
tl
49, 2
1.3
Clin
50, 6
1.6
Clin, lab, rad
2.9
5
38
P
4. 5 15
3-l 1-15
100
tl
39
P
5
l-3
102
65
P
1
l-4
54
55
66
P
3
1421142
4. 5
57
10
k2.5
Rad
Rad *37
Clin, lab Clin, rad
67
R
1
1-6
56
5, 7-10
68
P
1
l-2
57
32 < 1
69
P
2-4
1-4
33
cl
35 f 9
1.8
70
P
2
1-4
33
<1
35 + 9
1.8
71
P
Once
46
*8
+55
Once
30190
12
50. 7
3.0
Clin, lab, rad
132
+13
56
2.5
Clin, rad
Lab, rad Clin, rad
t2.5
Clin, lab, rad Clin, lab Lab, rad
73
P
74
R
75
P
2
Once
103
12
61
3.0
Rad
76
P
5
1-4
54
<5
38, 5
3.0
Lab, rad
15
>15
51, 2
3.0
77
P
1
l-2
64
6+6
47
2.2
Clin, lab, rad
10
O-5-10
112
5. 2
45, 2
2.7
Clin, lab, rad
1
Abbreviations: P/R.
prospective/retrostive;
F/M, female/male ratio: clin, clinical; lab, laboratory; rad, radiological.
266
VAN DER HEIJDE ET AL
Studies with a follow-up period of tl year or insufficient patient information were excluded. Table 1 presents a summary of the study designs of the reports on demographic, clinical, and laboratory features used in our review.
with RA onset at an older age than in younger RA controls matched for disease duration. A more favorable outcome of RA at an older age also has been reported in other studies.“23 Some investigators *.+**found little difference in course and prognosis in older patients, or even a more unfavorable outcome.9*13*15*29-32 In one study,i9 48% of the older-onset patients and 72% of the younger-onset patients were seropositive. This tendency of less seropositivity among RA patients with disease onset at an older age also has been established in other studies.*O**’The data on seropositivity in other studies are insufficient. The difference in outcome may be attributable to the fact that older-onset patients suffer from a distinct subset of RA. Masi et al’ noticed that especially patients with a combination of younger age and acute onset showed a more favorable outcome. Various studies failed to show any difference in the course and prognosis of RA between an acute, subacute, or gradual onset.14*20~33-35 Other studies reported a better outcome in patients with an acute onset.8,9,15*27~36 The studies of Feigenbaum et al6 and Fleming et al” showed the influence of malaise and weight loss at onset on the overall prognosis. Patients with these general complaints showed a less favorable outcome. Sherrer et all3 did not find any difference between patients with or without weight loss at onset in terms of overall prognosis. Some investigators6~9~15*31*36,37 have suggested that a particular site or pattern of early
DEMOGRAPHIC AND CLINICAL FEATURES
Table 2 summarizes reports on variables in descriptive factors and disease manifestations. Little is known about different expressions of RA in the various races. A prospective study by Feigenbaum et al6 concluded that white patients had a less favorable outcome. There are many contradictory reports about the relation of sex to the subsequent course. Most studies”‘3 claim that males show a more favorable outcome, although some studies”” found no correlation or even a more unfavorable outcome.‘* Obviously, patient selection seriously influences such studies, and one might expect less sex difference in outcome if one starts with patients who already have established disease. In fact, one report’ concluded that initial attacks of RA occur with practically the same frequency in both sexes, but that the chronic form is more often found in females. A positive family history does not seem to influence the course of RA, as shown in different studies.‘3-‘5 Characteristics at Onset Studies of the prognosis with regard to age at onset are inconclusive. Deal et ali9 found significantly greater clinical improvement in patients Table
Summary of Reports Inveetigating the InRuence of Clinical and Demogrephic Variables on the Final Outcome in RA
2.
Better
Variables
WOrSe
NOlIe 6 (White)
Race 6, 7. 6, 9, 10, 11, 12, 13
Male sex
14. 15, 16, 17
18
13, 14, 15
Family history Older age at onset
19, 20, 21, 22, 23, 24
Younger age at onset
7 (acute) 9
Acute onset
8. 9, 15. 27, 36
25, 26, 27, 28
13, 15, 29, 30, 31, 32
14. 20, 33. 34. 35 6
Malaise at onset Weight loss at onset Site of onset Nodules
13
15
14
6’ 9T$ 15T§ 31$3611
Raynaud-like symptoms
6
Early radiological features
13, 15,31,39
NOTE. Numbers in table refer to referenced studies. *More than two joints of the upper extremity affected. tNumber of joints affected. SSymmetrical involvement. §Metatarsophalangeal
)IHand
and feet.
37s
7. 10, 11, 13, 16, 17, 16,36,38
joints and wrist.
PROGNOSTIC FEATURES IN RA: A REVIEW
287
lation, but also in families, has a positive correlation between RA and DR4 been reported? In Indian immigrants to the United Kingdom, DRl was correlated with RA.42 Recently, an association between the HLA class 2 determinant MC1 on the one hand and DRl, DR4, and RA on the other has been described by Duquesnoy.47 A primary association of RA with this antigen would explain the increase in both DRl and DR4. Moreover, it has been suggested that DR4 is not primarily associated with susceptibility to Clinical and Radiological Features RA, but correlates more with the clinical The existence of subcutaneous nodules is concourse.48*49de Jongh et a148found no difference in incidence of DR4 between RA patients and sidered by many investigators7*10~11*13*16~18~36~38 to non-RA patients in the outpatient Dutch populapredict an unfavorable prognosis. Feigenbaum et tion: in most studies the higher incidence of DR4 al6 found Raynaud-like symptoms in 24% of their is established in hospital populations of RA patients, and this was associated with an unfapatients and therefore biased for more severe vorable outcome. According to some investigators 13*15*31,39 patients with early radiological feaRA. Perhaps, therefore, DR4 is correlated only tures show an unfavorable outcome. This is not with severe RA and not with milder subsets as present in the outpatient population. Young et surprising because erosive disease as such is a a149 found more erosions in patients with DR4 sign of a more aggressive subgroup of RA, and than in RA patients without DR4. Other investithese patients therefore have a less favorable prognosis. gators found only a relation between DR4 and seropositivity and not between DR4 and HLA AND PROGNOSIS seronegativity. S&5’Dequeker et a14’ found a high The human genes concerned with expression incidence of DR4 among RA patients, especially of the histocompatibility antigens are situated on in men under 40 years, but they found no correlathe sixth chromosome. The four principal loci are tion between DR4 and seropositivity, nodules, A, B, C, and D. The gene products, or histocomantinuclear antibody (ANA), or keratoconjuncpatibility antigens A, B, and C, are detected tivitis. The same observations were made by serologically and the D antigen either serologiMaeda et a1.3 Queiros et alSo established a positally (HLA-DR) or by mixed lymphocyte reactive correlation between DR4 and RF, but not tion (HLA-D). In the mouse, genes concerned with regard to radiological erosions, nodules, with the immune response are found in associasicca syndrome, or age at onset. Silman et a153did tion with the H-2 complex (analogue to the HLA not find DR4 a useful predictor of poor outcome complex in man), close to the D/DR locus. at 3 years. On the other hand, Walker et al& Perhaps this can be extrapolated to humans. An reported less favorable clinical outcome in famiassociation has been noted between particular lies with members with DR4 positivity. In this HLA groups and certain diseases. study, homozygotes for DR4 did not differ in outcome from heterozygotes. Brackertz and HLA-DR4 Wernet54 described a rapidly progressive, At the seventh International Histocompatibilextraordinarily severe course of RA in homozyity Workshop, Stastny et al4 reported for the gotes for DR4 in families. first time a relation between the HLA-DR4 HLA-DR2 and HLA-DR7 antigen and RA. This has been confirmed, especially in whites, by many investigators.3p41-““. The While the frequency of DR4 among RA relative risk of developing RA for DR4-positive patients is significantly high as compared individuals as compared with DR4-negative indiwith the population, DR2 is significantly viduals ranges from 2.5 to 6.0 according to low.3*43*“*46,55*56 There is no agreement on the role various investigators. 2*3*4’*45 Not only in the popuof DR2 either: does DR2 play a role in susceptijoint disease is of prognostic significance, but there is no agreement as to which site or pattern. On the other hand, Jacoby et alI4 could not find a relationship between the site of onset and the prognosis. Some studies9*15 have found a relationship between the number of affected joints and overall prognosis: the more joints affected, the worse the final outcome. A few studies9v31*37mention early symmetrical involvement of the joints as a predictor of poor prognosis.
288
VAN DER HEIJDE ET AL
bility to RA or does it protect against an unfavorable outcome? Maeda et al3 could not establish an association between DR2 and RF or clinical outcome. In a study by Panayi et a15’ DR2 was correlated with seronegative RA. Griffin et als7 found a correlation with a more favorable outcome: DR2-positive patients showed less nodules, Sjijgren syndrome, and erosions, and a better outcome in the Ritchie index. These patients also showed a better response to second-line drugs. Less erosions in patients with DR2 were also reported by Young et a1.4gBarger et a156noticed in a group of patients treated with gold a significantly lower frequency of DR2 than in RA patients never treated with gold. They presumed that DR2-positive patients have a better prognosis and are therefore less in need of treatment with gold (and perhaps other second-line drugs?). There are only a few reports about DR7 in relation to RA. Queiros et alSo found a lower frequency of DR7 in RA patients and Barger et als6 noted a lower frequency in RA patients treated with gold. This last observation indicates that DR7 possibly protects against a more aggressive course of RA. Many studies established a correlation between certain HLA determinants (mostly DR3) and the response to and toxicity of secondline drugs. This item is left undiscussed here, but will be considered in detail in another report. HLA-B27
In Northern Sweden and in Finland a significant association between B27 and RA has been Rantapaa et a16’ found signifiobserved.5g*60V61 cantly more radiological sacroiliitis in RA patients with HLA-B27 than in HLA-B27-negative RA patients. The patients with HLA-B27 also showed more subcutaneous nodules and a less favorable functional class. They concluded that the presence of B27 is an indicator of a less favorable final outcome. This has also been stated by other investigators, although there is a difference of opinion about the degree to which HLA-B27 influences the prognosis.62*63 LABORATORY VARIABLES
RF and Prognosis
In Table 3 the laboratory features correlated with the course and prognosis of RA are sum-
Table 3. Summary
of Reports Investigating Correlations of
Laboratory Variables With the Course of RA Variables RF
18, 67, 68
6, 7, 6, 9, 12, 13, 14, 15, 16, 31, 32, 35, 36, 38, 64,’ 65,’ 66, 69, 70 69,70
IgA-RF IgM-RF
69 70
IgG-RF ANA
31
73
APF
74
IgA-IC
66
Clq > 250 pg/L
76 75
Haptoglobulin Ceruloplasmin
75 36, 67, 75
CRP ESR
7. 14. 75
10, 11, 13, 15,31, 32, 67. 77
H8
15, 31, 36
LDH
18
AST
8. 18
Cholesterol
18
NOTE. Numbers in table refer to referenced studies. Abbreviations: ANA,
antinuclear antibody; APF, antiperinu-
clear factor; CRP. C-reactive protein; H8, hemoglobulin: LDH, lactic dehydrogenase. *Combined with DR4.
marized. Since the introduction of the tests for RF, most reports state that the presence of RF is (strongly) predictive of an unfavorable course of the disease. Most investigatorS6-9,12-16,31,32,35,36,38,64-66 agree about an unfavorable final outcome correlated to seropositivity, although different variables are used, eg, the number of American Rheumatism Association (ARA) criteria and Ritchie index,14 total joint count,” and erosive changes.6 The studies of Amos et a1,6’ Dawes et a1,18and Jacoby et all4 revealed no correlation between seropositivity and radiological changes. A few studies mention a correlation between the height of the RF titer and disease activity6* or final outcome.14*35V38 A longitudinal study performed by Westedt et al68 revealed that the RF correlated significantly with clinical indices for joint activity, but had no prognostic value for the final outcome. The authors postulated that the RF is an epiphenomenon and not a pathogenetic factor. The same investigators found IgA-containing immune complexes (IgA-IC) to be a good marker of the development of erosive disease. IgA-IC corre-
PROGNOSTIC FEATURES IN RA: A REVIEW
lated significantly with the appearance of extraarticular features, but not with the joint disease activity. A combined effect of RF and DR4 on the severity of the disease is also suggested by some investigators.64*65 Subtypes of RF
With the further development of the enzymelinked immunoabsorbent assay (ELISA) and the radioimmunoassay technique, it is now possible to measure IgA, IgG, and IgM isotypes of RF. The group of Teitsson and Withrington29*69*70 found that fluctuations in IgA-RF levels correlated significantly with corresponding fluctuations in grip strength, ESR, and a composite index of disease activity. The IgG-RF was associated with changes in ESR and grip strength, but IgM-RF showed only a weak association with fluctuations in ESR and none with any other clinical parameter. In the same prospective study they established that all seven patients who presented with increased IgA-RF developed erosions of the hands and wrists. By contrast, none of the five patients who presented with isolated increased IgM-RF developed erosive disease. These data suggest that detection of IgA-RF in early RA indicates a poor prognosis. Some patients who are conventionally seronegative have high serum IgA-RF values.” Recently, Gioud-Paquet et al’* published a study comparing the Rose-Waaler test with the ELISA test for RF: 70% of patients seronegative in the RoseWaaler test were seropositive in the ELISA test for one or more subtypes of the RF (IgM-, IgA - , IgG -, IgE- RF). Perhaps this is one of the reasons why different investigators disagree on the clinical and radiological pattern of “seronegative” RA: in fact they are comparing heterogeneous groups of patients. Other Laboratory Variables
Many laboratory features are considered to be of prognostic significance by one or a few research groups. ANA was considered to indicate a poor prognosis by Linn et a1,73but this autoantibody did not add any value to a positive RF: RF-ANA - patients showed significantly fewer nodules and less destructive disease than RF-ANA+, RF+ANA-, and RF+ANA+ patients (which groups were identical). Kaarela)’ could not establish a relationship between ANA
and final outcome. Westgeest et a174 found a strong relationship between RF and ANA and no relation between antiperinuclear factor (APF) and ANA. They divided seronegative patients into two different groups, dependent on the APF status: the RF-APF+ patients showed a more severe form of disease as measured by the ARA functional class, and more extra-articular features and radiological destruction than RF-APFpatients. In a report by Sjiiblom et al,” serum haptoglobulin was significantly related to the rate of destruction, whereas serum ceruloplasmin (and other conventional acute-phase proteins) did not show such a relation. In a Japanese study by Ochi et a1,76serum Clq levels turned out to be a prognostic guide to articular erosions. They found no correlation between the Clq level and ESR or C-reactive protein (CRP). However, patients with initial Clq levels >250 pg/mL showed a persistently increased CRP over a period of years and had more erosive changes than patients with Clq levels ~250 pg/mL. The authors suggested that active RA can be classified into two subsets by Clq levels: one with persistent inflammation and a large number of joints with erosions, and another without persistent inflammation and with a small number of joints with erosions. A study by Dawes et all8 on the prediction of progressive joint damage in patients with RA receiving gold or D-penicillamine therapy, showed that patients without radiological destruction had lower serum antistreptolysine titer (AST) levels and serum lactic dehydrogenase (LDH) levels but higher serum cholesterol levels. The study included only a 6-month period to establish the radiological progression. Otten et al* noticed a poor prognosis in seronegative patients with a positive serum AST. ESR, CRP, and Hemoglobin
In various studies, an increase in ESR’0,11*67,77 or CRP36*67*75 is correlated with a less favorable course of the disease. An increased ESR was one of the parameters predicting a poor prognosis according to Fleming et al,” but Masi et al’ and Jacoby et al’* established that the level at an early point in the course did not seem to have prognostic implications. Patients with persistently low hemoglobin (Hb) values had a less
290
VAN DER HEIJDE ET AL
favorable prognosis according to WawrzynskaPagowska et a136 and Sjijblom et a1.75 Most laboratory features such as CRP, ESR, or Hb are just epiphenomena of active disease and in most studies correlate with clinical disease activity: an association with radiological progression is less often established. Whether CRP, ESR, or Hb levels at the beginning of RA have any predictive value for the final outcome remains uncertain due to the lack of prospective studies of recently developed RA over a long follow-up period. DISCUSSION
RA is a heterogeneous disease in which subgroups show different courses. As described above, there have been many attempts to discover predictive features in order to select those subgroups with different prognoses. Most investigators studied patients with well-established RA in a retrospective way. Only a few based their results on a prospective study with newly diagnosed RA patients. This may be the main reason for the contradictory results of the different reports. Moreover, it is difficult to untangle independent correlations. Many studies, for instance, regard subcutaneous nodules as predictive of a poor prognosis; but subcutaneous nodules are strongly associated with a positive RF.7 Nearly all studies indicate a strong correlation between a positive RF and more aggressive disease. The question therefore arises whether there is a real correlation between subcutaneous nodules and prognosis or whether there is only a seeming correlation due to the association of
subcutaneous nodules and RF. On the other hand, DR4 is associated with seropositive RA and a less favorable clinical course: is DR4 the cause of a poor prognosis and RF just one of the measurable parameters? To summarize, only a few factors are known as predictors of an unfavorable clinical outcome in early RA: female sex and a positive RF, especially IgA - RF. During the course of the disease patients with the following characteristics show a more unfavorable clinical course: persistently increased ESR or CRP, decreased Hb, or the appearance of subcutaneous nodules. A prospective study of newly diagnosed RA patients over a long period of time with frequent measurements has to be undertaken to assess the importance of the other characteristics as predictive features in the early course of the disease. With this knowledge patients can be selected to ensure the best risk-benefit ratio of therapy with second-line drugs.
SUMMARY
The literature on prognostic features on the final outcome in RA is reviewed. It is generally agreed that female sex and a positive RF are variables indicating a poor prognosis. Longstanding increased ESR and CRP values, decreased Hb, or the appearance of subcutaneous nodules are indicators of a less favorable clinical course. No conclusions can be drawn regarding other factors due to the incomplete and heterogeneous study designs.
REFERENCES 1. Stockman A, Zilko PJ, Major GAC, et al: Genetic markers in rheumatoid arthritis relationship to toxicity from n-penicillamine. J Rheumatol 13:269-273, 1986 2. Perrier P, Raffoux C, Thomas Ph, et al: HLA antigens and toxic reactions to sodium aurothiopropanol sulphonate and rqe.nicillamine in patients with rheumatoid arthritis. Ann Rheum Dis 44:621-624,1985 3. Maeda H, Juji T, Mitsui H, et al: HLA DR4 and rheumatoid arthritis in Japanese people. Ann Rheum Dis 40:299-302,198l 4. Brook A, Corbett M: Radiographic changes in early rheumatoid disease. Ann Rheum Dis 36:71-73, 1977 5. Pinals RS: Survival in rheumatoid arthritis. Arthritis Rheum 30:473-475, 1987 6. Feigenbaum SL, Masi AT, Kaplan SB: Prognosis in rheumatoid arthritis. A longitudinal study of newly diagnosed younger adult patients. Am J Med 66:377-384, 1979 7. Masi AT, Maldonado-Cocco JA, Kaplan SB, et al:
Prospective study of the early course of rheumatoid arthritis in young adults: Comparison of patients with and without rheumatoid factor positivity at entry and indentification of variables correlating with outcome. Semin Arthritis Rheum 5~299-326, 1976 8. Otten HA, Westendrop Boerma F: Significance of the Waaler-Rose test, streptococcal agglutination, and antistreptolysin titer in the prognosis of rheumatoid arthritis. Ann Rheum Dis 18:24-28,1959 9. Bywaters EGL, Curwen M, Dresner E, et al: Ten-year follow-up of rheumatoid arthritis. Lancet 2: 138 1, 1960 10. Duthie JJR, Brown PE, Truelove LH, et al: Course and prognosis in rheumatoid arthritis. Ann Rheum Dis 23:193-204, 1964 11. Cats A, Hazevoet HM: Significance of positive tests for rheumatoid factor in the prognosis of rheumatoid arthritis. Ann Rheum Dis 29:254-259,1964 12. Hill A, Greenbury CL: Clinical interpretation of
PROGNOSTIC FEATURES IN RA: A REVIEW
serological tests in rheumatoid arthritis, in Dixon ADJ (ed): Progress in Clinical Rheumatism. Boston, Little Brown, 1965, pp 42-55 13. Sherrer YS, Bloch DA, Mitchell DM, et al: The development of disability in rheumatoid arthritis. Arthritis Rheum 28:494-500, 1986 14. Jacoby RK, Jayson MIV, Cosh JA: Onset, early stages, and prognosis of rheumatoid arthritis: A clinical study of 100 patients with II-year follow up. Br Med J 2:96-100, 1973 15. Fleming A, Crown JM, Corbett M: Prognostic value of early features in rheumatoid disease. Br Med J 1:12431245,1976 16. Ragan C, Farrington E: The clinical features of rheumatoid arthritis. JAMA 181:663-667, 1962 17. Sharp JT, Calkins E, Cohen AS, et al: Observations on the clinical, chemical, and serological manifestations of rheumatoid arthritis, based on the course of 154 cases. Medicine (Baltimore) 43:41-58, 1964 18. Dawes PT, Fowler PD, Jackson R, et al: Prediction of progressive joint damage in patients with rheumatoid arthritis receiving gold or D-penicillamine therapy. Ann Rheum Dis 451945949, 1986 19. Deal CL, Meenan RF, Goldenberg DL, et al: The clinical features of elderly-onset rheumatoid arthritis. Arthritis Rheum 28:987-994, 1985 20. Terkeltaub R, Esdaile J, D&cary F, et al: A clinical study of older age rheumatoid arthritis with comparison to a younger onset group. J Rheumatol 10:418-424, 1983 21. Ehrlich GE, Katz WA, Cohen SA: Rheumatoid arthritis in the aged. Geriatrics 25:103-l 13, 1970 22. Adler E: Rheumatoid arthritis in old age. Isr J Med 2:607-613, 1966 23. Isemein L, Redon M: La polyarthrite chronique evolutive apres Page de 55 ans. Rev Rhum Mal Osteoartic 20:877-881, 1953 24. Cecil RL, Kammerer WH: Rheumatoid arthritis in the aged. Am J Med 10:439-445,195l 25. Brown JW, Somes DA: The onset of rheumatoid arthritis in the aged. J Am Geriatr Sot 10:873-881, 1967 26. Dordick JR: Rheumatoid arthritis in the elderly. J Am Geriatr Sot 4:588-591, 1956 27. Moesmann G: Clinical features in subacute rheumatoid arthritis in old age. Acta Rheum Stand 14:285-297, 1968 28. Corrigan AB, Robinson RG, Terenty TR, et al: Benign rheumatoid arthritis of the aged. Br Med J 1:444-446, 1974 29. Schnell A: The clinical features of rheumatic infection in the old. Acta Med Stand 106:345-350, 1941 30. Evers A: Die im Alter auftretende primar chronische polyarthritis. Z Rheumaforsch 24:280-283, 1965 3 1. Kaarela K: Prognostic factors and diagnostic criteria in early rheumatoid arthritis. Stand J Rheumatol 57:1-54, 1985 (suppl) 32. Scott DL, Coulton BL, Symmons DPM, et al: Longterm outcome of treating rheumatoid arthritis: Results after 20yearsLancet 1:1108-1111, 1987 33. Luukkainen R, Isomaki H, Kajander A: Prognostic value of the type of onset of rheumatoid arthritis. Ann Rheum Dis 42:274-275, 1983
291
34. Hart FD: Presentation of rheumatoid arthritis and its relation to prognosis. Br Med J 2:621-624, 1977 35. Reeback J, Silman A: Predictors of outcome at two years in patients with rheumatoid arthritis. J R Sot Med 77:1002-1005, 1984 36. Wawrzynska-Pagowska J, Brzezinska B, Brzozowska M, et al: Observations on the symptoms and signs of “early” rheumatoid arthritis in a prospective study. Acta Rheum Stand 16:99-105, 1970 37. Fleming A, Crown JM, Corbett M: Early rheumatoid disease. I. Onset. Ann Rheum Dis 35:357-360, 1976 38. Rasker JJ, Cosh JA: The natural history of rheumatoid arthritis: A fifteen year follow-up study. The prognostic significance of features noted in the first year. Clin Rheumato1 3:l l-20, 1984 39. Brook A, Fleming A, Corbett M: Relationship of radiological change to clinical outcome in rheumatoid arthritis. Ann Rheum Dis 36:274-275, 1977 40. Stastny P, Batchelor J, Morris P: HLA and disease, in Bodmer WR, Batchelor JR, Bodmer JC, et al (eds): Histocompatibility Testing 1977. Report of the seventh Histocompatibility Workshop and Conference. Copenhagen, Munksgaard, 1978, p 205 4 1. Dequeker J, van Wanghe P, Verdickt W: A systematic survey of HLA-A, B, C, and D antigens and drugs toxicity in rheumatoid arthritis. J Rheumatol 11:282-286, 1984 42. Woodrow JC, Nichol FE, Zaphiropoulos G: DR antigens and rheumatoid arthritis: A study of two populations. Br Med J 283:1287-1288, 1981 43. Speerstra F, Reekers P, van de Putte LBA, et al: HLA-DR antigens and proteinuria induced by aurothioglucase and D-penicillamine in patients with rheumatoid arthritis. J Rheumatol 10:948-953, 1983 44. Panayi GS, Griffin AJ, Wooley PH, et al: Genetic predisposition to gold and o-penicillamine toxicity. Arthritis Rheum 22:645,1979 45. Roitt IM, Brostoff J, Male D: Immunology. London, Gower Medical, 1985, pp 23-24 46. Walker DJ, Griffiths M, Dewar P, et al: Association of MHC antigens with susceptibility to and severity of rheumatoid arthritis in multicase families. Ann Rheum Dis 44:519525,1985 47. Duquesnoy RJ, Marrari M, Hackbarlh S, et al: Serological and cellular definition of a new HLA-DR associated determinant, MCl, and its association with rheumatoid arthritis. Hum Immunol 10:165-176, 1984 48. de Jongh BM, van Romunde LKJ, Valkenburg HA, et al: Epidemiological study of HLA and GM in rheumatoid arthritis and related symptoms in an open Dutch population. Ann Rheum Dis 43:613-619, 1984 49. Young A, Jaraquemada D, Awad J, et al: Association of HLA-DR4/Dw4 and DR2/Dw2 with radiologic changes in a prospective study of patients with rheumatoid arthritis. Arthritis Rheum 27:20-25, 1984 50. Queiros MV, Sancho MRH, Caetano JM: HLA-DR4 antigen and IgM rheumatoid factors. J Rheumatol 9:370373,1982 51. Dobloug JH, Forre 0, Kass E, et al: HLA antigens and rheumatoid arthritis. Arthritis Rheum 23:309-313, 1980 52. Walton K, Dyer PA, Grennan DM, et al: Clinical
292
features, autoantibodies and HLA-DR antigens in rheumatoid arthritis. J Rheumatol 12:223-226, 1985 53. Silman AJ, Reeback J, Jaraquemada D: HLA-DR4 as a predictor of outcome three years after onset of rheumatoid arthritis. Rheumatol Int 6:223-235, 1986 54. Brackertz D, Wernet P: A hereditary basis for rheumatoid arthritis. HLA-D/DR-alloantigens and disease severity: Population and family studies, in Paulus HE, Erlich GE, Linenlaub E (eds): Controversies in the Clinical Evaluation of Analgesic-Anti-inflammatory-Antirheumatic Drugs. Stuttgart, Schattauer, 1981, pp 453-462 55. Panayi GS, Wooley P, Batchelor JR Genetic basis of rheumatoid disease: HLA antigens, disease manifestations, and toxic reactions to drugs. Br Med J 2:1326-1328, 1978 56. Barger BO, Acton RT, Koopman WJ, et al: DR antigens and gold toxicity in white rheumatoid arthritis patients. Arthritis Rheum 27:601-605, 1984 57. Griffin AJ, Wooley P, Panayi GS, et al: HLA DR antigens and disease expression in rheumatoid arthritis. Ann Rheum Dis 43:218-221,1984 58. Rantapaa Dahlqvist S, Strom H, Bjelle A, et al: HLA antigens in rheumatoid arthritis patients with and without a family history of polyarthritis. Stand J Rheumatol 14:375380,198s 59. Bjelle A, Cedergren B, Rantapaa Dahlqvist S: HLAB27 in the population of northern Sweden. Stand J Rheumato1 11:23, 1982 60. Isomaki H, Koota K, Nissila MJ, et al: HLA-B27 and arthritis. Ann Clin Res 7:138, 1975 61. Rantapaa Dahlqvist S, Nordmark LG, Bjelle A: HLA-B27 and involvement of sacroiliac joints in rheumatoid arthritis. J Rheumatol 11:27-32,1984 62. Elhabali M, Scherak 0, Seidl G, et al: Tomographic examinations of sacroiliac joints in adult patients with rheumatoid arthritis. J Rheumatol6:417-425, 1979 63. Martel W, Duff IF: Pelvo-spondylitis in rheumatoid arthritis. Radiology 77:744-756, 1961 64. Roitt IM, Corbett M, Festenstein, et al: HLA-DRW4 and prognosis in rheumatoid arthritis. Lancet 1:990, 1978 65. G’Duffy JD, O’Fallon WM, Hunder GG, et al: An attempt to predict the response to gold therapy in rheumatoid arthritis. Arthritis Rheum 27:1210-1217, 1984
VAN DER HEIJDE ET AL
66. Isomaki H, Martio J, Sarna S, et al: Predicting the outcome of rheumatoid arthritis. Stand J Rheumatol 13:3338,1984 67. Amos RS, Constable TJ, Crockson RA, et al: Rheumatoid arthritis: Relation of serum C-reactive protein and erythrocyte sedimentation rates to radiographic changes. Br Med J 1:195-197, 1977 68. Westedt ML, Daha MR, Baldwin WM III, et al: Serum immune complexes containing IgA appear to predict erosive arthritis in a longitudinal study in rheumatoid arthritis. Ann Rheum Dis 45:809-815, 1986 69. Teitsson I, Withrington RH, Seifert MH et al: Prospective study of early rheumatoid arthritis. I. Prognostic value of IgA rheumatoid factor. Ann Rheum Dis 43:673-678, 1984 70. Withrington RH, Teitsson I, Valdimarsson H, et al: Prospective study of early rheumatoid arthritis. II. Association of rheumatoid factor isotypes with fluctuations in disease activity. Ann Rheum Dis 43:679-685, 1984 71. Withrington RH, Seifert MH: Predominant wrist disease in rheumatoid arthritis associated with high concentration of IgA rheumatoid factor. Br Med J 291:1388, 1985 72. Gioud-Paquet M, Auvinet M, Raffin T, et al: IgM rheumatoid factor (RF), IgA RF, IgE RF, and IgG RF detected by ELISA in rheumatoid arthritis. Ann Rheum Dis 46:65-71, 1987 73. Linn JE, Hardin JG, Halla JT: A controlled study of ANA+RFarthritis. Arthritis Rheum 21:645-651, 1978 74. Westgeest AAA, Boerbooms AMTh, Jongmans M, et al: Antiperinuclear factor: Indicator of more severe disease in seronegative rheumatoid arthritis. J Rheumatol 14:893-897, 1987 75. Sjoblom KG, Saxne T, Pettersson H, et al: Factors related to the progression of joint destruction in rheumatoid arthritis. Stand J Rheumatol 13:21-27,1984 76. Ochi T, Yonemasu K, Iwase R, et al: Serum Clq levels as a prognostic guide to articular erosions in patients with rheumatoid arthritis. Arthritis Rheum 27:883-887, 1984 77. Scott DL, Grindulis KA, Struthers GR et al: Progression of radiological changes in rheumatoid arthritis. Ann Rheum Dis 43:8-17, 1984