Palindromic rheumatism: Part of or apart from the spectrum of rheumatoid arthritis

Palindromic Rheumatism: Part of or Apart From the Spectrum of Rheumatoid Arthritis PIERRE-ANDREGUERNE,M.D., Geneva, Switzerland,

MICHAEL Ii. WEISMAN, M.D.,

San Diego, California

Palindromic rheumatism (PR), or&ally described in 1944,is characterized by recurrent episodeaof mostly 0ligoarticuIar arthritis with peri- and para-articular tissue infkmmation, leavhq no residual clinical and radiographic changea It appears that palindromic syndrome is entity, encompassingother . a heterogeneous mfhnmatory conditions at early stagesof their evolution, and whose relationship with rheumatoid arthritis (RA) is evident but still unclear. Evolution of up to SO%of these cases intO otherwise typical RA, mtuuonly accompaniedby the conversion to rheumatoid factor seropositivity, the frequent occurrence of nodules,the reported responseto RA treatment, and the observation of familial aggregation of the two conditions suggestthatPRispartofthespe&um,ora stagein the evolutkm of RA. However, justification for the distinct exknce of PR corneafrom reports that identity well-defined and recognkable clinical manifestations such as descriptions of the acute attach the frequent peri-articular manifestations, the absenceof bone and cartilage destruction even after extended periods of time, and the generally goodlong-term prognosis. Immunogenetic studies with HLA-DR phenotyping and the absenceof female preponderance tend to add additional support for the separate identity of PR.

From the Department of Medicine, Division of Rheumatology (MHW). University of California at San Diego Medical Center, San Diego, California, and the Division of Rheumatology (PAG). Department of Medicine, HOpital Cantonal Universitaire, Geneva, Switzerland. Requests for reprints should be addressed to Pierre-Andre Guerne, M.D., Division of Rheumatology, Hbpital Cantonal Universitaire. 26 avenue Beau-Sejour. 1211 Geneve-4, Switzerland. Manuscript submitted November 9. 1990. and accepted in revised form June 11,1992.

he term palindromic rheumatism (PR) was T first usedby Hench and Rosenberg[l] in 1944, for their descriptionof 34patients presentingwith a unique set of signs and symptoms. The condition wasmarked by multiple recurring afebrile episodes of acute arthritis, periarthritis, and sometimes para-arthritis, lasting from a few hours to several days before disappearing completely, and apparently producing no residual articular effect. The term palindromic, derived from the Greek “palindromos,” which literally means“to run back,” was chosento expressthis concept of recurrence.The authors separatedthis condition from rheumatoid arthritis (RA) becauseof the totally different pattern of distribution of the arthritis, the frequent attacks of para-arthritis, the generalabsenceof significant constitutional symptoms, and the lack of progression, chronicity, and/or radiographic changes. Fifteen yearsafter the original report, Ansell and Bywaters [2] reexamined the concept of PR by studying 28 patients who fulfilled the criteria defined by Hench and Rosenberg[11.They concluded that PR wasmerely a variant or a modeof presentation of RA, noting that the majority of the patients, if followed long enough,will eventually display residual signsof chronic (rheumatoid) arthritis. Subsequently, a number of series [3-71 followed patients longitudinally and noted that about onehalf of the casesevolveinto RA. Neither clinical, immunologic, nor genetic studies could predict reliably which caseremains PR or evolvesinto chronic RA. The controversyoverwhetherPR deservesa special nosologicplace or servesas a prodroma of RA has not beenresolved. We recently observeda patient whose clinical manifestations consisted mainly of recurrent episodesof severeperi- and para-articular inflammation in addition to arthritis. The special clinical featuresof this inflammation suggestthat at least somecasesof PR truly representan entity different from RA. We therefore believe it would be important to reviewthe literature concerningthis intriguing syndrome, with emphasis on the elements, among its clinical, serologic,etiologic, and genetic features,that can help to retain it in its rightful nosologicplace.

October 1992 The American Journal of Medicine

Volume 93

451

PALINDROMIC RHEUMATISM / GUERNE AND WEISMAN

Figure

peri-articular, 1. Articular, the hypothenar and, to a lesser sents a para-articular manifestation. articular swelling and erythema.

and para-articular extent, the thenar Right. Synovitis

manifestations of palindromic rheumatism. eminences, without clear-cut involvement of the second proximal interphalangeal

CASEREPORT A 51-year-old patient, a native of Java, presented to our clinic with a 5-year history of episodic and recurrent pains in the back and the extremities, involving mainly the hands, feet, knees, and ankles. These attacks lasted from 1 to 2 days, sometimes up to a week. They were often, but not always, associated with redness and swelling, both in the joints and in the extra-articular sites (Figure l), including the palms, soles, lateral aspects of the heels, and the forearms. Episodes progressively worsened, both in frequency (one or two per week) and in intensity. The patient had to quit his job in 1986. He saw a variety of doctors, but no diagnosis was established. The different treatments that he received, including analgesics, nonsteroidal anti-inflammatory agents (NSAIDs), and benzodiaxepines, did not improve his condition. A skin biopsy specimen, obtained from the left hypothenar eminence during an 452

October 1992 The American Journal of Medicine

left. Swelling and erythema of of joint structures. This reprejoint in addition to marked peri-

episode of swelling and redness, had shown a positive lupus band, but was otherwise histologically normal. Because of this finding, the diagnosis of systemic lupus erythematosus (SLE) was considered and the patient began to undergo a treatment trial of hydroxychloroquine for 4 months without any benefit. At the first arthritis clinic visit, results of physical examination were entirely unremarkable except for some bilateral tenderness of the trapezius insertion points, pes anserinus insertions, and epicondyles, consistent with the diagnosis of fibromyalgia syndrome. Results of laboratory tests were all normal: erythrocyte sedimentation rate (ESR) 16 mm/h; white blood cell (WBC) count of 14.7 X log/L with 68% neutrophils, 3% bands, 21% lymphocytes, 7% monocytes, and 1% eosinophils; hemoglobin, 13.2 g/dL (mean corpuscular volume, 90 fL); rheumatoid factor (RF), negative times 2; antinuclear antibody (ANA) negative; anti-SSA and anti-SSB nega-

Volume 93

PALINDROMIC

tive; total hemolytic complement (CH&, 135 U/mL; VDRL negative; muscle enzymes, not elevated; serum protein electrophoresis normal. Radiographic studies including the knees, hands, and the chest were negative. Treatment with amitriptyline only partially relieved the pain. At the second visit, the patient presented with very localized swelling and redness on the medial aspect of the left knee, with a diameter of about 7 cm; synovial fluid aspirated from this knee was mildly inflammatory, containing a WBC count of 2,100/mm3 (56% segmented, 41% monocytes, and 3% lymphocytes). The peri-articular swelling and arthritis disappeared spontaneously. On the third visit, the patient presented with localized swelling, redness, and extreme tenderness of the lateral aspect of the right ankle that was clearly para-articular. Because of the pattern of recurring episodes of arthritis and para-arthritis over a 5-year period, disappearing completely between episodes, and the exclusion of other forms of chronic arthritis, we strongly considered the diagnosis of PR. During the subsequent 3-year period of observation, the patient received complete therapeutic courses of gold, colchicine, sulfasalaxine, methotrexate, dapsone, and systemic steroids. No effect on the patient’s condition was observed; he continued to exhibit even more frequent episodes of peri- and para-arthritis, especially in his pelvis, soles, lateral aspects of his ankles, and heels, and occasional arthritis. On almost all occasions his ESR remained at baseline levels; during one episode, his ESR was elevated to 45 mm/h.

CLINICALFEATURESOF PR The nature of PR arthritis remains its most distinctive feature. Joint involvement is commonly noted to be monoarticular with maximal intensity reached in a matter of hours [1,7], at any time of the day, usually without any obvious inciting event, although, in a few cases, attacks have been noted to relate to another event such as a respiratory tract infection, weather change, childbirth, or overexercise [1,3,7]. The disease can strike almost any joint, although the spine and the hips are rarely affected. When the spine is involved, it is usually at the cervical level. Table I displays the distribution of the joints involved, based on five large series totaling 227 patients [1,3-5,781. The joints involved in initial attacks appear to be mostly knees, fingers, or shoulders. In the vast majority of the patients, the attacks are confined to a single articulation, but in rare cases more than four joints can be affected at a time [l]. The pain is generally described as very severe or excruciating, mimicking gout or septic arthritis in some cases,

RHEUMATISM

/ GUERNE

AND WEISMAN

TABLE I Distribution of Joints Involved in AttacksBased Upon a Cumulative ExperienceWith 227 Patients Joint Involvement

Mean% Patients

of

Rangeof % of Patients

MCP & PIP

91

74-100

Wrists

78

54-82

Knees

64

41-94

Shoulders

65

33-75

Ankles

50

lo-67

Feet

43

15-73

Elbows

38

13-60

Hips

17

O-40

Temporo-mandibular

8

O-28

Spine

4

o-11

Sternoclavicular

2

O-6

27

20-29

Para-articular sites

CP = metacarpophalangeal;PIP = proximal interphalangeal

often confining the patient to bed [1,5,7]. A few patients have been described as having only mild pains. In addition, extra-articular tissue inflammation is observed in about 30% of the patients [1,7]. This involvement may be inflammation of a tendon or swelling of undefined para-articular structures, generally close to a joint but often remote enough to allow the patient to clearly distinguish it from joint involvement. The size of these para-articular swellings varies from 2 cm to much larger areas. The PR episodes last less than 2 days in most of the cases, sometimes as short as 2 hours [l]. However, in a few cases, they can extend to more than 2 weeks [1,5]; their frequency is highly variable, from a few to more than 250 a year [1,5,8]. Persons who have frequent attacks (more than one a month) may be more likely to develop frank RA [6]. Only a few patients display constitutional reactions such as fever or weight loss. Of interest is the observation that patients from the Hannonen [7] series (7 of 60) who had such systemic manifestations were all RF-positive and later progressed to erosive RA. A peculiar feature of PR is the occasional occurrence of nodules, usually the size of a pea and overlying the tendons in the hands or fingers or located on the thumb pads. Their appearance parallels the arthritis attacks and generally lasts a week at most [1,3-5,9,10]; however, typical PR may also be associated with more sustained nodules. The presence of nodules has been used to link PR to other forms

October

1992

The American

Journal

of Medicine

Volume

93

453

PALINDROMIC RHEUMATISM / GUERNE AND WEISMAN TABLE II Percentagesof Patients RF-Positivein Three Series: RF Determination at Follow-Up in Patients Who Have DevelopedChronic Arthritis (PR-RA)Compared With Patients Who Remained Palindromic (PR-PR) Reference

ii [51

PR(%)

PR-RA(%)

1 l/25 (44) 3116 (19) 2122(9)

29135 (83) 16125 (64) 16/17 (94)

I

I

RF = rheumatoid factor; PR = palindromicrheumatism; RA = rheumatoid arthritis.

of chronic rheumatic diseases. For example, the clinical entity of rheumatoid nodulosis [ll], combining rheumatoid nodules, subchondral bone cysts, and recurrent joint symptoms and considered a variant of RA with a good prognosis, closely resembles PR. A recent review of all of the published cases of rheumatoid nodulosis [12] reveals that 24 of the 26 reported patients were described as having a clinical presentation consistent with PR. Kaye et al [13] place PR with nodules in the group IIa of their proposed classification of conditions associated with rheumatoid nodules. To add further to the confusion, there is a report of a patient with PR and pulmonary nodules [14], and two patients with PR and Felty’s syndrome [7,15]. Finally, in the 60 patients with PR reported by Hannonen et al [7], the criteria for fibromyalgia syndrome were fulfilled by 8 of their 60 patients. In contrast to the female preponderance noted in RA, the incidence of PR is relatively equal in each sex, with a female:male ratio that varies between 1.7:1 and 0.55:1 [1,3-5,7,8]. The mean age of onset of PR is about 45 years, with cases reported from the early 20s to the 80s. The prevalence of this disease is difficult to establish, and certainly may be underestimated. In the most recent studies on PR, the prevalence is estimated to be 1 of 8 [7] to 1 of 20 [8] that of RA.

LABORATORYFEATURESOF PR ESR is frequently increased during or just after the episodes, but rarely between them [1,3,5], and the ESR might be more likely elevated in the patients who subsequently develop RA. There is generally no abnormality of the WBC and red blood cell counts or miscellaneous chemical constituents [l], except for a few patients who were reported to have transient anemia [3]. RF status, associated with the disclosure that a significant number of patients with PR developed chronic arthritis identical to RA, has become an important diagnostic and nosologic issue. A number of reports show that the percentage of seropositive patients among those with PR that has evolved into 454

October 1992 The American Journal of Mediclne

chronic arthritis is not different from the percentage of seropositivity in RA. Among the patients observed in different series [5-71, about one third were found to convert from negative to positive, usually shortly before their evolution to chronic arthritis. The patients who remained palindromic seemed, in one series [7], to be more frequently positive (11 of 25, including 2 followed prospectively for more than 5 years) than the normal population, and in addition, patients with PR and positive RF seemed more susceptible to develop RA [7]. Finally, seropositivity is more frequent in PR patients after they have developed RA than in patients who remained palindromic [5-71 (Table II). Because of the limited follow-up time periods and the dynamic changes in the status of RF and arthritis patterns, the meaning of the seropositivity among patients who remain palindromic is less clear-cut. However, it seems that seroconversion almost always shortly precedes the appearance of chronic arthritis [5]. It is likely, therefore, that there is a true increase in RF positivity in PR patients who remain palindromic in addition to the definite increase in those who convert to RA. ANA studies, on the other hand, are usually negative in PR cases [6,16]. Radiographic studies, as pointed out originally by Hench and Rosenberg [l], are always normal in patients with PR [3,5,6]. Absence of radiographic signs is thus among the six diagnostic criteria proposed by Pasero and Barbieri [8]. In the cases that have evolved into chronic arthritis, erosions not different from the ones seen in classic RA usually develop [3,5,6]. The rare synovial fluid analyses that have been reported [17] show variable leukocyte counts (150 to 12,700/mm3) and differentials (2% to 66% polymorphonuclear cells) that do not correlate with the severity of the symptoms or attacks.

PATHOLOGICFINDINGS Not surprisingly, there are very few pathologic examinations of tissues in this disease. The pathologic studies of tissues obtained by Hench and Rosenberg [l] during an attack showed periarticular edema and gross thickening of the joint capsule. There was infiltration by large numbers of polymorphonuclear leukocytes without the presence of significant numbers of eosinophils. Between episodes, gross and histologic examination of joint tissues revealed no significant evidence of inflammation. Similar changes were described in tendon sheaths. No pannus formation or cartilage destruction was observed in any subject studied. More recently, Schumacher [17] performed a thorough examination, including electron microscopy, of two PR patients before they developed RA. These spec-

Volume 93

PALINDROMIC

RHEUMATISM

/ GUERNE

AND WEISMAN

TABLE III Percentageof Patients With HLA-DR4or -5* f201

El

[al

No. of Subjects

DR4

DR5

No. of Subjects

Controls

20

36.6

14

150

PR

20

60+

5

26

20

70’

10

158

DR4

DR5

DR4

DR5

27

9

99

30.3

32.3

38.5

3.8

43

34.9

48.8

:sI

36 31.2

PR-PR PR-RA RA

1221

No. of Subjects

64.6$

No. of Subjects

DR4

44

20.4

34 10

ii

3.2

‘atients with PR are comoared with normal controls and with RA. In the last two series (L6.221). patients with PR are subduded mto ones who remalned palmdromlc (PR-PRI and ones who developf chronic arthritis (Pi-RA). r ‘p co.05. *p t0.005. sp
imens showedminimal superficial reactionand only mild lining cell hyperplasia (one to three layers), with moreneutrophils than aregenerallyseenin the synovitis of chronic RA, small numbers of perivascular lymphocytes and absenceof plasmacells were noted. Fibrin and cellular thrombi are noted in somesmall blood vesselsalongwith congestionand someperivascularfibrosis. Thesechangescould not be differentiated from other arthritides such as gout, Reiter’s syndrome, or early RA. Examination of a transient subcutaneousnodule by Hench and Rosenberg[l] showed nonspecific inflammation, without the necrotic zonesor palisading cells seenin classicrheumatoid nodules.On the other hand, Schreiber et al [18] studied the ultrastructure of a nodulethat had beendeveloping for 6 months in a patient with typical PR, findings revealedcentral fibrinoid necrosissurrounded by palisading mononuclearcells,asdescribedtypically in classicalRA. HLA ANTIGEN ASSOCIATIONS A family incidencehas beennoted for PR beginning in the reports from the 1960s[3,5]. Since the pioneerwork of Stastny [19] showingan association of the B-cell alloantigen DRw4 with RA, many studies have analyzedpatients with PR in an attempt to better classify this diseaseand to define more clearly its relationship with RA. The data, summarized in Table III, do not provide a clear answerto this question.Somestudiesreveala weak associationof PR with HLA-DR4 [20], or with DR5 [6], whereasothers concedean absenceof association with any of theseantigens [6,21,22].Deserving mention is an interesting report of two brothers with PR, sharing HLA-DR4 and followed for more than 12yearsafter the diagnosis;arthritis remained palindromic in the first brother, but the seconddeveloped chronic deformities and classicalRA early

in the courseof the disease[23]. The variability in the different seriesprobably arisesfrom a lack of homogeneityamongthe groupsdue in part to different diagnostic criteria. In view of the debate aboutautonomyof PR from RA, the issueof genetic associationassumesgreat importance.The association of RA itself with MHC classII antigensis now clearly established: 70% of the patients express HLA-DR4 [19]and 93%either DR4 or DRl in some studies [24]. The susceptibility to RA has been mappedto five amino acids in the third hypervariableregion of the HLA-DR B-1 chain [25];this area distinguishes HLA-DRl and the HLA-DR4 subtypes associatedwith RA (DW4) from the onesthat are not (DWlO and DW13). In PR, an association with HLA-DRl wasalsofound to be negativewhen examined [6,20,21].Therefore, the absence of a clear associationwith HLA-DR4 and -DRl may indicatethat PR is either not simply a merevariant of RA, or that the different HLA backgroundleadsto an alternate expressionof the disease.Additional studies would be necessaryto settle this question, including onesexamining largerand better-defined populations.Thesefuture studiesshould include an analysis of HLA expression,including HLA-DR4 subtypes,in PR patients accordingto their ultimate evolution and the presenceor absenceof RF. TREATMENT A number of different treatments havebeenused with inconsistent results. However,none of the interventions have been submitted to rigorous controlled studies, likely due to the relative rarity of the condition. NSAIDs appearedto be effective for a partial relief of joint pain in somepatients and can rarely inducea long-lasting remission.In a questionnairebasedstudy, Grattan et al [26] reportedthat in 68% of 38 patients,NSAIDs or analgesicsweresaid to be

October

1992

The American

Journal

of Medlclne

Volume

93

455

0

5 10 15 20 25 Duration of follow-up (years)

30

Figure 2. The natural

history of PR evolution into chronic arpooled from two series [5,7] are represented horizontal bars. For each patient, the left-hand end of the bar represents the onset of PR and the right-hand end the duration of follow-up in years. The positive-negative signs represent the status of rheumatoid factor (RF) at follow-up. RF was measured by latex [5] and sensitized sheep erythrocyte agglutination [7] testing. Titers were considered positive when greater than or equal to 1:80 [5] and greater than or equal to 1:64 [7]. thritis. by the

Patients

[3]. In another series [6], however, the disease resolved after gold treatment in only 20% of the patients, and the authors pointed out that spontaneous remissions occurred in three among seven untreated patients (42%). Other disease-modifying antirheumatic drugs used in RA have had varying successes: D-penicillamine was reported, in one series, to be very effective, with complete remission in four of five patients [27]. Of note is the fact that the four patients who experienced remissions were RF-negative whereas the one who did not was mildly seropositive. Good responses, however, were much rarer in two other series [6,7]. Response to antimalariala varies in different studies, from 15% [6] to almost 80% in a recent series of 71 patients [28]. Interestingly, in one series, effectiveness was much lower in the group of patients who remained palindromic than in the ones who developed chronic arthritis [7,29]. Sulfasalazine has also been shown to induce a favorable response in about half the patients [7,30]. Other drugs, such as chlorambucil and dapsone, employed more rarely, reportedly induced prolonged remissions in a few patients [6,31]. Finally, colchicine has recently been evaluated at prophylactic doses in five patients, with remission or good response in all subjects [32]. The variable evolution of PR and its frequent spontaneous remissions together with the absence of well-defined criteria for diagnosis as well as for treatment outcome all contribute to the difficulty of evaluating any drug intervention. The diversity of the drugs employed is an indication of the sometimes disabling nature of PR and a less than satisfactory therapeutic experience with nonspecific treatment. However, many PR patients remain free of progressive or severe disease. Therefore, it is quite reasonable to reassure these patients of the likely possibility of a favorable outcome regardless of the generally negative published experience with antirheumatic drugs.

EVOLUTIONAND PROGNOSIS effective; however, in a more recent series, only 2 cases out of 60 responded to NSAIDs [7]. Chrysotherapy, because of the observed relationship between PR and RA, has been widely used, with frequent, although highly variable, testimonials for favorable results. Good responses were observed in about two thirds of the patients treated with injectable gold in two series [3,7]. The injections appeared to stop the acute attacks, often within a few weeks, and usually by the time 200 to 500 mg had been reached. An unusually high percentage of cutaneous side effects was noted in one series 456

October 1992

The American Journal of Medlclne

Since the description by Hench and Rosenberg [l], numerous authors have observed that between one third and one half of PR patients developed chronic arthritis, most of the time not distinguishable from RA [2-71. Figure 2 illustrates the detailed evolution of 54 patients pooled from 2 of these series [5,7], who developed chronic diseases. The interval before the conversion to chronic disease took place varied from a few months to more than 20 years. Among these chronic conditions, 50 (92.6%) patients had definite, or classical RA, including 45 (83.3%) who were seropositive. Two patients developed seronegative pauci-articular

Volume 93

PALINDROMIC

chronic arthritis, with one caseeach evolving into Wegener’sgranulomatosisand SLE. In these same series,45 patients remained palindromic with follow-ups of up to 31 years, and we recently have describeda patient with persistent PR followed for 47 years(Gabayet al, submitted for publication). It can certainly be argued, however,that if patients arefollowed for a long enoughperiod of time, all will developchronic arthritis. Among patients who developedchronic arthritis, 10 were observedto convert to RF seropositivity; most of the time this conversion closely precededthe evolution to chronic arthritis. Among the 45patients who remainedpalindromic, only 1 had such a conversion. PR is allegedto evolvetowards a variety of other illnessesincluding nonspecificseronegativearthritis, Whipple’s disease,and ankylosing spondylitis. Nevertheless,in most of the series,a significant number of PR patients do not changeat all, except to enter into prolongedremission.Table IV shows the results of 9 series,describing the outcomesof 653 patients. It reveals that 48% of the patients continue to have PR, 33%progressto chronic arthritis that usually is seropositiveRA, 4% progress to other diseases,and 15%finally havea prolonged remissionor are cured.This latter number includes patients who had their remission after progression through RA, and patients who respondedto treatment. The results, however,are variable from one seriesto another, and probably reflect heterogeneity of the populationsand differencesin the criteria applied for diagnosisand remission. The character and severity of the RA that evolvesfrom PR do not seemto be different from that of typical RA. Some investigators have attempted to define clinical or biologic parametersthat might predict the differential evolution toward RA [5,21,26]. Analysis of this question, however,is problematic considering that a few patients had PR for more than 10yearsbeforeevolution to RA [32],rendering the notion of persistingPR impreciseand questionable.Nevertheless,someparametershavebeenproposed,although they do not seemto have definite prognosticsignificance: in one series,frequencyof HLA-DR4 was higher in patients who developed RA. IncreasedESR and positive serology [5] and morning stiffness and pain in severaljoints at once [26] were more common among the patients who ultimately developedRA. Interestingly, skin erythema and localized para-articular swelling, features generallynot encounteredin establishedRA, occurred in both groups. A family history of RA doesnot appearto be of definite prognosticsignificanceto predict evolution, with the occurrenceof both persistent PR and PR evolving to RA within the same family [5,23].

RHEUMATISM

/ GUERNE

AND WEISMAN

TABLE IV Evolution of PR Patients in Nine Series Totaling 653 Patients*

I

Reference

Total/average

Other No.of Remissionor Persistent PR-RA Diseases Cases Cure(%) PR(%) (%) (%I

653

15

48

4

33

PR = palrndromic rheumatism;.RA = rheumatoid arthritis. “In each serves, the numbers ot patlents undergomga remIssron or a cure, remarnmgpalmdromrc, evolving toward RA (PR-RA), or developinganother drsease are expressed m percent.

PATHOGENESIS Not surprisingly, given the relatively low incidence of the diseaseand the many controversies concerningits definition, the etiopathogenesisof PR remains unknown. Due to the peculiar clinical presentationof their patients with PR, Hench and Rosenberg[l] attempted, without success,to find an allergic origin to this disease.Not discounting the importance of this idea in recent times, Panush [33] describedthree patients with arthritis displaying a palindromic pattern in whom immunologic sensitivity to certain foodsand reappearanceof the symptoms after challengewere demonstrated.The relationship of food sensitivity to intermittent episodes of arthritis remains an area of continued investigation. The possibility that the palindromic attacks might be related to a deficiency of Cl e&eraseinhibitor has been explored and ruled out [4,5]. No evidenceof reduction in CHss,C3, or factor B could be detected[5]. Moderately elevatedlevelsof circulating immune complexes,as assessedby Clq-binding test [10,34],were detected in a few patients, were generally seropositive,and sometimes paralleled diseaseactivity. Schumacher[17], using electron microscopy,described finely granular electron-dense material within vacuolesin synovial cells and venules,suggestingthat immune complexesmight play a role in PR pathogenesis.The composition of this material is unknown. Electron microscopyof PR synovium also revealed tubular structures within the rough endoplasmicreticulum of the capillary endothelial cells, similar to nonspecific structures already reported by others in the tissuesof patients with different rheumatic diseases,neoplasms,infections, and various other conditions [35]. The origin of thesestructures is controversialand has been pro-

October

1992

The American

Journal

of Medicine

Volume

93

457

PALINDROMIC

RHEUMATISM

/ GUERNE

AND WEISMAN

TABLE V Proposed DiagnosticCriteria for Palindromic Rheumatism* l

Recurrent Arthritis and Periarthritic ConditionsThat Can Mimic PR

A 6-mo history of brief sudden-onset and recurrent episodes of monoarthritis or rarely polyarthritis or of soft tissue inflammation.

l

Direct observation of one attack by a physician.

l

Three or more joints involved in different attacks.

l

Absence of erosions on radiographs.

+ Microcrystalline tis

arthri-

Presence of crystals in SF Typical radiologic signs frequently present

Reactive arthritis

Frequent history of diarrhea or urethritis Frequent ocular or mucocutaneous manifestations Usually longer episodes

Intermittent hydrarthrosis

Absence of inflammatory signs Generally painless

Familial Mediterranean fever

Onset usually in childhood Fever (but might be low-grade)* Abdominal +I- chest pains*

Type II hyperlipidemia

Xanthomas Hvoerlioidemia

I

0 Exclusion of other arthritides. \dapted from Pasemand Barbieri [El, and Hannoneneta/ 171.

posed to be incomplete viruses, immunoglobulins, or incomplete virus-immunoglobulin complexes. No other report has been forthcoming to support this finding in patients with PR.

Her$itap

Eosinophilic synovitis

DISCUSSIONAND CONCLUSIONS In view of the information related to long-term outcome from many different series, there is reasonable certainty to establish a palindromic syndrome that previews a variety of heterogeneous entities at the early stages of their evolution including, but not limited, to SLE, seronegative spondyloarthropathies, other connective tissue diseases, and, most probably, still undefined entities. Further, there appears to be good evidence that an entity exists that should be called PR whose relationship with classical RA is certain, but not entirely defined. Some observations suggest that PR is a stage in the evolution or part of the spectrum of RA, as evidenced by progression of approximately 50% of the patients to seropositive RA, the occurrence of nodules, the reported, although inconstant, response to RA treatments such as gold, the observation of familial aggregation of the two conditions [5,23], and the high prevalence of RF in some series. Alternatively, several special features of PR speak for its autonomy from RA: they include para-articular inflammation, frequent remissions, the absence of bone and joint destruction, even after very long evolution, the lack of association with HLA-DR4 and/or -DRl, and the absence of female preponderance. These recognizable clinical manifestations have permitted a proposal of diagnostic criteria to be validated by future prospective studies (Table V). Given the fact that in several series [5-71, a certain number of patients with persistent PR are latex-positive, we agree with Hannonen et al [7] that a negative RF is not a requirement among the diagnostic criteria. For the same reason, the presence of acute-phase reactants should not hinder the

458

October

1992

The American

Journal of Medicine

angio-

I

I

Nonpainful, nonpruritic swelling Abdominal pain History of allergy Pain usually minimal Elevated IgE in blood Eosinophilia in SF usually >lO% I

‘R = palindromicrheumatism; SF = synovialfluid. Can beabsent at the beginningof the disease.

establishment of a diagnosis of PR. These criteria include the exclusion of other arthritides, as noted in Table VI. A number of clinical conditions can mimic PR, but generally demonstrate distinctive features that usually make them easily recognizable. The etiology of RA is still unknown. It is becoming increasingly clear at the present time, however, that RA is due to the combination of an unknown arthritogenic stimulation and a genetic susceptibility that includes at least an MHC class II antigen association [36]. Conceivably, the links between RA and PR could derive from shared elements, whereas dissimilarities are explained by differences among this combination. The expression of inflammatory diseases is largely determined by the cytokine network activation, whose role is beginning to be well characterized in the regulation of the clinicopathologic manifestations of rheumatic diseases [36,37]. It is likely that some of the features that distinguish PR from RA reflect differences in the expression of these mediators. These differences might involve: interleukin-1 and tumor necrosis factor, and/or their inhibitors, implicated in articular destruction through the induction of proteases [37]; interleukin-6, the main inducer of the acute-phase reaction in RA [38] and a B-cell activator involved in RF production [39,40]; granulocyte-macrophage colony-stimulating factor, which might play a role in

Volume 93

PALINDROMIC RHEUMATISM / GUERNE AND WEISMAN

inflammation maintenance and chronicity [41]; interleukin-4 and tumor growth factor+, which have the potential to downregulate certain immuno-inflammatory mechanisms [42,43]. Although it appears likely that there is a link between RA and PR, it is not possible, at this time, to define it with great precision. In any case, the analysis of the relationship between these two entities and the study of the conversion from one to the other may provide important clues for the pathogenesis of either or both conditions. Examining differences in the expression of cytokines and mediators, together with further study of their immunogenic associations, should be a fruitful area for future investigation. If the diseases differ by external arthritogenic stimuli, examination of the conversion may provide substrate for discovery of the triggering agents. Although its recognition is still very high in some centers [7,8,28], it is probable that PR, dating from its original description, may be diminishing in overall prevalence, likely due to the more accurate diagnosis and proper nosologic placement of other rheumatic diseases in the modern era. Nevertheless, there appears to be a core of evidence, almost 50 years after the original description, that some patients continue to have PR.

ACKNOWLEDGMENT We thank Dr. R. Terkeltaub for helpful discussions, Denise Smith for typing assistance, and Dr. Tuska for kindly providing the photographs.

REFERENCES 1. Hench PS, Rosenberg

EF. Palindromic rheumatism. A “new” oft recurring disease of joints (arthritis, periarthritis, para-arthritis) apparently producing no articular residues. Report of thirty-four cases: its relation to “angio-neural arthrosis.” ” allergic rheumatism” and rheumatoid arthritis. Arch Intern Med

1944; 73: 293-321. 2. Ansell BM, Bywaters

EGL. Palindromic

rheumatism.

Ann Rheum Dis 1959; 18:

331-2. 3. Mattingly S. Palindromic rheumatism. Ann Rheum Dis 1966; 25: 307-17. 4. Wtlliams MH. Sheldon PJ. Torrigiani G, Eisen V. Mattingly S. Palindromic rheumatism. Clinical and immunological studies. Ann Rheum Dis 1971; 30: 37E-BG. 5. Wajad MA. Brown DL. Currey HL. Palindromic rheumatism. Clinical and serum complement study. Ann Rhaum Dis 1977; 36: 56-61. 6. Bregeon C. Dajon JL, Renier L. eta/. Palindromic rheumatism. Rev Rhum Mal Osteoartic 1986; 53: 441-9. 7. Hannonen P, M6tt6nen T, Oka M. Palindromic rheumatism. A clinical survey of sixty patients. Stand J Rheumatol 1987; 16: 413-20. 8. Pasero G, Barbieri P. Palindromic rheumatism: you just have to think about it. Clin Exp Rheumatol 1986; 4: 197-9. 9. Ward LE, Okihiro MM. Palindromic rheumatism: a follow up study. Archives of

Rheumatology 1959; 2: 208-9. 8, Mohammed I, Holborow EJ. Currey HL. Palindromic rheumatism. II. Failure to detect circulating immune complexes during acute episodes. Ann Rheum Dis 1979; 381 329-31. 11. Ginsberg MH, Genant HK, Yu TF, McCarty DJ. Rheumatoid nodulosis: an unusual variant of rheumatoid disease. Arthritis Rheum 1975; 18: 49-58. 12. Couret M, Combe B. Chuong VT, Leroux JL. Blotman F, Sany J. Rheumatoid nodulosis: report of two new cases and discussion of diagnostic criteria. J Rheumatol 1988; 15: 1427-30. 13. Kaye BR, Kaye RL, Bobrove A. Rheumatoid nodules. Review of the spectrum Interamerican

10. Thompson

of associated conditions and proposal of a new classification, with a report of four seronegative cases. Am J Med 1984; 76: 279-92. 14. Franck JL, Durroux R. Condouret J, Arlet J. Seropositive palindromic rheumatism and pulmonary rheumatoid nodule: apropos of a case [letter]. Rev Rhum Mal Osteoartic 1982; 49: 309-10. 15. Davies PG. Thompson PW. Palindromic rheumatism and Felty’s syndrome. Ann Rheum Dis 1985; 44: 640-l. 16. Barbieri P, Benedettini G. Ciompi ML, Campa M. Failure to detect a lymphocyte imbalance during acute attacks of palindromic rheumatism [letter]. Clin Rheumatol 1987: 6: 1023. 17. Schumacher HR. Palindromic onset of rheumatoid arthritis. Clinical, synovial fluid, and biopsy studies. Arthritis Rheum 1982; 25: 361-9. 18. Schreiber S, Schumacher HR, Cherian PV. Palindromic rheumatism with rheumatoid nodules: a case report with ultrastructural studies. Ann Rheum Dis

1986; 45: 78-81. 19. Stastny P. Association of the B-cell alloantigen DRw4 with rheumatoid arthritis. N Engl J Med 1978; 298: 869-71. 20. Fisher LR. Kirk A, Awad J, et a/. HLA antigens in palindromic rheumatism and palindromic onset rheumatoid arthritis. Br J Rheumatol 1986; 25: 345-8. 21. Gran JT. Husby G, Thorsby E. HLA antigens in palindromic rheumatism, nonerosive rheumatoid arthritis and classical rheumatoid arthritis. J Rheumatol

1984; 11: 136-a 22. Barbieri P, Ciompi ML, Menicucci A, Pasero G. HLA antigens in palindromic rheumatism. An Italian study. Clin Rheumatol 1988; 7: 470-3. 23. Hannonen P, Hakola M. Oka M. Palindromic rheumatism in two non-identical brothers with identical HLA including DR4. Ann Rheum Dis 1985; 441 2024. 24. McDermott M, McDevitt H. The immunogenetics of rheumatic diseases. Bull Rheum Dis 1988; 38: l-10. 25. Gregersen P, Shen M, Song QL. eta/. Molecular diversity of HLA DR4 haplotypes. Proc Natl Acad Sci U S A 1986; 83: 2642-6. 26. Grattan CE. Kennedy TD. Yates DB. Prognostic factors in palindromic rheumatism. Bristol Med Chir J 1984; 99: 51-4. 27. Huskisson EC. Treatment of palindromic rheumatism with Dpenicillamine. BMJ 1976; 2: 979. 28. Youssef W, Yan A, Russell AS. Palindromic rheumatism: a response to chloroquine. J Rheumatol 1991; 18: 35-7. 29. Hannonen P. Mbtt6nen T, Oka M. Treatment of palindromic rheumatism with chloroquine [letter]. BMJ 1987; 294: 1289. 30. Golding DN. Sulphasalazine for palindromic rheumatism [letter]. Br J Rheumatol 1988; 27: 79. 31. Meyer 0, Ryckewaert A. Palindromic rheumatism: efficacy of dapsone [letter]. Rev Rhum Mal Osteoartic 1985; 52: 596-7. 32. Schwartzberg M. Prophylactic colchicine therapy in palindromic rheumatism [letter]. J Rheumatol 1982; 9: 3413. 33. Panush RS. Food induced (“allergic”) arthritis: clinical and serologic studies. J Rheumatol 1990; 17: 2914. 34. Shingu M. Yasuda M, Nobunaga M. A case of palindromic rheumatism suggestive of immune complex disease [letter]. J Rheumatol 1982; 9: 480-l. 35. Molnar 2, Metzger AL, McCarty DJ Jr. Tubular structures in endothelium in palindromic rheumatism. Arthritis Rheum 1972; 15: 553-6. 36. Harris ED. Rheumatoid arthritis. Pathophysiology and implications for therapy. N Engl J Med 1990; 322: 1277-89. 37. Arend WP. Dayer J-M. Cytokines and cytokine inhibitors or antagonists in rheumatoid arthritis. Arthritis Rheum 1990: 33: 305-15. 38. Guerne P-A, Zuraw BL, Vaughan JH, Carson DA, Lots M. Synovium as a source of interleukin 6 in vitro. Contribution to local and systemic manifestations of arthritis. J Clin Invest 1989; 83: 585-92. 39. Hirano T. Matsuda T, Turner M. et al. Excessive production of lLb/B cell stimulatory factor-2 in rheumatoid arthritis. Eur J lmmunol 1988; 18:

October 1992 The American Journal of Medlclne

Volume 93

459

PALINDROMIC RHEUMATISM / GUERNE AND WEISMAN 1797-801. 40. Lipsky PE. The control of antibody production by immunomodulatory mole cules. Arthritis Rheum 1989; 32: 1345-55. 41. Alvaro-Gracia J. Firestein GS, Xu WD, Teetle R, Zwaifler NJ. GMCSF is a major macrophage activating factor (MAF) in rheumatoid synovitis [abstract]. Arthritis Rheum 1988; 31: S27. 42. Brandes ME. Allen JB, Ogawa Y, Wahl SM. Transforminggrowth factor beta 1 suppresses acute and chronic arthritis in experimental animals. J Clin Invest

460

October 1992 The American Journal of Medicine

1991; 87: 1108-13. 43. Miossec P. Naviliat M, Dupuy D’Angeac A, Sany J, Banchereau J. Low levels of interleukin-4 and high levels of transforming growth factor beta in rheumatoid synovitis. Arthritis Rheum 1990; 33: 1180-7. 44. Renier JC, Bregeon C, Besson J. Palindromic rheumatism (contribution of 16 cases and general review). Rev Rhum Mal Osteoartic 1969; 36: 583-96. 45. Mattingly S. Jones DW, Robinson WM, Williams RA, Dunn EC. Palindromic rheumatism. J R Coll Physicians Lond 1981; 15: 119-23.

Volume 93