A-36-year-old female office worker presented in the Arthritis Unit with an 8-month history of intermittent joint pain and swelling. The family physician had prescribed naproxen during flares without significant improvement. The first blood test disclosed a rheumatoid factor (RF) of 40UI (NV <20) and a negative antinuclear (ANA) test, without other significant changes and with normal levels of CRP, ESR and uricemia. No familial occurrence, arthritis or other immune-mediated disease was found and no comorbidities were recorded. A detailed clinical history of the joint symptoms revealed that the patient had joint attacks that initiated abruptly (maximum pain in 2 h) and disappeared after 24–36 h. Between attacks, the patient was symptom free. The interval between flares was irregular but the patient reported 2–3 episodes per month, some of which were disabling. The most frequent joints affected were the wrists and metacarpophalangeal joints, but also the shoulders and ankle on a few occasions. Swelling was observed in several but not all attacks, especially in the hand joints, accompanied by erythema. No fever was documented. All but one of the attacks were monoarticular; the other affected two joints simultaneously. Laboratory data in our hospital confirmed RF positivity and anti–citrullinated protein antibody (ACPA) (CCP2 test) positivity at high titers (456 UI). Treatment with hydroxychloroquine was begun.

The most probable diagnosis of this patient is palindromic rheumatism, due to the typical presentation, but what other diagnoses should be considered? Can the evolution or prognosis be ascertained? Can the evolution to persistent arthritis, such as rheumatoid arthritis (RA) or other rheumatic diseases be predicted? What are the best management and therapeutic strategies?

In this article, we try to analyze these questions in accordance with scientific evidence and our own experience, focusing on the enigmatic relationship between PR and RA.

The first patient was recognized in 1928 by Philip S Hench in the Mayo Clinic, Rochester, USA. A 21-year-old women presented with hundreds of attacks of pain and swelling in different joints of very short duration (between 12 and 36 h), in general involving one joint at a time, with the patient being symptom-free between attacks. Hench observed 34 similar cases and, together with Edward Rosenberg, reported them in 1944 (1). The authors stated the attacks were unpredictable and, in some cases, periarticular/para-articular inflammation was observed. The clinical course varied considerably but no residual joint damage was observed. The authors analyzed entities with a similar clinical presentation and concluded that this was a unique, undescribed entity. Since no etiologic factor was found, and terms such as intermittent, recurrent or remitting rheumatism seemed to be unsatisfactory because they were too unspecific, the authors proposed “palindromic rheumatism” as a name. Palindromic is derived from the Greek palin dromein and means “to run back” or simply returning or recurring. The term was used by Hippocrates to denote erysipelas and other conditions that tend to repeat in the same individual. One of the article's advisors said, “What was good enough for Hippocrates ought to be good enough for you.” Since this first description, the designation of PR has remained until the present.

Hench and Rosenberg separated PR from RA due to the differentiating clinical relapsing/remitting course. Surprisingly, none of the patients in the original series evolved to persistent, chronic arthritis even after a long-term follow-up. By contrast, all subsequent series found that a significant proportion of patients evolved to RA or other chronic rheumatic disease. In 1959, Ansell and Bywaters, in a clinical meeting of the Heberden Society, presented a study in 28 patients with PR that showed that 18 patients evolved to chronic polyarthritis (13 were RF positive), and suggested that PR was merely a form of pre-RA, since almost all patients would progress to RA in the long term (2). However, the same authors also found that one patient remained as PR after 24 years of follow up.

The discovery that most, but not all, patients evolved to RA in the long-term (2, 3) and that a significant proportion had the same autoantibody profile: RF (4) and, more recently, ACPA (5), reinforced the close relationship between PR and RA. However, this relationship is not clear, and it uncertain whether PR is simply part of the RA continuum, or represents a pre-rheumatoid stage or is a separate entity (6–8). From today's perspective, PR probably represents a syndrome rather than a disease as cases with a similar phenotype of palindromic-like arthritis may display different mechanistic pathways with a different evolution and prognosis, although a high proportion of patients evolve to RA.

The diagnosis of PR is clinical, with a presentation of very short-lasting joint attacks with a relapsing/remitting course and no other explanation of the clinical symptoms. However, as there are no validated classification or diagnostic criteria in PR, there are wide variations in the clinical phenotype described in different series, although in all series a characteristic clinical presentation is described: patients are symptom free between attacks, joint swelling frequently presents with periarticular inflammation, and the recurrent episodes last <3 days in most cases (between 12–48 h in the majority) (6). The differential diagnosis of PR is quite wide due to the diseases that can cause an intermittent pattern of arthritis (9). In our opinion, the very spontaneous short-lasting episodes are a hallmark of PR and may differentiate it from other arthritis with an intermittent course (10) (Table 1). Whipple disease and autoinflammatory disorders such as familial Mediterranean fever (FMF) may have a similar clinical presentation of very short-lasting attacks (11, 12). The attacks only occasionally last for more than a week, a finding that prompted Barbero and Pasieri to determine this as the maximum permissible to consider in the diagnosis of PR (13). The distribution of the joint involved is similar to that seen in RA, with some differences between studies (6, 14–17) (Table 2).

Laboratory data show no evidence of biological parameters of inflammation such as CRP or ESR. Some authors suggest that these parameters are frequently increased during the flare, although evidence is scarce and not well-studied (6).

Imaging findings corroborate the absence of articular damage on conventional radiographs. Ultrasound studies between attacks have demonstrated a lack of significant subclinical synovitis in almost all patients (16, 18), confirming the intermittent, not persistent, nature of PR. Interestingly, an extracapsular, periarticular inflammation pattern was observed during clinical attacks, often without accompanying synovitis, which differs from that observed in RA, where intrasynovial inflammation is encountered (18). Other studies mainly found joint synovitis during flares (16, 19). However, the presence of this intrasynovial inflammation in PR during flares has been associated with the presence of ACPA (20) and higher risk for RA progression (18).

The four proposed classification criteria for PR are summarized in Table 3 (6, 13, 15, 21). All suggest that no other diagnosis explains the clinical symptoms and the observation of a clinical episode by a physician in a clinical examination is mandatory. This second statement is of interest, although it may be substituted, in some instances, by a good current mobile imaging photograph. No set of criteria has been validated in clinical studies.

It is established that a significant number of patients with PR show a similar serum autoantibody profile to that of RA. RF and ACPA are present in 39–68% of patients in different series, frequently at high titers (5, 16–18, 22–24). Only two Asian studies reported a surprisingly-low prevalence of the two antibodies, suggesting different inclusion criteria or ethnic differences (20, 25). Recently, anti-carbamylated protein (anti-CarP) antibodies have been documented in a quarter of longstanding PR patients (26). An analysis by our group of the antibody immune response to posttranslational modified antibodies (citrullinated or homocitrullinated) showed a more restricted response in PR than in established RA, with less use of the IgM or IgG isotypes (26, 27). This antibody profile resembles that observed in the preclinical phases of RA (28) or in unaffected relatives of patients with RA (29), suggesting the differing B cell immune response, at least in some patients with PR, may play a pathogenetic role, with less probability of evolving to persistent polyarthritis. Whether this antibody profile changes over the PR disease course, with more antigenic expansion or greater isotype distribution, as occurs in RA, in patients who evolve to RA is unclear.

A possible role of autoinflammation in PR has been suggested in view of the clinical presentation of diseases with this relapsing/remitting course that resemble those observed in autoinflammatory disease, such as FMF. MEFV mutations have been observed in 12% of PR patients in a multicenter study, an unexpectedly high prevalence that is more evident in patients without ACPA (23%) (30). This study showed that some cases with a clinical presentation compatible with PR without autoantibodies may be related to autoinflammatory genes, even without other diagnostic criteria.

Observational and retrospective series have confirmed the high rate of disease evolution to other chronic rheumatic diseases in PR, with RA by far the most common. The rate of progression to RA is between 10 and 66%, depending on the inclusion criteria, follow-up time and possible therapy (5, 22–24, 31–33). The highest percentage was found in a Scandinavian study of patients followed for more than 20 years (33), although the highest ratio was observed during the first years of evolution (31, 33). It is suggested that 18% of patients with established RA had a history of PR (34). In these patients, RA developed a mean of 1.2 years after initial palindromic symptoms (Figure 1). No particular phenotype emerged in these patients, although palindromic attacks may persist during the disease course of established RA (34). Lower levels of other diseases have been described in the follow-up of patients with PR, including Sjogren syndrome, psoriatic arthritis, vasculitis, myositis, Behçet disease and, most commonly, systemic lupus erythematosus (SLE) (9).

Several prognostic factors for RA progression have been reported, including demographics (female sex) (32), genetics (homozygosity for the shared epitope) (35), the clinical phenotype (hand involvement) (23, 24, 32) and autoantibodies (ACPA and RF) (22–24). ACPA, the most specific serological marker for RA, has been associated with the highest risk for progression to RA, as expected, although we have shown that a significant proportion of ACPA-positive PR patients do not evolve to RA after a long term follow-up (31). Almost all studies were designed to find associations rather than predictive factors, and the limitations included different geographic populations, a small number of cases, a short follow-up time, and different therapies, while some were carried out decades ago. All these factors limit the interpretation of the results and make it difficult to establish the risk for individual progression to RA.

A significant number of antirheumatic drugs have been used in PR, all in case series or observational studies, but no randomized clinical trials have been carried out (36). Therefore, the management of PR is empirical. NSAIDs are used to treat acute attacks with variable results (36). Glucocorticoids may be useful but have been shown to be efficacious only in anecdotal reports (37). Gold salts were effective in reducing the number and intensity of attacks in older studies (3). The same is observed with antimalarials, probably the most widely investigated and used drugs to treat PR to date (36). The efficacy on clinical symptoms is demonstrated in observational studies (38), and currently hydroxychloroquine is probably the drug of choice for PR. One study found that antimalarials may increase the time of progression to RA but not avoid it (15). Methotrexate, the most widely-used drug in the treatment of RA, has not been tested in PR, although a recent observational, uncontrolled study used methotrexate as part of a treat-to-target strategy that resulted in substantial clinical benefits, with low rates of progression to RA (39). A surprisingly high rate of clinical response was observed with rituximab in an Indian survey (40). Colchicine is effective in seronegative PR patients with MEFV mutations (30). A multicenter randomized clinical trial comparing the efficacy of abatacept vs. hydroxychloroquine in recent-onset seropositive PR is ongoing (ClinicalTrials.gov Id: NCT03669367).

Several unmet needs in PR may be identified: (1) concerns about the clinical case definition and diagnostic criteria, (2) the enigmatic and not well-understood relationship with RA, and (3) the best management and therapeutic strategy to improve acute attacks and prevent the evolution to RA.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s.

RS, JC, and IH contributed to the conception and manuscript design and wrote the first version of the manuscript. RC-M, SC-V, JR, GS, BF-S, and RM takes part of the different studies referenced in the manuscript and revised it critically. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.