Reactive arthritis secondary to gastrointestinal infection is sometimes bracketed with spondyloarthritis associated with inflammatory bowel disease as enteropathic arthritis.¹¹ However, although links between the gut and arthritis are very important pathologically, particularly in spondyloarthritis, enteropathic arthritis is really an ill-defined overlap term, which often includes other forms of arthritis that do not have classic features of spondyloarthritis but occur in relation to gastrointestinal disorders. Examples include Whipple’s disease and celiac disease.

Undifferentiated Spondyloarthritis

Obvious overlap has been noted between different members of the spondyloarthritis family. By definition, patients with undifferentiated spondyloarthritis have arthritis that fails to satisfy diagnostic or classification criteria for one of the other forms. Thus, it may not be legitimate to regard undifferentiated spondyloarthritis as a separate “disease” because over time, patients often develop new features, which means that their disease is no longer “undifferentiated.” The commonest of these is development of radiographic changes in the sacroiliac joints that allow the patient to satisfy criteria for ankylosing spondylitis. This occurs in approximately 60% of patients in many series,¹² and all ankylosing spondylitis patients will pass through an “undifferentiated” phase (often termed axial spondyloarthritis) if seen before developing structural changes on radiographs. Likewise, patients may develop psoriasis at some point after the onset of their spondyloarthritis. In cases where there is a family history of psoriasis or features such as dactylitis, which is very common in psoriatic arthritis, the term psoriatic arthritis sine psoriasis is sometimes used, but if patients fail to meet diagnostic criteria for psoriatic arthritis (such as the recently devised Classification Criteria for Psoriatic Arthritis [CASPAR]¹³), they should be classified as having undifferentiated spondyloarthritis. Note that in the absence of a personal or family history of skin or nail psoriasis, patients fulfill the CASPAR criteria for psoriatic arthritis only if they have dactylitis and juxta-articular new bone formation on hand or foot radiographs, and lack rheumatoid factor. In addition, underlying inflammatory bowel disease may declare itself clinically only at some point after the onset of undifferentiated spondyloarthritis.

Finally, as discussed in the following section, it may be difficult to make a certain diagnosis of reactive arthritis, particularly when evidence of a triggering infection is absent or incomplete, resulting in significant overlap between reactive arthritis and undifferentiated spondyloarthritis—hence their being considered together in this chapter.

Classification Criteria for Reactive Arthritis and Undifferentiated Spondyloarthritis

Because both reactive arthritis and undifferentiated spondyloarthritis are members of the spondyloarthritis group, patients can first be identified as having spondyloarthritis by means of rather wide classification criteria. It must be recognized that classification criteria are devised to allow homogeneous sets of patients with particular features to be defined. This is critical for research studies, ensuring that different investigators report on the same set of patients. Thus classification criteria are not, and should not be used as, diagnostic criteria; clinicians will certainly make confident clinical diagnoses for patients who fail to fulfill particular classification criteria. Nevertheless, such criteria often serve as a convenient checklist of the features that are usually present in a given disease.

The two best known classification criteria for all forms of spondyloarthritis considered together are those devised by Amor ¹⁴ (Table 76-2) and the later European Spondyloarthropathy Study Group (ESSG) criteria¹⁵ (Table 76-3). The Amor criteria allocate “points” to 12 features characteristic of spondyloarthritis, and classification is based on reaching a specified total score of 6 points. The ESSG criteria are applied to patients with inflammatory back pain or oligoarthritis, with the presence of additional features required for classification. The sensitivity and specificity achieved by the Amor and ESSG criteria vary in different series and inevitably depend on the population to which they are applied, but these values are usually around 80% and 90%, respectively. A more recent development in the classification of spondyloarthritis has been the separation of axial (spine and sacroiliac) inflammation from peripheral arthritis, with recognition that in different forms of spondyloarthritis, one or another of these may predominate, although with time many patients will develop both. To aid classification of axial spondyloarthritis, which traditionally requires radiographic sacroiliitis, as exemplified by the modified New York criteria for ankylosing spondylitis (AS),¹⁶ magnetic resonance imaging (MRI) changes of sacroiliitis have been included in the recently published Assessment of SpondyloArthritis international Society [ASAS] criteria¹⁷) (Table 76-4). MRI sacroiliitis is carefully defined as bone marrow edema on short tau inversion recovery (STIR) sequences, or as osteitis on T1 images with contrast medium, localized to subchondral or periarticular bone marrow. The changes are also required to be multiple or present in at least two consecutive slices. Therefore, patients younger than 45 years of age, with at least 3 months’ history of back pain or sacroiliitis on MRI (or on radiographs), plus at least one feature from the list of spondyloarthritis-associated signs and symptoms shown in Table 76-4, fulfill the criteria. In the absence of sacroiliitis, HLA-B27⁺ individuals can still be classified as having axial spondyloarthritis when two or more of these features are present. These ASAS criteria represent a recent and welcome development, but it remains to be determined what the specificity and sensitivity of MRI changes in sacroiliac joints will be when used in routine practice in nonresearch settings.

Table 76-2 Amor Criteria for Spondyloarthritis

Clinical Symptoms or Past History of:	Points
Lumbar or dorsal pain during the night, or morning stiffness of lumbar or dorsal spine	1
Asymmetric oligoarthritis	2
Buttock pain	1
Alternating buttock pain	2
Dactylitis of finger or toe	2
Heel pain or other well-defined enthesopathy	2
Iritis	2
Nongonococcal urethritis or cervicitis within 1 mo of arthritis onset	1
Acute diarrhea within 1 mo of arthritis onset	1
Psoriasis, balanitis, or inflammatory bowel disease	2
Radiology
Sacroiliitis (grade ≥2 if bilateral; grade ≥3 if unilateral)	3
Genetic Background
HLA-B27⁺ or family history of ankylosing spondylitis, reactive arthritis, uveitis, psoriasis, or inflammatory bowel disease	2
Response to Treatment
Good response to NSAIDs within 48 hr, or relapse within 48 hr if NSAIDs withdrawn	2

For a definitive diagnosis of spondylarthritis, ≥6 points is required; 5 points indicates probable spondyloarthritis.

NSAIDs, nonsteroidal anti-inflammatory drugs.

Table 76-3 European Spondyloarthropathy Study Group Criteria for Spondyloarthritis

Table 76-4 Assessment of SpondyloArthritis international Society Classification Criteria for Axial Spondyloarthritis (SpA)

For patients with back pain for ≥3 mo and aged <45 yr:
Sacroiliitis on imaging	+	≥1 SpA feature
or
HLA-B27	+	≥2 other SpA features
Sacroiliitis on imaging defined as: Active acute inflammation on magnetic resonance imaging highly suggestive of sacroiliitis associated with SpA or Definitive radiographic sacroiliitis according to the modified New York criteria SpA features comprising: Arthritis Enthesitis (heel) Inflammatory back pain Dactylitis Uveitis Psoriasis Inflammatory bowel disease Good response to nonsteroidal anti-inflammatory drugs Family history of SpA HLA-B27 Elevated C-reactive protein

The main driver for developing the ASAS criteria for axial spondyloarthritis has been the need to define early ankylosing spondylitis without requiring radiologic changes in the sacroiliac joints. Because these often take years to develop, new criteria were required to identify, investigate, and treat patients with early ankylosing spondylitis. Nevertheless, not all patients who fulfill the criteria for axial spondyloarthritis go on to develop ankylosing spondylitis, that is, early ankylosing spondylitis is a subset of axial spondyloarthritis. Patients who do not progress to radiographic changes or other clinical features of ankylosing spondylitis will continue to be classified as having undifferentiated spondyloarthritis.

The new ASAS criteria for axial disease are clearly useful for identifying patients with undifferentiated spondyloarthritis. In reactive arthritis, acute inflammatory back pain is common, but no studies have established the frequency of MRI-defined sacroiliitis in this disease, or of MRI-defined changes in the spine, although the latter are not used in the ASAS classification. Thus the ASAS criteria for axial spondyloarthritis are not likely to be very useful in reactive arthritis.

More relevant to reactive arthritis are the ASAS criteria for classification of peripheral spondyloarthritis, which have also been published and discussed recently ^18,¹⁹ (Table 76-5). These are applied to patients who present with arthritis, particularly asymmetric oligoarticular lower limb arthritis OR enthesitis OR dactylitis. These patients are classified as having peripheral spondyloarthritis if they have one additional feature from the list shown in Table 76-5. In the absence of any of these features, two features from an additional list are required: arthritis, enthesitis, dactylitis, inflammatory back pain, or family history of spondyloarthritis. Note that because these criteria are applied to patients who present with arthritis or enthesitis or dactylitis, this second list gives weight to the occurrence of these symptoms in combination. These criteria achieved a sensitivity of 78% and a specificity of 83%, representing a compromise between lower sensitivity and greater specificity of the Amor or ESSG criteria. How the ASAS criteria will fare in more general use remains to be seen; they were developed in a population in whom 85% were younger than age 45, and 30% had radiographic changes in sacroiliac joints that met modified New York criteria for ankylosing spondylitis, even though they did not have back pain at recruitment.

Table 76-5 Assessment of SpondyloArthritis international Society Classification Criteria for Peripheral Spondyloarthritis

Used together, ASAS criteria for axial and peripheral spondyloarthritis should classify all those patients who have some form of spondyloarthritis (again recalling that some patients will be diagnosed with spondyloarthritis even though they do not meet these classification criteria). Of those who meet ASAS criteria, some will have documented psoriasis or inflammatory bowel disease, or radiologic sacroiliitis, and therefore can be classified as having a particular form of spondyloarthritis. The remainder will have undifferentiated spondyloarthritis or reactive arthritis. In relation to diagnosis, checking the features included in the classification criteria will allow the clinician to identify patients in whom some form of spondyloarthritis is likely, and will prompt additional clinical examination and investigations that may clarify the diagnosis or even elicit additional criteria.

In relation to reactive arthritis, there is no wholly satisfactory definition of the disease.²⁰ One practical proposal is shown in Table 76-6. In the setting of an outbreak of food poisoning, when all those known to have been infected with Salmonella or Campylobacter can be followed up, all who develop new joint symptoms can be regarded as having reactive arthritis, but inevitably this also includes patients with only arthralgias.^21,²² Joint symptoms have a high background prevalence in the community, and it is difficult to be certain that those reported by patients following infection are genuinely related to the infection and not to pre-existing conditions that assume new prominence in the context of an infection, or to questionnaires seeking such symptoms.

Table 76-6 Proposed Definition of the Diagnosis of Reactive Arthritis