Eurofever Criteria Bring Structure to Autoinflammatory Recurrent Fevers
Hereditary recurrent fevers have long challenged clinicians because the syndromes overlap at the bedside while diverging sharply in genetics, trigger patterns, and long-term risk. Familial Mediterranean fever, TRAPS, mevalonate kinase deficiency, and cryopyrin-associated periodic syndromes can all present with episodic inflammation, yet each disorder carries distinct implications for diagnosis, counseling, and treatment selection. PFAPA adds another layer of complexity because it is periodic, inflammatory, and recurrent, but not hereditary in the same sense as the monogenic syndromes. In that setting, classification rules are not a paperwork exercise; they are the framework that determines which patients are grouped together for research and which signals are lost in the noise of symptom overlap.
The Eurofever effort set out to build evidence-based classification criteria for hereditary recurrent fevers and PFAPA by combining clinical reasoning with statistical validation. The work moved through several sequential steps: candidate clinical, laboratory, and genetic variables were selected; 360 randomly chosen patients from the Eurofever Registry were reviewed by a blinded expert panel of 25 clinicians and 8 geneticists; statistical methods were then used to identify the strongest candidate criteria; a larger consensus conference with 33 panellists refined the final rules; and the resulting criteria were tested again in a cross-sectional validation set. The design matters because recurrent fever syndromes are not diagnosed from one isolated marker. They require pattern recognition anchored by genetics, age of onset, attack features, and the inflammatory phenotype itself.
Consensus Classification Across Four Hereditary Syndromes and PFAPA
Among the 360 registry patients reviewed in the consensus phase, 281 patients, or 78%, were ultimately classified by the panel. The cohort included 32 patients with CAPS, 36 with FMF, 56 with MKD, 37 with PFAPA, 39 with TRAPS, and 81 with undefined recurrent fever. Those numbers matter because they show both the reach and the limits of current clinical labeling. Even among a curated registry population, a substantial group remained unclassified, which reflects the real-world difficulty of assigning a single diagnosis when fever syndromes present with partial phenotypes or incomplete genetic clarity.
Two sets of criteria were generated for each hereditary recurrent fever syndrome: one incorporating genetic and clinical variables, and another relying on clinical variables only. PFAPA received newly developed criteria as well. In practice, that dual-track structure is important because genetic confirmation is not always available, not always definitive, and not always fast enough to guide the initial classification of a patient with recurrent inflammatory attacks. A clinician at the bedside often faces the first question before sequencing results arrive: does this child or adult fit a monogenic periodic fever syndrome, PFAPA, or something else entirely?
The answer now has a more disciplined evidence base. The four hereditary recurrent fever criteria achieved sensitivity ranging from 0.94 to 1 and specificity ranging from 0.95 to 1. PFAPA criteria reached sensitivity of 0.97 and specificity of 0.93. Those are strong operating characteristics for disorders that often look similar during fever flares, when laboratory inflammation can be broad and symptoms can blur together. High sensitivity reduces the chance of missing true cases, while high specificity helps avoid sweeping unrelated febrile illnesses into the wrong classification bucket.
Why Clinical and Genetic Tracks Were Kept Side by Side
The decision to build both genetic-inclusive and clinical-only versions reflects the biological reality of autoinflammatory disease. Some patients carry recognizable pathogenic variants, and in those cases genotype can sharpen the classification. Others may have suggestive phenotypes without immediately informative genetics, especially when testing is unavailable, delayed, or inconclusive. By preserving a clinical-only pathway, the criteria remain usable in settings where access to molecular testing is limited, while the genetic-supported version can tighten classification when genomic data are present.
This is not a minor methodological detail. In autoinflammatory disease, genotype and phenotype do not always move in perfect lockstep. Expressivity varies, penetrance varies, and some patients present with a clinical picture that is sufficiently distinct to classify even before molecular confirmation. Conversely, a genetic result without the right clinical context can mislead. The Eurofever structure acknowledges both realities and builds them into the same framework instead of forcing one domain to dominate the other.
Validation Strength Across an Independent Cohort
The final test came in an independent data set of 1,018 patients, where validation accuracy ranged from 0.81 to 0.98. External validation is the real stress test for any classification system because it shows whether a rule set can survive outside the cohort in which it was crafted. Here, the performance remained high across a much larger group, suggesting that the criteria are not merely tailored to a single registry or a narrow expert consensus.
The validation result is particularly relevant for recurrent fever syndromes because the field has long depended on criteria that vary across diseases and across centers. A patient with episodic fever, serositis, aphthous ulcers, or pharyngitis may drift between labels depending on which specialist evaluates them first. A common validated set of criteria does not eliminate diagnostic uncertainty, but it gives research cohorts a more consistent foundation and gives clinicians a better way to sort patients who likely belong to a specific autoinflammatory category.
What the New Criteria Change at the Bedside
For clinicians, the immediate value lies in triage and classification. A patient who fits one of these criteria is more likely to be studied in the right disease group, which improves the quality of future genetic, mechanistic, and therapeutic research. For patients, the downstream consequence is more precise counseling about recurrence patterns, inheritance risk, and the rationale for targeted therapy. For PFAPA, a validated set of criteria also helps separate a common non-hereditary periodic fever syndrome from monogenic disorders that may need different treatment strategies and different long-term follow-up.
The broader biological lesson is straightforward. Autoinflammatory recurrent fevers are driven by dysregulated innate immune pathways, not by a single generic fever mechanism. Classification criteria work best when they reflect that biology rather than flattening it into a nonspecific fever syndrome. By combining clinical features with genetic information where available, Eurofever created a structure that mirrors how these illnesses actually present in practice: recurrent, overlapping, and often ambiguous until the pattern is assembled carefully.
The remaining challenge is implementation. These criteria are validated, but they still depend on consistent clinical data capture, access to relevant laboratory and genetic information, and thoughtful use by clinicians who understand where one syndrome ends and another begins. In patients with undefined recurrent fever, the criteria will not replace judgment, and they do not erase the need for longitudinal observation when the phenotype is incomplete. They do, however, give the field a more reliable standard for separating hereditary recurrent fevers and PFAPA from the broader pool of recurrent inflammatory illness.