PoM No 3,2013

Publication of the Month

March 03/13: Classification of anti-neutrophil cytoplasmic autoantibody vasculitides – towards a more useful system

 

PART 1:

Key messages:  

  • Current classification is controversial and clinically unhelpful
  • Usefulness would be significantly improved by addition of specific ANCA test results.  

Lionaki S, Byth E, Hogan S, Hu Y, Senior B, Jennette C, Nachman P, Jennette J, Falk, R
Classification of Antineutrophil Cytoplasmic Autoantibody Vasculitides
Arthritis Rheum 2012; 64: 3452-3462  

Background: There are currently two options for the nomenclature and definition of vasculitis – the 1994 Chapel Hill Consensus Conference (CHCC) and the 2007 European Medicines Agency (EMA) which refined and expanded the 1990 American College of Rheumatology classification system.  However, neither system provides diagnostic criteria that enable practicing physicians to differentiate microscopic polyangiitis (MPA) from granulomatosis with polyangiitis (GPA – formerly known as Wegener’s).  Nor do they offer any clues as to the prognosis and possible outcomes of treatment in patients.  This retrospective study looks at the characteristics of a large (502) cohort of patients where access to individual outcomes allows an analysis of which aspects proved more useful in distinguishing phenotypes and prognosis.  Pre-selection of the cohort on the basis of identifiable anti-neutrophil cytoplasmic antibody (ANCA) positivity meant that the antibody specificity (either for myeloperoxidase – MPO ANCA or proteinase -3 – PR3 ANCA) could also be assessed as a potential tool.  

Summary: The study cohort was statistically assessed for a multitude of characteristics including demographics, organ involvement, treatment, response to treatment, biochemical and serological markers and eventual outcome.  The authors found a surprisingly substantial discrepancy in the allocation of patients into specific clinical pathologic categories between the two systems and confirmed that neither could predict patient outcome.  Classification using ANCA specificity alone had the best predictive model fit with PR3 ANCA positive patients almost twice as likely to relapse as those with MPO ANCA positivity.  There was also a strong correlation between ANCA specificity and manifestations of vasculitis involving multiple anatomic sites and tissues.  

Conclusions: The existing classification systems for AAV result in significant differences in categorization of patients.  The use of a classification system to aid clinicians in treatment decisions would be enhanced by the use and analysis of specific ANCA testing.  

Comment: The group sought to improve the AAV classification system to standardise describing of patient groups, not only to facilitate categorization for clinical trials and comparisons but ”above all to assist in patient care”. They were open to consideration of many characteristics but their finding that “Because ANCA specificity not only is implicated in pathogenesis and correlates with clinical symptoms, but more importantly helps predict the outcome of disease, it is appropriate to include ANCA specificity in the diagnostic classification”.  

 

PART 2:

 Key messages:   

  • A new method of classifying AAV patients using clinical variables and ANCA specificity is suggested. This method would divide patients into clinically relevant sub-groups and give an indication of likely outcome.  

Mahr A, Katsahian S. Varet H, Guillevin L, Hagen EC, Höglund P, Merkel P, Pagnoux C, Rasmussen N, Westman K, Jayne D.
Revisiting the classification of clinical phenotypes of anti-neutrophil cytoplasmic antibody-associated vasculitis: a cluster analysis.
Ann Rheum Dis, 2012, doi: 10.1136/annrheumdis-2012-201750  

Background: Another group concerned with the usefulness of the available systems for the classification of AAV decided on a different approach to the issue.  Recognising that the overlapping features of MPA and GPA have meant that they are becoming frequently combined under the term AAV, the authors sought to divide patients with AAV into subgroups that would allow a better standardisation in clinical trials but would also give more accurate indicators as to prognosis and outcome for the patients.  They decided to use a data-driven cluster analysis.  

Summary: The study also used a large (673) group of subjects who had been newly diagnosed with either GPA or MPA. (Note that the diagnosing clinicians appear to have used both the Chapel Hill and ACR classification systems).  Cluster analysis was based on two models. One model included nine clinical baseline variables as input variables while a second model additionally included ANCA specificities.  Although the statistical processing is complex, the models were able to correctly assign nearly 97% of patients to the five categories on the basis of 6 variables (Model 1) or just 4 variables (Model 2).  

Conclusions: The categories, named:
Renal AAAV with Proteinase 3 (PR3)-ANCA (40% of subjects)
Renal AAV without PR3-ANCA (32%)
Non-renal AAV (12%)
Cardiovascular AAV (9%) and
Gastrointestinal AAV (7%)
-      allowed patients to be separated into clinically relevant groups which also had distinct death and relapse rates.   

Comment: Using just four clearly defined variables, it is possible to classify AAV patients into five classes associated with distinct phenotypes of the disease and giving an indication of likely outcome.  

 

As in all diagnostic testing, the diagnosis is made by the physican based on both test results and the patient history.