Objectives To investigate the prognostic value of the lymphocytic focus score (LFS) and the percentages of IgA+, IgM+ and IgG+ plasma cells for disease severity of primary Sjögren syndrome (pSS).
Methods Medical charts of 174 pSS patients were retrospectively analysed, comparing histology results (LFS and percentages of IgA+, IgM+ and IgG+ plasma cells) with disease outcomes as non-Hodgkin lymphoma (NHL) and clinical scores including cumulative EULAR (European League against Rheumatism) Sjögren syndrome disease activity index (ESSDAI) and the total number of extraglandular manifestations.
Results The mean LFS was significantly higher in patients developing NHL (3.0±0.894 vs 2.25±1.086; p=0.021). The threshold of ≥3 foci has a positive predictive value of 16% for lymphoma, and a negative predictive value of 98%. Only LFS ≥3 contributed significantly and independently to NHL development in a standard multiple regression model. Ig class distribution of plasma cells did not help to identify patients developing lymphoma. Patients with LFS ≥3, ≤40% IgA+ or ≥25% IgM+ plasma cells in salivary gland biopsy specimens had significantly enhanced systemic disease.
Conclusions Routine histopathological minor salivary gland assessment has important prognostic value. The LFS might help to identify patients with an increased risk for lymphoma.
- Sjøgren's Syndrome
- Outcomes Research
- Disease Activity
Statistics from Altmetric.com
Primary Sjögren syndrome (pSS) is a chronic autoimmune disease with symptoms of dryness due to inflammation of lacrimal and salivary glands, which may be accompanied by extraglandular manifestations (EGM) like arthritis and vasculitis. The disease is characterised by B cell hyperactivity, as shown by the presence of autoantibodies, enhanced serum immunoglobulin levels, and robust focal infiltration of exocrine glands with B lymphocytes in addition to large numbers of T cells. The presence of Sjögren syndrome (SS) antibodies (anti-SSA/SSB) and lymphocytic focus score (LFS) are both part of the pSS classification criteria and are of diagnostic relevance.1 ,2
The prognostic relevance of analysis of salivary gland infiltrates has recently been highlighted by the observation that organisation of focal lymphocytic infiltrates into germinal centres (GCs) is associated with higher risk of non-Hodgkin lymphoma (NHL) and systemic disease.3 The number of foci in labial salivary glands has been shown to be positively associated with GC formation.4 These observations underscore the importance of routinely acquiring minor salivary gland biopsies.5 Routine histology includes assessment of the number of lymphocytic foci. In addition, several labs apply quantitative immunohistology (QIH) to assess the percentages of immunoglobulin specific plasma cells, with IgA+ plasma cells <70% and/or IgM+ plasma cells >10% as positive criteria for SS.6 ,7 This latter assessment has been given modest attention.8 ,9
The aim of the present study was to investigate the prognostic value of minor salivary gland assessments (LFS and plasma cell Ig class distribution) which are routinely performed in daily practice for pSS diagnosis.
We retrospectively analysed, according to the 2002 classification criteria, the medical charts of all patients with pSS who attended our outpatient clinic for at least 1 year in the period 1998 to July 2011.1 Of the 195 patients described in more detail elsewhere,10 ,11 174 patients had documented minor salivary gland histology and were included in this analysis. The retrospective and anonymous nature of this report made ethical approval unnecessary, following the ethics guidelines of our hospital.
Assessment of disease severity
Disease manifestations were scored with the ‘cumulative ESSDAI’ (cumulative EULAR (European League against Rheumatism) Sjögren syndrome disease activity index), summating the maximum ESSDAI scores achieved per organ domain at any time point during disease.10 The ESSDAI scores 12 organ domains on the severity of involvement, ranging from 0 to 3 points. These points are then multiplied by an assigned weight factor, ranging from 1 to 6. Total scores can range from 0 to 123 points.12 Additionally, total EGM, using the same definitions as the ESSDAI but without a weight factor, occurring at any time point during disease were summated. For calculating ESSDAI/EGM scores, missing laboratory data were interpreted as being normal/absent if no value was available for any time point, the last value was carried forward, or the first value was carried backward. Since it can be questioned whether lymphoma and manifestations occurring during its course should be scored under pSS,11 scores of the cumulative ESSDAI and EGM are displayed with and without counting those manifestations most likely related to NHL. The diagnosis of NHL was made according to the WHO classification with staging following the Ann Arbor system.
Minor salivary gland assessments
Lymphocytic foci were scored on H&E stained sections.13 IgA+, IgM+ and IgG+ plasma cells were stained with their respective antibodies and their percentages were manually analysed in the same, representative part of the specimen.6 ,8 ,9
Analyses were carried out with SPSS V.17.0, using the Mann–Whitney U, χ2 test or Spearman's rank correlation when appropriate. Standard multiple regression was used to assess the ability of diagnostic parameters to predict pSS disease severity. For histology, parameter cut-offs were made based on group means (LFS ≥3, IgA+ plasma cells ≤40%, IgM+ plasma cells ≥25%). A two-sided p value <0.05 was considered significant.
Correlation between LFS and percentage of IgA+ plasma cells
In total, 174 pSS patients with minor salivary gland assessments were included (table 1).
LFS ≥1 were seen in 99%, <70% IgA+ plasma cells in 90% and >10% IgM+ plasma cells in 71% of patients. The number of foci correlated to a decrease in the percentage of IgA+ plasma cells (R=0.315, p=0.0001), but not to the percentages of IgM+ or IgG+ plasma cells. Additionally, patients positive for either anti-SSA, anti-SSB, rheumatoid factor (RF) or antinuclear antibodies (ANA) had significantly higher LFS (mean 2.4–2.7 vs 1.6–2.0; p≤0.002) and lower percentages of IgA+ plasma cells (mean 37–42% vs 48–55%; p≤0.007). Percentages of IgM+ or IgG+ plasma cells did not differ between antibody-positive and antibody-negative patients. Increased IgG correlated with an increased LFS (R=0.377, p<0.0001) and a decreased percentage of IgA+ plasma cells (R=−0.265, p=0.002), but not with percentages of IgM+ or IgG+ plasma cells.
Lymphoma development is predicted by lymphoid neogenesis in minor salivary glands
NHL developed in 16 patients (9%) a median of 68 months after salivary gland biopsy. The mean LFS was significantly higher in patients with NHL (3.0±0.9 vs 2.3±1.1; p=0.021). Mean percentages of Ig specific plasma cells did not significantly differ between patients with or without NHL development. The threshold of ≥3 foci had a positive predictive value of 16% for development of lymphoma, and a negative predictive value of 98% (OR 7.9; p=0.008). With a total of 1169 person-years at risk, the incidence rate of NHL was 3.7 (LFS=1), 2.4 (LFS=2), and 12.8 (LFS≥3) per 1000 person-years, respectively. Only LFS≥3 contributed significantly and independently to NHL development in a standard multiple regression model (β=0.244; p=0.017) (table 2).
Correlation between clinical scores and minor salivary gland assessments
Cumulative ESSDAI and EGM were significantly correlated to LFS, a decreased percentage of IgA+ and an increased percentage of IgM+ plasma cells (R=0.166–0.284; p≤0.04), but not to the percentage of IgG+ plasma cells. Patients with ≥3 foci, ≤40% IgA+ plasma cells or ≥25% IgM+ plasma cells in salivary gland specimens had significantly more systemic disease (table 3, only LFS and percentage of IgA+ plasma cell results are shown).
Lymphoma and systemic disease are important clinical problems complicating the course of pSS. Our study shows that results of routinely performed minor salivary gland assessments are useful to predict these adverse outcomes at the time of diagnosis.
The positive and negative predictive value for lymphoma development of LFS ≥3 was in the same order as the results for GC positivity (16% and 99%, respectively) recently presented in a study with comparable patient numbers.3 We retrospectively assessed H&E stained minor salivary glands of 13 pSS–NHL patients and observed GC morphology in only 3 (23%). This is much lower than the 6/7 (86%) patients with lymphoma positive for GCs reported previously.3 Until now, the technique to be used to detect germinal centres and how they should be defined remains unclear and open for debate.4 This makes comparison between results from different studies difficult and causes confusion with respect to the application of GC positivity as a marker in routine biopsy work-up. International guidelines on GC assessment are therefore necessary before implementation into clinical practice will be possible. In the meantime, using the LFS not only for diagnosis but also for prognosis can be implemented directly into daily practice and seems just as informative.
Plasma cell Ig class distribution was less informative for prognosis than the number of lymphocytic infiltrates (LFS). This might be due to the direct importance of the interplay between B cells and other inflammatory cells (eg, T cells, myeloid cells, dendritic cells) in lymphocytic foci contributing to lymphoma development, whereas plasma cells are less likely to be important for this process.
A limitation of our study is its retrospective design, which increases the number of missing data. Several patients were eventually lost to follow-up; the Netherlands do not have national registries to retrieve lost patient outcomes like lymphoma. Additionally, patients with EGM and evident SS are less likely to have a salivary gland biopsy performed, because it is deemed unnecessary for their diagnosis.10 ,11 This group of patients is therefore underrepresented in histology studies. Immunomodulating medication was used in a minority of patients, but likely had no influence on salivary gland histology. Although the evidence is sparse, the number of lymphocytic foci seems to remain stable when using immunomodulating therapy. Prednisone usage of 30 mg on alternate days for 6 months showed no effect on histology in a randomised, double-blind, placebo-controlled study.14 For hydroxychloroquine no evidence is available. Patients on immunomodulating medication and thus with higher disease activity have probably a higher focus score and increased risk of lymphoma based on their disease process, irrespective of medication usage. Prospectively designed studies comprising a large and complete Sjögren cohort will be necessary to reveal the prognostic strength of our findings.
In conclusion, routinely performed minor salivary gland assessments have important prognostic value for clinical practice in addition to their diagnostic value. The number of lymphocyte foci can be used to identify patients with increased or very low risk of lymphoma.
We would like to thank JM van Woerkom MD PhD, and the pathology department of the University Medical Center Utrecht for their work on the minor salivary gland assessments.
Handling editor Tore K Kvien
Contributors APR: design, data collection, analyses, manuscript preparation; AAK: design, correction of analyses and manuscript; RG: design, data collection (histology), correction of analyses and manuscript; FPJGL: design, correction of analyses and manuscript; JWJB: design, correction of analyses and manuscript; JAGvR: design, correction of analyses and manuscript.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement The complete dataset is available to fellow researchers on request.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.