Risk of Non-Hodgkin's Lymphoma in Primary Sjögren's Syndrome: A Population-Based Study




Primary Sjögren's syndrome (SS) is associated with an increased risk of non-Hodgkin's lymphoma (NHL), but the reported prevalence and risk vary considerably. The objective of this study was to determine the risk of NHL in a well-defined population-based primary SS cohort in Norway.


The authors examined all patients fulfilling the American–European Consensus Group criteria for primary SS from 2 Norwegian counties and compared the data to the Cancer Registry of Norway to identify the primary SS patients who had lymphoma. In addition, lymphoma patient files from the same period were reviewed for undiagnosed primary SS to ensure the quality of registry data.


As of July 1, 2009, 443 living subjects with primary SS were identified in an area with 896,840 inhabitants, which is 18.6% of the total population of Norway. Seven cases of NHL (1.6%) were found during a total followup of 3,813 person-years, resulting in a standardized incidence ratio of 9.0 (95% confidence interval 7.1–26.3) for NHL in primary SS patients.


The risk of NHL in patients with primary SS in Norway is increased 9 times compared with the general population. This is in accordance with recent studies, and the quality and completeness of the registries and strict use of diagnostic criteria support the validity of the results.


Primary Sjögren's syndrome (SS) is a chronic autoimmune disease histopathologically characterized by focal lymphocytic infiltration and destruction of the exocrine glands and clinically characterized by dry mouth, dry eyes, and systemic and focal extraglandular features (1). Grading the lymphocytic cell infiltration in labial salivary gland biopsies is part of the classification criteria for primary SS (2). The number of mononuclear cell aggregates containing >50 cells per 4 mm2 is the focus score, and a score ≥1 is considered a positive finding. Ectopic lymphoid germinal centers in labial salivary gland biopsies, autoantibodies to SSA/Ro and SSB/La ribonucleic particles, hypergammaglobulinemia, cryoglobulinemia, and rheumatoid factor are common findings indicating a general increase in B cell activity (3).

A number of different criteria for primary SS have been used during the past few decades, resulting in a prevalence range from 0.02–4% depending on the type of criteria used and the population studied. The American–European Consensus Group (AECG) criteria have been widely accepted since 2002 and include ocular and oral symptoms and signs, autoantibodies, and labial salivary gland abnormalities (2). Applying these criteria in an epidemiologic study of 2 Norwegian counties, we recently found that the prevalence of primary SS in white subjects was 0.05% (95% confidence interval [95% CI] 0.048–0.052%) (4). This indicates that primary SS, at least in Scandinavia, may be less frequent than previously assumed.

Several studies have highlighted the increased risk of non-Hodgkin's lymphoma (NHL) in primary SS. A multicenter study from 1999 found malignant lymphomas in 4.3% of primary SS patients, with the most common subtype being extranodal marginal-zone B cell lymphoma of the mucosa-associated lymphoid tissue (MALT) (5). A meta-analysis from 2005 found a pooled standardized incidence ratio (SIR) of 18.8 (95% CI 9.5–37.3), but the studies included in this meta-analysis used different diagnostic criteria and, consequently, a broad variation of SIRs was noted (6). A Swedish study in 2006 reported an SIR of 15.5 for NHL in primary SS patients classified according to the AECG criteria (7).

The current state of knowledge concerning the increased risk of NHL in primary SS is affected by uncertainty regarding factors such as the use of different diagnostic criteria, paucity of population-based recruitment, and examination of different ethnic groups. Therefore, the objective of the current study was to investigate the risk of NHL based on a well-defined open primary SS cohort in 2 Norwegian counties (4).

Significance & Innovations

This is a population-based study of the risk of non-Hodgkin's lymphoma in primary Sjögren's syndrome conducted in an area including 896,840 inhabitants (18.6% of the total population of Norway).

Seven of 443 patients with primary Sjögren's syndrome developed lymphoma during a total followup of 3,813 person-years.

After adjusting for age and sex, this corresponds to a standardized incidence ratio of 9.0 compared to the general population in Norway.

Significance & Innovations

  • This is a population-based study of the risk of non-Hodgkin's lymphoma in primary Sjögren's syndrome conducted in an area including 896,840 inhabitants (18.6% of the total population of Norway).

  • Seven of 443 patients with primary Sjögren's syndrome developed lymphoma during a total followup of 3,813 person-years.

  • After adjusting for age and sex, this corresponds to a standardized incidence ratio of 9.0 compared to the general population in Norway.

Patients and methods


Rogaland and Hordaland are 2 counties on the west coast of Norway comprising 896,840 inhabitants (as of July 1, 2009), which constitutes 18.6% of the total Norwegian population (8) (see Supplementary Figure 1, available in the online version of this article at http://onlinelibrary.wiley.com/doi/10.1002/acr.21887/abstract). A population-based registry was created with the intention to identify all inpatients and outpatients with primary SS still living in these regions (4). All subjects registered between January 1, 1980 and January 1, 2009 with a diagnosis of primary SS at Stavanger University Hospital, Haukeland University Hospital, and Haugesund Rheumatism Hospital were identified and included in an open cohort study. These hospitals are responsible for all secondary health service in the area, including rheumatology and oncology. The medical records were reviewed and patients with incomplete data were invited for a screening visit. Patients with a diagnosis of primary SS visiting 2 of the 3 private rheumatologists in this area were also identified. New patients with primary SS were included until July 1, 2009. A total of 443 patients met the AECG criteria for primary SS and were included in the present study.

Cancer Registry of Norway.

In Norway, all inhabitants receive a unique 11-digit personal identification number at birth or permanent residency, allowing easy and reliable identification and tracking of individuals. The Cancer Registry of Norway is a national population-based registry responsible for recording and reporting all cancer cases since 1951. Clinical data are collected from structured templates and submitting data to the registry is required by law.

Study design.

By comparing data from the Cancer Registry of Norway with our primary SS registry, we identified myeloid and lymphoid neoplasms defined according to the World Health Organization classification criteria (International Statistical Classification of Diseases and Related Health Problems, Tenth Revision [ICD-10] codes C81–C96). The type and location of any lymphoma and nonsolid hematologic neoplasms were recorded. To ensure the completeness and quality of the Cancer Registry of Norway data, 212 primary SS patient files (47.9%) were randomly chosen and reviewed without discovering any unreported hematologic neoplasms. Similarly, we searched the Cancer Registry of Norway for all subjects living within the 2 counties who were diagnosed with any lymphoproliferative neoplasm between 2003 and 2008. This search identified 761 cases, and 416 (54.7%) were randomly chosen and reviewed completely for the possibility of unreported primary SS. The files were searched for summaries, symptoms, signs, and immunologic tests, or referrals to rheumatologists, ophthalmologists, and otolaryngologists without discovering any new cases. Finally, we searched the Cancer Registry of Norway for the combination of lymphoproliferative diseases and a secondary ICD-10 diagnosis of primary SS (code M35.0). This search identified 40 patients, and all patient files, immunologic test results, and anatomic pathology department records were screened without discovering any patients not originally included in our primary SS registry (Figure 1).

Figure 1.

Flowchart demonstrating how patients were included and excluded from the search for a lymphoproliferative disease (C81–C96) and primary Sjögren's syndrome (pSS; M35.0) in the Cancer Registry of Norway. All correctly identified patients had already been included in the pSS registry. RA = rheumatoid arthritis; SLE = systemic lupus erythematosus; UC = ulcerative colitis; AECG = American–European Consensus Group; MGUS = monoclonal gammopathy of undetermined significance.

Statistical analysis.

The data were described with the median and range for continuous variables and the proportion with percentage for categorical variables. The associations between pairs of categorical variables were assessed with Fisher's exact test and the differences in continuous variables with the Mann-Whitney-Wilcoxon test. For the calculation of disease duration, we consistently used the time point when primary SS was first diagnosed according to the patients' files. Only the year of diagnosis was available for some patients, in which case July 1 was set as the time of diagnosis. In order to quantify the association between primary SS and NHL, we compared the observed number of NHL cases diagnosed among primary SS patients with the number expected in the general population. The expected number of NHL cases was calculated using the data from the Cancer Registry of Norway (9). Indirect age and sex standardization was used to provide age-specific rates and SIRs were calculated by dividing the observed number of cases by the expected number of cases. All NHL cases diagnosed in our primary SS sample were included in the SIR calculation and 95% CIs were constructed assuming a Poisson distribution. The crude cumulative incidence (with 95% CI) of NHL was calculated using the Kaplan-Meier method and depicted graphically with a Kaplan-Meier curve. For the purpose of the cumulative incidence calculation, the observation times were set to 0 for NHL patients diagnosed prior to or simultaneously with their primary SS diagnosis. No patients were lost to followup. All P values were 2-sided and significance was defined as P less than 0.05.


The relevant clinical and demographic data are shown in Table 1. Most patients presented with some symptoms of dryness (426 [96.2%] of 442 patients). Autoantibodies or positive minor salivary gland biopsies were present in all patients (319 [74.7%] of 427 patients and 299 [82.8%] of 361 patients, respectively) and both were present in 178 (40.2%) of 443 patients. The median duration since the primary SS diagnosis was 8.0 years (range 0.5–35.0 years) and the total followup was 3,813 person-years. NHL occurred in 7 patients (1.6%) and the median time between diagnosis of primary SS and NHL was 9.3 years (range −6.8 to 18.2 years). We found higher focus scores, longer primary SS disease duration, and younger age at primary SS diagnosis in patients developing lymphoma, although these differences did not reach statistical significance (Table 1).

Table 1. Selected clinical data for patients with primary SS*
 All patients (n = 443)Without NHL (n = 436)With NHL (n = 7)P
  • *

    Categorical variables were analyzed by Fisher's exact test and continuous variables by the Mann-Whitney U test. SS = Sjögren's syndrome; NHL = non-Hodgkin's lymphoma; NS = not significant (P ≥ 0.05).

  • Patients with NHL compared to patients without NHL.

Age, total/median (range) years443/63.0 (16–91)436/63.0 (16–91)7/57.0 (40–76)NS
Females and males, total/no. (%)443/417 (94.1) females, 26 (5.9) males436/411 (94.3) females, 25 (5.7) males7/6 (85.7) females, 1 (14.3) maleNS
Age at primary SS diagnosis, total/median (range) years443/53.0 (11–85)436/54.0 (11–85)7/44 (29–62)NS
Duration of primary SS, total/median (range) years443/8.0 (0–35)436/8.0 (0–35)7/11 (3–32)NS
Ocular dryness symptoms, total/no. (%)438/379 (86.5)431/372 (86.3)7/7 (100.0)NS
Oral dryness symptoms, total/no. (%)437/400 (91.5)430/393 (91.4)7/7 (100.0)NS
Positive Schirmer's test (<1 mm/minute), total/no. (%)412/272 (66.0)406/267 (65.8)6/5 (83.3)NS
Positive histopathology (focus score ≥1), total/no. (%)361/299 (82.8)354/292 (82.5)6/6 (100.0)NS
Focus score, total/median (range)347/1.0 (0.0–12.0)343/1.0 (0.0–12.0)4/2.5 (2.0–4.0)NS
Unstimulated salivary flow <0.1 ml/minute, total/no. (%)348/221 (63.5)343/217 (63.3)5/4 (80.0)NS
Anti-SSA and/or anti-SSB antibody, total/no. (%)427/319 (74.7)420/313 (74.5)7/6 (85.7)NS
Other hematologic malignancies, total/no. (%)443/3 (0.7)436/3 (0.7)7/0 (0.0)NS

The lymphoid neoplasms in the primary SS patients were all marginal zone lymphomas, and 6 of these were MALT-type lymphomas (Table 2). Four of the MALT-type lymphomas were located in the parotid gland, the others in the labial salivary glands, the thymus gland, and the lingual tonsil.

Table 2. Clinical data for patients with primary SS and NHL*
 Patient 1Patient 2Patient 3Patient 4Patient 5Patient 6Patient 7
  • *

    SS = Sjögren's syndrome; NHL = non-Hodgkin's lymphoma; WHO = World Health Organization; MZL = marginal zone lymphoma; CLL = chronic lymphatic leukemia; MALT = mucosa-associated lymphoid tissue; B-CLL = B-cell chronic lymphocytic leukemia.

  • Patient presented with 2 lymphomas.

Age, years76706553574054
Age at primary SS diagnosis, years44576242482938
Age at NHL onset, years62506253443947
Positive Schirmer's testYesYesYesYesYes
Reduced salivary flow (<1.5 ml/minute)YesYesYesYes
Time between primary SS and NHL, months218−81−3127−57121112
Positive histopathologyYesYesYesYesYesYesYes
Focus score3.
Anti-SSA and/or anti-SSB antibodyYesYesNoYesYesYesYes
Hypergammaglobulinemia (>15 gm/liter)NoNoNoYesYesYesYes
Low complement (C3 and/or C4)NoNoNoYesYes
Palpable skin purpuraNoNoNoNoNo
Parotid enlargementYesNoNoYesYesYes
Leukopenia (<4 × 10−9/liter)NoNoYesYesNoYesNo
Lymphoma localizationParotid glandLingual tonsilBone marrowMinor salivary glandsParotid glandParotid gland, thymusParotid gland

The relative risk of NHL adjusted for the risk in the general population of Norway is expressed as an SIR. The primary SS patients were almost 9 times more likely to have NHL compared to the general population (SIR 9.0 [95% CI 7.1–26.3]). The cumulative incidence of NHL was estimated to be 0.7% (95% CI 0.2–2.1%) within the first 5 years, 1.2% (95% CI 0.4–3.3%) within 10 years, and 7.4% (95% CI 1.9–26.8%) within 20 years. The cumulative incidence with 95% CI is shown using a Kaplan-Meier curve in Figure 2.

Figure 2.

Kaplan-Meier plot showing the cumulative incidence (horizontal lines) and the 95% confidence interval (shaded area) of non-Hodgkin's lymphoma in patients with primary Sjögren's syndrome. The number of patients at risk is given for each time point.


In a well-defined geographic area comprising 18.6% of the total population of Norway, 7 (1.6%) of 443 primary SS patients developed NHL during a total followup of 3,813 person-years. This corresponds to a 9-fold increase of NHL risk in patients with primary SS compared to the general population. To ensure the completeness and quality of data, we studied all available registries, including patient files from hospitals and private practices, anatomic pathology department registries, and available information from the Cancer Registry of Norway. This study is probably as close to being a population-based study as is practically possible for a rare disease such as primary SS, and the results indicate that the risk of lymphoma development in primary SS may have been estimated to be too high.

The inconsistencies regarding NHL risk in primary SS patients have several possible explanations; the use of different primary SS classification criteria is probably the most obvious. Potential limitations to community-based studies of primary SS prevalence are nonresponders and dropouts. A bias in patient selection with a tendency to overrepresent patients with more severe disease at major hospital centers could also play a role. To avoid this bias, we sought to retrieve patients from all sources of rheumatology and oncology services in the specified area.

Concerns have been raised about the nonreferral of patients with mild symptoms of primary SS by general practitioners to specialized health care (4). One reason for this nonreferral could be that the therapeutic options for primary SS patients are limited. The proportion of patients with primary SS that actually present with symptoms so mild that they do not seek medical attention is not known, but could be a considerable proportion according to a recent epidemiologic study (10). To correct for this, we reviewed 416 lymphoma patient files from the same period and region without revealing any suspected or undiagnosed cases of primary SS.

The demographic and disease characteristics of the study population were comparable to those in recent studies (5, 7, 11). The majority of subjects in the 2 counties were white (98.0%) and the demographics and clinical presentation of the white and nonwhite patients were homogenous (see Supplementary Table 1, available in the online version of this article at http://onlinelibrary.wiley.com/doi/10.1002/acr.21887/abstract). Also, based on inquiries at neighboring hospitals, the likelihood of patients being lost to living outside of the geographic area was low (4).

Lymphoma distribution and incidence trends are identical across Northern Europe. The age-adjusted SIR for NHL per 100,000 inhabitants for women in Norway was 8.8 (from 2005–2009), which is identical to the counties investigated (9). Regional differences in lymphoma incidence are therefore not likely to be an explanation, and we are not aware of any patients with primary SS and lymphoma who died during the past 5 years in this region. Another study from a Nordic region conducted in 1997 reported an SIR of 8.7 for NHL in primary SS (11), and a recent study found an SIR of 6.1 for primary SS in patients with NHL (12).

Marginal zone lymphomas, especially MALT-type lymphomas, are reportedly the most common subtype of lymphoma in primary SS (5, 12). MALT arises in various sites under conditions of chronic antigen-driven inflammation. The expected chronological order of presentation would be chronic inflammation followed by secondary lymphoma development. Three of the patients were first diagnosed with lymphoma, which is usually considered an exclusion criterion for primary SS. In retrospect, we learned that these patients had presented with symptoms strongly indicative of primary SS years before lymphoma was found. All 3 had reduced tear flow and mesoepithelial lesions in the salivary glands, and 2 of 3 presented with reduced salivary flow and autoantibodies (Table 2). It is therefore most likely that primary SS preceded NHL in all patients, but primary SS did not reach enough clinical significance to come to medical attention before the lymphoma became evident.

The strengths of our study are the large number of primary SS patients, the strict use of AECG criteria, and the long observation time. The registries are valid, and no misclassifications or missed cases were discovered when reviewing a large number of NHL and primary SS patient files.

A limitation of the study is the restricted clinical information available on the individual patients. Further, the low number of lymphoma cases implies low precision in the statistical analysis, and subgroup analyses and evaluation of clinical risk factors for lymphoma development were therefore not possible. The cumulative incidence after 15 years of observation was based on ∼10% of the total number of patients, and the calculations were accordingly imprecise. In addition, primary SS patients with subclinical symptoms and signs may have been missed, and the vague initial presentation of primary SS in some subjects made it difficult to determine the exact time points for the start of the disease. There were no data available concerning how many primary SS patients died during the observation period and this may have led to an overestimation of the lymphoma risk. In addition, it seems unlikely that any patient with a suspected lymphoma would not be referred to specialist health care; thus, it appears more likely that patients with NHL and primary SS would be overrepresented in our study.

In conclusion, based on our study of 443 primary SS cases in 2 Norwegian counties, patients with primary SS have a 9-fold increased risk for NHL compared to the general population. This is in accordance with recent studies and the completeness of the registries and use of strict criteria support the validity of this finding.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Johnsen had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Johnsen, Brun, Gøransson, Småstuen, Jonsson, Meyer, Omdal.

Acquisition of data. Johnsen, Brun, Gøransson, Småstuen, Johannesen, Haldorsen, Harboe, Omdal.

Analysis and interpretation of data. Johnsen, Småstuen, Meyer, Omdal.


The authors thank Dr. Kjell H. Kjellevold for helpful advice.