Article Text
PDF
Abstract
Objectives Rheumatic diseases were previously associated with increased incidence of monoclonal gammopathy (MG) and its malignant transformation. The present study aimed to investigate the prevalence, malignant transformation risk, clinical correlates and prognostic impact of MG in SLE.
Methods A retrospective cohort study based on the medical records of n=1039 patients with SLE fulfilling the 1997 American College of Rheumatology (ACR), the 2019 European Alliance of Associations for Rheumatology (EULAR)/ACR and/or the 2012 Systemic Lupus International Collaborating Clinics (SLICC) criteria managed at two tertiary care departments of the University Hospital (Krakow, Poland) from January 2012 until November 2019.
Results SLE+MG cases were older at SLE diagnosis compared with non-MG SLE controls (53±15 years vs 37±15 years, respectively, p<0.01), had higher rates of lymphopenia, anaemia, haemolysis, serous effusions and interstitial lung disease (all p<0.05), and were more frequently treated with cyclophosphamide (57% vs 28%, p<0.01) or rituximab (13% vs 3%, p<0.01). Most MG cases were detected within a year after SLE diagnosis (Q25, Q75: 0, 12 years). With the median follow-up of 11 years (Q25, Q75: 6, 19 years), 34.8% (8 cases) of the SLE+MG cohort were diagnosed with malignancy, compared with 8.1% (82 cases) among the SLE controls (p<0.001). MG was associated with the relative hazard of death of HR 2.99 (95% CI 1.26 to 7.06, p<0.05) and a median survival time from SLE diagnosis to death of 5 years (Q25, Q75: 1, 14; range 0–41) for SLE+MG cases, as compared with 12 years (Q25, Q75: 6, 19; range 0–62) for the controls. The effect was non-independent on antimalarial medication use.
Conclusions Our study emphasises heightened malignancy and mortality rates in SLE+MG cases. The association between immunosuppression, MG incidence and progression warrants further research.
- Systemic Lupus Erythematosus
- Hematology
- Risk Factors
- Cause of Death
- B-Lymphocytes
Data availability statement
Data are available upon reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Monoclonal gammopathy (MG) is detected in approximately 3% of the population over the age of 50, usually incidentally, and may progress to plasma cell dyscrasia or lymphoproliferative disorders.
Its incidence and the risk of malignant transformation appear to be increased among rheumatic disease cases, including Sjoegren’s syndrome, rheumatoid arthritis and SLE.
WHAT THIS STUDY ADDS
Patients with SLE and MG exhibit distinct clinical features including higher rates of lymphopenia, haemolytic anaemia, serous effusions and interstitial lung disease compared with SLE controls.
If biopsied, lupus nephritis cases with MG display higher prevalence of International Society of Nephrology/Renal Pathology Society (ISN/RPS) class IV and more frequently require the use of cyclophosphamide and rituximab to manage the renal disease flare.
Patients with SLE and MG showed a higher incidence of malignancies, both lymphoproliferative and solid, suggestive of MG being not only a premalignant condition but also a marker of immunodeficiency.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Our study indicates SLE activity as a contributor to the pathophysiology of MG within patients with SLE, potentially informing the design of future studies on MG progression.
The findings may influence screening recommendations for early detection of MG among patients with SLE to help identify patients at risk of unfavourable course of the disease and malignancy.
Introduction
Monoclonal gammopathy (MG) is defined by the presence of circulating monoclonal protein (the M-protein) produced by the clonally expanded plasma cells or B-cells, which may or may not be related to an overt haematological malignancy.1 The M-protein may be a complete immunoglobulin of any class which co-occurs with corresponding immunoglobulin light chains or consist exclusively of the light chains and only seldom the immunoglobulin heavy chains.2 The presence of different M-proteins is possible and points to co-existence of separate cell clones (eg, biclonal gammopathy) or rarely to polymeric forms of the same M-protein.3
MG occurs in approximately 3% of those over the age of 50 years,4 as compared with <1% of the younger population.5 It is usually detected incidentally by serum protein electrophoresis (SPEP) and free light chain (FLC) assays ordered as screening tests or due to the clinical suspicion of a related disorder. The type of monoclonal protein is determined with immunofixation electrophoresis (IFE).
The type of M-protein is correlated with the risk of malignant transformation and pathogenicity:
Non-IgM (IgG (50% cases), IgA (9%) or IgD (<1%)), the most common type, may progress to multiple myeloma (MM), lymphoma, amyloid light chain (AL) amyloidosis or light chain deposition disease, and either class may result in cryoglobulinaemia; IgD is almost always associated with an underlying malignancy.6
IgM (14%), with the additional risk of progression to Waldenström’s macroglobulinaemia.6 7
Light chain (19%), with the additional risk of idiopathic Bence-Jones proteinuria and progression to MM and light chain amyloidosis.8
Heavy chain, α, γ or µ (<1%), associated with B-cell malignancies of variable presentation.9
Detection of MG necessitates the exclusion of MM (with serum M-protein <30 g/L, fewer than 10% bone marrow plasma cells and absence of MM-defining events) and B-cell lymphomas, with predictive models developed to establish the need for bone marrow biopsy.10 If non-malignant, the disorder is classified into MG of undetermined significance (MGUS, in the absence of clear pathogenetic effects) or MG of clinical significance (MGCS, defined by the end-organ dysfunction or multisystem disorders attributed to the M-protein).1 For example, monoclonal gammopathy of renal significance (MGRS) results from isolated damage to the glomeruli or renal tubules and includes light chain deposition disease and secondary membranoproliferative glomerulonephritis, among others.6 7 Multisystem disorders associated with MGCS include, but are not limited to, cryoglobulinaemia syndromes, light chain amyloidosis and rare conditions such as POEMS (polyneuropathy, organomegaly, endocrinopathy, monoclonal plasma cell disorder, skin changes), CANOMAD (chronic ataxic neuropathy, ophthalmoplegia, immunoglobulin M (IgM) paraprotein, cold agglutinins, and disialosyl antibodies) or Schnitzler’s syndromes.11
MG was previously associated with numerous chronic inflammatory rheumatic diseases (RDs), especially Sjogren’s disease and SLE.12 In SLE, the reported incidence of MG ranges from 2.2% of the paediatric cases13 to 5.4% of adult cases.14
A recent report15 points to the increased risk of malignant transformation in such cases. For non-antibody-mediated RDs, the risk was up to twofold compared with non-RD MG cases, with the incidence of 4% for non-RD MG up to 10% for non-antibody-mediated RDs and 2% for antibody-mediated RDs cases.
While these reports focused on the risk factors of MG occurrence in RDs as well as RDs as the risk factors of malignant transformation of MG, few authors have investigated the association between clinical course of the RD with the presence of MG. One report indicates increased severity of rheumatoid arthritis16 but scarce data pertain to the clinical differences between SLE cases with and without MG.14
Given the broad range of pathogenicity, reported association with malignancy rates and widely available testing methods, further evaluation of MG as a prognostic factor in SLE is warranted to inform screening policy and possibly guide clinical decisions regarding treatment options.
The aims of our study were to explore (1) the prevalence of MG and (2) the risk of its malignant transformation in a cohort of patients with SLE, (3) determine the clinical differences in SLE cases with and without MG and (4) assess the significance of MG as a risk factor of unfavourable outcomes in SLE.
Methods
Study sample and data acquisition
To prepare the retrospective cohort study, we have reviewed the medical records of all n=1039 patients with SLE registered at two tertiary rheumatology departments in Krakow, Poland (former Department of Allergy and Clinical Immunology and the Department of Rheumatology, Immunology and Internal Medicine, University Hospital) from January 2012 to November 2019, with a follow-up of at least 1 year. All the patients included in our study fulfilled either the 1997 American College of Rheumatology (ACR),17 the 2019 European Alliance of Associations for Rheumatology (EULAR)/ACR18 and/or the 2012 Systemic Lupus International Collaborating Clinics (SLICC) criteria,19 owing to the wide study timeframe, and were not reclassified with an alternative diagnosis in the course of the follow-up.
The definitions of MGUS, MM and related plasma cell disorders used in this study were in concordance with the International Myeloma Working Group updated criteria.6
The primary endpoint measured was death from any cause in the course of the follow-up. The secondary endpoint was the development of malignancy. The available sample size allowed for detection of OR of the endpoints’ occurrence of at least 1.5 at the confidence level of 95% and statistical power of 90% (required sample size=883).
The medical records were searched for gender, age at SLE diagnosis, age at death, duration of the disease (calculated from disease diagnosis to last visit or patient’s death), family history of SLE and other autoimmune diseases, clinical and laboratory disease manifestations, comorbidities (including hypertension, diabetes, dyslipidaemia, thyroid dysfunction, atrial fibrillation, coronary artery disease, peripheral artery disease, heart failure and chronic kidney disease), medications used to treat SLE, smoking status (defined according to Levy et al20), diagnosis of a neoplastic disease and cause of death. Clinical and laboratory manifestations of SLE (including leucopenia, lymphopenia, haemolytic anaemia and thrombocytopenia) were defined according to the ACR classification criteria.21 Anaemia was defined according to WHO, haemoglobin (Hb) concentration <130.0 g/L for male and <120.0 g/L for female.22
The studied sample was screened for monoclonal protein using SPEP and serum FLC assays as part of routine diagnostic workup for SLE or on the initial department visit in case of patients diagnosed beforehand and referred for continued management and the diagnostic workup was repeated in cases suspected of disease flare, signs of immunodeficiency or symptoms suggestive of haematological malignancy (plasma cell dyscrasia or lymphoma). The suspicion of monoclonal protein was confirmed and the paraprotein identified using IFE. Urine protein electrophoresis with urine IFE were not routinely done in case of negative serum studies. Cases diagnosed with MG were evaluated for signs of plasma cell dyscrasias (the CRAB symptoms, ie, hypercalcaemia, renal disease, anaemia and bone lesions, the latter screened using skeletal radiographs, if not redundant to previously obtained imaging studies) and lymphoma (with focused physical examination, chest X-ray, abdominal ultrasound, if not redundant to previously obtained imaging studies), with bone marrow trephine biopsy performed if deemed necessary. Patients diagnosed with MG were followed-up with yearly SPEP studies and additional workup for SLE (including complete blood count, serum calcium and creatinine concentrations and urinalysis with urine sediment analysis). Patients were allocated to the MG cohort based on singular positive MG results. Only cases with completed workup for MG were included in the study.
Mortality data were based on the Polish personal identification number database and validated with the electronic beneficiary entitlement verification system.
The laboratory parameters acquired for the study encompassed standard haematological, biochemical and immunological profile and included complete blood count, 24-hour urinary protein, urinary sediment, ANA (screened with indirect immunofluorescence (IIF) using Hep-2 cell as substrate), anti-dsDNA antibodies (assayed with Crithidia luciliae as substrate and by standard ELISA), ANA immunoblot (for anti-Ro, anti-La, anti-histone, anti-nucleosome, anti-ribonucleoprotein (RNP) antibodies; assayed by standard ELISA), rheumatoid factor (assayed by standard ELISA), anti-neutrophile cytoplasmic antibodies (screened with IIF), anti-proteinase 3 (anti-PR3) and anti-myeloperoxidase antibodies (assayed by standard ELISA), bilirubin (direct and indirect), direct Coomb’s test, haptoglobin, lupus anticoagulant, anti-cardiolipin IgG and IgM antibodies, anti-β2 glycoprotein I IgG and IgM antibodies (assayed with standard diagnostic measures). Serum IFE was carried out on agarose gels with specific antisera (anti-gamma, anti-alpha, anti-mu, anti-kappa and anti-lambda). Urine protein electrophoresis and urine IFE were performed on aliquot from 24-hour urine collection.
Statistical analysis
Statistical analysis of the data was performed using TIBCO Statistica (TIBCO Software, Palo Alto, California, USA). Continuous variables were tested for distribution normality with the Shapiro-Wilk test and presented as means±SD or medians with Q25 and Q75 quartiles (Q25, Q75) and compared using unpaired t-tests, the Mann-Whitney U test or the Kruskal-Wallis test, as appropriate. Discrete variables were expressed as the number of cases and relative frequency (percentage of the sample) and compared using the χ2 or Fisher’s exact test. Patients with MG were compared with their matched controls through generalised linear models (namely, linear regression for continuous variables and logistic regression for categorical variables) where the grouping of matched sets was retained.
The analyses excluded cases with lymphoproliferative disorders diagnosed within 1 year of the MG. Survival analysis was conducted with Cox proportional hazards model to adjust for covariates and calculate adjusted HR of death from any cause. The odds of secondary endpoint was calculated using multiple logistic regression to control for confounders. The confounders controlled for in our analyses included variables differing between SLE cases diagnosed with MG and SLE controls and included patient’s age at SLE diagnosis and selected comorbidities (hypertension, diabetes mellitus and peripheral artery disease), unless stated otherwise.
The results were considered significant at p value <0.05. Missing data were deleted pairwise. No loss of follow-up occurred due to the retrospective cohort design. The raw data are available on request.
Results
Clinical presentation of the patients
Cases with MG were older at systemic lupus diagnosis
Of the 1039 patients with SLE followed up at our departments (table 1), 23 (2.2%) were diagnosed with monoclonal gammopathy (SLE+MG). The remaining 1016 SLE cases constituted the controls in our study (SLE controls). The males comprised 21.7% of the SLE+MG cohort and 10.9% of the controls (difference p>0.05). Most SLE+MG cases were older than the controls at SLE diagnosis (53±15 years vs 37±15 years, respectively, p<0.01), and the mean age at MG detection was 59±10 years. Consequently, by the end of the follow-up, SLE+MG cases were older than SLE controls (average 82±20 years vs average 56±21 years; p<0.001). As the older group, SLE+MG cohort had higher prevalence of age-related comorbidities including diabetes mellitus, arterial hypertension and peripheral artery disease (all p<0.05). More of the patients with SLE+MG were smokers (n=12; 55%), compared with n=255 (36%) smokers among SLE controls (p>0.05).
Sample demographics
Cases with MG had higher rates of lymphopenia, haemolytic anaemia, serous effusions and interstitial lung disease, but similar ANA antibody profiles
Both patient groups presented with various and overlapping clinical manifestations of the disease (table 2). General symptoms were frequent among both SLE+MG and controls (91% vs 75%; p>0.05), especially fatigue (87% vs 62%; p<0.05) and arthralgias (74% vs 88%, p<0.05). Lymphadenopathy unrelated to lymphoma occurred at comparable rates of 22% and 18%, respectively (p>0.05).
Clinical manifestations and antibody profile of SLE sample
SLE+MG cases were more frequently diagnosed with lymphopenia (83% vs 74%; p<0.01), anaemia (100% vs 69%; p<0.01) and signs of haemolysis (45% vs 19%; p<0.05), while macrophage activation syndrome was rare and occurred only in n=9 (1%) of the controls.
Renal involvement occurred in n=10 (43%) of the SLE+MG cases (90% of whom had only one renal flare), as compared with 31% of the controls, in which 19% had only one, 6% had two, 4% had three and 1% had four or more renal flares. In those with biopsy-proven disease, SLE+MG cases were mostly diagnosed with International Society of Nephrology/Renal Pathology Society (ISN/RPS) class IV (5/6 cases, 83%), comparable to the SLE controls (89/177, 50%), p>0.05. Consequently, both groups had a high rate of proteinuria exceeding 500 mg/day or defined as urine protein-creatinine ratio >500 mg/g (SLE+MG 48% vs SLE controls 32%) as well as history of nephrotic-range proteinuria exceeding 3.5 g/day or urine protein-creatinine ratio >3500 mg/g (23% vs 17%, respectively), with no statistically significant differences (all p>0.05).
Serous effusions unrelated to malignancy were disproportionately more frequent among SLE+MG cases, both pleural (39% vs 16%; p<0.01) and pericardial (30% vs 13%; p<0.05), but frank pericarditis was absent in SLE+MG cases while present in 4% of the SLE controls.
Interstitial lung disease was more prevalent among SLE+MG cases (n=4, 17.4%), as opposed to the controls (n=47, 4.6%); p<0.01.
As much as n=4 (17.4%) of the SLE+MG cases were diagnosed with pulmonary hypertension defined as systolic pulmonary artery pressure ≥46 mm Hg, compared with only n=27 (3%) of the controls; p<0.001. Three of the four such SLE+MG cases had no interstitial lung disease or signs of heart failure and none had a history of pulmonary embolism or antiphospholipid syndrome. However, two had detectable rheumatoid factor and each of them was a smoker. Only one, n=1 (4%), of the SLE+MG cases was diagnosed with diffuse alveolar bleeding, as compared with n=10 (1%) of the controls (p>0.05).
Cutaneous manifestations were less frequent among SLE+MG cases (52% vs 80%, p<0.01) and included a variety of lesions (table 2).
The highest ANA titre recorded was similar in both groups (median SLE+MG ANA titre 1:7680 vs median controls ANA titre 1:5120; p>0.05) and they had comparable ANA immunoblotting results (no differences in the relative frequency of anti-SS-A, SS-B, -Ro-52, -Sm, anti-dsDNA, anti-nucleosome, anti-histone or anti-RNP antibodies). Anti-dsDNA, if present, had similar Crithidia luciliae immunofluorescence test (CLIFT) titre among SLE+MG and the controls (median 1:40 in both groups; p>0.05).
Patients with systemic lupus had similar antiphospholipid syndrome rate and thromboembolic events, regardless of MG
Of the SLE+MG cases, n=3 (13%) were diagnosed with antiphospholipid syndrome, as compared with n=126 (12%) of the controls (p>0.05). Antiphospholipid antibodies were variably prevalent among SLE+MG cases and controls (table 3.).
Clinical manifestations and antibody profile of antiphospholipid syndrome among the SLE sample
Among those with diagnosed APS, SLE+MG cases had significantly higher median concentrations of the highest recorded aCL IgM (163 vs 55 MPL; p<0.05) and aB2GP IgM antibodies (56 vs 0 MPL; p<0.05).
Miscarriages occurred in two of the SLE+MG cases (11% of the women), as compared with 118 of the SLE controls (16% of the women, p>0.05, of which 62% had not more than one episode and 82% not more than two episodes).
Deep vein thrombosis rates were comparable between SLE+MG cases and controls (p>0.05), leading to pulmonary embolism in a minority of patients. The rate of arterial thromboembolism showed no significant differences between the two groups (p>0.05), including the incidence of ischaemic stroke, peripheral arterial thromboembolism and myocardial infarction.
Of note, any manifestation of atherosclerosis (arterial thromboembolism or visualised atherosclerotic lesions) was present in n=12 (52%) of the SLE+MG cases, as compared with only n=277 (22%) controls (p<0.001), likely resulting from older age of the SLE+MG cohort.
Cases with MG had higher incidence of family history positive for RDs
Among patients with SLE+MG, there were more cases with positive history of first-degree relatives diagnosed with SLE (8.7% vs 4.4%, p<0.05) and rheumatoid arthritis (8.7% vs 5.7%, p>0.05). Additionally, the first-degree relatives of SLE control cases had an incidence of psoriasis of 2.5% (as compared with none in the SLE+MG group).
Cases with MG were more frequently treated for systemic lupus with cyclophosphamide and rituximab
Nearly all of the study sample had documented history of treatment with systemic steroids (91% SLE+MG vs 95% SLE controls, p>0.05) and nearly half of the patients were treated with chloroquine (CQ) (43% vs 52%, p>0.05) or hydroxychloroquine (HCQ) (17% vs 53%, p<0.001). SLE+MG cases were more frequently treated with cyclophosphamide (57% vs 28%, p<0.01) or rituximab (13% vs 3%, p<0.01), mostly used to manage lupus nephritis (LN). No significant differences were noted between the use rate of azathioprine (30% vs 39%, p>0.05), methotrexate (17% vs 20%, p>0.05), cyclosporine A (13% vs 8%, p>0.05), mycophenolate mofetil (39% vs 30%, p>0.05), sulfasalazine (4% vs 5%, p>0.05) or plasmapheresis (9% vs 3%, p>0.05). Single cases were treated with belimumab (n=41), intravenous immune globulins (n=29), anifrolumab (n=10), leflunomide (n=7), gold salts (n=6) or dapsone (n=3), and only among the controls.
MG, malignancy and mortality rate
MG characteristics
Most MG cases were detected up to a year after SLE diagnosis (Q25, Q75: 0, 12 years). The most common M-protein was IgG (16 cases; 70%), followed by IgM (4 cases; 17%) and lambda light chain (1 case; 4%) or remained undisclosed. On MG diagnosis (excluding readily diagnosed MM and WM cases), the average gamma-globulin concentration was 20 g/L (range 5.93–59.73 g/L), M-protein concentration was 6 g/L (range 0.64–29.4 g/L) and the extrema of the free light chain kappa/lambda ratios in the serum equalled 0.32 and 39. IgG levels were decreased in five cases (average 4.8 g/L, range 3.99–5.82 g/L), IgA levels in three cases (average 0.48 g/L, range 0.2–0.69 g/L) and IgM levels in six cases (average 0.22 g/L, range 0.1–0.39 g/L).
Cases with MG had higher incidence of malignancy
With the median follow-up of 11 years (Q25, Q75: 6, 19 years, range 0–62 years), a total of 90 patients (8.7%) were diagnosed with malignancy, 34.8% (8 cases) in the SLE+MG group and 8.1% (82 cases) in the SLE control group (p<0.001). The median time from MG detection to malignancy diagnosis was 6 months (Q25, Q75: 0, 2 years, range 0–32 years). After adjustment for age at SLE diagnosis and selected comorbidities (hypertension, diabetes mellitus and peripheral artery disease), the diagnosis of MG remained an independent predictor of malignancy, with adjusted OR (aOR) of 4.18 (95% CI 1.67 to 10.49, p<0.05).
Of the neoplastic diagnoses in the SLE+MG group, six were a lymphoproliferative disease (two cases of MM, IgG kappa and IgG lambda, one case of Waldenstroem’s macroglobulinaemia, one case each of chronic lymphocytic leukaemia, diffuse large B-cell lymphoma, follicular lymphoma) and two a solid tumour (breast cancer and planoepithelial lung cancer). Six (26%) of the patients with SLE+MG had bone marrow examination, positive for MM in two cases. Four had skeletal surveys (whole-body low-dose CT scans), three of which were positive for osteolytic lesions in two MM cases and one Waldenstroem’s macroglobulinaemia case. One case of MM and one case of Waldenstroem’s macroglobulinaemia were diagnosed at the time of first MG detection, while one MM case was diagnosed 7 years later.
These results are summarised in table 4.
Malignancy and mortality rate
Cases with MG had higher incidence of death but did not differ in age at death from any cause
Six (26%) of the patients with SLE+MG died in the course of the follow-up, as compared with 45 (4%) of the controls (p<0.001). Median time from SLE diagnosis to death from any cause was 6.5 years in the SLE+MG group (Q25, Q75: 1, 11 years; range 1–41 years) and 13.5 years for the controls (Q25, Q75: 6, 22 years; range 0–47 years); p>0.05.
There was no significant difference in the age at death between patients with SLE+MG and the controls (average 62±12 and median of 64 years, Q25, Q75: 54, 70 years vs average 60±10 and a median of 64.5 years, Q25, Q75: 54, 67 years; p>0.05).
The diagnosis of MG was associated with the relative hazard (HR) of death of HR 2.99 (95% CI 1.26 to 7.06, p<0.05) and a median survival time from SLE diagnosis to death of 5 years (Q25, Q75: 1, 14 years; range 0–41 years) for SLE+MG cases, as compared with 12 years (Q25, Q75: 6, 19 years; range 0–62 years) for the controls; p>0.05 (figure 1, figure 2). The effect remained significant after adjustment for age at SLE diagnosis and selected comorbidities (hypertension, diabetes mellitus and peripheral artery disease): adjusted HR 3.44 (95% CI 1.38 to 8.58, p<0.05).
Observed life duration in SLE+MG and SLE controls. The analysis excluded patients with readily diagnosed lymphoproliferative disorder at the time of MG detection. HR, relative hazard of death from any cause; MG, monoclonal gammopathy.
Monoclonal gammopathy and survival by Cox proportional hazard model estimated using the SLE+MG cohort and SLE controls. The analysis excluded patients with readily diagnosed lymphoproliferative disorder at the time of MG detection. Baseline hazard function estimated using Breslow’s method. HR, relative hazard of death from any cause; MG, monoclonal gammopathy.
However, the effect was diminished when the use of CQ or HCQ was considered, with comparable survival of SLE+MG cases treated with (H)CQ, SLE+MG cases not treated with (H)CQ and SLE controls not treated with (H)CQ (p>0.05), all inferior to SLE controls treated with (H)CQ (p<0.05; figure 3).
Observed life duration in SLE+MG and SLE controls depending on (H)CQ use. The analysis excluded patients with readily diagnosed lymphoproliferative disorder at the time of MG detection. HCQ, hydroxychloroquine; HR, relative hazard of death from any cause; MG, monoclonal gammopathy.
Causes of death for the SLE+MG group included (1) infectious endocarditis with sepsis, (2) pulmonary embolism (two non-malignant cases), (3) progression of diffuse large B-cell lymphoma or remained unknown but likely malignancy-related in the three remaining cases (diagnosed with follicular lymphoma, planoepithelial lung cancer and breast cancer). None of the two MM or one MW cases proved fatal in the course of the short follow-up (0, 2 and 6 years, respectively).
Among the controls, the non-malignant causes of death were either severe infections (9 cases, 27%), SLE aggravation (4 cases, 12%), remained undisclosed (14 cases, 42%) or varied (including acute pancreatitis, pulmonary embolism, aortic dissection, subarachnoid haemorrhage, gastrointestinal bleeding and anorexia). Among the 12 fatal cases of malignancy, all 12 were due to solid tumours.
Discussion
The presented study contributes insights into the clinical associations and prognostic significance of MG in SLE and adds to the otherwise scarce data available on the subject, identifying MG as an important risk factor for developing malignancy among SLE cases.
The SLE+MG cases displayed higher rates of lymphopenia, anaemia and signs of haemolysis, but had fewer cutaneous manifestations. If biopsied, SLE+MG displayed high prevalence of ISN/RPS class IV LN (83% in our study, as compared with 50% in SLE controls), and frequently required the use of cyclophosphamide and rituximab. While no MGRS cases were identified in our SLE+MG cohort, 4 out of 10 patients with LN had no pathological confirmation and might have had undiagnosed features of MGRS as well. Moreover, recognising MGRS in the setting of LN is inherently challenging as it is considered a diagnosis of exclusion.23 The efficacy of plasma cells-depleting daratumumab in LN24 further suggests the possible contribution of monoclonal plasma cells in the disease, as the medication was originally developed for the treatment of MM.
Distinctive patterns in malignancy rates and mortality between the SLE+MG group and the controls warrant further attention. Patients with SLE+MG showed a higher incidence of malignancies, both lymphoproliferative and solid, suggestive of MG being not only a premalignant condition but also a marker of immunodeficiency (either related to SLE treatment or inherent to the disease) resulting in impaired immune surveillance contributing to higher rate of neoplasia in general.
The association between immunodeficiency and MG is further indicated by an increased infection rate among patients with SLE and MG.25 Furthermore, an association between immune system dysregulation and MG was previously demonstrated for B-cell immunodeficiency26 and HIV infection.27
Higher incidence of SLE and rheumatoid arthritis among the relatives of patients with SLE+MG indicates primary immune system dysregulation as another contributor to MG. This is supported by shared pathogenetic features of SLE and MG, such as T-cell senescence, present in both conditions.28 29
In this regard, the regulatory CD4+CD25high T-cells might have a double-edge role in the pathogenesis and progression of MG is systemic lupus. Their role as regulatory cells involved in inducing immune tolerance to autoantigens is recognised in SLE, where decreased activity of the T-regulatory cells is related to increased SLE activity.30 On the contrary, the CD4+CD25high Foxp3+ T-cells are considered the key cells allowing for the progression of MGUS to MM by reducing immune surveillance in tumour microenvironment.31 Treatment of SLE flares restores the circulating CD4+CD25high Foxp3+ T-cells numbers, likely as an epiphenomenon of SLE remission32 rather than the direct result of the immunosuppressant use. The interplay between the CD4+CD25high T-cells, systemic lupus activity, immunosuppressant use and MGUS progression requires further studies.
While the probability of MGUS progression to an overt malignant or clinically significant disease is estimated at 1% per year,33 the SLE+MG cases in our cohort were diagnosed with MM or lymphoproliferative disease early since MG detection, suggesting an increased progression risk in patients with SLE+MG. However, it needs to be noted that our patients were screened with SPEP at non-standardised intervals and inferences about time to progression can only be estimated.
The mortality rate was notably higher among patients with SLE+MG, predominantly due to infectious endocarditis and pulmonary embolism in the non-malignant setting, as well as malignant diseases, with a shorter median survival time compared with the controls. Despite patients with SLE+MG being older both at the SLE onset and at MG diagnosis, a surprising lack of differences in the age at death between MG cases and the controls indicates a worse prognosis as a result of the disease course and its complications rather than older age.
The study limitations include retrospective design (resulting in reliance on medical records with occasionally incomplete data and lack of standardisation of care), sampling bias (with the SLE population being limited to the patients of only two medical centres), necessarily long study timeframe (which might bias comparisons between patients due to changing laboratory techniques and management guidelines) and reliance on agarose gel electrophoresis for MG detection (rather than currently available more sensitive tools, such capillary electrophoresis or mass spectrometry-based methods). The strengths of our study include large sample size allowing the detection of sufficient absolute count of MG cases among patients with SLE and case-control comparisons as well as longitudinal design survival analysis.
In conclusion, despite the differences in sample sizes and demographic disparities observed between groups, these findings warrant clinical consideration of MG as a significant factor associated with the SLE course and overall prognosis, especially in regard to high incidence of lymphoproliferative diseases. The interplay between immune dysregulation and the development of MG in autoimmune disorders merits further exploration.
Data availability statement
Data are available upon reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
The Bioethics Committee of the Jagiellonian University Medical College has approved the research (approval decision no. N41/DBS/000936) as adhering to the ethical principles outlined in the Declaration of Helsinki.
References
Footnotes
AS-K and PK-S contributed equally.
Contributors Concept and design: JK-W, MK, AS-K, PK-S. Acquisition of data: AS-K, MS. Analysis and interpretation of data: AS-K, PK-S, JK-W, MS, MK. Drafting the article: AS-K, PK-S. Critical revision of the article for important intellectual content: MK, JK-W. Final approval of the version to be submitted: AS-K, PK-S, JK-W, MS, MK. Guarantor: PK-S.
Funding This work was supported by the Research Grant of Jagiellonian University Medical College No. N41/DBS/000936 (to JK-W).
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.