To estimate the prevalence and investigate the clinical features and the outcome of clinically significant renal involvement in a large cohort of patients with primary Sjögren's syndrome (SS).
To estimate the prevalence and investigate the clinical features and the outcome of clinically significant renal involvement in a large cohort of patients with primary Sjögren's syndrome (SS).
Among 715 patients who met the American–European Consensus Group criteria for primary SS, those with clinically significant renal involvement were identified and their clinical and immunologic features were recorded. The prognosis in patients with primary SS with renal involvement was assessed by the clinical appearance of any of the following major outcomes: death, hemodialysis, chronic renal failure (CRF), and lymphoma. Kaplan-Meier analysis was applied to compare death rates between patients without and those with renal involvement.
Thirty-five patients with primary SS (4.9%) had clinically significant renal involvement, representing a total followup time after renal diagnosis of 252.2 person-years. Thirteen patients (37.1%) had interstitial nephritis alone, 17 patients (48.6%) had glomerulonephritis (GN) alone, and 5 patients (14.3%) had both entities. Nine patients died (25.7%), 11 developed CRF (including 4 requiring chronic hemodialysis) (31.4%), and 9 developed lymphoma (25.7%). The overall 5-year survival rate was 85%. Kaplan-Meier analysis showed statistically significant reduced survival for patients with primary SS with renal involvement compared to those without renal involvement (P < 0.0001 by log rank test), with GN patients displaying increased mortality. Eight of 9 reported deaths (89%) and 8 of 9 lymphomas (89%) were observed among patients with GN.
The long-term prognosis varies for patients with primary SS who have clinically significant renal involvement. Patients with interstitial nephritis display a favorable prognosis, while patients with GN are at high risk of developing lymphoma and have poor survival.
Primary Sjögren's syndrome (SS) is a systemic autoimmune disease that mainly affects the exocrine glands (). During the disease course, almost one-half of patients develop extraglandular manifestations attributed to periepithelial lymphocytic infiltration of the affected organs such as the kidneys, the lungs, and the liver (). Additionally, the disease may be complicated by immune complex–mediated manifestations including purpura, glomerulonephritis (GN), and peripheral neuropathy. Hence, 2 types of renal involvement are observed in patients with primary SS, representing the 2 different underlying immunologic mechanisms: interstitial nephritis and GN ([2-4]).
Renal involvement has been well documented in primary SS. In large series, the occurrence of interstitial nephritis ranges from 12% to 48%, but it appears that this entity evolves slowly with a rather benign course ([5-10]). In contrast, clinically overt GN has been described in few cases in the literature ([6, 7, 9, 11-15]). However, severe renal disease in primary SS, requiring kidney biopsy, is relatively rare (). The prognosis and outcome in these patients remain unclear, since no clinical study has addressed this issue so far. The aim of the present study was to investigate the clinical picture and the outcome in patients with primary SS with clinically significant renal involvement, including both interstitial nephritis and GN.
In 2012, from a cohort of 715 patients who fulfilled the American–European Consensus Group criteria for primary SS (), we identified those with clinically significant renal involvement. Clinically significant renal involvement in primary SS, either interstitial nephritis or GN, was defined by one or more of the following criteria ():
The clinical, histologic, and immunologic features of all patients along with their prognosis and outcome were investigated. The primary end points for clinical outcome were death, hemodialysis, chronic renal failure (CRF) (defined as creatinine serum levels >1.6 mg/dl or creatinine clearance <50 ml/minute) not requiring hemodialysis, and lymphoma. The duration of primary SS as well as age and the duration of renal disease were recorded until the last followup visit or the appearance of any of the above-mentioned end points. When 2 different clinical outcomes occurred in the same patient, the following order of clinical importance was followed: death > hemodialysis > CRF > lymphoma. To compare the overall death rates as well as the death rates due to lymphoma between patients with primary SS without renal involvement and those with renal involvement, a 6:1 randomization was performed after adjusting for disease duration and sex, and Kaplan-Meier analysis was applied. The study was approved by the Ethics Committee of the School of Medicine, National University of Athens.
Corresponding rates for hemodialysis, CRF, and lymphoma per 100 person-years were calculated. Kaplan-Meier analysis was performed to estimate the survival after renal diagnosis and the 10-year proportion without renal failure; 2 × 2 contingency tables for categorical data were analyzed with Fisher's exact test. All statistical analyses were conducted using GraphPad Prism version 6.0 and SPSS version 20.0 software packages.
Thirty-five of the 715 patients with primary SS (4.9%) had clinically significant renal involvement, with a cumulative followup time after renal diagnosis of 252.2 person-years (Table 1). All patients were women, their median age at entry was 52 years (range 25–76 years), and their median disease duration was 11 years (interquartile range 8–18 years). Seventeen patients (48.6%) had GN alone and 13 patients (37.1%) had interstitial nephritis alone. Five patients (14.3%) developed both entities and were counted in the group with GN. Two of the 35 patients with primary SS with renal disease refused renal biopsy. Of the 22 patients with GN, the renal biopsy findings in 21 included membranoproliferative GN (10 patients), mesangial GN (7 patients), membranous GN (2 patients), proliferative GN (1 patient), and focal segmental glomerulosclerosis (1 patient). In 12 of the 13 patients with interstitial nephritis alone, lymphocytic infiltration of the interstitium and renal tubules was noted. Nineteen of the 35 patients had already been identified in 1998 (), while 16 additional patients diagnosed between 1998 and 2012 were included in the present study.
|Patient/agea||Primary SS disease duration, yearsb||Type of renal involvement||Renal disease duration, yearsc||Death||Hemodialysis||CRF||Lymphoma|
|20/74||27||IN, MP GN||13||+||–||–||+|
|31/75||31||IN, MS GN||23||–||–||–||–|
|32/55||14||IN, MS GN||2.5||+||–||–||+|
|33/43||15||IN, proliferative GN||0||–||–||+||–|
In 1998, from a cohort of 471 patients who fulfilled the European criteria for primary SS (), 20 patients were found to have renal involvement severe enough to warrant the performance of kidney biopsy. All these patients also fulfilled the American–European Consensus Group criteria for primary SS. Ten of them had interstitial nephritis alone, 8 had GN alone, and 2 fulfilled the criteria for both entities (patients 1–20 according to the published list in 2000) (). Patient 18, who had GN alone, was lost during followup and excluded from the present study due to insufficient data. The characteristics of renal involvement in all 35 patients are shown in Table 2.
|Patient||Clinical manifestation||Creatinine, mg/dl||Urine abnormality||Imaging abnormality||Biopsy finding|
|1||None||1.0||SG <1,010, pH >7||None||IN|
|2||Renal colics||0.8||SG <1,010, pH >7||Urolithiasis||IN|
|3||Renal colics||1.2||SG <1,010, pH >7||Urolithiasis||IN|
|4||None||2.2||SG <1,010, pH >7||None||IN|
|5||Polydipsia, polyuria, nocturia||1.8||SG <1,010, pH >7||None||IN|
|6||None||1.1||SG <1,010, pH >7||None||IN|
|7||None||1.0||SG <1,010, pH >7||None||IN|
|8||None||NS||SG <1,010, pH >7||None||IN|
|9||None||NS||SG <1,010, pH >7||None||IN|
|10||Renal colics||2.4||SG <1,010, pH >7||Urolithiasis||ND (presumably IN)|
|11||Hypertension, periorbital edema||1.4||Proteinuria, hematuria||None||MP GN|
|12||None||0.9||Proteinuria, hematuria||None||MS GN|
|13||None||4.9||Proteinuria, hematuria||None||MP GN|
|14||None||0.8||Proteinuria, hematuria||None||MP GN|
|16||None||1.1||Proteinuria, hematuria, RBC casts||None||MS GN|
|19||Renal colics||6.2 (at GN onset)||SG <1,010, pH >7, proteinuria, hematuria, RBC casts||Urolithiasis||ND (presumably IN + GN)|
|20||Hypertension||1.6 (at GN onset)||SG <1,010, pH >7, proteinuria||None||IN, MP GN|
|21||Nephrogenic diabetes insipidus||0.7||Mild proteinuria (0.5 gm/24 hours), SG <1,010, low urine osmolality||None||IN|
|22||None||2.3||Impaired renal function, alkaline urinary Ph||None||IN|
|23||None||2.6||Impaired renal function||None||IN|
|24||None||1.1||Mild proteinuria (0.5 gm/24 hours), microscopic hematuria of glomerulus origin||None||MS GN|
|25||Nephrotic syndrome, edema of lower extremities, hypertension||0.9||Proteinuria (6.4 gm/24 hours), microscopic hematuria of glomerulus origin||None||MP GN|
|26||Nephrotic syndrome, edema of lower extremities, hypertension||0.7||Proteinuria (3.5 gm/24 hours), microscopic hematuria of glomerulus origin, RBC casts||None||MP GN|
|27||Edema of lower extremities, periorbital edema, hypertension||1||Mild proteinuria (1 gm/24 hours), microscopic hematuria of glomerulus origin||None||MP GN|
|28||None||1.5||Mild proteinuria (1.2 gm/24 hours)||None||MP GN|
|29||None||0.8||Proteinuria (2 gm/24 hours)||None||MP GN|
|30||None||1||Mild proteinuria (1 gm/24 hours), microscopic hematuria||None||MP GN|
|31||Renal colics||0.8||SG <1,010, microscopic hematuria of glomerulus origin||Urolithiasis||IN, MS GN|
|32||None||1.2||Mild proteinuria, macroscopic hematuria, alkaline urinary pH||None||IN, MS GN|
|33||None||1.52||Impaired renal function, SG <1,010, alkaline urinary pH||Urolithiasis||IN, proliferative GN|
|34||Nephrotic syndrome, edema of lower extremities, periorbital edema, generalized edema||1.9||Impaired renal function, proteinuria (3.5 gm/24 hours), RBC casts||None||FSGS|
|35||None||0.9||Microscopic hematuria||None||Membranous GN|
|36||Hypertension||0.9||Proteinuria (3 gm/24 hours)||None||Membranous GN|
Analysis of primary outcomes revealed that 9 patients died (25.7%), 9 developed non-Hodgkin's lymphomas (25.7%), and 11 developed CRF (31.4%), of whom 4 required hemodialysis (Table 1). The corresponding rates for lymphoma, CRF, and hemodialysis were estimated at 3.6 (95% confidence interval [95% CI] 1.7–6.7), 4.4 (95% CI 2.2–7.7), and 1.6 (95% CI 0.4–4.0), respectively, per 100 person-years. Of the 9 reported deaths, 3 were attributed directly to lymphomas (33.3%), 2 patients died of cardiac arrest related to either hemodialysis or CRF (22.2%), 1 died of infection (11.1%), 1 died of stroke, and 1 died of cerebral hemorrhage. In 1 patient the cause of death was unknown. The estimated 5-year overall survival rate after renal diagnosis was 0.85 (95% CI 0.73–0.97). However, after adjusting for age, type of renal involvement (interstitial nephritis versus GN), lymphoma, and hemodialysis, lymphoma remained the sole significant adverse predictor for survival (adjusted hazard ratio 6.49 [95% CI 1.24–34.48], P = 0.026).
To compare the death rates between patients with primary SS without renal involvement and those with renal involvement, a 6:1 randomization was performed in the cohort of 715 patients with primary SS, after adjusting for disease duration and sex. The mean ± SD age and disease duration at last followup in patients with primary SS with renal involvement were 52.8 ± 13.85 years and 12.62 ± 7.77 years, respectively, compared to 64.60 ± 12.62 years and 14.84 ± 7.05 years, respectively, in patients with primary SS without renal involvement (P < 0.0001 and P = 0.062, respectively, by Mann-Whitney test). Among the 210 patients with primary SS without renal involvement, 12 deaths (5.7%) were recorded; 4 patients (1.9%) died because of lymphoma, 3 patients (1.43%) due to solid tumor (breast, colon, and urinary bladder), 2 patients (0.95%) due to heart failure, and 1 patient (0.48%) because of infection and septicemia, while for 2 patients (0.95%) the cause of death was not known. Death rates for patients with primary SS with renal involvement and those without renal involvement are presented in Table 3. Kaplan-Meier analysis revealed reduced survival of patients with renal involvement compared to those without (P < 0.0001 by log rank [Mantel-Cox] test) (Figure 1A). A marginal difference was found for the death rates due to lymphoma between patients with primary SS with renal involvement and those without (3 of 35 versus 4 of 210, respectively; P = 0.0625 by Fisher's exact test).
|Cause of death||Renal involvement (n = 35)||No renal involvement (n = 210)|
|Lymphoma||3 (8.6)||4 (1.9)|
|Infections||1 (2.86)||1 (0.48)|
|Cardiac arrest related to renal involvement||2 (5.7)||–|
|Heart failure||–||2 (0.95)|
|Solid tumors||–||3 (1.43)|
|Unknown||1 (2.86)||2 (0.95)|
The comparison of clinical, immunologic, and laboratory features between GN and interstitial nephritis patients is presented in Table 4. GN developed later during the disease course compared to interstitial nephritis (mean ± SD disease duration 7.08 ± 5.28 years versus 2.75 ± 3.33 years; P = 0.008). The proportion of patients with glandular and extraglandular manifestations as well as the autoantibody profile was not statistically different between the 2 groups. Interestingly, immune complex–related manifestations such as purpura and peripheral neuropathy occurred in similar frequencies between the 2 groups. In contrast, low serum C3 levels and mixed monoclonal cryoglobulins of either type II or type III were detected more frequently in patients with GN compared to patients with interstitial nephritis (P = 0.013 for both). It is noteworthy that 8 of 22 patients with GN had previous measurement of cryoglobulins 1–10 years prior to the onset of renal disease, and 5 of them had been found to be positive. No difference was found in C4 serum levels between the 2 groups. Statistically significantly higher rates of active urine sediment and proteinuria were observed in GN patients compared to interstitial nephritis patients (P = 0.0001 for both).
|Clinical, immunologic, and laboratory parameters||Interstitial nephritis (n = 13)||GN (n = 22)||Pa|
|Age at onset of renal involvement, mean ± SD years||42.38 ± 15.16||47.55 ± 10.35||0.374|
|Primary SS disease duration prior to renal involvement, mean ± SD years||2.75 ± 3.33||7.08 ± 5.28||0.008|
|Active urine sedimentb||0/13||15/22||0.0001|
The percentage of reported deaths was 36.4% (8 of 22) in those with GN versus 7.7% (1 of 13) in those with interstitial nephritis. There was a tendency for GN patients to exhibit increased mortality compared to interstitial nephritis patients, as 8 of 9 recorded deaths were observed in GN patients (P = 0.092 by log rank [Mantel-Cox] test) (Figure 1B). Lymphoma development occurred in 8 of 22 GN patients (36.4%) versus 1 of 13 interstitial nephritis patients (7.7%). Of the 9 patients with primary SS with renal involvement who developed lymphoma, 8 (89%) had GN and 1 (11%) had interstitial nephritis. Of the 8 patients with GN and lymphoma, 2 developed lymphoma prior to GN onset, 5 developed lymphoma within 3 months of GN occurrence, and 1 developed lymphoma after 6 months of GN. CRF was observed in 11 patients (31.4%); 7 had interstitial nephritis, 2 had GN, and 2 had both and were excluded from the Kaplan-Meier analysis concerning the development of CRF. At 10 years, the estimated proportion without CRF was 0.61 (95% CI 0.30–0.92) in those with interstitial nephritis versus 0.94 (95% CI 0.81–1) in those with GN, and patients with interstitial nephritis were significantly more likely to develop CRF over time compared to GN patients (P = 0.037 by log rank [Mantel-Cox] test) (Figure 1C). Finally, end-stage renal disease requiring hemodialysis developed in 2 of 22 patients with GN (9.1%) and in 2 of 13 patients with interstitial nephritis (15.4%).
All patients with interstitial nephritis received potassium bicarbonate supplements. For patients with GN, the management was instituted based on clinical judgment. Thus, different approaches were applied as induction therapy, as follows: monthly intravenous pulses of cyclophosphamide at 1 gm/m2 body surface area plus 1 gram methylprednisolone, oral methylprednisolone as monotherapy (0.5–0.75 mg/kg daily) or in combination with another immunosuppressant (2 mg/kg azathioprine daily or 2–3 mg/kg cyclosporine daily), and B cell depletion therapy (3 cycles of rituximab; in each cycle 2 infusions of 1 gram of rituximab were administered 2 weeks apart and each cycle was repeated every 6 months). Seven of the 22 patients with GN (patients 11, 12, 13, 14, 19, 25, and 33) (Table 1) received long-term (for at least 6 months) treatment with intravenous pulses of cyclophosphamide plus methylprednisolone, and remission was achieved in 5 of them, while the other 2 patients (patients 13 and 19) required hemodialysis. In 2 patients (patients 26 and 28), short courses (1 or 2 pulses monthly) of intravenous pulse cyclophosphamide plus methylprednisolone were administered, resulting in complete remission. Five patients (patients 16, 17, 27, 29, and 35) received moderate to high doses of oral methylprednisolone with very good clinical and laboratory response. Methylprednisolone plus azathioprine or cyclosporine was given to 2 patients (patients 15 and 36), leading to remission of the features of GN. Patient 24 received 3 cycles of rituximab every 6 months, and complete remission was achieved.
Three patients (patients 30, 32, and 34) also had lymphoma. Treatment for lymphoma containing rituximab (rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone) was administered in 2 of them; in patient 34 renal disease was ameliorated, but in patient 30 proteinuria was sustained. Patient 32 received chemotherapy without rituximab with good response regarding renal function. Finally, no treatment was prescribed in 2 patients (patients 20 and 31), 1 of whom (patient 20) developed lymphoma.
The kidney is a potential target of 2 distinct immunopathologic phenomena in primary SS: either activated lymphocytes infiltrate tubular epithelium, resulting in interstitial nephritis, or an immune complex–mediated process causes GN ([3, 4, 18, 19]). Interstitial nephritis affects almost one-third of patients with primary SS but is considered a clinically silent entity that may escape clinical attention ([9, 20]). In a previous study at our institution, it has been shown that interstitial nephritis in primary SS occurs early in the disease course, in the context of other extraglandular manifestations which are all characterized by the presence of periepithelial lymphocytic infiltration of the involved organs (). Interstitial kidney disease may produce latent or complete distal tubular acidosis that is manifested by persistent alkaline urinary pH and hyposthenuria, hyperchloremic hypokalemic metabolic acidosis with anion gap, and nephrolithiasis/nephrocalcinosis along with recurrent renal colics ([9, 20, 21]). Rarely, interstitial kidney disease may affect the proximal renal tubules, which is clinically expressed as Fanconi's syndrome (type II renal tubular acidosis) ([14, 22, 23]).
In contrast, GN is less frequent but is manifested by more apparent clinical and laboratory features including hypertension, periorbital edema, active urine sediment, proteinuria, and renal insufficiency ([11, 19]). In the majority of patients with GN and primary SS, IgMκ-containing cryoglobulins are detected, along with low C4 levels. The presence of C3 and IgM deposits in biopsy specimens from these patients further supports the notion that cryoglobulinemia might possibly be the pathogenetic mechanism of GN in primary SS, and it appears that this type of renal involvement occurs as a late sequela of the disease (). However, clinically significant renal involvement in primary SS that may lead the clinician to perform kidney biopsy is rare, with interstitial nephritis and/or GN documented in <5% of the population of patients with primary SS (). The long-term prognosis of these patients and the major adverse events during the disease course have not been studied yet, considering the rarity of overt renal disease in primary SS. However, previous studies provide some evidence about the outcome of GN in primary SS. Skopouli et al reported that GN and lymphoma in primary SS tend to coexist in the same patients (), while almost half of patients with primary SS with GN have also developed lymphoma (), indicating that GN is associated with a less favorable prognosis.
Although the number of patients with severe renal involvement and primary SS included in this study is rather small, to our knowledge it remains the largest series with such patients studied. The total followup time was ∼252 person-years, and the estimated prevalence of clinically significant renal involvement was found to be 5% among patients with primary SS. Almost 60% of patients had GN clinically expressed by proteinuria, microscopic hematuria, impaired renal function, or nephritic syndrome, while 40% developed interstitial nephritis characterized by persistent alkaline urinary pH or hyposthenuria and less frequently by renal colics. The main histologic types of GN were membranoproliferative and mesangial GN. The majority of patients with primary SS with GN had low serum C4 levels and cryoglobulinemia, mainly of type II IgMκ. Additionally, GN patients had statistically significantly lower serum C3 levels compared to interstitial nephritis patients. The above data further support the main hypothesis that the pathogenesis of GN is associated with cryoprecipitable immune complexes that may deposit in the renal vascular filter, eliciting an inflammatory reaction. The fact that cryoglobulins had been measured in some patients with primary SS and GN years before the onset of renal disease and were found to be present in the majority of them suggests that cryoglobulinemia could be a possible predictor of GN in the population of patients with primary SS. However, lack of relative data concerning the detection of cryoglobulins does not allow any further clinical inferences.
In the present study, it was also shown that patients with primary SS with renal involvement had significantly reduced survival compared to those without renal involvement, with lymphoma being the leading cause of death for both groups. Patients with primary SS with renal disease exhibit higher death rates due to lymphoma than patients with primary SS without renal disease. The fact that death and lymphoma occurred more frequently in the GN patients than in the interstitial nephritis patients indicates that the increased mortality among patients with primary SS with renal involvement should be attributed to the proportion contributed by those with GN. Interestingly, the majority of GN patients developed lymphoma within 3 months of the appearance of renal disease. This short interval implies a direct association between lymphoma and GN, suggesting that GN might be a paraneoplastic phenomenon of an underlying lymphoproliferative process. However, all patients with primary SS who developed GN were evaluated for the presence of lymphoma at the time of diagnosis, with negative findings. In addition, the majority of them had cryoglobulinemia, which appears to play a pathogenetic role in a series of clinical manifestations such as peripheral neuropathy and purpura. Thus, it seems more possible that lymphoma and GN are closely related by sharing common pathogenetic mechanisms.
Indeed, the different impact of GN and interstitial nephritis on survival and morbidity of patients with primary SS reflects the 2 distinct underlying immunologic mechanisms mentioned above. Interstitial kidney disease is considered part of the extraglandular manifestations that are usually mild and evolve slowly without significant impairment of organ function over time. In contrast, purpura, GN, and peripheral neuropathy are considered extraepithelial manifestations of the disease that are attributed to a cryoprecipitable immune complex–mediated process possibly due to cryoglobulinemia ([2, 24]). Extraepithelial manifestations appear to occur later in the disease course, although cryoglobulinemia as a marker of monoclonality can be an early feature of the disease ([25, 26]). This subset of patients has the systemic form of the disease and is characterized by high mortality and morbidity rates as well as a tendency to develop lymphoma (). Considering that mortality in primary SS is slightly increased because of lymphoma development and that cryoglobulinemia is an independent risk factor for death ([27, 28]), the prognosis for patients with GN is expected to be less favorable than that for those with interstitial nephritis, as is strongly supported by the present study.
Unexpectedly, of the 11 patients who developed CRF, 7 had interstitial nephritis alone and 2 had GN alone, while 2 had both entities. The 10-year proportion without CRF was 61% for the interstitial nephritis patients and 94% for the GN patients. These findings imply that CRF in this series of patients with primary SS with renal disease is more likely to occur in interstitial nephritis patients than in GN patients. This observation is inconsistent with the notion that interstitial nephritis evolves slowly without causing renal damage, but it can be interpreted as a consequence of the reduced survival of GN patients and the fact that GN was mild in severity and was diagnosed and treated early, yielding a better outcome for renal function. In contrast, interstitial nephritis does not draw clinical attention and in some cases may lead to CRF. The small number of patients does not allow any conclusions about treatment efficacy, but it seems that for severe forms of GN intravenous pulses of cyclophosphamide plus methylprednisolone should be considered, whereas other regimens including oral methylprednisolone alone or in combination with azathioprine or cyclosporine and B cell depletion may be administered according to clinical judgment, histologic features, disease severity, and comorbidities.
In summary, the long-term prognosis varies for patients with primary SS with renal disease that is severe enough to warrant the performance of kidney biopsy. Although patients with primary SS with renal involvement have reduced survival, it seems that this fact is closely related to the type of renal disease and the propensity to develop lymphoma. Patients with interstitial nephritis have a better prognosis and clinical outcome, reflecting the benign nature and the clinical course of all the extraglandular manifestations characterized by the presence of periepithelial lymphocytic infiltration. In contrast, GN displays a less favorable prognosis and is associated with lymphoma development, as a part of extraepithelial manifestations mediated by cryoprecipitable immune complexes possibly due to cryoglobulinemia, which is considered to be an independent risk factor for mortality and morbidity in primary SS.
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. Tzioufas 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. Goules, Tatouli, Moutsopoulos, Tzioufas.
Acquisition of data. Goules, Tatouli, Tzioufas.
Analysis and interpretation of data. Goules, Tatouli, Moutsopoulos, Tzioufas.