Our first patient with LN was treated with RTX in 2002, and subsequently a total of 62 patients with SLE have been treated with RTX at some time point during the course of their disease. For this cross-sectional study 23 female patients (25 cases) with SLE (age range: 14 – 72 yr, mean: 31 yr) with active or relapsing disease, fulfilling inclusion criteria were enrolled. The median disease duration was 6 yr (range: 2 mo-27 yr) at the time of the first RTX treatment infusion. Our patient population demographics and clinical features are presented in Table 1.

Table 1 Demographic and clinical features of patients at the time of RTX administration

The serological profile of the patients enrolled is presented in detail in Table 2. All patients tested positive for ANA. Anti-dsDNA auto Ab were positive in 7/23 patients, anti-Ro were positive in 10/23 patients, anti-RNP in 9/23 patients, and anti-Sm in 6/23 patients. Depressed circulating concentrations of C3 and/or C4 were found in 14/23 patients at the time of RTX treatment initiation.

Table 2 Serological profile of SLE patients

Clinical manifestations at the time of RTX introduction included LN in 8 patients, arthritis in 6, neuropsychiatric involvement in 4, vasculitis in 2, lung involvement in 3 and hematological abnormalities in 3. RTX was also administered in one corticosteroid-plus-azathioprine (AZA) resistant case of lupus hepatitis and in a plasmapheresis-plus-steroid resistant case of thrombotic thrombopenic purpura. Some patients either simultaneously belonged into more than one of the above clinical manifestations groups or they had different clinical features at different time points during the course of their disease.

Previous treatments included corticosteroids in all patients, hydroxychloroquine (HCQ), mycophenolate mofetil (MMF), AZA, methotrexate (MTX), cyclophosphamide (CYC), intravenous immunoglobulin (IVIG) and cyclosporine A (CSA). The commonly used RTX regimens: “the RA protocol”, which was administered at a dose of 1000 mg repeated after 2 weeks or “the lymphoma protocol” (4-weekly infusions of 375 mg/m2) were randomly employed.

Response was assessed 1 yr after the first cycle of RTX treatment and during the latest visit (≥ 10 yr following first cycle of RTX). The response rate was 68.75% at 1 yr. The response rates at the ≥ 10 yr timepoint was 75%. Subsequently it was asked the question if certain SLE manifestations were more or less prone to respond to RTX treatment. The responses seen at 1 yr and at ≥ 10 yr after RTX treatment initiation according to isolated SLE manifestations is analyzed below.

Lupus nephritis

This study aimed to identify the effect(s) of RTX treatment in patients with LN after 1 yr and after ≥ 10 yr enrollment. The laboratory parameters of the patients with LN are presented in Table 3. Among patients with LN (8 cases), proteinuria ranged from 1,062 mg/24 h to 11,500 mg/24 h. An active urinary sediment was evident in 3 out of 8 patients before RTX treatment introduction.

Table 3 Laboratory parameters of the patients with lupus nephritis at baseline and 1-yr outcomes

A renal biopsy was not performed (although recommended) in 5 patients for other-than-medical reasons. Renal biopsy was compatible with Class III LN in 1 patient, Class V in a second patient, and Class IV & Class V in a third. The latter patient achieved a partial response after RTX introduction and underwent a second kidney biopsy. Histology of her second biopsy was compatible with predominately Class V LN with the appearance of sclerotic lesions. Among patients with LN, evidence of interstitial nephritis was present in all 3 biopsies. Finally, one patient developed end-stage renal disease, another one developed chronic kidney disease while the third patient responded clinically to RTX treatment.

The effects of RTX treatment on renal function were examined by evaluating the eGFR before RTX introduction (Mean ± SD: 92.62 ± 38.95), 1 yr after RTX treatment initiation (Mean \(\pm\) S.D.: 92.50 \(\pm\) 42.85, p = 0.99 vs. baseline) and finally after ≥ 10 yr (Mean \(\pm\) S.D.: 75.62 \(\pm\) 52.78, p = 0.47 vs. baseline) (Fig. 1). Therefore, eGFR did not decline significantly after 1 or after more than 10 years of follow-up. It is of note that non-responders (2/8) that developed end-stage renal disease (ESRD) had an already seriously limited kidney function at baseline (eGFR: 18 ml/min and 50 ml/min, respectively).

Fig. 1

Renal function (eGFR) of patients with lupus nephritis at baseline, after 1 year and after ≥ 10 years following RTX treatment introduction compared to baseline values [X axis: patients’ numbers, Y axis: eGFR (ml/min)]

After 1 yr, an overall renal response was observed in 75% (6 of 8); a complete response in 50% (4 of 8) and a partial response in 25% (2 of 8). This beneficial renal outcome dropped to 62.5% during the ≥ 10 yr visit. More specifically, 1/2 patients achieving an initial (1 yr) partial response was finally classified as having a complete response at the ≥ 10 yr visit. Of note, 2 out of the 3 patients who did not eventually benefit from RTX treatment developed ESRD necessitating hemodialysis; these were the 2/8 patients who did not respond at the earlier (1 yr) timepoint. Those 2 patients along with a third patient who finally developed chronic kidney disease were defined as non-responders. Thus, patients with LN of this study did benefit from RTX treatment because their eGFR apparently stabilizes early after RTX initiation and this effect lasts long. It should be noted herein that the long-term analysis of kidney function of patients with LN treated with the standard-of-care has been associated with less favorable long-term results. In contrast, RTX did not prevent the development of ESRD in those patients having an eGFR < 60 ml/min at the time of RTX treatment introduction.

Non-renal SLE

To evaluate the potential effect(s) of RTX treatment in refractory non-renal SLE manifestations the short and long-term efficacy of B cell depletion in 17 additional cases were analyzed.

Neuropsychiatric manifestations

This study included 4 patients with neuropsychiatric involvement to estimate the benefit(s) of RTX treatment in SLE patients with neuropsychiatric SLE (NPSLE). Their clinical manifestations included: benign intracranial hypertension, diplopia, camptocormia due to diffuse dystonia and transverse myelitis. The overall clinical response was 100% after 1 yr of RTX treatment introduction. The one patient with transverse myelitis relapsed at a later-than-1 yr timepoint. Therefore, the overall clinical response was 75% at ≥ 10 yr. Even though the number of patients is small, our long-term observations suggest that RTX may be of benefit in patients with NPSLE, a manifestation of SLE with obscure pathogenesis and incompletely defined treatment approach.

Hematological manifestations

The efficacy of RTX treatment in SLE patients with hematological abnormalities was also examined. Cytopenias were recorded in 3 patients (neutropenia in 1 and autoimmune hemolytic anemia in 2). An increased neutrophil count from 663/mm3 to 1,188/mm3 at 1 yr, and to 1,581/mm3 after ≥ 10 yr was recorded in the patient with neutropenia. The first patient with autoimmune hemolytic anemia (Ht: 30%, Hb: 9.4 g/dL) improved at 1 yr (Ht: 37.8%, Hb: 12.8 g/dL), but relapsed afterwards (Ht: 27%, Hb: 8.4 g/dL). The second patient with autoimmune hemolytic anemia, that was accompanied by arthritis, (Ht: 22.0%, Hb: 7.2 g/dL) also improved after RTX treatment introduction (Ht: 32.3%, Hb: 9.5 mg/dL and Ht: 32.6%, Hb: 10.2 mg/dL at 1 yr and ≥ 10 yr, respectively). Therefore, the hematological response were 100% at 1 yr and 66.66% after ≥ 10 yr, which could be considered as encouraging, despite the small number of patients enrolled in the study.

Vasculitic manifestations

To address the efficacy of RTX treatment in other severe non-renal manifestations of SLE 2 patients with vasculitis were included. There was 1 patient with intestinal vasculitis presenting with abdominal pain and vomiting. Contrast enhanced CT-scans of the abdomen demonstrated bowel wall thickening and abnormal wall enhancement (“double-halo” or “target” sign). A relapse occurred at 1 yr after RTX treatment initiation as well as at a later time point before 10 yr. Both relapses were successfully managed with IV pulsed corticosteroids. Another patient presenting with livedo reticularis resistant to corticosteroid treatment displayed complete resolution not only after 1 yr but also after ≥ 10 yr. Overall, the clinical response was 50% at 1 yr and 50% at the ≥ 10 yr timepoint.

Pulmonary manifestations

Aiming to examine the efficacy of RTX treatment in SLE pulmonary disease, 3 patients with lung involvement were enrolled. Among those patients 1 out of 3 had chronic lupus pneumonitis manifesting as interstitial lung disease. High-resolution computed tomography (HRCT) of the chest showed that a previously mild and restricted interstitial fibrosis changed to / evolved into organizing pneumonia at 1 yr after RTX treatment introduction; however, after ≥ 10 yr all CT scan findings had cleared. RTX was also administered in 2 patients with “shrinking lung” syndrome. The first patient did not benefit from RTX treatment at any timepoint (FVC: 67.2% predicted at baseline, 57.4% predicted at 1 yr and 58% predicted at ≥ 10 yr), but the second patient improved not only at 1 yr but also at the ≥ 10 yr visit (FVC: 50% predicted at baseline, 62% at 1 yr and 60% at ≥ 10 yr). The overall clinical response was 33.33% at 1 yr and 66.67% at the ≥ 10 yr timepoint. Thus, RTX treatment might represent an alternative therapeutic approach in patients with SLE and lung involvement, although the number of patients enrolled in the study is limited.

Mucocutaneous manifestations

This study evaluated 4 patients with mucocutaneous involvement, including erythematosus lesions in 3/4 and oral ulcers in 1/4 in order to determine the short- and long-term results of RTX treatment on skin and mucosal involvement. In all cases cutaneous disease was persistent and refractory to corticosteroids, topical therapies and antimalarials. A relapse occurred at 1 yr after RTX treatment in 1 patient despite an initial impressive improvement. However, this same patient had no signs of active skin disease at the ≥ 10 yr timepoint. There was no benefit from RTX treatment in one patient at all. Thus, the clinical response rates were 50% and 75% at 1 yr and at the ≥ 10 yr timepoint, respectively, rendering RTX treatment another potentially useful therapeutic option in the management of refractory mucocutaneous disease in patients with SLE, although the number of patients included in the study is small.

Articular manifestations

In order to evaluate the impact of RTX treatment in SLE patients with refractory articular involvement 6 patients with treatment-resistant arthritis were enrolled. All of them had persistent polyarthritis principally affecting the wrists, metacarpophalangeal and proximal interphalangeal joints, causing significant functional impairment. A combination of previously administered DMARDs such as HCQ and MTX were of no benefit. An initial improvement was recorded in all 6 patients; however 2/6 patients relapsed at 1 yr after RTX administration; both patients improved following a short-term corticosteroid course. Therefore, the clinical response was 66.67% at 1 yr after RTX treatment initiation. However, none of our 6 patients had active arthritis at the ≥ 10 yr timepoint. Thus, the long-term clinical response for SLE arthritis was 100%. In conclusion, RTX treatment may represent a reliable long-term treatment solution for patients with SLE and refractory arthritis, despite the limited number of patients enrolled in the study.

Miscellaneous manifestations

There were 2 patients with distinct manifestations that could not be classified into previous groups. A patient with lupus hepatitis [aspartate aminotransferase (AST): 308, alanine aminotransferase (ALT): 187, gamma-glutamyl transferase (GGT): 101, alkaline phosphatase (ALP): 221 at baseline] improved after a single course of RTX treatment showing complete normalization of serum values of the liver enzymes (AST: 33, ALT: 41, GGT: 45, ALP: 65 at 1 yr and AST: 18, ALT: 10, GGT: 13, ALP: 76 after ≥ 10 yr).

A patient with thrombotic thrombopenic purpura [hematocrit (Ht): 8.8%, hemoglobin (Hb): 3.9 g/dL, presence of schistocytes in the peripheral blood, platelets (PLT): 46,000/mm3 at baseline) responded favorably to RTX treatment (Ht: 35.8%, Hb: 11.8 g/dL, PLT: 252,000/mm3 at 1 yr, and Ht: 30.2%, Hb: 10.2 g/dL, PLT: 296,000/mm3 at the ≥ 10 yr timepoint). The overall short- and long-term therapeutic response for those isolated miscellaneous manifestations was 100%.

The cSLEDAI-2 k index response

It was decided that 16 cases were suitable for cSLEDAI-2 k calculation. According to the results the median cSLEDAI-2 K score decreased from 5.83 ± 3.70 at baseline to 1.95 ± 2.40 (p < 0.001) at 1 yr after RTX treatment initiation and to 2.37 ± 3.00 (p < 0.001) at the ≥ 10 yr time point of follow-up (Fig. 2).

Fig. 2

Reduction of cSLEDAI-2 k at 12 months after RTX treatment initiation and at last visit compared to baseline (p < 0.05)

Follow-up

Over the course of our at least 10 yr of follow-up, 10 out of 23 of our patients relapsed (43.48%) (Table 4). The earliest relapse was observed at 6 months and the latest one at 13 years after initial RTX treatment introduction. Among patients that relapsed 8 were retreated with RTX and 3/8 (37.50%) re-responded. Notably, 1 patient developed diffuse cervical lymphadenopathy after 1 month following the first RTX infusion. Lymph node biopsy excluded malignancy and DNA analysis disclosed clonal TCR-rearrangements but not malignancy; B cells were practically absent from the lymph node, a finding attributed to the preceding treatment.

Multiple cycles of RTX treatment were employed successfully for remission maintenance in 6 out of 23 patients. However, 3/23 (13%) patients achieved a sustained remission after the administration of one single course of RTX treatment.

Causes for RTX treatment discontinuation were: lack of efficacy (9/23, 39.13%) and adverse events (4/23, 17.39%). Three patients were lost to follow-up after completing a 10 yr follow-up. However, it should be noted that RTX treatment was discontinued in 6/23 (23.08%) because of long-term SLE remission, based on clinical judgement.

Steroid sparing effect

The potential effects of RTX treatment on the reduction of the mean daily dose of corticosteroids were analyzed in 22 out of 25 cases where data were available. The mean daily dose of corticosteroids was reduced both at the1yr and at the ≥ 10 yr time points; however, such decreases were not but statistically significant (p = 0.07 and p = 0.17, respectively) (Fig. 3).

Fig. 3

Median corticosteroid dose at 12 months after RTX treatment initiation and at last visit compared to baseline

In conclusion, RTX treatment in our cohort was not associated with a significant steroid sparing effect. The importance of long-term low doses of steroids and the contribution of other-than-steroids agents to achieve this has been recently recognized in all therapeutic trials in SLE.

Safety

No cases of mycobacterial or systemic fungal infections occurred during the ≥ 10 yr follow-up. Uncomplicated herpes zoster infection developed in 2 patients. A patient developed CMV-related retinitis resulting in single-eye blindness 3 yr following the single RTX infusion cycle she received. This same patient developed ecthyma gangrenosum due to Pseudomonas aeruginosa and Fusobacterium necrophorum co-infection 4 yr after RTX treatment, leading to hospitalization and a prompt response after prolonged IV antibiotic treatment. Due to the occurrence of unusual and severe infections despite a long time-interval from RTX treatment in this patient, laboratory tests were performed that revealed diffuse hypogammaglobulinemia. Detailed evaluation of previously unavailable to us medical records found that she had long-established diffuse hypogammaglobulinemia and a flow-cytometrically documented circulating B cell deficiency. Severe infections were not unknown to her previously, although she was never offered IVIG treatment. Despite treating her with regular IVIG substitution infusions for years afterwards in our center, the patient finally succumbed due to SARS-CoV-2 infection 10 yr after her one and only RTX infusion cycle.

Mild infusion reactions were recorded in 4 patients. A single case of late-onset neutropenia was identified in our cohort, with the absolute neutrophil count dropping to < 500/mm3 2 months after the second course of RTX treatment. This patient presented with low-grade fever without any other signs of infection and was not hospitalized; she was successfully treated with a short course of granulocyte-colony stimulating factor.