Long-term follow-up analysis after rituximab salvage therapy in adult patients with immune thrombocytopenia


  • Francesco Zaja,

    Corresponding author
    1. Clinica Ematologica, Centro Trapianti e Terapie Cellulari “Carlo Melzi”, DISM, Azienda Ospedaliero Universitaria S. M. Misericordia, Udine, Italy
    • Clinica Ematologica, Centro Trapianti e Terapie Cellulari “Carlo Melzi”, DISM, Azienda Ospedaliero Universitaria S. M. Misericordia, p.le S. Maria Misericordia 15, 33100 Udine, Italy
    Search for more papers by this author
  • Stefano Volpetti,

    1. Clinica Ematologica, Centro Trapianti e Terapie Cellulari “Carlo Melzi”, DISM, Azienda Ospedaliero Universitaria S. M. Misericordia, Udine, Italy
    Search for more papers by this author
  • Marianna Chiozzotto,

    1. Clinica Ematologica, Centro Trapianti e Terapie Cellulari “Carlo Melzi”, DISM, Azienda Ospedaliero Universitaria S. M. Misericordia, Udine, Italy
    Search for more papers by this author
  • Simona Puglisi,

    1. Clinica Ematologica, Centro Trapianti e Terapie Cellulari “Carlo Melzi”, DISM, Azienda Ospedaliero Universitaria S. M. Misericordia, Udine, Italy
    Search for more papers by this author
  • Miriam Isola,

    1. Cattedra di Statistica, Università degli Studi, Udine, Italy
    Search for more papers by this author
  • Silvia Buttignol,

    1. Clinica Ematologica, Centro Trapianti e Terapie Cellulari “Carlo Melzi”, DISM, Azienda Ospedaliero Universitaria S. M. Misericordia, Udine, Italy
    Search for more papers by this author
  • Renato Fanin

    1. Clinica Ematologica, Centro Trapianti e Terapie Cellulari “Carlo Melzi”, DISM, Azienda Ospedaliero Universitaria S. M. Misericordia, Udine, Italy
    Search for more papers by this author


We report the long-term outcome results of 57 consecutive adult patients with immune thrombocytopenia after being treated with rituximab. According to the different period of therapy, patients received either standard dose (SD) rituximab (i.e., 375 mg/m2 weekly for 4 weeks) or low dose (LD) rituximab (i.e., 100 mg flat dose weekly for 4 weeks). Overall (OR) and complete response (CR) rates were 60 and 40%, respectively. Patients' median follow-up was 52 months, 82 months in the SD, and 44 months in the LD group; 15 out of 34 responsive patients (44%) relapsed, with median response duration of 24 months (range 3–120). The estimated 4-years event-free survival (EFS, considering events the non response status at month 2 or relapses in responders) was 30%. Patients who received SD vs. LD rituximab had better outcome with regard to short term response (OR 66 vs. 52%, CR 50 vs. 28%), relapse rate (38 vs. 54%), probability to achieve and maintain long-term response (41 vs. 24%) and estimated 4-years EFS (35 vs. 23%). Patients with a longer interval between diagnosis and rituximab therapy had worse EFS [HR = 1.005; 95%IC: (1.002–1.009), P = 0.019]. Three patients developed short-term adverse events, two-serum sickness, and one interstitial pneumonia. Four cases of malignancies and two herpes zoster reactivations were registered during long-term follow-up; one patient died for cerebral bleeding. Rituximab SD appears a safe and active agent allowing in nearly 40% of cases to achieve long-term response and splenectomy sparing effect. Am. J. Hematol. 2012. © 2012 Wiley Periodicals, Inc.


Immune thrombocytopenia (ITP) in adults is an autoimmune disorder characterized by increased platelet destruction and inadequate platelet production. Quite recently, a consensus conference redefined the terminology, definitions, and outcome criteria in ITP [ 1].

Treatment of patients with ITP is generally reserved for patients with active bleeding or platelet count <20–30 × 109/L [ 2]. Steroid is the front line treatment [ 2] but, unfortunately, only 20–30% of patients achieve durable response, while the majority require further treatment. Splenectomy remains the standard salvage therapy for patients with active symptomatic disease [ 2], with 60% cure rate. However, surgery is not always feasible owing to medical comorbidities or patient unwillingness.

Treatment of relapsed or refractory symptomatic ITP, still lacks a gold standard therapy. Rituximab is an anti-CD20 chimeric monoclonal antibody which targets B-cell and induces prolonged yet reversible B-cell depletion [ 3]. Since this effect and its good safety profile, rituximab has been also used for treating autoimmune anemia, thrombocytopenia, and other non-hematological disorders. To date, there is not yet a gold standard regimen based on evidence, despite different schedules have been proposed, i.e., the standard dose (SD, 375 mg/m2 weekly for four administrations), the low dose (LD, 100 mg flat dose weekly for four administrations) or 1,000 mg flat dose on day 1 and 15. Several reports addressed the therapeutic activity of rituximab in pediatric and adults patients with chronic refractory or steroid dependent ITP, with nearly 40–70% short term overall (OR) and 20–50% complete response (CR) according to different cohort selection [ 4–11]. Younger age and shorter period from diagnosis to rituximab have been pointed as possible good prognostic factors for response [ 7–8]. Because of the yet relatively short experience with this agent in the ITP treatment, evidence concerning its efficacy and toxicity in the long term is still scarce.

In a previous report, we already evaluated the long-term effect of rituximab in 27 patients with ITP showing that 9 out of initial 27 patients (33%) maintained such response, with an estimated 61% 5-years relapse-free survival [ 12]. Recently, Patel et al. [ 13], analyzing data of 72 adults and 66 children with ITP who had at least an initial response to rituximab, calculated an estimated 5-years response rate of 21 and 30%, respectively.

We report, in this article, the results of our updated single institution experience with rituximab salvage therapy in adults with ITP, focusing on the long-term outcome.

Patients and Methods


This survey included consecutive adult patients with a diagnosis of relapsed-refractory, primary, or secondary ITP, according to the recent standardized criteria [ 1], who were started on rituximab at our Institution from October 1999 to March 2011. All patients signed an informed consent.

Treatment with rituximab

According to the different period of therapy and without any correlation with clinical features, patients received either standard dose (SD) rituximab (i.e., 375 mg/m2 from October 1999 to August 2005 and again from November 2009 up to date) or low dose (LD) rituximab (i.e., 100 mg flat dose from February 2006 to September 2009) both on days +1+8+15+22. Patients with steroid-dependency owing to low platelet number <20 × 109/L or active bleeding were allowed to continue steroids at the minimal individual effective dosage sufficient to maintain a safe number of platelets. However, only patients who reached steroid discontinuation during or soon after rituximab were considered responders.

Response criteria

Parameters of response were retrospectively evaluated according to standardized criteria [ 1]. Overall response (OR) was considered in all those patients achieving platelet count ≥30 × 109/L and at least two-fold increase from the baseline count and absence of bleeding; complete response (CR) in case of platelet count ≥100 × 109/L and absence of bleeding; no response (NR) in case of platelet count <30 × 109/L or less than two-fold increase from baseline; time-to-response (TTR) was the time elapsed from treatment start to achievement of OR or CR; loss of response was declared when platelet count fell below 30 × 109/L or less than two-fold increase of baseline platelet count; response duration (RD) was measured from the response achievement to its loss. Patients who needed steroid administration during rituximab therapy were considered responders only if steroid independency was achieved. To better evaluate the short and long-term impact of therapy we evaluated also the event-free survival (EFS), where failure to respond at 2 months and the loss of response are the events considered.


Rituximab-related toxicity was assessed during the treatment period (short-term toxicity; from baseline to week 6) and during the follow-up (long-term toxicity). Clinical and laboratory side effects were evaluated and graded according to the NCI common toxicity criteria (CTC) [ 14].

Statistical analysis

Characteristics of the study population were described using median and range for continuous variables and percentages for categorical variables. Data were tested for normal distribution using the Shapiro–Wilk test. T-test or Mann–Whitney test, as appropriate, was used to compare continuous variables. For categorical variables, cross-tabulations were generated, and chi-square or Fisher exact test was used to compare distributions. EFS was estimated according to Kaplan–Meier method. Cox regression analysis, after the proportional hazards assumption had been verified, was used to explore the association between EFS and clinical or laboratory variables. Differences were considered significant if the two-sided P value was less than 0.05.


Fifty-seven consecutive Caucasian adult patients were included into this survey; the results observed in the first 27 patients have already been reported [ 12] but had been in this article updated with longer follow-up. Table I summarizes patients' main characteristics. Median age was 47 years (range 14–80) and median platelet count before rituximab was 23 × 109/L. All patients had active disease that had relapsed or was refractory to at least one full course of steroid therapy or was steroid-dependent. Other treatments before rituximab included splenectomy (three patients), intravenous immunoglobulin (25 patients), azathioprine or cyclosporin-A (seven patients). Forty-six patients had a primary ITP and 11 a secondary ITP; this group included eight patients with more complex autoimmune disorders (undifferentiated connectivitis, systemic lupus erythematosus, positivity for antiphospholipid antibodies), two patients with ITP associated with hepatitis C virus, and one patient with immune thrombocytopenia and neutropenia (ITN). The median time from diagnosis to rituximab was 24 months (range 2–324).

Table I. Main Clinical and Laboratory Features of Patients Before Rituximab Treatment
 All patientsStandard dose rituximabLow dose rituximab
  1. UCTD, undifferentiated connectivitis; SLE, systemicus lupus erythemathosus; APLA, anti phospholipid antibodies; ITN, immune thrombocytopenia and neutropenia; IVIg, intravenous immunoglobulin; AZA, azathioprine; CSA, cyclosporin-A.

Median age, years (range)47 (14–80)51 (16–80)43 (14–74)
Primary ITP (%)46 (80)24 (75)22 (88)
ITP + UCTD/SLE/APLA (%)8 (14)6 (19)2 (8)
ITN (%)1 (2)1 (3)0
ITP + HCV (%)2 (4)1 (3)1 (4)
Median interval from diagnosis and rituximab, months (range)26 (1–324)31 (3–264)24 (2–324)
Previous number of therapy
1 (%)57 (100)32 (100)25 (100)
2 (%)26 (46)14 (44)12 (48)
3 (%)4 (7)1 (3)3 (12)
Previous treatments
Steroids (%)57 (100)32 (100)25 (100)
IVIg (%)23 (40)10 (31)13 (52)
AZA/CSA (%)7 (12)4 (12.5)3 (12)
Splenectomy (%)3 (5)2 (6)1 (4)
Median number of platelets (×109/L)231832
Period of treatmentOctober 1999 to March 2011October 1999 to March 2011February 2006 to September 2009


Rituximab was administered at SD in 32 patients and at LD in 25. Fifty-five patients (93%) completed the therapeutic program receiving the four scheduled doses of rituximab. Two patients interrupted treatment with SD rituximab after the second administration because of the development of serum sickness.

Short-term response

OR and CR were achieved in 34 (60%) and 23 (40%) patients, respectively. There was a trend toward higher OR and CR for patients treated with SD (OR 66 vs. 52%, P = 0.298; CR 50 vs. 28%, P = 0.093; Table II). The median TTR after SD and LD rituximab was 10.5 days (range 7–120 days) and 30 days (range 7–100 days), respectively (P = 0.141). The patient with ITN showed complete normalization of platelet and neutrophils count upon rituximab treatment.

Table 2. Response Rate and Outcome after Rituximab Therapy
 All patientsStandard dose rituximabLow dose rituximab
  1. OR, overall response; CR, complete response; RD, response duration. OR includes CR + PR.

OR (%)34 (60)21 (66)13 (52)
CR (%)23 (40)16 (50)7 (28)
NR (%)23 (40)11 (34)12 (48)
Median time to response, days (range)14 (7–120)10.5 (7–120)30 (7–100)
Median time to CR, days (range)45.5 (7–150)40 (7–150)30 (7–150)
Relapse (%)15 (44)8 (38)7 (54)
Relapse after CR (%)7 (30)6 (37.5)1 (14)
Median RD, months (range)24 (3–120)26.5 (3–120)22 (3–52)
Median follow-up, months (range)52 (3–144)82 (3–144)44 (13–63)
Long-term response (%)19 (33)13 (41)6 (24)

Short-term toxicity

Three patients (5%) experienced relevant short-term side effects. Two young female patients (19 and 28 years old, respectively) developed after the second administration of SD rituximab, a serum sickness with rapid improvement following treatment discontinuation and a brief course of steroids. A 74-year-old patient developed nearly 1 month after the fourth administration of LD rituximab an interstitial pneumonia with quick recovery after empiric antibiotic and steroid therapy.

Long-term response

The median time of observation in patients who received SD or LD was 82 months (range 3–144 months) and 44 months (range 13–63 months), respectively. Among responders, the rate of relapse was higher, even if statistically not significant, in patients treated with LD rituximab (54 vs. 38%, P = 0.369) and in those who achieved only partial response; in particular, all six patients in partial response after LD rituximab eventually relapsed (Table II). The median RD was 21 months (range 3–120) in the SD group and 22 months (range 3–52) in the LD group (P = 0.148). The probability to achieve and maintain long-term response was 41% in the SD and 24% in the LD group (Table II), with an estimated 4-year projected EFS of 35 and 23%, respectively (P = 0.1228; Fig. 1). Cox analysis showed time from diagnosis to rituximab therapy was associated with EFS HR = 1.005 [95%IC: (1.002–1.009), P = 0.019]. Therefore, loss of response was more probable when the interval from diagnosis and rituximab therapy was longer.

Figure 1.

Estimated event-free survival curves (which included, as negative events, the absence of response within 2 months from rituximab start and the lost of response in patients who responded) in patients who received standard dose rituximab and low dose rituximab.

Patients unresponsive or relapsed after rituximab received different salvage therapy according to clinical status and previous treatments. Fourteen patients underwent splenectomy, nine of which achieved a durable response; 20 patients were treated with dapsone, obtaining OR in 11. Two patients, one primary ITP and the patient with ITN who had a medium to long-term response to rituximab (14 and 45 months, respectively) were retreated with rituximab. Both regained a CR; the ITP patient further relapsed in 8 months time, while the ITN subject is still in remission (both for platelets and neutrophils count) after 36 months.

Long-term toxicity

During the follow-up period no cases of opportunistic infections, progressive multifocal leukoencephalopathy, or other severe infectious complications were observed. Two patients developed abdominal Herpes Zoster 8 and 12 months after rituximab therapy, promptly cured by anti-viral therapy. Four cases of malignancies were registered; these included one atypical ductal hyperplasia and one tubular adenoma in transformation, developed 8 and 6 years after SD rituximab, respectively and single cases of lung cancer and carcinoid of appendix, diagnosed 1 year after LD rituximab. No other late hematological or extra-hematological toxicities were registered.


The ideal agent for the treatment of a benign yet often insidious, chronic, and potentially life-threatening disease such as ITP, had to combine the achievement of long-lasting response with a good-safety profile. Currently, splenectomy remains the best therapeutic option for patients with symptomatic chronic ITP with approximately 60–70% long-term effect [ 2]. However, it is not uncommon to find patients not keen on surgery and physicians who either delay or omit to propose splenectomy because they are worried about treatment unresponsiveness, possible surgical risks, or delayed life-threatening infections; furthermore some patients may bear surgical contraindications. Eltrombopag and romiplostim are two thrombopoietin receptor agonists which share high-therapeutic activity and good mid-term safety profile in ITP [ 15–18]. However these agents are not given with curative intent and there is still concern as to the possible effect of chronic bone marrow stimulation, particularly in younger patients.

The results of our analysis confirm the short to mid-term effect of rituximab salvage therapy in adult patients with ITP reported earlier by our group and other authors [ 4–12]. Our data might suggest SD is the preferable therapeutic regimen due to the shorter timing of response, the higher response rate, the lower incidence of relapse, and the similar safety profile. Unfortunately, such a conclusion is hampered by the lack of statistical significance, probably secondary to the small patients' sample, and by the uncontrolled observational study design. Nevertheless, according to our experience, SD rituximab is capable to induce a long-term response and a splenectomy-sparing effect in nearly 40% of ITP patients who previously received and failed at least a full course of steroid therapy.

The effect of rituximab to safely avoid splenectomy has been also prospectively investigated by Godeau et al. [ 19] in 60 adults with chronic ITP. The authors indicated that, at 2 years, 24 (40%) patients responded to rituximab (i.e., platelet counts 30 × 109/L or more off treatment) avoiding splenectomy. Among the 36 non-responders, 25 underwent splenectomy with response in 15 (60%) after a median follow-up of 18 months.

Recently, we have updated the long-term results of the prospective Italian study, which compared dexamethasone vs. dexamethasone plus rituximab as front line therapy for adults with ITP [ 20]. Fifty-four patients who achieved sustained response (i.e., platelet counts 50 × 109/L or more at month six from the beginning of treatment) including 13 out of initial 52 of the dexamethasone monotherapy arm, 27 out of initial 49 of the dexamethasone plus rituximab arm, and 14 out of 27 of the dexamethasone plus rituximab salvage therapy arm were subsequently followed for a median observation period of 34 months (range 4–54 months). The relapse rate in the three groups was 23, 26, and 14%, with a calculated long-term effect (i.e., probability to achieve and maintain response) of 19, 41, and 44%, respectively [ 21].

Taken together, all these data suggest that rituximab SD may be associated with long-term response in nearly 40% of patients and could be considered as a pre-splenectomy therapeutic option particularly for those patients at risk of surgical complication or those refusing intervention. Nevertheless, there is still insufficient evidence to support the replacement of splenectomy with rituximab as a second-line treatment of chronic ITP.

Less impressive long-term outcome has been on the contrary reported by Patel et al. [ 13] who analyzed data from 17 published studies which included 376 adults showing 5 years response rate only in 21%. This discrepancy could be at least partially explained by ours and Godeau's selection of patients with a lower degree of treatments received before rituximab [ 19].

Unfortunately, to date there are not clinical, laboratory, or biological parameters, which could be utilized to predict the response to rituximab. Several studies did not found any correlation between response to rituximab and splenectomy. On the contrary, younger age and a shorter interval between diagnosis and rituximab therapy appeared to be positive prognostic indicators of sustained response [ 7, 8]; this analysis has also confirmed the negative effect delayed rituximab administration induces on outcome. This observation could be important to understand the best timing of rituximab therapy in the management of patients with ITP. On these grounds, we can hypothesize a role for rituximab in an earlier phase of the disease, when its administration might guarantee better short and long-term results and possibly lower risk of infectious complications.

Regarding the safety profile, no late serious infections including progressive multifocal leukoencephalopathy were observed. Four patients developed malignancies, but the real cause–effect role of rituximab in these complications remains uncertain because of the heterogeneity in type of neoplasia and the too short interval elapsed from rituximab treatment at least in two of these cases (i.e., only 1 year). Overall, responder patients experienced a good quality of life, in terms of long-time off therapy and of reduced medical care needs.

In conclusion, our results indicate that rituximab therapy may induce long-term response, avoidance of splenectomy, and of further medical therapies in approximately 40% ITP adult patients, together with a good-safety profile. This favorable effect seems more marked with rituximab SD and with early treatment.