REVIEW
`
`ASBM1
`
`American Society for Blood
`and Marrow Transplantation
`
`Efficacy of Rituximab in the Setting
`of Steroid-Refractory Chronic Graft-versus-Host
`Disease: A Systematic Review and Meta-Analysis
`Mohamed A. Kharfan-Dabaja, 1
`2 Asmita R. Mhaskar, 3 Benjamin Djulbegovic,2
`•
`Corey Cutler,5 Mohamad Mohty, 6 Ambuj Kumar3
`4
`•
`
`4
`•
`
`3
`•
`
`Increased insight into the role of B lymphocytes in the pathophysiology of graft-versus-host disease has led to
`a number of studies assessing the efficacy of the anti-CD20 monoclonal antibody (mAb) rituximab in treating
`steroid-refractory chronic graft-versus-host disease (cGVHD). Findings vary greatly among these studies,
`however. We conducted a systematic review to summarize the totality of evidence on the efficacy of ritux(cid:173)
`imab in steroid-refractory cGVHD. We performed a PubMed search and contacted experts in the field to
`identify relevant studies. Endpoints included overall response rate (including organ-specific) and ability of rit(cid:173)
`uximab to allow dosage reduction of immunosuppressive therapies. Data were pooled under a random-ef(cid:173)
`fects model. Seven studies (3 prospective and 4 retrospective, with a total of I I I patients) met the inclusion
`criteria. The pooled proportion of overall response was 0.66 (95% confidence interval = 0.57 to 0.74). There
`was no heterogeneity among the pooled studies. Response rates were 13% to I 00% for cGVHD of the skin,
`0 to 83% for cGVHD of the oral mucosa, 0 to 66% for cGVHD of the liver, and Oto 38% for cGVHD of the
`lung. Common adverse events were related to infusion reactions or infectious complications. The relatively
`small number of patients and the varying criteria for reporting organ response and dosage reduction of ste(cid:173)
`roids, among other limitations, hinders our ability to reach definitive conclusions on the overall efficacy of
`rituximab for cGVHD involving other organs.
`Biol Blood Marrow Transplant 15: 1005-1013 (2009) © 2009 Amei-ican Society for Blood and Mai7·ow Transplantation
`
`KEY WORDS: Rituximab, Chronic graft-versus-host disease, Systematic review, Meta-analysis
`
`INTRODUCTION
`
`Chronic graft-versus-host disease (cGVHD) rep(cid:173)
`resents one of the most challenging sequelae of alloge(cid:173)
`neic hematopoietic cell transplantation (allo-HCT),
`resulting in significant long-term morbidity and
`mortality. The continuous growth in the number of
`
`From the 1 Department of Blood and Marrow Transplantation;
`2Hematology; Health Outcomes and Behavior, Moffitt Can(cid:173)
`cer Center, Tampa, Florida; 4Center for Evidence-Based
`Medicine, University of South Florida, College of Medicine,
`Tampa, Florida; 5Department of Medical Oncology, Dana(cid:173)
`Farber Cancer Institute, Boston, Massachusetts; and 6Hema(cid:173)
`tology Clinic, Hospital Center of the University of Nantes,
`Nantes, France.
`Financial disclosure: See Acknowledgments on page 1012.
`Correspondence and reprint requests: Ambuj Kumar, MD, MPH,
`Moffitt Cancer Center & Research Institute, College of Medi(cid:173)
`cine, University of South Florida, 12902 Magnolia Drive,
`MRC 232A, Tampa, FL 33612 (email: Ambuj.kumar@moffitt.
`org).
`Received February 25, 2009; accepted April 6, 2009
`© 2009 American Society for Blood and Marrow Transplantation
`1083-8791/09/159-0001$36.00/0
`doi: 10.1016/j.bbmt.2009.04.003
`
`allo-HCTs using alternate donors, particularly
`HLA-mismatched donors, is further increasing the in(cid:173)
`cidence of cGVHD [1,2]. The increasing use of mobi(cid:173)
`lized peripheral blood stem cells (PBSC) as the graft
`source also is contributing to the increasing prevalence
`of cGVHD [1-3].
`A recently released National Institutes of Health
`(NIH) working group report on criteria for clinical tri(cid:173)
`als in cGVHD provides standardized criteria for diag(cid:173)
`nosis of cGVHD and an improved scoring system that
`better describes the extent and severity of cGVHD for
`each organ, taking into account the importance of pre(cid:173)
`serving function [4]. Similarly, expert consensus opin(cid:173)
`ion has resulted in the establishment of more practical
`criteria aimed at assessing the therapeutic response in
`patients with cGVHD more objectively [5]. Unfortu(cid:173)
`nately, however, treatment outcomes for cGVHD
`remain disappointing. Systemic corticosteroid therapy
`is the most commonly used first-line treatment for
`patients with established cGVHD, but long-term cor(cid:173)
`ticosteroid use is limited by the increased risk of infec(cid:173)
`tion, which remains the leading cause of death in
`cGVHD [ 6]. There is no consensus regarding the
`best treatment option for patients with cGVHD who
`
`Pharmacyclics Exhibit 2052
`Sandoz v. Pharmacyclics ·
`IPR2019-00865
`
`1 005
`
`

`

`1006
`
`M. A. Kharfan-Dabaja et al.
`
`Biol Blood Marrow Transplant 15:/005-1013, 2009
`
`do not respond to or progress after corticosteroid ther(cid:173)
`apy. Encouraging responses have been reported using
`extracorporeal photopheresis [7,8], mycophenolate
`mofetil (MMF) [9, 10], and low-dose methotrexate
`(MTX), among other modalities [11].
`Advances in the understanding of cGVHD have
`implicated B lymphocytes in the pathophysiology of
`cGVHD. Miklos et al. [12] demonstrated a correlation
`between cGHVD and development of antibody re(cid:173)
`sponses to H-Y minor histocompatibility antigens in
`cases of sex-mismatched (male recipients with female
`donors) allo-HCT. These findings provided the scien(cid:173)
`tific rationale for a number of studies exploring rituxi(cid:173)
`mab to treat patients with steroid-refractory cGVHD.
`These studies yielded varying findings regarding or(cid:173)
`gan-specific responses, however. Consequently, we
`performed a systematic review to evaluate the totality
`of evidence regarding the efficacy of rituximab in treat(cid:173)
`ing steroid-refractory cGVHD.
`
`METHODS
`
`Literature Search
`We searched the Medline (Pubmed) database us(cid:173)
`ing a broad search strategy to identify prospective or
`retrospective studies evaluating the efficacy of ritux(cid:173)
`imab in patients with steroid-refractory cGVHD.
`The search was conducted using following terms:
`"Rituximab"[Substance Name] AND "Graft vs. Host
`Disease"[MeSH]. Relevant references in each ob(cid:173)
`tained article were scanned to identify other relevant
`studies. In addition, experts in the field were ap(cid:173)
`proached for unpublished data or to identify addi(cid:173)
`tional studies in the subject area. No search limits
`were applied.
`
`Inclusion Criteria
`All prospective studies evaluating the efficacy of
`rituximab in cGVHD were included regardless of the
`number of patients enrolled. Retrospective studies
`were included if tl1ey evaluated tl1e efficacy of rituxi(cid:173)
`mab in cGVHD in a minimum of 5 patients. Single
`case reports were excluded.
`
`Study Selection, Quality Assessment, and Data
`Extraction
`Two reviewers (A.M. and A.K.) appraised tl1e list
`of references and selected the studies in consultation
`witl1 other reviewers (M.K.D. and C.C.). Disagree(cid:173)
`ments were resolved by consensus. Two reviewers
`(A.M. and A.K.) independently extracted the data
`from selected articles. Data were extracted on specific
`clinical outcomes (benefits and harms), as well as on
`the methodological quality of the stu dies.
`
`Data Analysis and Statistical Methods
`For tl1e purpose of meta-analysis, the proportions
`were first transformed into a quantity according to
`the Freeman-Tukey variant of tl1e arcsine square
`root transformed proportion [13] . T he pooled propor(cid:173)
`ti on was calculated as the back-transform of the
`weighted mean of tl1e transformed proportions, using
`a random-effects model [14] .
`A forma l statistical test for heterogeneity using an
`I 2 test was performed [ 15]. The heterogeneity and ro(cid:173)
`bustness of tl1e findings also were evaluated tlrrough
`additiona l sensitivity analyses. T he possibility of pub(cid:173)
`li cation bias was assessed using the Begg and Egger
`fum1el plot metl1od [16] ; although tl1is method has
`some limitations, it is widely used to assess publication
`bias [17]. The meta -analysis was performed using
`StatsDirect software (StatsDirect Ltd , Altrincham,
`Cheshire, UK). The work was performed ion accor(cid:173)
`dance with the guidelines promulgated at tl1e Quality
`of Reporting of Meta-Analyses conference [1 5).
`
`RESULTS
`
`Identification of Studies
`Figure 1 summarizes tl1e process used to identify
`and select the studies for the systematic review. The
`initial search yielded 3 7 articles, of which 31 were
`excluded for tl1e reasons shown in Figure 1. Of the 6
`studies tl1at met the inclusion criteria, 3 were catego(cid:173)
`rized as prospective studies and 3 were retrospective
`analyses. One retrospective case series was identified
`tlrrough expert contact [1 8). We found no randomized
`controlled trials evaluating the efficacy of rituximab
`versus other therapeutic alternatives for treating ste(cid:173)
`roid-refractory cGVHD.
`
`Methodological Quality of Studies
`We conducted a critical appraisal of the methodo(cid:173)
`logical quality of all studies.
`
`Prospective Studies
`Unclear reporting of sampling procedures makes it
`difficult to determine whether the study sample con(cid:173)
`sisted of consecutive series of patients or a convenient
`sampling method was used , possibly introducing a se(cid:173)
`lection or an ascertaimnent bias that could potentially
`plague observational studies.
`
`Retrospective Studies
`Whether an analysis addressed a priori hypotl1esis
`or was a result of some post hoc observation was un(cid:173)
`clear. In addition, the relatively small sample sizes
`(range, 8 to 38 patients) limited our ability to draw de(cid:173)
`finitive conclusions from these studies.
`
`

`

`Biol Blood Marrow Transplant 15:/005-/0/3, 2009
`
`Rituximab in Steroid-Refractory Chronic GVHD
`
`1007
`
`Number of studies
`retrieved from
`PubMed and
`manually
`N=37
`
`1'
`
`Studies
`included in the
`fina l analysis
`N=7
`(one added as
`per suggestion
`from an expert)
`
`Excluded N=31
`
`Case reports <5 patients
`(N=6)
`Editorial (N=1)
`Not GVHD (N=13)
`Not refractory GVHD (N=3)
`Not rituxan therapy (N=3)
`Review article (N=5)
`
`l
`
`Prospective
`studies (N=3)
`
`l
`
`Retrospective
`studies with>5
`patients (N=4)
`
`Figure I. Identification and selection of studies.
`
`Publication Bias
`A Begg and Egger funnel plot showed a symmetric
`distribution, indicating the absence of a publication
`bias for all of the outcomes assessed here (results not
`shown).
`
`Outcomes
`
`Mortality
`Mortality data were extractable from all 7 studies
`[1 8-24]. The pooled proportion of mortality from 7
`studies involving 111 patients was 0.158 (95 % confi(cid:173)
`dence interval [CI]= 0.083 to 0.253) (Fig 2A). There
`was no statistically significant heterogeneity among
`the studies (J2 = 32 .7%; P = .1 78). The pooled pro(cid:173)
`portion of mortality was 0.122 (95% CI= 0.034 to
`0.253) in 3 prospective studies involving 37 patients
`[19,20,22] and 0.158 (95% CI= 0.08 to 0.252) in 4 ret(cid:173)
`rospective studies evaluating 74 patients [1 8,21,23 ,24].
`
`Overall Response Rate
`Overall response rate (ORR) data were extract(cid:173)
`able from 6 studies involving a total of 108 patients
`
`(Fig 2B) [18-21,23,24]. The pooled proportion of
`ORR was 0.66
`(95 % CI= 0.57
`to 0.74), and
`there was no statistically significant heterogeneity
`among the studies (J2 = 0%; P = .90). The pooled
`proportion of ORR was 0.70 (95% CI= 0.54 to
`0.83) in 2 prospective studies evaluating 34 pa(cid:173)
`tients [1 9,20] and 0.64 (95 % CI= 0.53
`to 0. 74)
`in 4 retrospective studies
`involving 74 patients
`[1 8,21,23,24].
`
`Organ-Specific Response
`
`Skin cGVHD
`Data on cutaneous cGVHD were extractable from
`6 studies involving a total of 67 patients [18, 19,21 -24].
`The pooled proportion ORR was 0.60 (95% CI= 0.41
`to 0.78) (Fi_g 2C). There was a statistic~lly s\gnificant
`heterogeneity among the pooled studies (J· = 60%;
`P = .03). The pooled proportion ORR was 0.85
`(95 % CI= 0.59 to 0.98) in 2 prospective studies en(cid:173)
`rolling 9 patients [1 9,22] and 0.51 (95% CI= 0.308
`to O. 717) in 4 retrospective studies involving 5 8 pa(cid:173)
`tients [18,21,23,24] .
`
`

`

`1008
`
`M. A. Khorfon-Dobojo et al.
`
`Bio l Blood Morrow Transplant 15:/005- 10/ 3, 2009
`
`A
`
`Proportion meta-analysis plot for the outcome of survival
`
`Proportion (95% confidence interval)
`
`0.333 (0.118, 0.616)
`Mohty et. al. , 2008
`-+a--
`0.211 (0.096, 0.373)
`Zaja et. al., 2007
`Okamoto et. al., 2006 , _ _ _ _ _ _ _ _ 0.333 (0.008, 0.906)
`Cutler et. al., 2006 -a+--
`Canninga-van Dijk et al; 2004
`Ratanatharathom et. al; 2003 a-~ -
`van Bonin et. al; 2008 1-a-..;....-
`combined ,_+~·~~-~~~~0.158 (0.083, 0.253)
`1.0
`00 02 04 06 08
`
`0.107 (0.023, 0.282)
`0.000 (0.000, 0.459)
`0.000 (0.000, 0.369)
`0.077 (0.002, 0.360)
`
`C
`
`Proportion meta-analysis plot for the outcome of skin response
`Proportion (95% confidence interval)
`
`B
`Proportion meta-analysis plot for the outcome of overall response
`Proportion (95% confidence interval)
`
`Mohty et. al., 2008
`
`Zaja et. al. , 2007
`
`Cutler et. al., 2006
`
`Canninga-van Dijk et al; 2004
`
`Ratanatharathom et. al; 2003
`
`van Bonin et. al; 2008
`
`combined
`
`--(cid:173)•
`
`- ----+- -----
`
`0.67 (0.38, 0.88)
`
`0. 66 (0.49, 0. 80)
`
`0. 68 (0.48, 0. 84)
`
`0.83 (0.36, 1.00)
`
`0.50 (0.16, 0.84)
`
`0.69 (0.39, 0.91)
`
`0.66 (0.57, 0.74)
`
`1-.-.-~~~-~-~-~
`4
`
`0.0
`
`0.2
`
`0.4
`
`0.6
`
`0. 8
`
`1.0
`
`D
`Proportion meta-analysis plot for the outcome of Liver response
`Proportion (95% confidence interval)
`
`• ■
`
`Mohty et. al., 2008
`
`Zaja et. al., 2007
`
`Okamoto et. al., 2006
`
`Canninga-van Dijk et al; 2004
`
`Ratanatharathom et. al; 2003
`
`van Bonin et. al; 2008
`
`combined
`
`0.0
`
`0.2
`
`0.4
`
`0.692 (0.386, 0.909)
`
`0.607 (0.406, 0. 785)
`
`1.000 (0.292, 1.000)
`
`Mohty et. al., 2008
`
`Zaja et. al., 2007
`
`Okamoto et. al., 2006
`
`0.833 (0.359, 0.996)
`
`Canninga-van Dijk et al; 2004
`
`■
`
`0.67 (0.30, 0.93)
`
`0.21 (0.05, 0.51)
`
`0.00 (0.00, 0.84)
`
`0.40 (0.05, 0.85)
`
`Ratanatharathom et. al; 2003
`
`- - - - - - - - - - 0.00 (0.00, 0.98)
`
`van Bonin et. al; 2008
`
`combined
`
`6
`i--.-~-~-~-~~
`
`0.00 (0.00, 0.71)
`
`0.29 (0.12, 0.51)
`
`0.0
`
`0.2
`
`0.4
`
`0.6
`
`0.8
`
`1.0
`
`.!
`~
`0.6
`
`0.125 (0.003, 0.527)
`
`0.556 (0.212, 0.863)
`
`0.606 (0.412. 0.784)
`
`0.8
`
`1.0
`
`Figure 2. Forest plot for the outcomes of survival, overall response and organ specific response (skin and liver). The summary effect estimate (pro(cid:173)
`portion) for individual studies are in dicated by black rectangles (the size of t he rectangle is proportional co the study weight). with the li nes representing
`95% confidence intervals (Cls). The overall summary effect estimate (proportion) and 95% Cl are indicated by the diamond.
`
`Mucosa (Oral) cGVHD
`Data on cGVHD of the oral mucosa were extract(cid:173)
`able from 5 studies involving a total of 46 patients
`[18-20,22,24] . The pooled proportion of ORR was
`0.36 (95 % CI = 0.12 to 0.65). There was a statistically
`significant heterogeneity among the included studies
`for this outcome (J2 = 73 % ; P = .0046). The pooled
`proportion of oral cG'i/HD response was 0.26 (95%
`CI= 0.007 to 0.84) in 3 prospective studies involving
`15 patients [19,20,22] and 0.45 (95 % CI= 0.29 to
`0.62) in 2 retrospective studies involving 31 patients
`[18,24].
`
`Liver cGVHD
`ORR data for the outcome of liver cGVHD were
`extractable from 6 studies involving a total of 34 pa(cid:173)
`tients [18,19,21 -24]. The pooled proportion of ORR
`was 0.29 (95 % CI = 0.12 to 0.51) (Fig 2D). There
`was no statistically significant heterogeneity among
`the pooled studies for this outcome (I2 = 41.8% ;
`P = .126). The pooled proportion ORR was 0.28
`(95 % CI= 0.03 to 0.64) in 2 prospective studies enroll(cid:173)
`ing 7 patients [l 9,22] and 0.29 (95 % CI= 0.06 to 0.59)
`from 4 retrospective studies involving 27 patients
`[18,21,23,24] .
`
`Gastrointestinal cGVHD
`Data on response rate for gastrointestinal (GI)
`cGVHD were reported in 4 studies (I prospective
`
`and 3 retrospective) involving a total of 12 patients
`[18,21,22,24]. The pooled proportion ORR was 0.31
`(95 % CI= 0.07 to 0.62). There was no statistically sig(cid:173)
`nificant heterogeneity among the pooled studies
`(I2 = 3 5. 7%; P = . I 9). One prospective study showed
`no response to rituximab treatment in 1 patient with
`steroid-refractory gastrointestinal cGVHD [22]. The
`pooled proportion ORR in 3 retrospective studies in(cid:173)
`volving 11 patients was 0.346 (95% CI= 0.05 to
`0.72) [18,21,24].
`
`Lung cGVHD
`Data on response rates in cases of steroid-refrac(cid:173)
`tory cGVHD involving the lung were extractable
`from 4 studies involving a total of 15 patients [18,22 -
`24]. In 1 prospective study, rituximab produced no re(cid:173)
`sponse [22]. The pooled proportion of lung cGVHD
`response in 3 retrospective studies involving 14 pa(cid:173)
`tients was 0.30 (95% CI= 0.11 to 0.53) [18,23,24].
`There was no significant heterogeneity among the
`pooled studies (I2 = 0%; P = .58).
`
`Other Organs with cGVHD
`Reponses to rituximab also were reported in pa(cid:173)
`tients with steroid-refractory ocular cGVHD, with
`rates ranging from 13 % (1/8) to 38% (6/16) [23,24],
`and in patients with steroid-refractory cGVHD of
`the musculoskeletal system, with response rates of
`100% (1/1) and 75 % (3/4) [18,23].
`
`

`

`Biol Blood Marrow Transplant 15: /005-10/ 3, 2009
`
`Rituximab in Steroid-Refractory Chronic GVHD
`
`1009
`
`Does Administration of Rituximab Allow
`Reduction (or Discontinuation)
`of lmmunosuppressive Therapies, Including
`Corticosteroids?
`Administration of rituximab facilita tes dosage
`reduction of previous immunosuppressive therapies
`in patients with refractory cGVHD. Zaja et al. [24)
`reported a median dosage reduction of immunosup(cid:173)
`pressive therapy (including corticosteroids) of 82 %
`(range, 0 to 100%), mostly in cases of steroid-refrac(cid:173)
`tory cGVHD involving the skin and oral mucosa.
`Two studies specifically addressed tl1e glucocorti(cid:173)
`coid-sparing effect of rituximab in patients with ste(cid:173)
`roid-refractory cGVHD. Mohty et al. [21) reported
`a median glucocorticoid dosage reduction of 86%
`(range, 0 to 100%) in 11 of 15 patients (73 % ) treated
`with rituximab; tl1is steroid sparing-effect also was
`more pronounced in skin and oral mucosa] cGVHD,
`consistent with a previous report [24) . Similarly, Cut(cid:173)
`ler et al. [20) reported a 75 % median dosage reduction
`of prednisone (from 40 mg to 10 mg) in more tl1an
`two-thirds of their patients. These and otl1er studies
`are summarized in Table 1.
`
`Treatment-Related Morbidity and Mortality
`(TRM)
`Rituximab appears to be relatively well tolerated,
`with side effects related mainly to infusion reactions
`(5 % to 11 % ) and infectious complications, including
`sepsis (3% to 33 %), pneumonia (8 % to 33 %), viral
`conjunctivitis (5 %), diarrhea (14%), and herpes zoster
`reactivation (33 %; 1/3), among otl1ers (18-24]. Long(cid:173)
`term toxicities related to treatment were not reported
`(18-24).
`None of the studies, prospective or retrospective,
`reported mortality attributable to rituximab treatment.
`
`Sensitivity Analyses
`Because of the limited number of prospective stud(cid:173)
`ies available, as well as the relatively small number of
`
`patients for each cGVHD manifestation, we could
`not perform a sensitivity analysis to explore tl1e reasons
`behind the heterogeneity in tl1e outcomes of organ(cid:173)
`specific responses related to cGVHD of the skin and
`mucosa. This heterogeneity can be attributed to sev(cid:173)
`eral clinical factors, however. The patients enrolled
`in these studies had a wide range of diseases and previ(cid:173)
`ous interventions (eg, differing conditioning regimens,
`number of treatments for cGVHD before rituximab
`tl1erapy, concomitant treatment with corticosteroids
`or other immunosuppressive treatments), as well as
`differing criteria for assessing response rates. All of
`tl1ese factors may possibly contribute to the heteroge(cid:173)
`neity for some of tl1e outcomes.
`
`DISCUSSION
`
`The totality of the evidence on the efficacy of ritux(cid:173)
`imab for treating steroid-refractory cGVHD demon(cid:173)
`strates that the skin is the most responsive organ
`(Table 2) [19-22 ,24) . Responses were impressive in
`cases of sclerodermatous or lichenoid cutaneous
`cGVHD (20,22] . In the prospective study of Okamoto
`et al. (22), 3 patients (100%) with sclerodermatous
`cGVHD had responses occurring between 60 and 90
`days from initiation of tl1erapy. Similarly, Cutler
`et al. (20) reported a decrease in median body surface
`area (BSA) involved with sclerodermatous cGVHD
`from 35% to 25 % after 2 cycles of therapy, followed
`by a further decrease to 20% at 1 year after the initia(cid:173)
`tion of rituximab. Cases of lichenoid cutaneous
`cGVHD also responded to rituximab tl1erapy, show(cid:173)
`ing a decrease in median BSA involvement from 20%
`to 5 % after 2 cycles and a furtl1er decrease to 3 % after
`1 year (20). It is important to keep in mind tl1at clinical
`responses may continue to improve several montl1s
`after the start of rituximab. In summary, these findings
`suggest that rituximab is effective in treating cutaneous
`cGVHD .
`
`Table I. Dose Reduction of lmmunosuppressive or Corticosteroid Therapy after Initiation of Rituximab
`
`Author. Year
`
`✓flkamoto et al., 2006 (22]
`
`Median Dose
`Reduction(Range)
`
`NR
`86% (3 3%- 100%)
`82% (0-1 00%)t
`NE:j:
`75% (N E)
`NE
`NR (6875%-87. 5%)§
`
`/2:
`on Bonin et al.. 2008 [ 18]
`J /',ohty et al., 2008 (21]
`J _taja et al.. 2007 (24]
`f' Cutler et al., 2006 (20]
`/
`/Canninga-Van Dij ketal., 2004 (1 9]
`J Ratanatharathorn et al., 2003 (23]
`NR indicates not reported; NE, not extractable; NA. not available.
`*Updated data were provided by the authors (Cutler et al., CS).
`t Zaja et al. reported organ-specific dose reduction; the numbers given here are for median % dose reduction of CS for IO patients evaluable for skin
`response.
`:j:Okamoto et al. reported no change in the dose of immunosuppressive drugs during rituximab therapy.
`§Ratanatharathorn et al. reported data for dose reduction of CS extractable for 2 of the 4 patients who showed a response to rituximab; the val ues for%
`CS dose red uction are 68.75% for one patient an d 87.5% for the other patient.
`
`Proportion of Patients Discontinuing
`lmmunosuppressive Therapy(n/N)
`
`23% (3/ 13)*
`NR
`NE
`NA
`11 % (3 /28)*
`67% (4/6)
`NR
`
`

`

`c:,
`c:,
`~
`
`:, ...
`-er a
`:, ..
`Q
`:;t
`'l:
`a
`Q -,
`~
`Q.
`0
`
`~
`[
`!XI
`
`Q i
`1:5
`?>-
`~
`
`,....
`Q
`...
`ID
`Q
`.S.
`0"
`Q
`C
`I
`:,
`
`0
`0
`
`c:, ...
`.!N ..,
`"' !.
`
`c:,
`
`c:,
`
`organ provided
`responses for each
`aggregate
`
`weeks) (I)
`intervals for 4
`
`Differed by organ;
`
`375 mg/m2 (weekly
`
`11% (3/2811)
`
`Infusion reaction, 5%
`Nephrolithiasis, 5%
`GI hemorrhage, S%
`Septic arthritis, 5%
`reactivation, 5%
`Hepatitis B
`5%
`Viral conjunctivitis,
`Infant diarrhea, 14%
`
`NR
`
`patient-based
`
`weeks) (I)
`intervals for 4
`
`Herpes zoster, 33%
`Sepsis, 33%
`
`Subjective; individual
`
`375 mg/m2 (weekly
`
`33% (1/3)
`
`60-90 days Pneumonia, 33%
`
`weeks) (I)
`intervals for 4
`
`Differed by organ
`
`375 mgim'{weekly
`
`21% (8/38)
`
`Sepsis, 3%
`system, 3%
`Central nervous
`Renal failure, 3%
`Pneumonia, 8%
`Infusion reaction, I I%
`
`78 days
`46 days
`60 days
`NR
`49 days
`138 days
`57 days
`
`responses
`specific objective
`CR, PR. and organ-
`
`weeks) (I)
`intervals for 4
`
`375 mg/m2 (weekly
`
`33% (5/15)
`
`Negligible
`
`NR
`
`criteriat
`Health consensus
`
`National Institutes of
`
`weeks) (3)*
`intervals for 3
`50 mg/m2 (weekly
`
`8% (1/13)
`
`Infectious
`
`NR
`
`complications, 2
`
`(0/7)
`Oral mucosa, 0%
`Lungs, NA
`Gut, NA
`Liver, NR (NR/1)
`Eyes, 0% (0/8)
`Skin, (§/17)
`ORR, 68%
`(0/2)
`Oral mucosa, 0%
`Lungs, 0% (0/ I)
`Gut, NR
`Liver, 0% (0/2)
`Eyes, 0% (0/3)
`Skin, 100%
`ORR, NE
`Musculoskeletal, 80%
`Oral mucosa, 30%
`Lungs, 38%
`Gut, 75%
`Liver, 25%
`Eyes, 43%
`Skin, 63%
`ORR, 65%
`Oral mucosa, NA
`Lungs, 0% (0/2)
`Gut, 20%
`Liver, 66%
`Eyes, NR
`Skin, 69%
`ORR, 66%
`Muscles, 75% (3/4)
`(4/8)
`Oral mucosa, 50%
`Lungs, 0% (0/2)
`Gut, 0% (0/2)
`Liver, 0% (0/3)
`Eyes, 0% (0/4)
`Skin, 56% (5/9)
`ORR, 69%
`
`Assessment
`
`(Cycles)
`
`Mortality, % (n/N)
`
`Morbidity, %
`
`Median
`
`(Responders/Total)
`
`42 (21-62)
`
`28t
`
`( noncontrolled)
`Prospective
`
`2006 [20]
`Cutler et al.,
`
`35 (33-42)
`
`3
`
`( noncontrolled)
`Prospective
`
`2006 [22]
`
`Okamoto et al.,
`
`48 (22-61)
`
`38
`
`Retrospective
`
`50 (20-67)
`
`15
`
`Retrospective
`
`2007 [24]
`
`Zaja et al.,
`
`2008 [21]
`Mohty et al.,
`
`60 (40-67)
`
`(Range)
`
`13
`
`Enrolled
`
`Retrospective
`
`Von Bonin et al.,
`
`2008 [18]
`
`Study Design
`
`Author, Year
`
`Criteria for Response
`
`Number of Doses)
`
`Dose (Median
`
`TTR Days,
`
`By Organ
`
`Median Age, Years
`
`Number of Patients
`
`ORR,%
`
`Table 2. Clinical Studies Evaluating the Efficacy of Rituximab in the Setting of Steroid-Refractory cGVHD
`
`

`

`Biol Blood Marrow Transplant 15:/005-/0/3, 2009
`
`Rituximab in Steroid-Refractory Chronic GVHD
`
`1011
`
`..r::: :;
`
`Q)
`'-
`..f!
`" ""
`~ .g
`"%~re,
`8 ~ 0
`
`V>
`
`2-(cid:173).,,
`
`Q)
`Q)
`
`....
`'-
`
`~..f! :=:-
`
`i
`E
`e: C
`0
`~
`~
`~
`Q)
`>
`0
`a:
`Cl'.
`0
`
`vi' u
`
`Cl'. z
`
`Cl'. z
`
`co
`V>
`d,
`!::',
`....
`-.0
`
`CX)
`
`Cl'. z
`
`Cl'. z
`
`0
`V>
`r.:.
`
`V>
`
`~
`
`'o'
`~
`" e
`> ~
`'8 g
`Q) u
`Cl. C
`~ 0 e C
`o...-
`
`C M
`
`'- N o ~
`..r:; M
`:;; 0
`'- 0
`"'N
`~ _;
`o,
`"' ~
`:;; Q)
`Cl'.
`
`C
`
`The response to rituximab appears to be less pro(cid:173)
`nounced in other organs. Cases of steroid-refractory
`cGVHD of the oral mucosa showed ORRs ranging
`from O [20,22] to 83 % [1 9]. Similarly, clinical re(cid:173)
`sponses to rituximab in cases of hepatic cGVHD
`ranged from O [18,22 ,23 ] to 66% [21]. N ote that these
`differing hepatic cGVHD response rates might be ex(cid:173)
`plained by intrinsic differences in the degree of severity
`of hepatic damage resulting from cGVHD before the
`initiation of rituximab, differing durations of therapy,
`or the use of different criteria to measure responses
`to treatment, among other factors. The time to achieve
`clinical response in hepatic cGVHD was 49 days in 1
`study [24] . The responses to rituximab in patients
`with lung cGVHD were even less impressive, ranging
`from 0% [21 ,22] to 38 % [23 ,24]. It is unclear whether
`the limited efficacy of rituximab in organs other than
`skin is due to the treatment per se, or whether those or(cid:173)
`gans are more susceptible to irreversible cGVHD-im(cid:173)
`mune mediated damage. The underuse of more
`aggressive diagnostic tools to assess organ damage be(cid:173)
`fore initiation of treatment, coupled with the lack of
`more objective clinical tools to better quantitate clini(cid:173)
`cal improvement in certain organs, limit clinicians'
`ability to assess the true efficacy of rituximab, or lack
`thereof, in those organs.
`The bona fide corticosteroid-sparing effect of rit(cid:173)
`uximab treatment in cGVHD resulted in a shortened
`duration, or sometimes even discontination, of corti(cid:173)
`costeroid treatment. It is possible that this effect ulti(cid:173)
`mately may translate into improved quality of life for
`these patients. Quality of life assessment tools should
`be incorporated into prospective clinical trials to ob(cid:173)
`jectively address this issue.
`There are several limitations to the evidence pre(cid:173)
`sented herein, including (1) the absence of randomized
`trials; (2) the inclusion of only 3 prospective studies
`that enrolled a very small number of patients
`(n = 37), with none of the prospective studies clearly
`reporting either the design or phase of these prospec(cid:173)
`tive studies; (3) the poor methodological quality of the
`prospective studies (eg, none of these studies provided
`details on patient enrollment, limiting our ability to as(cid:173)
`certain whether the study subjects were derived
`through convenient sampling or consecutive enroll(cid:173)
`ment, to avoid selection bias); and (4) the lack of a ho(cid:173)
`mogeneous response criteria among studies, which
`limits the ability to assess responses as a composite
`or to compare responses among studies. In addition,
`the retrospective studies also were limited by the rela(cid:173)
`tively small sample size, and none of the reports in(cid:173)
`cluded in this systematic review indicated whether
`the analysis was intended to address an a priori hypoth(cid:173)
`esis or to generate a hypothesis. Several questions re(cid:173)
`main unanswered: What is the relationship between
`response rates (and time to maximum response) and
`depletion of B cell levels in relation to the number of
`
`

`

`1012
`
`M. A. Kharfan-Dabaja et al.
`
`Biol Blood Marrow Transplant /5:/005-10/3, 2009
`
`doses of rituximab administered? Is there an additional
`benefit to using higher doses of rituximab? Future clin(cid:173)
`ical trials should evaluate response rates at different
`times in the course of cGVHD to provide a better
`understanding of the pathophysiologic evolution of
`cGVHD over time.
`Limited understanding of the complex pathophys(cid:173)
`iology of GVHD remains the main barrier to develop(cid:173)
`ment of universally effective prophylaxis against acute
`GVHD (aGVHD) and successful treatment of both
`aGVHD and cGVHD. Development of antibody re(cid:173)
`sponses, such as allogeneic H-Y, in association with
`important
`cGVHD implicates B lymphocytes as
`players in the pathogenesis of cGVHD [20]. B lym(cid:173)
`phocytes play an essential role in antigen presentation
`to T cells and allogeneic antibody induction, among
`other mechanisms [2 5]. Several studies already have in(cid:173)
`corporated rituximab into the conditioning regimen
`for patients with CD20+ -expressing malignancies un(cid:173)
`dergoing allo-HCT with encouraging results [26,27].
`A BMT Clinical Trials Network is currently evaluat(cid:173)
`ing a multicenter phase II clinical trial (BMT CTN
`protocol 0701) that combines rituximab with fludara(cid:173)
`bine (Flu) plus cyclophospharnide (Cy) as a nonmye(cid:173)
`relapsed
`for patients with
`regimen
`loablative
`follicular non-Hodgkin lymphoma (NHL).
`Earlier intervention with anti B cell therapy could
`potentially reduce cGVHD-mediated irreversible or(cid:173)
`gan damage. Investigators at Stanford University are
`currently evaluating a strategy of adding rituximab to
`prednisone as front-line therapy for newly diagnosed
`cGVHD, aiming to improve response while allowing
`more rapid tapering of corticosteroid dosage (Clinical(cid:173)
`Trials.gov; identifier NCT003 50545); similar studies
`are underway in Europe.
`It is important to understand that the totality of the
`evidence generated through this systematic review
`demonstrates the gaps in the existing evidence base re(cid:173)
`lated to the efficacy of rituximab in treating patients
`with steroid-refractory cGVHD. This systematic re(cid:173)
`view underscores the need for well-designed and ade(cid:173)
`quately powered prospective studies to conclusively
`address this issue.
`
`ACKNOWLEDGEMENTS
`
`Financial disclosure: The authors have nothing to
`disclose.
`
`REFERENCES
`1. Karanes C, Nelson GO, Chitphakdithai P, et al. Twenty years of
`unrelated donor hematopoietic cell transplantation for adult re(cid:173)
`cipients facilitated by the National Marrow Donor Program.
`Biol Blood Man·ow Transplant. 2008;14:8-15.
`2. MacMillan ML, Davies SM, Nelson GO, et al. Twenty years of
`unrelated donor bone marrow transplantation for pediatric acute
`
`leukemia facilitated by the National Marrow Donor Program.
`Biol Blood Ma,7·ow Transplant. 2008;14:16-22.
`3. Stem Cell Trialists' Collaborative Group Allogeneic peripheral
`blood stem cell compared with bone marrow transplantation in
`the management of hematologic malignancies: an individual pa(cid:173)
`tient data meta-analysis of nine randomized trials. J Clin Oncol.
`2005;23 :5074-5087.
`4. Filipovich AH, v\Teisdorf D, Pavletic S, et al. National Institutes
`of Health consensus development project on criteria for clinical
`trials in chronic graft-versus-host disease, I: Diagnosis and Stag(cid:173)
`ing Working Group report. Biol Blood Mmrow Transplant. 2005;
`11:945-956.
`5. Pavletic SZ, Martin P, Lee SJ, et al. Measuring therapeutic re(cid:173)
`sponse in chronic graft-versus-host disease. National Institutes
`of Health Consensus Development Project on Criteria for Clin(cid:173)
`ical Trials in Chronic Graft-versus-Host Disease, IV: Response
`Criteria Working Group report. Biol Blood Maffow Transplant.
`2006;12:252-266.
`6. Lee SJ, Vogelsang G, Flowers ME. Chronic graft-versus-host
`disease. Biol Blood Mm7·ow Transplant. 2003;9:215-233.
`7. Courie! D, Hosing C, Saliba R, et al. Extracorporeal photophe(cid:173)