- Email: [email protected]
Extracorporeal Photopheresis Improves Survival in Hematopoietic Cell Transplant Patients with Bronchiolitis Obliterans Syndrome without Significantly Impacting Measured Pulmonary Functions
Accepted Manuscript Title: Extracorporeal Photopheresis Improves Survival in Hematopoietic Cell Transplant Patients with Bronchiolitis Obliterans Syndrome Without Significantly Impacting Measured Pulmonary Functions Author: Mehrdad Hefazi, Kimberly J. Langer, Nandita Khera, Jill Adamski, Vivek Roy, Jeffrey L. Winters, Dennis A. Gastineau, Eapen K. Jacob, Justin D. Kreuter, Manish J. Gandhi, William J. Hogan, Mark R. Litzow, Shahrukh K. Hashmi, Hemang Yadav, Vivek N. Iyer, J.P. Scott, Mark E. Wylam, Rodrigo Cartin-Ceba, Mrinal M. Patnaik PII: DOI: Reference:
S1083-8791(18)30193-9 https://doi.org/10.1016/j.bbmt.2018.04.012 YBBMT 55097
To appear in:
Biology of Blood and Marrow Transplantation
Received date: Accepted date:
31-1-2018 9-4-2018
Please cite this article as: Mehrdad Hefazi, Kimberly J. Langer, Nandita Khera, Jill Adamski, Vivek Roy, Jeffrey L. Winters, Dennis A. Gastineau, Eapen K. Jacob, Justin D. Kreuter, Manish J. Gandhi, William J. Hogan, Mark R. Litzow, Shahrukh K. Hashmi, Hemang Yadav, Vivek N. Iyer, J.P. Scott, Mark E. Wylam, Rodrigo Cartin-Ceba, Mrinal M. Patnaik, Extracorporeal Photopheresis Improves Survival in Hematopoietic Cell Transplant Patients with Bronchiolitis Obliterans Syndrome Without Significantly Impacting Measured Pulmonary Functions, Biology of Blood and Marrow Transplantation (2018), https://doi.org/10.1016/j.bbmt.2018.04.012. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
TITLE PAGE TITLE: Extracorporeal Photopheresis Improves Survival in Hematopoietic Cell Transplant Patients with Bronchiolitis Obliterans Syndrome without Significantly Impacting Measured Pulmonary Functions
SHORT TITE: ECP Improves Survival in HCT-Related BOS
AUTHORS: Mehrdad Hefazi1, Kimberly J. Langer1, Nandita Khera2, Jill Adamski3, Vivek Roy4, Jeffrey L. Winters5, Dennis A. Gastineau5, Eapen K. Jacob5, Justin D. Kreuter5, Manish J. Gandhi5, William J. Hogan1, Mark R. Litzow1, Shahrukh K. Hashmi1, Hemang Yadav6, Vivek N. Iyer6, J. P. Scott6, Mark E. Wylam6, Rodrigo Cartin-Ceba7, and Mrinal M. Patnaik1* 1
Division of Hematology, Mayo Clinic, Rochester, MN; 2Division of Hematology, Mayo Clinic,
Scottsdale, AZ, 3Department of Laboratory Medicine and Pathology, Mayo Clinic, Scottsdale, AZ; 4
Division of Hematology, Mayo Clinic, Jacksonville, FL; 5Division of Transfusion Medicine, Mayo Clinic,
Rochester, MN; 6Division of Pulmonary and Critical Care Medicine, Mayo Clinic, Rochester, MN; 7
Division of Pulmonary and Critical Care Medicine, Mayo Clinic, Scottsdale, AZ
*
CORRESPONDING AUTHOR:
Mrinal M. Patnaik, MBBS Mayo Clinic 200 First St SW, Rochester, MN 55905 Tel: 507-538-0591 Fax: 507-266-4972 [email protected] 1 Page 1 of 31
MANUSCRIPT DETAILS: Word count (body): 3995 Word count (abstract): 250 Tables: 5 Figures: 2 References: 30
Keywords: Hematopoietic Cell Transplantation Bronchiolitis Obliterans Syndrome Extracorporeal Photopheresis Graft-versus-Host Disease
HIGHLIGHTS:
ECP favorably impacts overall survival in HCT patients with BOS.
This survival benefit is independent of the ECP effect on measured pulmonary function.
Corticosteroid-sparing effect of ECP may be responsible for the improved survival.
2 Page 2 of 31
Abstract
We carried out the first matched retrospective cohort study aimed at studying the safety and efficacy of extracorporeal photopheresis (ECP) for bronchiolitis obliterans syndrome (BOS) following allogeneic hematopoietic cell transplantation (HCT). Medical records of 1325 consecutive adult patients who underwent HCT between 2005 and 2015 were reviewed. Seventyfour patients (median age, 51 years) with a diagnosis of BOS were included in the study. After propensity-score matching for BOS severity, 26 patients who underwent ≥3 months of ECP were matched to 26 non-ECP-treated patients, who were assigned an index date corresponding to the ECP start date for their matched pairs. The rate of decline in FEV1 percentage predicted (FEV1PP) decreased after ECP initiation (and after index date in the non-ECP group), with no significant difference between the two groups (p=0.33). On a multivariable analysis that included baseline transplant and pulmonary function test variables, matched related donor HCT (HR 0.1 [0.03-0.5]; p=0.002), ECP (HR 0.1 [0.01-0.3]; p=0.001) and slower rate of decline in FEV1PP before ECP/index date (HR 0.7 [0.6-0.8]; p=0.001) were associated with a better overall survival. At last follow-up, non-ECP-treated patients were more likely to be on >5 mg daily dose of prednisone (54% vs. 23%; p=0.04) and had a greater decline in their Karnofsky performance score (mean difference −9.5 vs. −1.6; p=0.06) compared to ECP-treated-patients. In conclusion, compared to other BOS-directed therapies, ECP was found to improve survival in HCT patients with BOS, without significantly impacting measured pulmonary functions. These findings need prospective validation in a larger patient cohort.
3 Page 3 of 31
Introduction
Bronchiolitis obliterans syndrome (BOS) is an increasingly recognized cause of morbidity and mortality in patients undergoing allogeneic hematopoietic cell transplantation (HCT). BOS affects 3–14% of patients, usually within the first 2 years of HCT1–3. BOS is intricately linked to graft-versus-host disease (GVHD) and often occurs following a reduction in immunosuppressive therapies (IST)4. In spite of augmented IST, prognosis in affected patients is dismal with a reported overall survival (OS) of 10–20% at 5 years5,6. Although combination ISTs, mainly steroids and calcineurin inhibitors (CNIs), have attempted to slow disease progression and prolong survival, these regimens are associated with significant risks and toxicities. Likewise, results of combination fluticasone, azithromycin, and montelukast (FAM) have shown conflicting results, with a recent clinical trial suggesting azithromycin may worsen OS when used pre-emptively for BOS prevention after HCT7.
Thus, due to suboptimal responses to medical therapies, extracorporeal photopheresis (ECP) has been used as an immunomodulatory alternative treatment for BOS. A few small cohort studies on BOS after lung transplantation have indicated that ECP may stabilize pulmonary function test (PFT) results and improve survival 7–9. However, to the best of our knowledge, there are only three published single-center retrospective studies on the utility of ECP for BOS after HCT10–12. In 2011, Lucid et al.10 reported the experience on nine patients with rapidly declining PFTs, who were treated with ECP for a maximum of 12 months.
4 Page 4 of 31
Approximately two-thirds of their patients showed PFT stabilization, and were able to taper corticosteroids. In the second report11, ECP was used in 13 chronic graft-versus-host (cGVHD) patients with BOS refractory to standard treatment. Seven of these patients had symptomatic BOS before starting ECP, whereas the remaining six patients developed BOS while on ECP treatment for steroid-refractory cGVHD. Preliminary results from this study supported the efficacy of ECP for slowing PFT decline, even in patients with advanced stages of BOS. Similarly, Brownback et al.12 demonstrated stability in % predicted FEV1 (FEV1PP) values in 7 of 8 BOS patients within three months after the initiation of ECP, a finding that was limited to only 2 of the 8 patients at the one year mark. They also noted a significant decrease in the median dose of prednisone per patient throughout the 12 months of ECP treatment.
Taken together, these retrospective studies, although encouraging, have been limited by small patient numbers, varying BOS severities, lack of control groups, and limited follow-up data on clinical outcomes. Therefore, we conducted this large multi-site matched retrospective cohort study to evaluate the safety and efficacy of ECP for the treatment of BOS after HCT.
Methods Patients and study design Medical records of 1,325 consecutive patients, ≥18 years old, who underwent HCT at the Mayo Clinic sites in Rochester, MN (n = 810; 2005–2015), Scottsdale, AZ (n = 397; 2011–2015), and Jacksonville, FL (n = 118; 2011–2015) were retrospectively reviewed. Patients were excluded if they died before day 100 after HCT (n = 172), did not have consistent PFT monitoring (pre5 Page 5 of 31
HCT, day 100, and yearly thereafter) (n = 118), had a baseline obstructive PFT pattern (n = 59), or underwent a second HCT (n = 17) (Figure 1). PFT data were interpreted according to the previously published guidelines for all 959 eligible patients.13,14 For patients with an obstructive, mixed, or non-specific PFT pattern after HCT, clinical notes, chest computed tomography, bronchoscopy, lung biopsy, and microbiology results were reviewed, if available. The study was approved by the Mayo Clinic Institutional Review Board.
BOS diagnosis BOS was diagnosed according to the NIH 2014 criteria15 with an additional requirement that the obstructive findings on PFT be present on two consecutive tests within 3 months, to ensure PFT abnormalities were not secondary to other reversible causes. For patients who met the PFT criteria but did not have other supporting features, such as distinctive cGVHD manifestations, air-trapping (by PFT or expiratory CT), or unequivocal ruling out of confounders, such as respiratory tract infections, a complete review of the post-transplant course was performed by the research team’s pulmonary and transplant physicians. Patients were excluded (n = 49) if their BOS diagnosis remained indeterminate.
BOS Treatments Patients treated with ≥3 months of ECP within 1 year of BOS diagnosis comprised the ECP group. Patients were excluded if they started ECP prior to (n = 2) or >12 months after (n = 9) BOS diagnosis. The UVAR XTS or the CELLEX system (Mallinckrodt Pharmaceuticals [formerly Therakos, Inc.], West Chester, PA) was used at all three sites to administer ECP according to previously described methods.8 Vascular access was obtained either through 6 Page 6 of 31
peripheral veins (n = 12), tunneled central venous catheters (n = 4), or implanted vascular access devices (n = 12). In three patients, vascular access had to be changed from peripheral venous to a central venous catheter during the course of treatment, due to issues with sustained peripheral venous access. Methoxypsoralen was administered at a dose of 0.35 µg per ml of collected buffy coat. The planned therapeutic course consisted of one treatment cycle on two consecutive days per week for 1 month, followed by one treatment cycle every other week for 2 months, followed by one treatment cycle every third week for 3 months, and followed by one treatment cycle monthly thereafter (Mayo Clinic standard operating protocol). Treatment frequency could be modified at the physician’s discretion. Non-ECP treatments primarily consisted of systemic steroids, CNIs (cyclosporine or tacrolimus), and other ISTs that were started at varying time points chosen by treating physicians.
Changes in FEV1 over time Each FEV1 was compared to the predicted normal value for that particular patient and a percentage predicted FEV1 (FEV1PP) was used for PFT analyses. Similar percentage predicted values were calculated for the FVC (FVCPP) and DLCO (DLCOPP). The rate of change in FEV1PP was calculated for each patient by using the slope of a linear regression line for FEV1PP vs time, as described previously6,8. This method was chosen because PFTs were performed at varying intervals in each patient. The rate of change in FEV1PP was calculated separately for the periods from BOS diagnosis to ECP initiation and for the subsequent 12 months. For non-ECP treated patients, these intervals were based on the establishment of an index date corresponding to the ECP start date for their matched pairs (see Statistical methods section). 7 Page 7 of 31
Outcome variables PFT data, including FEV1PP, FEV1/FVC ratio (%), and DLCOPP (after adjustment for hemoglobin) were compared between the ECP and non-ECP groups at baseline (pre-HCT), at BOS diagnosis, and at ECP/index data before and after propensity score matching (PSM). The rate of change in FEV1PP before and after ECP/index date was compared as paired variables. Acute and cGVHD were graded according to the Glucksberg15 and NIH 2014 criteria14, respectively. Since all patients were diagnosed with lung GVHD, only non-pulmonary manifestations were included in the overall grading of cGVHD. The Karnofsky performance score (KPS)16 at BOS diagnosis and at last outpatient follow-up were compared between the two groups by using matched paired data analysis. Respiratory tract infections were classified according to the implicated pathogens. Bacterial infections were diagnosed based on clinical and/or microbiological data, whereas viral and fungal infections were diagnosed solely based on microbiological tests.
Statistical analysis Since ECP users did not start ECP immediately after the diagnosis of BOS, there is an immortal time bias in favor of the ECP group. To account for this, the “prescription time distribution matching” (PTDM) method was used, as previously described17,18. Non-ECP users were assigned an index date corresponding to the length of time from BOS diagnosis to ECP start date for their matched ECP-treated pairs. Non-ECP users were excluded if they died before their assigned index date plus 90 days. The 90-day period was added to the index date because all ECP users had received ≥3 months of ECP, during which they were considered immortal. 8 Page 8 of 31
Propensity score matching was used to balance the distribution of potential confounders between the ECP and non-ECP groups. FEV1PP and DLCOPP at ECP/index time were included in the model because their distribution was significantly different between the two groups. A multivariable analysis was then used to account for other potential confounders and for possible residual imbalance in PFT variables after PSM. Propensity scores were computed by using the logistic regression model, and the matching tolerance was set at 0.2, as previously suggested 19. The chi-square (or Fisher’s exact) and Wilcoxon tests were used to compare categorical and continuous variables, respectively, between independent groups. The Wilcoxon signed-rank test was used for paired data analysis.
The Kaplan–Meier method was used to estimate OS from BOS diagnosis, and the log-rank test was used comparisons. A Cox proportional hazard model was used to analyze the association of ECP and PFT variables with OS in the entire and propensity score-matched cohorts. All variables tested in the univariable analysis were entered into a multivariable model, and were retained in the final model using backwards selection method with a significance level of p< 0.05. All tests were two-sided. SPSS software version 22 (SPSS, Inc., Chicago, Illinois) was used for PSM. All other analyses were performed by using JMP software version 10 (SAS Institute Inc., Cary, NC, USA).
Results
9 Page 9 of 31
Across the three Mayo Clinic sites, 88 patients were diagnosed with BOS between January 2005 and December 2015. Of these, 74 patients met the study’s eligibility criteria and were included in the study. Twenty-eight patients received ECP, and 46 were treated without ECP. After PSM, 26 ECP-treated patients were matched to 26 non-ECP-treated patients (Figure 1). Baseline characteristics and PFT data for the matched and unmatched cohorts are provided in tables 1 and 3, whereas data on treatments, clinical outcomes and adverse events are shown only for the PSM cohorts and described in more details in the following sections.
Baseline characteristics Demographics and baseline transplant characteristics for the matched ECP-treated and non-ECPtreated cohorts are shown in Table 1. There were no significant differences in patients’ age, year of HCT, transplant center, HCT comorbidity index, distribution of underlying disease, conditioning regimen, donor type, ABO compatibility, CMV status, and immunosuppression at HCT between the two groups. Severity grades of acute GVHD before BOS diagnosis and nonpulmonary cGVHD at the time of BOS diagnosis were comparable between the two groups.
BOS was diagnosed after a median of 14 months (range, 3-45 month) from HCT in the matched ECP group and after 13 months (range, 3-36 months) in the non-ECP group (p = 0.45). Median time from BOS diagnosis to ECP initiation was 1.5 months (range, 0–11 months). This was comparable to the time interval from BOS diagnosis to the assigned index dates for the non-ECP users (median, 1.5; range, 0–11 months), indicating proper matching based on the PTDM method.
10 Page 10 of 31
ECP and Non-ECP Treatments In the matched ECP group, 26 patients underwent a median of 31 cycles of ECP (range, 6–60 cycles) over a median of 15 months (range, 3–58 months). Of these, 7 (27%) patients received ≤20 cycles of ECP, 11 (42%) received 21−40 cycles of ECP, and the remaining 8 (31%) patients received ≥41 cycles of ECP. Immunosuppressive therapies at the time of BOS diagnosis, during the interval from BOS diagnosis to ECP/Index date, and following ECP/Index date are summarized in Table 2. At the time of BOS diagnosis, approximately half of the matched ECP and non-ECP-treated patients were on corticosteroids (61% vs. 46%; p = 0.40) and/or CNIs (54% vs. 46%; p = 78). The rate of corticosteroid use was comparable between the ECP and non-ECP treated groups during the interval from BOS diagnosis to ECP/Index date (88% vs. 73%; p = 0.29), and also after the ECP/Index date (81% vs. 81%; p = 1.0). Likewise, the rates of CNI use and treatment with other IST agents were comparable between the two groups during both periods (Table 2). The majority of patients (88% of the ECP-treated and 81% of non-ECP-treated patients) received combination therapy with fluticasone, azithromycin, and montelukast (FAM protocol)20 after ECP/Index date.
PFT data before and after BOS diagnosis PFT variables, including FEV1PP, FEV1/FVC ratio, and DLCOPP, were comparable between the matched ECP and non-ECP groups before HCT, at BOS diagnosis, and at ECP/index date (Table 3). FEV1PP values at last follow-up were comparable between the matched ECP and non-ECP groups (median 43% vs. 46 % predicted; p = 0.43), whereas there was a trend toward lower DLCOPP in the matched ECP group (48% vs. 62 % predicted; p = 0.08). The rate of decline in FEV1PP was faster before rather than after the ECP/index date in both the ECP (−4.5 vs −0.2; p < 11 Page 11 of 31
0.0001) and non-ECP groups (−3.6 vs −0.5; p < 0.001). Comparison between the two groups, however, did not reveal a significant difference (p = 0.33) (Table 3).
Clinical outcomes and adverse events With a median follow-up of 38 months (range, 1–135 months), OS was significantly better in the matched ECP group than in the non-ECP group (estimated median not reached vs 32 months; p = 0.01) (Figure 2). Six (23%) patients in the ECP group and 13 (50%) in the non-ECP group had expired at last follow-up. Respiratory tract infection was the leading cause of death for both ECP- and non-ECP-treated patients (75% vs 69%, p = 1.0). One (4%) patient in each group died from underlying disease relapse. Six (23%) patients in the ECP group and 4 (15%) in the nonECP group developed chronic respiratory failure requiring home oxygen therapy at last followup (Table 4).
In each group, 14 (50%) patients had moderate or severe non-pulmonary cGVHD at the last follow-up. Despite comparable severities of non-pulmonary cGVHD between the two groups, non-ECP-treated patients were almost twice as likely to be on >5 mg daily dose of prednisone at the last outpatient visit (54% vs 23%; p = 0.04). The use of CNIs and other ISTs was comparable between the two groups. ECP- and non-ECP-treated patients had comparable KPS at BOS diagnosis (mean, 78 vs 82; p = 0.34). However, non-ECP-treated patients had a greater decline in KPS over the course of treatment (mean difference, −9.5 vs. −1.6; p = 0.06).
12 Page 12 of 31
Univariable survival analysis in the matched cohort demonstrated better OS in HCT recipients from matched related donor (MRD) compared to matched unrelated donor (MUD) (HR 0.3 [0.10.8]; p = 0.02) and in patients treated with ECP (HR 0.3 [0.1-0.8]; p = 0.01) (Table 5). On multivariable analysis that included baseline transplant and PFT variables, HCT from MRD (HR 0.1 [0.03-0.5]; p=0.002), ECP (HR 0.1 [0.01-0.3]; p = 0.003) and slower rate of decline in FEV1PP before ECP/index date (HR 0.7 [0.6-0.8] per unit change; p = 0.001) were independently and favorably prognostic with regards to OS.
Respiratory tract infections due to bacterial, viral, and fungal pathogens were observed in 15 (58%), 9 (35%), and 7 (27%) patients in the ECP group, and in 13 (50%), 7 (27%), and 4 (15%) of the non-ECP-treated patients (p = 0.78, 0.76, and 0.49, respectively). Complications related to central venous access were observed in 2 (33%) of 6 patients with a tunneled central catheter and in 9 (75%) of 12 patients with an implanted vascular access device. These were most commonly due to central venous access malfunction/malposition requiring catheter or device exchange (n = 7), followed by central line associated blood stream infection (CLABSI) (n = 4), bleeding, and venous thrombosis (each in one patient). No cases of bacteremia were documented when ECP was administered through peripheral intravenous catheters.
Discussion We performed a large multi-site matched retrospective cohort study evaluating the safety and efficacy of ECP for patients with post-HCT BOS. Eligible patients met the NIH 2014 criteria for BOS and had a new obstructive PFT pattern on two consecutive tests within 3 months, prior to inclusion; so as to ensure that PFT abnormalities were not secondary to acute reversible causes.
13 Page 13 of 31
We excluded patients with an obstructive PFT pattern at baseline and those for whom confounding factors could not be excluded after review of the post-transplant course. The ECP group consisted of a relatively homogenous cohort of patients treated for ≥3 months with ECP within 1 year of BOS diagnosis, and the non- ECP group was selected by PSM.
In our study, FEV1PP values progressively declined from BOS diagnosis to last follow-up in the ECP- and non-ECP-treated cohorts. Although the rate of FEV1PP decline was faster before rather than after ECP initiation, a similar trend was also noted in the non-ECP-treated group; indicative of a comparable impact on PFT changes. Slowing of the rate of FEV1PP decline has been reported in HCT-related BOS after ECP therapy 7,8 and also in patients treated without ECP21; however, only one study has compared PFT results between ECP- and non-ECP-treated patients. In that particular post lung transplantation study, Pecoraro et al.9 reported that FEV1PP values at 12 months after BOS diagnosis were higher for the ECP-treated patients (n = 15) than for nonECP-treated patients (n = 39). The former group, however, had higher FEV1PP values at BOS diagnosis, which, in the absence of matched pair data analysis, may have confounded the results. Furthermore, there is a possibility that earlier initiation of ECP could be more effective in mitigating the decline in pulmonary functions, and in preventing the irreversible changes that ensue. Therefore, closer monitoring of PFTs after HCT, and dedicated prospective studies to establish the relationship between the timing of ECP and PFT changes are thus warranted.
There are currently no published reports on the impact of ECP on survival in BOS patients following HCT. In our study, ECP-treated patients had significantly better OS compared to non14 Page 14 of 31
ECP-treated patients (estimated median not reached vs 32 months). After PSM, the only independent prognostic factors for OS were ECP therapy, HCT from MRD, and the rate of decline in FEV1PP before ECP/index date. Rate of decline in FEV1 after ECP/index date did not appear to be different between the two groups. Hence, there may be extra-pulmonary aspects to ECP that are likely more important factors in survival.
To explore the reasons for survival difference between the ECP and non-ECP groups, we compared the patients’ KPS, ISTs, and non-pulmonary cGVHD status at BOS diagnosis and at last outpatient follow-up. ECP and non-ECP groups had comparable rates of moderate/severe non-pulmonary cGVHD at BOS diagnosis (65% for both groups) and at last follow-up (54% for both groups). Thus, the ECP effect on non-pulmonary cGVHD was unlikely to be responsible for the survival difference. While the use of CNI at last follow-up was comparable between the two groups, non-ECP-treated patients were more likely to be on a >5 mg daily prednisone dose and had a greater decline in KPS during treatment. This is consistent with the results of several prior studies 22–24 that have demonstrated corticosteroid-sparing effects of ECP and correlations between corticosteroid reduction and improved survival in patients with GVHD. Similarly, in a retrospective study of 77 BOS patients, Bergeron et al. demonstrated that intentional treatment with corticosteroids was associated with a shorter OS, without significantly affecting FEV1 values25. Another possible explanation for the greater decline in KPS in non-ECP-treated patients could be the higher toxicity from prolonged corticosteroid use. In our study, we chose a cut-off value of 5 mg/day prednisone dose as a measure of successful corticosteroid tapering, based on the current evidence that long-term prednisone use of ≥5 mg/day correlates with adverse events
15 Page 15 of 31
in a dose-dependent fashion26. It should, however, be noted that our study was not specifically designed to determine the effect of ECP on corticosteroid tapering.
One of the intriguing aspects of ECP is its ability to induce two opposing effects: activation of the immune system (as in treatment of cutaneous T cell lymphoma) and down-regulation of the activity of T lymphocytes (as in treatment of allograft rejection and GVHD)27,28. The immunomodulatory effects of ECP may be additionally advantageous in patients with hematologic malignancies and GVHD after HCT, because ECP does not cause generalized immunosuppression and thus, does not increase the risk of infectious complications and disease relapse29,30. Further studies on the immunological mechanisms of ECP are much needed.
Adverse ECP-related events were mainly due to central venous access complications (central line malfunction, infection, and thrombosis). The incidence of CLABSI (22%) in our study was slightly higher than previously reported (13%)8. All CLABSI cases were successfully treated with catheter removal and antibiotics, without long-term sequelae. The incidence of respiratory tract infections after BOS diagnosis was comparable between the ECP and non-ECP groups. However, only 4 (21%) of 19 episodes of respiratory infections in the ECP group resulted in mortality, versus 9 (56%) of 16 episodes in the non-ECP group (p = 0.04). Although this could be because of the superior safety profile and corticosteroid-sparing effect of ECP, confirmation of a causal relationship is difficult to ascertain in a retrospective study. The financial toxicity and the logistical challenges of ECP administration are other areas of concern, especially for BOS, where patients undergo prolonged courses of therapy (a median of 31 cycles in this study)8. The 16 Page 16 of 31
evaluation of the cost-effectiveness of ECP and its impact on patients’ quality of life are thus urgently required.
Limitations of our study include its retrospective non-randomized design and that the decision to start and subsequently taper ECP was at the clinician’s discretion; a process that can potentially bias outcomes. Despite available diagnostic criteria15, BOS remains difficult to diagnose, and there are no consistent response-monitoring guidelines. The heterogeneity of concomitant ISTs, varying duration of ECP, and missing data also complicate the analysis. We tried to address these limitations by excluding indeterminate cases of BOS, restricting the ECP group to patients treated with ≥3 months of ECP within one year of diagnosis, and comparing potential confounders at baseline and at last follow-up. While the use of specific classes of ISTs after BOS diagnosis were comparable between the ECP and non-ECP groups in our study, a complete matching with regard to the type and duration of non-ECP therapies was not possible. In our opinion, this is a difficult task to achieve even in the setting of a prospective study, given the fact that the current therapeutic approach to BOS is largely empirical and is associated with various treatment crossovers depending on each individual’s response. It should also be noted that our study was not designed to compare the efficacy of different non-ECP therapies such as FAM or other novel ISTs because we did not specifically match patients for these other therapeutic modalities.
We used the percentage predicted slope of FEV1PP to incorporate PFT results from varying time points into a single analysis, and therefore decrease the problem of missing values at pre-defined
17 Page 17 of 31
time points. The slope of FEV1 decline was calculated for the periods between BOS diagnosis and ECP/index date, and also for the first 12 months after BOS diagnosis. Although the calculated rate of FEV1 decline provides a crude estimate of the trajectory of pulmonary function decline over a specific time period, accurate measurement of the rate of FEV1 decline in each individual will require consistent PFT monitoring at specific time intervals that are yet to be defined and prospectively validated.
Conclusion
In conclusion, we found that ECP improves survival in HCT patients with BOS without significantly impacting pulmonary functions. After adjusting for BOS severity, OS was significantly better for patients treated with ECP for ≥3 months as compared to that in a propensity matched non-ECP treated cohort. The survival benefit was independent of the effects on PFT results and may have been related to the corticosteroid-sparing effect of ECP or other yet to be defined immunological effects. Prospective randomized trials are needed to confirm the efficacy of ECP, determine the optimal schedule, and standardize the response evaluation. Immunological investigations should also be incorporated into future clinical studies to explore biomarkers for use before or during treatment to identify patients who will benefit from ECP.
Acknowledgements
18 Page 18 of 31
The authors thank the staff at apheresis units and pulmonary labs at the Mayo Clinic in Rochester, MN, Scottsdale, AZ, and Jacksonville, FL for their technical and logistical support.
Authorship Contributions M.H. contributed to study conception, methodology, data collection, data analysis, and manuscript writing; K.J.L: contributed to data collection; N.K., J.A., V.R., J.L.W., D.A.G., E.K.J., M.R.L, and W.J.H. provided logistical and administrative support, and contributed to patient care and manuscript writing; J.D.K., M.J.G., S.K.H, H.Y., and V.N.I contributed to data interpretation and manuscript writing; J.P.S., M.E.W., and R.C.C, contributed to patient care, data interpretation, and manuscript writing; M.M.P: contributed to study conception, methodology, patients care, manuscript writing, and supervision. All authors contributed substantially to this work and have approved the submission of this manuscript.
Conflict of Interest Disclosure The authors declare no conflict of interest.
19 Page 19 of 31
References 1.
Au BKC, Au MA, Chien JW. Bronchiolitis obliterans syndrome epidemiology after allogeneic hematopoietic cell transplantation. Biol Blood Marrow Transplant. 2011;17(7):1072-1078. doi:10.1016/j.bbmt.2010.11.018
2.
Dudek AZ, Mahaseth H, DeFor TE, Weisdorf DJ. Bronchiolitis obliterans in chronic graft-versus-host disease: analysis of risk factors and treatment outcomes. Biol Blood Marrow Transplant. 2003;9(10):657-666. doi:10.1016/S1083
3.
Sakaida E, Nakaseko C, Harima A, et al. Late-onset noninfectious pulmonary complications after allogeneic stem cell transplantation are significantly associated with chronic graft-versus-host disease and with the graft-versus-leukemia effect. Blood. 2003;102(12):4236-4242. doi:10.1182/blood-2002-10-3289
4.
Gazourian L, Rogers AJ, Ibanga R, et al. Factors associated with bronchiolitis obliterans syndrome and chronic graft-versus-host disease after allogeneic hematopoietic cell transplantation. Am J Hematol. 2014;89(4):404-409. doi:10.1002/ajh.23656
5.
Williams KM, Chien JW, Gladwin MT, Pavletic SZ. Bronchiolitis obliterans after allogeneic hematopoietic stem cell transplantation. JAMA. 2009;302(3):306-314. doi:10.1001/jama.2009.1018
6.
Williams KM. How I Treat How I treat bronchiolitis obliterans syndrome after hematopoietic stem cell transplantation. Blood. 2017;129(4):448-456. doi:10.1182/blood2016-08-693507.448
7.
Jaksch P, Scheed A, Keplinger M, et al. A prospective interventional study on the use of extracorporeal photopheresis in patients with bronchiolitis obliterans syndrome after lung transplantation. J Hear Lung Transplant. 2012;31(9):950-957. doi:10.1016/J.HEALUN.2012.05.002
8.
Morrell MR, Despotis GJ, Lublin DM, Patterson GA, Trulock EP, Hachem RR. The efficacy of photopheresis for bronchiolitis obliterans syndrome after lung transplantation. J Hear Lung Transplant. 2010;29(4):424-431. doi:10.1016/j.healun.2009.08.029
20 Page 20 of 31
9.
Pecoraro Y, Carillo C, Diso D, et al. Efficacy of Extracorporeal Photopheresis in Patients With Bronchiolitis Obliterans Syndrome After Lung Transplantation. Transplant Proc. 2017;49(4):695-698. doi:10.1016/j.transproceed.2017.02.035
10.
Lucid CE, Savani BN, Engelhardt BG, et al. Extracorporeal photopheresis in patients with refractory bronchiolitis obliterans developing after allo-SCT. Bone Marrow Transplant. 2011;46(3):426-429. doi:10.1038/bmt.2010.152
11.
Del Fante C, Galasso T, Bernasconi P, et al. Extracorporeal photopheresis as a new supportive therapy for bronchiolitis obliterans syndrome after allogeneic stem cell transplantation. Bone Marrow Transplant. 2016:1-4. doi:10.1038/bmt.2015.324
12.
Brownback KR, Simpson SQ, Pitts LR, et al. Effect of extracorporeal photopheresis on lung function decline for severe bronchiolitis obliterans syndrome following allogeneic stem cell transplantation. J Clin Apher. 2016;31(4):347-352. doi:10.1002/jca.21404
13.
Pellegrino R, Viegi G, Brusasco V, et al. Interpretative strategies for lung function tests. Eur Respir J. 2005;26(5):948-968. doi:10.1183/09031936.05.00035205
14.
Iyer VN, Schroeder DR, Parker KO, Hyatt RE, Scanlon PD. The nonspecific pulmonary function test: Longitudinal follow-up and outcomes. Chest. 2011;139(4):878-886. doi:10.1378/chest.10-0804
15.
Jagasia MH, Greinix HT, Arora M, et al. National Institutes of Health Consensus Development Project on Criteria for Clinical Trials in Chronic Graft-versus-Host Disease: I. The 2014 Diagnosis and Staging Working Group report. Biol Blood Marrow Transplant. 2015;21(3):389-401.e1. doi:10.1016/j.bbmt.2014.12.001
16.
Karnofsky DA. Determining the extent of the cancer and clinical planning for cure. Cancer. 1968;22(4):730-734. http://www.ncbi.nlm.nih.gov/pubmed/5212297.
17.
Li G, Holbrook A, Delate T, Witt DM, Levine MA, Thabane L. Prediction of individual combined benefit and harm for patients with atrial fibrillation considering warfarin therapy: a study protocol. BMJ Open. 2015;5(11):e009518. doi:10.1136/bmjopen-2015009518
21 Page 21 of 31
18.
Zhou Z, Rahme E, Abrahamowicz M, Pilote L. Survival bias associated with time-totreatment initiation in drug effectiveness evaluation: A comparison of methods. Am J Epidemiol. 2005;162(10):1016-1023. doi:10.1093/aje/kwi307
19.
Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat. 2011. doi:10.1002/pst.433
20.
Williams KM, Cheng G-S, Pusic I, et al. Fluticasone, Azithromycin, and Montelukast Treatment for New-Onset Bronchiolitis Obliterans Syndrome after Hematopoietic Cell Transplantation. Biol Blood Marrow Transplant. 2016;22(4):710-716. doi:10.1016/j.bbmt.2015.10.009
21.
Cheng G-S, Storer B, Chien JW, et al. Lung Function Trajectory in Bronchiolitis Obliterans Syndrome after Allogeneic Hematopoietic Cell Transplant. Ann Am Thorac Soc. 2016;13(11):1932-1939. doi:10.1513/AnnalsATS.201604-262OC
22.
Couriel DR, Hosing C, Saliba R, et al. Extracorporeal photochemotherapy for the treatment of steroid-resistant chronic GVHD. Blood. 2006;107(8):3074-3080. doi:10.1182/blood-2005-09-3907
23.
Foss FM, DiVenuti GM, Chin K, et al. Prospective study of extracorporeal photopheresis in steroid-refractory or steroid-resistant extensive chronic graft-versus-host disease: analysis of response and survival incorporating prognostic factors. Bone Marrow Transplant. 2005;35(12):1187-1193. doi:10.1038/sj.bmt.1704984
24.
Greinix HT, Worel N, Just U, Knobler R. Extracorporeal photopheresis in acute and chronic graft-versus-host disease. Transfus Apher Sci. 2014;50(3):349-357. doi:10.1016/j.transci.2014.04.005
25.
Bergeron A, Godet C, Chevret S, et al. Bronchiolitis obliterans syndrome after allogeneic hematopoietic SCT: Phenotypes and prognosis. Bone Marrow Transplant. 2013;48(6):819-824. doi:10.1038/bmt.2012.241
26.
Saag KG, Koehnke R, Caldwell JR, et al. Low dose long-term corticosteroid therapy in
22 Page 22 of 31
rheumatoid arthritis: an analysis of serious adverse events. Am J Med. 1994;96(2):115123. http://www.ncbi.nlm.nih.gov/pubmed/8109596. 27.
Fimiani M, Di Renzo M, Rubegni P. Mechanism of action of extracorporeal photochemotherapy in chronic graft-versus-host disease. Br J Dermatol. 2004;150(6):1055-1060. doi:10.1111/j.1365-2133.2004.05918.x
28.
Voss CY, Fry TJ, Coppes MJ, Blajchman MA. Extending the Horizon for Cell-Based Immunotherapy by Understanding the Mechanisms of Action of Photopheresis. Transfus Med Rev. 2010;24(1):22-32. doi:10.1016/j.tmrv.2009.09.008
29.
Greinix HT, Socié G, Bacigalupo A, et al. Assessing the potential role of photopheresis in hematopoietic stem cell transplant. Bone Marrow Transplant. 2006;38(4):265-273. doi:10.1038/sj.bmt.1705440
30.
Knobler R, Barr ML, Couriel DR, et al. Extracorporeal photopheresis: past, present, and future. J Am Acad Dermatol. 2009;61(4):652-665. doi:10.1016/j.jaad.2009.02.039
Figure 1: Flow diagram of the Study Cohort
Figure 2: Kaplan-Meier Survival Curves for the Matched ECP and Non-ECP-Treated Patients
23 Page 23 of 31
Table 1 Table 1: Demographics and Transplant Characteristics of the ECP and Non-ECP-Treated Groups before and after Propensity-Score Matching Unmatched Cohort ECP No ECP (n = 28) (n = 46)
p
ECP (n = 26)
Matched Cohort No ECP (n = 26)
p
Year of HCT 2005-2010 2011-2015
10 (36) 18 (64)
16 (35) 30 (65)
1.0
9 (35) 17 (65)
8 (31) 18 (69)
1.0
Transplant Center Rochester, MN Scottsdale, AZ Jacksonville, FL
20 (71) 6 (21) 2 (8)
37 (80) 5 (11) 4 (9)
0.48
19 (72) 5 (20) 2 (8)
19 (72) 4 (15) 3 (13)
0.91
46 (21-67)
51 (20-76)
0.64
50 (21-67)
51 (21-76)
0.76
Male, n (%)
11 (39)
26 (56)
0.23
10 (38)
14 (54)
0.40
HCT-CI, n 0-2 ≥3
24 (86) 4 (14)
38 (83) 8 (17)
0.76
22 (85) 4 (15)
23 (88) 3 (12)
0.99
Underlying Diagnosis, n (%) AML ALL MDS CML Other
10 (36) 8 (28) 4 (14) 4 (14) 2 (7)
23 (50) 8 (17) 6 (13) 2 (5) 7 (15)
0.33 0.38 1.0 0.19 0.46
9 (34) 8 (31) 4 (15) 3 (12) 2 (8)
14 (54) 6 (23) 1 (4) 1 (4) 4 (15)
0.26 0.75 0.34 0.60 0.66
Conditioning Regimen, n (%) MAC RIC
16 (57) 12 (43)
21 (46) 25 (54)
0.47
14 (54) 12 (46)
12 (46) 14 (54)
0.78
Donor Type, n (%) MRD MUD Other
12 (43) 15 (54) 1 (3)
21 (46) 19 (41) 6 (13)
0.36
11 (42) 14 (54) 1 (4)
13 (50) 9 (35) 4 (15)
0.19
ABO Compatibility, n (%) Compatible Major/bidirectional mismatch Minor Mismatch
18 (64) 7 (25) 3 (11)
36 (78) 7 (15) 3 (7)
0.47
16 (61) 7 (27) 3 (12)
21 (81) 3 (11) 2 (8)
0.30
GVHD Prophylaxis, n (%) MTX + CSA MTX + TAC Other regimens
9 (32) 16 (57) 3 (11)
23 (50) 17 (37) 6 (13)
0.22
8 (31) 15 (58) 3 (11)
14 (54) 8 (31) 4 (15)
0.17
CMV Status D+ or R+ D- and R-
23 (82) 5 (18)
37 (80) 9 (20)
1.0
21 (81) 5 (19)
21 (81) 5 (19)
1.0
Age at HCT, median (range)
Acute GVHD, n (%)
24 Page 24 of 31
20 (71) 8 (29)
26 (57) 8 (17)
0.22 0.38
18 (69) 6 (23)
15 (58) 3 (11)
0.56 0.46
Non-pulmonary chronic GVHD at BOS diagnosis None Mild Moderate Severe
1 (4) 9 (32) 10 (36) 8 (28)
7 (15) 13 (28) 19 (42) 7 (15)
0.29
1 (4) 8 (31) 9 (34) 8 (31)
4 (15) 5 (19) 11 (42) 6 (23)
0.45
Non-pulmonary chronic GVHD sites at BOS diagnosis, n (%) Skin (any stage) GI tract (any stage) Liver (any stage) Eyes (any stage)
23(82) 6 (21) 10 (36) 20 (86)
29 (63) 14 (30) 17 (37) 26 (54)
0.13 0.43 1.0 0.30
22 (85) 5 (19) 10 (38) 18 (69)
18 (69) 16 (62) 13 (50) 14 (54)
0.32 0.22 0.57 0.39
Overall grade II-IV Overall grade III-IV
Table 2 Table 2. Immunosuppressive Therapies (IST) in the Matched Cohort IST at the time of BOS Diagnosis
IST from BOS Diagnosis to ECP/Index Date
IST after ECP/Index Date
ECP (n = 26)
No ECP (n = 26)
p
ECP (n = 26)
No ECP (n = 26)
p
ECP (n = 26)
No ECP (n = 26)
p
Corticosteroids, n (%)
16 (61)
12 (46)
0.15
23 (88)
19 (73)
0.29
21 (81)
1.0
CSA/TAC, n (%)
14 (54)
12 (46)
0.78
15 (58)
11 (42)
0.40
14 (54)
0.39
1 (4)
3 (12)
0.60
6 (23)
7 (27)
1.0
11 (42)
0.57
1 (4) 0 (0) 0 (0) 0 (0) 1 (0-2)
2 (8) 0 (0) 1 (4) 0 (0) 1 (0-3)
0.31
3 (12) 1 (4) 1 (4) 2 (8) 2 (1-3)
4 (15) 0 (0) 1 (4) 2 (8) 1 (1-3)
0.16
21 (81) 18 (69) 14 (54) 5 (19) 5 (19) 4 (15) 3 (12) 2 (1-4)
5 (19) 3 (12) 3 (12) 3 (12) 2 (0-3)
0.20
3 (12)
0 (0)
0.23
21 (81)
18 (69)
0.52
Other IST Agents, n (%) MMF Sirolimus Rituximab Other agents
Number of IST agents, median (range)* FAM Protocol, n (%)
23 21 (81) (88) * Systemic immunosuppressive agents used concomitantly or sequentially during the specified time period.
0.70
25 Page 25 of 31
Abbreviations: CSA: Cyclosporine; MMF: Mycophenolate Mofetil; TAC: Tacrolimus
26 Page 26 of 31
Table 3 Table 3: PFT Data for the ECP and Non-ECP-Treated Groups before and after Propensity-Score Matching Unmatched Cohort
Matched Cohort
ECP (n = 28)
No ECP (n = 46)
p
ECP (n = 26)
No ECP (n = 26)
p
PFT data before HCT FEV1PP FEV1/FVC ratio DLCOPP
86 (64-109) 0.7 (0.6-0.9) 84 (55-106)
96 (68-124) 0.7 (0.6-0.9) 82 (55-116)
0.05 0.76 0.91
87 (64-109) 0.7 (0.6-0.9) 84 (55-106)
91 (68-110) 0.7 (0.6-0.9) 81 (55-116)
0.72 0.49 1.0
PFT data at BOS diagnosis FEV1PP FEV1/FVC ratio DLCOPP
56 (23-74) 0.6 (0.3-0.7) 63 (42-102)
63 (16-74) 0.6 (0.4-0.7) 66 (38-113)
0.22 0.40 0.24
56 (23-74) 0.7 (0.3-0.7) 63 (42-78)
54 (16-74) 0.6 (0.4-0.7) 67 (38-96)
0.87 0.37 0.18
PFT data at ECP/index date FEV1PP FEV1/FVC ratio DLCOPP
42 (20-79) 0.6 (0.3-0.8) 60 (43-84)
64 (14-94) 0.6 (0.3-0.9) 65 (20-113)
0.001 0.22 0.03
43 (23-79) 0.6 (0.3-0.8) 60 (46-84)
52 (14-94) 0.5 (0.3-0.9) 61 (20-96)
0.20 0.86 0.40
PFT data at last follow up FEV1PP FEV1/FVC ratio DLCOPP
40 (14-74) 0.5 (0.3-0.8) 53 (27-75)
54 (10-94) 0.6 (0.3-0.9) 62 (29-96)
0.007 0.17 0.05
43 (17-74) 0.5 (0.3-0.8) 48 (27-75)
46 (10-86) 0.5 (0.3-0.9) 62 (29-95)
0.43 0.74 0.08
-4.5 (-16 to
-3.1 (-15 to -
0.5) -0.3 (-7.5 to 1.3)
0.7)
-4.5 (-16 to 0.5)
-3.6 (-15 to -0.7)
Rate of decline in FEV1, percent predicted per month Before ECP/index date After ECP/index date
0.0 (-3.1 to 19)
1
0.83
1
0.33 -0.2 (-2.3 to 1.3)
-0.5 (-2.5 to 3)
All data presented as median (range) 1
Wilcoxon signed-rank test for comparison of paired data (before and after ECP/index date) between ECP and nonECP groups
27 Page 27 of 31
Table 4 Table 4: Clinical Outcomes and Adverse events in the Matched Cohort ECP (n = 26)
No ECP (n = 26)
p
34 (12-126)
44 (1-135)
0.63
Home oxygen at last follow up, n (%)
6 (23)
4 (15)
0.72
Non-pulmonary chronic GVHD at last follow-up None Mild Moderate Severe
5 (27) 7 (27) 7 (27) 7 (27)
4 (31) 8 (31) 5 (19) 9 (35)
0.84
22 (85)
20 (77)
0.72
13 (50) 7 (27) 6 (23) 10 (38) 8 (30)
9 (34) 3 (12) 14 (54) 8 (30) 4 (15)
0.06
Karnofsky Performance Score (KPS), mean ± SD At BOS diagnosis At last outpatient follow up
78 ± 11 76 ± 14
82 ± 6 73 ± 16
0.061
Respiratory infections after BOS diagnosis Bacterial2 Viral3 Fungal4
19 (73) 15 (58) 9 (35) 7 (27)
16 (65) 13 (50) 7 (27) 4 (15)
0.55 0.78 0.76 0.49
10 (55) [n = 18]
-
-
[n = 6] 4 (75) 0 (0) 1 (7.5) 1 (7.5)
[n = 13] 9 (69) 1 (8) 1 (8) 2 (15)
Follow-up months after BOS diagnosis, median (range)
Immunosuppression at last follow up Corticosteroids None Prednisone 1−5 mg daily Prednisone >5 mg daily Calcineurin inhibitors Other immunosuppressive agents
Central venous catheter complications5 Cause of death, n (%) Respiratory infection Other infections Disease relapse Other / unknown
0.77 0.32
1.0
1
Wilcoxon signed-rank test for comparison of paired data (at BOS diagnosis and at last outpatient follow-up) between the ECP and non-ECP groups 2
Clinical or microbiological diagnosis
3
Microbiological diagnosis via Influenza/RSV PCR (2005-2015) and/or Film Array Respiratory Panel (2010-2015)
4
Microbiological diagnosis
5
Central venous catheter malfunction/malposition (n = 7), infection (n = 4), and thrombosis (n = 1).
28 Page 28 of 31
Table 5
Table 5: Factors Associated with Overall Survival in the Matched Cohort Univariable Analysis HR (95% CI) p 1.9 (0.2-19) 0.54 0.6 (0.2-1.5) 0.29 0.3 (0.1-0.8) 0.02
Age at HCT2 MAC (vs. RIC) MRD (vs. MUD)3 ABO Compatibility Compatible Major/Bidirectional mismatch Minor CMV+ (vs. CMV-) aGVHD: overall grade II-IV (vs. 0-I) Non-pulmonary cGVHD: mod-severe (vs. mild-none) FEV1PP at ECP/index date
2
DLCOPP at ECP/index date2 Rate of decline in FEV1PP before ECP/index date ECP (vs. no ECP)
2
Multivariable Analysis1 HR (95% CI) p 0.1 (0.03-0.5) 0.002
Ref
Ref
-
-
1.0 (0.3-3.0) 2.0 (0.3-7.4)
0.89 0.31
-
-
2.8 (0.7-18) 0.5 (0.2-1.5)
0.12 0.26
-
-
0.7 (0.2-1.9) 0.9 (0.9-1.0)
0.4 0.40
-
-
0.9 (0.9-1.0)
0.15
-
-
0.9 (0.8-1.0)
0.24
0.3 (0.1-0.8)
0.01
0.7 (0.6-0.8) 0.1 (0.01-0.3)
0.001 0.003
1
All factors tested in the univariable analysis were entered into the multivariable model, and were retained in the final model using backward selection method with a cutoff significance level of p < 0.05 2
HR per unit change in regressor, as a continuous variable
3
HCT from cord or haploidentical donors were excluded due to small numbers (n = 6).
Abbreviations MAC: myeloablatrive conditioning; MRD: matched related donor; MUD: matched unrelated donor; RIC: reduced intensity conditioning
29 Page 29 of 31
Figure 1
30 Page 30 of 31
Figure 2
31 Page 31 of 31
S1083-8791(18)30193-9 https://doi.org/10.1016/j.bbmt.2018.04.012 YBBMT 55097
To appear in:
Biology of Blood and Marrow Transplantation
Received date: Accepted date:
31-1-2018 9-4-2018
Please cite this article as: Mehrdad Hefazi, Kimberly J. Langer, Nandita Khera, Jill Adamski, Vivek Roy, Jeffrey L. Winters, Dennis A. Gastineau, Eapen K. Jacob, Justin D. Kreuter, Manish J. Gandhi, William J. Hogan, Mark R. Litzow, Shahrukh K. Hashmi, Hemang Yadav, Vivek N. Iyer, J.P. Scott, Mark E. Wylam, Rodrigo Cartin-Ceba, Mrinal M. Patnaik, Extracorporeal Photopheresis Improves Survival in Hematopoietic Cell Transplant Patients with Bronchiolitis Obliterans Syndrome Without Significantly Impacting Measured Pulmonary Functions, Biology of Blood and Marrow Transplantation (2018), https://doi.org/10.1016/j.bbmt.2018.04.012. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
TITLE PAGE TITLE: Extracorporeal Photopheresis Improves Survival in Hematopoietic Cell Transplant Patients with Bronchiolitis Obliterans Syndrome without Significantly Impacting Measured Pulmonary Functions
SHORT TITE: ECP Improves Survival in HCT-Related BOS
AUTHORS: Mehrdad Hefazi1, Kimberly J. Langer1, Nandita Khera2, Jill Adamski3, Vivek Roy4, Jeffrey L. Winters5, Dennis A. Gastineau5, Eapen K. Jacob5, Justin D. Kreuter5, Manish J. Gandhi5, William J. Hogan1, Mark R. Litzow1, Shahrukh K. Hashmi1, Hemang Yadav6, Vivek N. Iyer6, J. P. Scott6, Mark E. Wylam6, Rodrigo Cartin-Ceba7, and Mrinal M. Patnaik1* 1
Division of Hematology, Mayo Clinic, Rochester, MN; 2Division of Hematology, Mayo Clinic,
Scottsdale, AZ, 3Department of Laboratory Medicine and Pathology, Mayo Clinic, Scottsdale, AZ; 4
Division of Hematology, Mayo Clinic, Jacksonville, FL; 5Division of Transfusion Medicine, Mayo Clinic,
Rochester, MN; 6Division of Pulmonary and Critical Care Medicine, Mayo Clinic, Rochester, MN; 7
Division of Pulmonary and Critical Care Medicine, Mayo Clinic, Scottsdale, AZ
*
CORRESPONDING AUTHOR:
Mrinal M. Patnaik, MBBS Mayo Clinic 200 First St SW, Rochester, MN 55905 Tel: 507-538-0591 Fax: 507-266-4972 [email protected] 1 Page 1 of 31
MANUSCRIPT DETAILS: Word count (body): 3995 Word count (abstract): 250 Tables: 5 Figures: 2 References: 30
Keywords: Hematopoietic Cell Transplantation Bronchiolitis Obliterans Syndrome Extracorporeal Photopheresis Graft-versus-Host Disease
HIGHLIGHTS:
ECP favorably impacts overall survival in HCT patients with BOS.
This survival benefit is independent of the ECP effect on measured pulmonary function.
Corticosteroid-sparing effect of ECP may be responsible for the improved survival.
2 Page 2 of 31
Abstract
We carried out the first matched retrospective cohort study aimed at studying the safety and efficacy of extracorporeal photopheresis (ECP) for bronchiolitis obliterans syndrome (BOS) following allogeneic hematopoietic cell transplantation (HCT). Medical records of 1325 consecutive adult patients who underwent HCT between 2005 and 2015 were reviewed. Seventyfour patients (median age, 51 years) with a diagnosis of BOS were included in the study. After propensity-score matching for BOS severity, 26 patients who underwent ≥3 months of ECP were matched to 26 non-ECP-treated patients, who were assigned an index date corresponding to the ECP start date for their matched pairs. The rate of decline in FEV1 percentage predicted (FEV1PP) decreased after ECP initiation (and after index date in the non-ECP group), with no significant difference between the two groups (p=0.33). On a multivariable analysis that included baseline transplant and pulmonary function test variables, matched related donor HCT (HR 0.1 [0.03-0.5]; p=0.002), ECP (HR 0.1 [0.01-0.3]; p=0.001) and slower rate of decline in FEV1PP before ECP/index date (HR 0.7 [0.6-0.8]; p=0.001) were associated with a better overall survival. At last follow-up, non-ECP-treated patients were more likely to be on >5 mg daily dose of prednisone (54% vs. 23%; p=0.04) and had a greater decline in their Karnofsky performance score (mean difference −9.5 vs. −1.6; p=0.06) compared to ECP-treated-patients. In conclusion, compared to other BOS-directed therapies, ECP was found to improve survival in HCT patients with BOS, without significantly impacting measured pulmonary functions. These findings need prospective validation in a larger patient cohort.
3 Page 3 of 31
Introduction
Bronchiolitis obliterans syndrome (BOS) is an increasingly recognized cause of morbidity and mortality in patients undergoing allogeneic hematopoietic cell transplantation (HCT). BOS affects 3–14% of patients, usually within the first 2 years of HCT1–3. BOS is intricately linked to graft-versus-host disease (GVHD) and often occurs following a reduction in immunosuppressive therapies (IST)4. In spite of augmented IST, prognosis in affected patients is dismal with a reported overall survival (OS) of 10–20% at 5 years5,6. Although combination ISTs, mainly steroids and calcineurin inhibitors (CNIs), have attempted to slow disease progression and prolong survival, these regimens are associated with significant risks and toxicities. Likewise, results of combination fluticasone, azithromycin, and montelukast (FAM) have shown conflicting results, with a recent clinical trial suggesting azithromycin may worsen OS when used pre-emptively for BOS prevention after HCT7.
Thus, due to suboptimal responses to medical therapies, extracorporeal photopheresis (ECP) has been used as an immunomodulatory alternative treatment for BOS. A few small cohort studies on BOS after lung transplantation have indicated that ECP may stabilize pulmonary function test (PFT) results and improve survival 7–9. However, to the best of our knowledge, there are only three published single-center retrospective studies on the utility of ECP for BOS after HCT10–12. In 2011, Lucid et al.10 reported the experience on nine patients with rapidly declining PFTs, who were treated with ECP for a maximum of 12 months.
4 Page 4 of 31
Approximately two-thirds of their patients showed PFT stabilization, and were able to taper corticosteroids. In the second report11, ECP was used in 13 chronic graft-versus-host (cGVHD) patients with BOS refractory to standard treatment. Seven of these patients had symptomatic BOS before starting ECP, whereas the remaining six patients developed BOS while on ECP treatment for steroid-refractory cGVHD. Preliminary results from this study supported the efficacy of ECP for slowing PFT decline, even in patients with advanced stages of BOS. Similarly, Brownback et al.12 demonstrated stability in % predicted FEV1 (FEV1PP) values in 7 of 8 BOS patients within three months after the initiation of ECP, a finding that was limited to only 2 of the 8 patients at the one year mark. They also noted a significant decrease in the median dose of prednisone per patient throughout the 12 months of ECP treatment.
Taken together, these retrospective studies, although encouraging, have been limited by small patient numbers, varying BOS severities, lack of control groups, and limited follow-up data on clinical outcomes. Therefore, we conducted this large multi-site matched retrospective cohort study to evaluate the safety and efficacy of ECP for the treatment of BOS after HCT.
Methods Patients and study design Medical records of 1,325 consecutive patients, ≥18 years old, who underwent HCT at the Mayo Clinic sites in Rochester, MN (n = 810; 2005–2015), Scottsdale, AZ (n = 397; 2011–2015), and Jacksonville, FL (n = 118; 2011–2015) were retrospectively reviewed. Patients were excluded if they died before day 100 after HCT (n = 172), did not have consistent PFT monitoring (pre5 Page 5 of 31
HCT, day 100, and yearly thereafter) (n = 118), had a baseline obstructive PFT pattern (n = 59), or underwent a second HCT (n = 17) (Figure 1). PFT data were interpreted according to the previously published guidelines for all 959 eligible patients.13,14 For patients with an obstructive, mixed, or non-specific PFT pattern after HCT, clinical notes, chest computed tomography, bronchoscopy, lung biopsy, and microbiology results were reviewed, if available. The study was approved by the Mayo Clinic Institutional Review Board.
BOS diagnosis BOS was diagnosed according to the NIH 2014 criteria15 with an additional requirement that the obstructive findings on PFT be present on two consecutive tests within 3 months, to ensure PFT abnormalities were not secondary to other reversible causes. For patients who met the PFT criteria but did not have other supporting features, such as distinctive cGVHD manifestations, air-trapping (by PFT or expiratory CT), or unequivocal ruling out of confounders, such as respiratory tract infections, a complete review of the post-transplant course was performed by the research team’s pulmonary and transplant physicians. Patients were excluded (n = 49) if their BOS diagnosis remained indeterminate.
BOS Treatments Patients treated with ≥3 months of ECP within 1 year of BOS diagnosis comprised the ECP group. Patients were excluded if they started ECP prior to (n = 2) or >12 months after (n = 9) BOS diagnosis. The UVAR XTS or the CELLEX system (Mallinckrodt Pharmaceuticals [formerly Therakos, Inc.], West Chester, PA) was used at all three sites to administer ECP according to previously described methods.8 Vascular access was obtained either through 6 Page 6 of 31
peripheral veins (n = 12), tunneled central venous catheters (n = 4), or implanted vascular access devices (n = 12). In three patients, vascular access had to be changed from peripheral venous to a central venous catheter during the course of treatment, due to issues with sustained peripheral venous access. Methoxypsoralen was administered at a dose of 0.35 µg per ml of collected buffy coat. The planned therapeutic course consisted of one treatment cycle on two consecutive days per week for 1 month, followed by one treatment cycle every other week for 2 months, followed by one treatment cycle every third week for 3 months, and followed by one treatment cycle monthly thereafter (Mayo Clinic standard operating protocol). Treatment frequency could be modified at the physician’s discretion. Non-ECP treatments primarily consisted of systemic steroids, CNIs (cyclosporine or tacrolimus), and other ISTs that were started at varying time points chosen by treating physicians.
Changes in FEV1 over time Each FEV1 was compared to the predicted normal value for that particular patient and a percentage predicted FEV1 (FEV1PP) was used for PFT analyses. Similar percentage predicted values were calculated for the FVC (FVCPP) and DLCO (DLCOPP). The rate of change in FEV1PP was calculated for each patient by using the slope of a linear regression line for FEV1PP vs time, as described previously6,8. This method was chosen because PFTs were performed at varying intervals in each patient. The rate of change in FEV1PP was calculated separately for the periods from BOS diagnosis to ECP initiation and for the subsequent 12 months. For non-ECP treated patients, these intervals were based on the establishment of an index date corresponding to the ECP start date for their matched pairs (see Statistical methods section). 7 Page 7 of 31
Outcome variables PFT data, including FEV1PP, FEV1/FVC ratio (%), and DLCOPP (after adjustment for hemoglobin) were compared between the ECP and non-ECP groups at baseline (pre-HCT), at BOS diagnosis, and at ECP/index data before and after propensity score matching (PSM). The rate of change in FEV1PP before and after ECP/index date was compared as paired variables. Acute and cGVHD were graded according to the Glucksberg15 and NIH 2014 criteria14, respectively. Since all patients were diagnosed with lung GVHD, only non-pulmonary manifestations were included in the overall grading of cGVHD. The Karnofsky performance score (KPS)16 at BOS diagnosis and at last outpatient follow-up were compared between the two groups by using matched paired data analysis. Respiratory tract infections were classified according to the implicated pathogens. Bacterial infections were diagnosed based on clinical and/or microbiological data, whereas viral and fungal infections were diagnosed solely based on microbiological tests.
Statistical analysis Since ECP users did not start ECP immediately after the diagnosis of BOS, there is an immortal time bias in favor of the ECP group. To account for this, the “prescription time distribution matching” (PTDM) method was used, as previously described17,18. Non-ECP users were assigned an index date corresponding to the length of time from BOS diagnosis to ECP start date for their matched ECP-treated pairs. Non-ECP users were excluded if they died before their assigned index date plus 90 days. The 90-day period was added to the index date because all ECP users had received ≥3 months of ECP, during which they were considered immortal. 8 Page 8 of 31
Propensity score matching was used to balance the distribution of potential confounders between the ECP and non-ECP groups. FEV1PP and DLCOPP at ECP/index time were included in the model because their distribution was significantly different between the two groups. A multivariable analysis was then used to account for other potential confounders and for possible residual imbalance in PFT variables after PSM. Propensity scores were computed by using the logistic regression model, and the matching tolerance was set at 0.2, as previously suggested 19. The chi-square (or Fisher’s exact) and Wilcoxon tests were used to compare categorical and continuous variables, respectively, between independent groups. The Wilcoxon signed-rank test was used for paired data analysis.
The Kaplan–Meier method was used to estimate OS from BOS diagnosis, and the log-rank test was used comparisons. A Cox proportional hazard model was used to analyze the association of ECP and PFT variables with OS in the entire and propensity score-matched cohorts. All variables tested in the univariable analysis were entered into a multivariable model, and were retained in the final model using backwards selection method with a significance level of p< 0.05. All tests were two-sided. SPSS software version 22 (SPSS, Inc., Chicago, Illinois) was used for PSM. All other analyses were performed by using JMP software version 10 (SAS Institute Inc., Cary, NC, USA).
Results
9 Page 9 of 31
Across the three Mayo Clinic sites, 88 patients were diagnosed with BOS between January 2005 and December 2015. Of these, 74 patients met the study’s eligibility criteria and were included in the study. Twenty-eight patients received ECP, and 46 were treated without ECP. After PSM, 26 ECP-treated patients were matched to 26 non-ECP-treated patients (Figure 1). Baseline characteristics and PFT data for the matched and unmatched cohorts are provided in tables 1 and 3, whereas data on treatments, clinical outcomes and adverse events are shown only for the PSM cohorts and described in more details in the following sections.
Baseline characteristics Demographics and baseline transplant characteristics for the matched ECP-treated and non-ECPtreated cohorts are shown in Table 1. There were no significant differences in patients’ age, year of HCT, transplant center, HCT comorbidity index, distribution of underlying disease, conditioning regimen, donor type, ABO compatibility, CMV status, and immunosuppression at HCT between the two groups. Severity grades of acute GVHD before BOS diagnosis and nonpulmonary cGVHD at the time of BOS diagnosis were comparable between the two groups.
BOS was diagnosed after a median of 14 months (range, 3-45 month) from HCT in the matched ECP group and after 13 months (range, 3-36 months) in the non-ECP group (p = 0.45). Median time from BOS diagnosis to ECP initiation was 1.5 months (range, 0–11 months). This was comparable to the time interval from BOS diagnosis to the assigned index dates for the non-ECP users (median, 1.5; range, 0–11 months), indicating proper matching based on the PTDM method.
10 Page 10 of 31
ECP and Non-ECP Treatments In the matched ECP group, 26 patients underwent a median of 31 cycles of ECP (range, 6–60 cycles) over a median of 15 months (range, 3–58 months). Of these, 7 (27%) patients received ≤20 cycles of ECP, 11 (42%) received 21−40 cycles of ECP, and the remaining 8 (31%) patients received ≥41 cycles of ECP. Immunosuppressive therapies at the time of BOS diagnosis, during the interval from BOS diagnosis to ECP/Index date, and following ECP/Index date are summarized in Table 2. At the time of BOS diagnosis, approximately half of the matched ECP and non-ECP-treated patients were on corticosteroids (61% vs. 46%; p = 0.40) and/or CNIs (54% vs. 46%; p = 78). The rate of corticosteroid use was comparable between the ECP and non-ECP treated groups during the interval from BOS diagnosis to ECP/Index date (88% vs. 73%; p = 0.29), and also after the ECP/Index date (81% vs. 81%; p = 1.0). Likewise, the rates of CNI use and treatment with other IST agents were comparable between the two groups during both periods (Table 2). The majority of patients (88% of the ECP-treated and 81% of non-ECP-treated patients) received combination therapy with fluticasone, azithromycin, and montelukast (FAM protocol)20 after ECP/Index date.
PFT data before and after BOS diagnosis PFT variables, including FEV1PP, FEV1/FVC ratio, and DLCOPP, were comparable between the matched ECP and non-ECP groups before HCT, at BOS diagnosis, and at ECP/index date (Table 3). FEV1PP values at last follow-up were comparable between the matched ECP and non-ECP groups (median 43% vs. 46 % predicted; p = 0.43), whereas there was a trend toward lower DLCOPP in the matched ECP group (48% vs. 62 % predicted; p = 0.08). The rate of decline in FEV1PP was faster before rather than after the ECP/index date in both the ECP (−4.5 vs −0.2; p < 11 Page 11 of 31
0.0001) and non-ECP groups (−3.6 vs −0.5; p < 0.001). Comparison between the two groups, however, did not reveal a significant difference (p = 0.33) (Table 3).
Clinical outcomes and adverse events With a median follow-up of 38 months (range, 1–135 months), OS was significantly better in the matched ECP group than in the non-ECP group (estimated median not reached vs 32 months; p = 0.01) (Figure 2). Six (23%) patients in the ECP group and 13 (50%) in the non-ECP group had expired at last follow-up. Respiratory tract infection was the leading cause of death for both ECP- and non-ECP-treated patients (75% vs 69%, p = 1.0). One (4%) patient in each group died from underlying disease relapse. Six (23%) patients in the ECP group and 4 (15%) in the nonECP group developed chronic respiratory failure requiring home oxygen therapy at last followup (Table 4).
In each group, 14 (50%) patients had moderate or severe non-pulmonary cGVHD at the last follow-up. Despite comparable severities of non-pulmonary cGVHD between the two groups, non-ECP-treated patients were almost twice as likely to be on >5 mg daily dose of prednisone at the last outpatient visit (54% vs 23%; p = 0.04). The use of CNIs and other ISTs was comparable between the two groups. ECP- and non-ECP-treated patients had comparable KPS at BOS diagnosis (mean, 78 vs 82; p = 0.34). However, non-ECP-treated patients had a greater decline in KPS over the course of treatment (mean difference, −9.5 vs. −1.6; p = 0.06).
12 Page 12 of 31
Univariable survival analysis in the matched cohort demonstrated better OS in HCT recipients from matched related donor (MRD) compared to matched unrelated donor (MUD) (HR 0.3 [0.10.8]; p = 0.02) and in patients treated with ECP (HR 0.3 [0.1-0.8]; p = 0.01) (Table 5). On multivariable analysis that included baseline transplant and PFT variables, HCT from MRD (HR 0.1 [0.03-0.5]; p=0.002), ECP (HR 0.1 [0.01-0.3]; p = 0.003) and slower rate of decline in FEV1PP before ECP/index date (HR 0.7 [0.6-0.8] per unit change; p = 0.001) were independently and favorably prognostic with regards to OS.
Respiratory tract infections due to bacterial, viral, and fungal pathogens were observed in 15 (58%), 9 (35%), and 7 (27%) patients in the ECP group, and in 13 (50%), 7 (27%), and 4 (15%) of the non-ECP-treated patients (p = 0.78, 0.76, and 0.49, respectively). Complications related to central venous access were observed in 2 (33%) of 6 patients with a tunneled central catheter and in 9 (75%) of 12 patients with an implanted vascular access device. These were most commonly due to central venous access malfunction/malposition requiring catheter or device exchange (n = 7), followed by central line associated blood stream infection (CLABSI) (n = 4), bleeding, and venous thrombosis (each in one patient). No cases of bacteremia were documented when ECP was administered through peripheral intravenous catheters.
Discussion We performed a large multi-site matched retrospective cohort study evaluating the safety and efficacy of ECP for patients with post-HCT BOS. Eligible patients met the NIH 2014 criteria for BOS and had a new obstructive PFT pattern on two consecutive tests within 3 months, prior to inclusion; so as to ensure that PFT abnormalities were not secondary to acute reversible causes.
13 Page 13 of 31
We excluded patients with an obstructive PFT pattern at baseline and those for whom confounding factors could not be excluded after review of the post-transplant course. The ECP group consisted of a relatively homogenous cohort of patients treated for ≥3 months with ECP within 1 year of BOS diagnosis, and the non- ECP group was selected by PSM.
In our study, FEV1PP values progressively declined from BOS diagnosis to last follow-up in the ECP- and non-ECP-treated cohorts. Although the rate of FEV1PP decline was faster before rather than after ECP initiation, a similar trend was also noted in the non-ECP-treated group; indicative of a comparable impact on PFT changes. Slowing of the rate of FEV1PP decline has been reported in HCT-related BOS after ECP therapy 7,8 and also in patients treated without ECP21; however, only one study has compared PFT results between ECP- and non-ECP-treated patients. In that particular post lung transplantation study, Pecoraro et al.9 reported that FEV1PP values at 12 months after BOS diagnosis were higher for the ECP-treated patients (n = 15) than for nonECP-treated patients (n = 39). The former group, however, had higher FEV1PP values at BOS diagnosis, which, in the absence of matched pair data analysis, may have confounded the results. Furthermore, there is a possibility that earlier initiation of ECP could be more effective in mitigating the decline in pulmonary functions, and in preventing the irreversible changes that ensue. Therefore, closer monitoring of PFTs after HCT, and dedicated prospective studies to establish the relationship between the timing of ECP and PFT changes are thus warranted.
There are currently no published reports on the impact of ECP on survival in BOS patients following HCT. In our study, ECP-treated patients had significantly better OS compared to non14 Page 14 of 31
ECP-treated patients (estimated median not reached vs 32 months). After PSM, the only independent prognostic factors for OS were ECP therapy, HCT from MRD, and the rate of decline in FEV1PP before ECP/index date. Rate of decline in FEV1 after ECP/index date did not appear to be different between the two groups. Hence, there may be extra-pulmonary aspects to ECP that are likely more important factors in survival.
To explore the reasons for survival difference between the ECP and non-ECP groups, we compared the patients’ KPS, ISTs, and non-pulmonary cGVHD status at BOS diagnosis and at last outpatient follow-up. ECP and non-ECP groups had comparable rates of moderate/severe non-pulmonary cGVHD at BOS diagnosis (65% for both groups) and at last follow-up (54% for both groups). Thus, the ECP effect on non-pulmonary cGVHD was unlikely to be responsible for the survival difference. While the use of CNI at last follow-up was comparable between the two groups, non-ECP-treated patients were more likely to be on a >5 mg daily prednisone dose and had a greater decline in KPS during treatment. This is consistent with the results of several prior studies 22–24 that have demonstrated corticosteroid-sparing effects of ECP and correlations between corticosteroid reduction and improved survival in patients with GVHD. Similarly, in a retrospective study of 77 BOS patients, Bergeron et al. demonstrated that intentional treatment with corticosteroids was associated with a shorter OS, without significantly affecting FEV1 values25. Another possible explanation for the greater decline in KPS in non-ECP-treated patients could be the higher toxicity from prolonged corticosteroid use. In our study, we chose a cut-off value of 5 mg/day prednisone dose as a measure of successful corticosteroid tapering, based on the current evidence that long-term prednisone use of ≥5 mg/day correlates with adverse events
15 Page 15 of 31
in a dose-dependent fashion26. It should, however, be noted that our study was not specifically designed to determine the effect of ECP on corticosteroid tapering.
One of the intriguing aspects of ECP is its ability to induce two opposing effects: activation of the immune system (as in treatment of cutaneous T cell lymphoma) and down-regulation of the activity of T lymphocytes (as in treatment of allograft rejection and GVHD)27,28. The immunomodulatory effects of ECP may be additionally advantageous in patients with hematologic malignancies and GVHD after HCT, because ECP does not cause generalized immunosuppression and thus, does not increase the risk of infectious complications and disease relapse29,30. Further studies on the immunological mechanisms of ECP are much needed.
Adverse ECP-related events were mainly due to central venous access complications (central line malfunction, infection, and thrombosis). The incidence of CLABSI (22%) in our study was slightly higher than previously reported (13%)8. All CLABSI cases were successfully treated with catheter removal and antibiotics, without long-term sequelae. The incidence of respiratory tract infections after BOS diagnosis was comparable between the ECP and non-ECP groups. However, only 4 (21%) of 19 episodes of respiratory infections in the ECP group resulted in mortality, versus 9 (56%) of 16 episodes in the non-ECP group (p = 0.04). Although this could be because of the superior safety profile and corticosteroid-sparing effect of ECP, confirmation of a causal relationship is difficult to ascertain in a retrospective study. The financial toxicity and the logistical challenges of ECP administration are other areas of concern, especially for BOS, where patients undergo prolonged courses of therapy (a median of 31 cycles in this study)8. The 16 Page 16 of 31
evaluation of the cost-effectiveness of ECP and its impact on patients’ quality of life are thus urgently required.
Limitations of our study include its retrospective non-randomized design and that the decision to start and subsequently taper ECP was at the clinician’s discretion; a process that can potentially bias outcomes. Despite available diagnostic criteria15, BOS remains difficult to diagnose, and there are no consistent response-monitoring guidelines. The heterogeneity of concomitant ISTs, varying duration of ECP, and missing data also complicate the analysis. We tried to address these limitations by excluding indeterminate cases of BOS, restricting the ECP group to patients treated with ≥3 months of ECP within one year of diagnosis, and comparing potential confounders at baseline and at last follow-up. While the use of specific classes of ISTs after BOS diagnosis were comparable between the ECP and non-ECP groups in our study, a complete matching with regard to the type and duration of non-ECP therapies was not possible. In our opinion, this is a difficult task to achieve even in the setting of a prospective study, given the fact that the current therapeutic approach to BOS is largely empirical and is associated with various treatment crossovers depending on each individual’s response. It should also be noted that our study was not designed to compare the efficacy of different non-ECP therapies such as FAM or other novel ISTs because we did not specifically match patients for these other therapeutic modalities.
We used the percentage predicted slope of FEV1PP to incorporate PFT results from varying time points into a single analysis, and therefore decrease the problem of missing values at pre-defined
17 Page 17 of 31
time points. The slope of FEV1 decline was calculated for the periods between BOS diagnosis and ECP/index date, and also for the first 12 months after BOS diagnosis. Although the calculated rate of FEV1 decline provides a crude estimate of the trajectory of pulmonary function decline over a specific time period, accurate measurement of the rate of FEV1 decline in each individual will require consistent PFT monitoring at specific time intervals that are yet to be defined and prospectively validated.
Conclusion
In conclusion, we found that ECP improves survival in HCT patients with BOS without significantly impacting pulmonary functions. After adjusting for BOS severity, OS was significantly better for patients treated with ECP for ≥3 months as compared to that in a propensity matched non-ECP treated cohort. The survival benefit was independent of the effects on PFT results and may have been related to the corticosteroid-sparing effect of ECP or other yet to be defined immunological effects. Prospective randomized trials are needed to confirm the efficacy of ECP, determine the optimal schedule, and standardize the response evaluation. Immunological investigations should also be incorporated into future clinical studies to explore biomarkers for use before or during treatment to identify patients who will benefit from ECP.
Acknowledgements
18 Page 18 of 31
The authors thank the staff at apheresis units and pulmonary labs at the Mayo Clinic in Rochester, MN, Scottsdale, AZ, and Jacksonville, FL for their technical and logistical support.
Authorship Contributions M.H. contributed to study conception, methodology, data collection, data analysis, and manuscript writing; K.J.L: contributed to data collection; N.K., J.A., V.R., J.L.W., D.A.G., E.K.J., M.R.L, and W.J.H. provided logistical and administrative support, and contributed to patient care and manuscript writing; J.D.K., M.J.G., S.K.H, H.Y., and V.N.I contributed to data interpretation and manuscript writing; J.P.S., M.E.W., and R.C.C, contributed to patient care, data interpretation, and manuscript writing; M.M.P: contributed to study conception, methodology, patients care, manuscript writing, and supervision. All authors contributed substantially to this work and have approved the submission of this manuscript.
Conflict of Interest Disclosure The authors declare no conflict of interest.
19 Page 19 of 31
References 1.
Au BKC, Au MA, Chien JW. Bronchiolitis obliterans syndrome epidemiology after allogeneic hematopoietic cell transplantation. Biol Blood Marrow Transplant. 2011;17(7):1072-1078. doi:10.1016/j.bbmt.2010.11.018
2.
Dudek AZ, Mahaseth H, DeFor TE, Weisdorf DJ. Bronchiolitis obliterans in chronic graft-versus-host disease: analysis of risk factors and treatment outcomes. Biol Blood Marrow Transplant. 2003;9(10):657-666. doi:10.1016/S1083
3.
Sakaida E, Nakaseko C, Harima A, et al. Late-onset noninfectious pulmonary complications after allogeneic stem cell transplantation are significantly associated with chronic graft-versus-host disease and with the graft-versus-leukemia effect. Blood. 2003;102(12):4236-4242. doi:10.1182/blood-2002-10-3289
4.
Gazourian L, Rogers AJ, Ibanga R, et al. Factors associated with bronchiolitis obliterans syndrome and chronic graft-versus-host disease after allogeneic hematopoietic cell transplantation. Am J Hematol. 2014;89(4):404-409. doi:10.1002/ajh.23656
5.
Williams KM, Chien JW, Gladwin MT, Pavletic SZ. Bronchiolitis obliterans after allogeneic hematopoietic stem cell transplantation. JAMA. 2009;302(3):306-314. doi:10.1001/jama.2009.1018
6.
Williams KM. How I Treat How I treat bronchiolitis obliterans syndrome after hematopoietic stem cell transplantation. Blood. 2017;129(4):448-456. doi:10.1182/blood2016-08-693507.448
7.
Jaksch P, Scheed A, Keplinger M, et al. A prospective interventional study on the use of extracorporeal photopheresis in patients with bronchiolitis obliterans syndrome after lung transplantation. J Hear Lung Transplant. 2012;31(9):950-957. doi:10.1016/J.HEALUN.2012.05.002
8.
Morrell MR, Despotis GJ, Lublin DM, Patterson GA, Trulock EP, Hachem RR. The efficacy of photopheresis for bronchiolitis obliterans syndrome after lung transplantation. J Hear Lung Transplant. 2010;29(4):424-431. doi:10.1016/j.healun.2009.08.029
20 Page 20 of 31
9.
Pecoraro Y, Carillo C, Diso D, et al. Efficacy of Extracorporeal Photopheresis in Patients With Bronchiolitis Obliterans Syndrome After Lung Transplantation. Transplant Proc. 2017;49(4):695-698. doi:10.1016/j.transproceed.2017.02.035
10.
Lucid CE, Savani BN, Engelhardt BG, et al. Extracorporeal photopheresis in patients with refractory bronchiolitis obliterans developing after allo-SCT. Bone Marrow Transplant. 2011;46(3):426-429. doi:10.1038/bmt.2010.152
11.
Del Fante C, Galasso T, Bernasconi P, et al. Extracorporeal photopheresis as a new supportive therapy for bronchiolitis obliterans syndrome after allogeneic stem cell transplantation. Bone Marrow Transplant. 2016:1-4. doi:10.1038/bmt.2015.324
12.
Brownback KR, Simpson SQ, Pitts LR, et al. Effect of extracorporeal photopheresis on lung function decline for severe bronchiolitis obliterans syndrome following allogeneic stem cell transplantation. J Clin Apher. 2016;31(4):347-352. doi:10.1002/jca.21404
13.
Pellegrino R, Viegi G, Brusasco V, et al. Interpretative strategies for lung function tests. Eur Respir J. 2005;26(5):948-968. doi:10.1183/09031936.05.00035205
14.
Iyer VN, Schroeder DR, Parker KO, Hyatt RE, Scanlon PD. The nonspecific pulmonary function test: Longitudinal follow-up and outcomes. Chest. 2011;139(4):878-886. doi:10.1378/chest.10-0804
15.
Jagasia MH, Greinix HT, Arora M, et al. National Institutes of Health Consensus Development Project on Criteria for Clinical Trials in Chronic Graft-versus-Host Disease: I. The 2014 Diagnosis and Staging Working Group report. Biol Blood Marrow Transplant. 2015;21(3):389-401.e1. doi:10.1016/j.bbmt.2014.12.001
16.
Karnofsky DA. Determining the extent of the cancer and clinical planning for cure. Cancer. 1968;22(4):730-734. http://www.ncbi.nlm.nih.gov/pubmed/5212297.
17.
Li G, Holbrook A, Delate T, Witt DM, Levine MA, Thabane L. Prediction of individual combined benefit and harm for patients with atrial fibrillation considering warfarin therapy: a study protocol. BMJ Open. 2015;5(11):e009518. doi:10.1136/bmjopen-2015009518
21 Page 21 of 31
18.
Zhou Z, Rahme E, Abrahamowicz M, Pilote L. Survival bias associated with time-totreatment initiation in drug effectiveness evaluation: A comparison of methods. Am J Epidemiol. 2005;162(10):1016-1023. doi:10.1093/aje/kwi307
19.
Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat. 2011. doi:10.1002/pst.433
20.
Williams KM, Cheng G-S, Pusic I, et al. Fluticasone, Azithromycin, and Montelukast Treatment for New-Onset Bronchiolitis Obliterans Syndrome after Hematopoietic Cell Transplantation. Biol Blood Marrow Transplant. 2016;22(4):710-716. doi:10.1016/j.bbmt.2015.10.009
21.
Cheng G-S, Storer B, Chien JW, et al. Lung Function Trajectory in Bronchiolitis Obliterans Syndrome after Allogeneic Hematopoietic Cell Transplant. Ann Am Thorac Soc. 2016;13(11):1932-1939. doi:10.1513/AnnalsATS.201604-262OC
22.
Couriel DR, Hosing C, Saliba R, et al. Extracorporeal photochemotherapy for the treatment of steroid-resistant chronic GVHD. Blood. 2006;107(8):3074-3080. doi:10.1182/blood-2005-09-3907
23.
Foss FM, DiVenuti GM, Chin K, et al. Prospective study of extracorporeal photopheresis in steroid-refractory or steroid-resistant extensive chronic graft-versus-host disease: analysis of response and survival incorporating prognostic factors. Bone Marrow Transplant. 2005;35(12):1187-1193. doi:10.1038/sj.bmt.1704984
24.
Greinix HT, Worel N, Just U, Knobler R. Extracorporeal photopheresis in acute and chronic graft-versus-host disease. Transfus Apher Sci. 2014;50(3):349-357. doi:10.1016/j.transci.2014.04.005
25.
Bergeron A, Godet C, Chevret S, et al. Bronchiolitis obliterans syndrome after allogeneic hematopoietic SCT: Phenotypes and prognosis. Bone Marrow Transplant. 2013;48(6):819-824. doi:10.1038/bmt.2012.241
26.
Saag KG, Koehnke R, Caldwell JR, et al. Low dose long-term corticosteroid therapy in
22 Page 22 of 31
rheumatoid arthritis: an analysis of serious adverse events. Am J Med. 1994;96(2):115123. http://www.ncbi.nlm.nih.gov/pubmed/8109596. 27.
Fimiani M, Di Renzo M, Rubegni P. Mechanism of action of extracorporeal photochemotherapy in chronic graft-versus-host disease. Br J Dermatol. 2004;150(6):1055-1060. doi:10.1111/j.1365-2133.2004.05918.x
28.
Voss CY, Fry TJ, Coppes MJ, Blajchman MA. Extending the Horizon for Cell-Based Immunotherapy by Understanding the Mechanisms of Action of Photopheresis. Transfus Med Rev. 2010;24(1):22-32. doi:10.1016/j.tmrv.2009.09.008
29.
Greinix HT, Socié G, Bacigalupo A, et al. Assessing the potential role of photopheresis in hematopoietic stem cell transplant. Bone Marrow Transplant. 2006;38(4):265-273. doi:10.1038/sj.bmt.1705440
30.
Knobler R, Barr ML, Couriel DR, et al. Extracorporeal photopheresis: past, present, and future. J Am Acad Dermatol. 2009;61(4):652-665. doi:10.1016/j.jaad.2009.02.039
Figure 1: Flow diagram of the Study Cohort
Figure 2: Kaplan-Meier Survival Curves for the Matched ECP and Non-ECP-Treated Patients
23 Page 23 of 31
Table 1 Table 1: Demographics and Transplant Characteristics of the ECP and Non-ECP-Treated Groups before and after Propensity-Score Matching Unmatched Cohort ECP No ECP (n = 28) (n = 46)
p
ECP (n = 26)
Matched Cohort No ECP (n = 26)
p
Year of HCT 2005-2010 2011-2015
10 (36) 18 (64)
16 (35) 30 (65)
1.0
9 (35) 17 (65)
8 (31) 18 (69)
1.0
Transplant Center Rochester, MN Scottsdale, AZ Jacksonville, FL
20 (71) 6 (21) 2 (8)
37 (80) 5 (11) 4 (9)
0.48
19 (72) 5 (20) 2 (8)
19 (72) 4 (15) 3 (13)
0.91
46 (21-67)
51 (20-76)
0.64
50 (21-67)
51 (21-76)
0.76
Male, n (%)
11 (39)
26 (56)
0.23
10 (38)
14 (54)
0.40
HCT-CI, n 0-2 ≥3
24 (86) 4 (14)
38 (83) 8 (17)
0.76
22 (85) 4 (15)
23 (88) 3 (12)
0.99
Underlying Diagnosis, n (%) AML ALL MDS CML Other
10 (36) 8 (28) 4 (14) 4 (14) 2 (7)
23 (50) 8 (17) 6 (13) 2 (5) 7 (15)
0.33 0.38 1.0 0.19 0.46
9 (34) 8 (31) 4 (15) 3 (12) 2 (8)
14 (54) 6 (23) 1 (4) 1 (4) 4 (15)
0.26 0.75 0.34 0.60 0.66
Conditioning Regimen, n (%) MAC RIC
16 (57) 12 (43)
21 (46) 25 (54)
0.47
14 (54) 12 (46)
12 (46) 14 (54)
0.78
Donor Type, n (%) MRD MUD Other
12 (43) 15 (54) 1 (3)
21 (46) 19 (41) 6 (13)
0.36
11 (42) 14 (54) 1 (4)
13 (50) 9 (35) 4 (15)
0.19
ABO Compatibility, n (%) Compatible Major/bidirectional mismatch Minor Mismatch
18 (64) 7 (25) 3 (11)
36 (78) 7 (15) 3 (7)
0.47
16 (61) 7 (27) 3 (12)
21 (81) 3 (11) 2 (8)
0.30
GVHD Prophylaxis, n (%) MTX + CSA MTX + TAC Other regimens
9 (32) 16 (57) 3 (11)
23 (50) 17 (37) 6 (13)
0.22
8 (31) 15 (58) 3 (11)
14 (54) 8 (31) 4 (15)
0.17
CMV Status D+ or R+ D- and R-
23 (82) 5 (18)
37 (80) 9 (20)
1.0
21 (81) 5 (19)
21 (81) 5 (19)
1.0
Age at HCT, median (range)
Acute GVHD, n (%)
24 Page 24 of 31
20 (71) 8 (29)
26 (57) 8 (17)
0.22 0.38
18 (69) 6 (23)
15 (58) 3 (11)
0.56 0.46
Non-pulmonary chronic GVHD at BOS diagnosis None Mild Moderate Severe
1 (4) 9 (32) 10 (36) 8 (28)
7 (15) 13 (28) 19 (42) 7 (15)
0.29
1 (4) 8 (31) 9 (34) 8 (31)
4 (15) 5 (19) 11 (42) 6 (23)
0.45
Non-pulmonary chronic GVHD sites at BOS diagnosis, n (%) Skin (any stage) GI tract (any stage) Liver (any stage) Eyes (any stage)
23(82) 6 (21) 10 (36) 20 (86)
29 (63) 14 (30) 17 (37) 26 (54)
0.13 0.43 1.0 0.30
22 (85) 5 (19) 10 (38) 18 (69)
18 (69) 16 (62) 13 (50) 14 (54)
0.32 0.22 0.57 0.39
Overall grade II-IV Overall grade III-IV
Table 2 Table 2. Immunosuppressive Therapies (IST) in the Matched Cohort IST at the time of BOS Diagnosis
IST from BOS Diagnosis to ECP/Index Date
IST after ECP/Index Date
ECP (n = 26)
No ECP (n = 26)
p
ECP (n = 26)
No ECP (n = 26)
p
ECP (n = 26)
No ECP (n = 26)
p
Corticosteroids, n (%)
16 (61)
12 (46)
0.15
23 (88)
19 (73)
0.29
21 (81)
1.0
CSA/TAC, n (%)
14 (54)
12 (46)
0.78
15 (58)
11 (42)
0.40
14 (54)
0.39
1 (4)
3 (12)
0.60
6 (23)
7 (27)
1.0
11 (42)
0.57
1 (4) 0 (0) 0 (0) 0 (0) 1 (0-2)
2 (8) 0 (0) 1 (4) 0 (0) 1 (0-3)
0.31
3 (12) 1 (4) 1 (4) 2 (8) 2 (1-3)
4 (15) 0 (0) 1 (4) 2 (8) 1 (1-3)
0.16
21 (81) 18 (69) 14 (54) 5 (19) 5 (19) 4 (15) 3 (12) 2 (1-4)
5 (19) 3 (12) 3 (12) 3 (12) 2 (0-3)
0.20
3 (12)
0 (0)
0.23
21 (81)
18 (69)
0.52
Other IST Agents, n (%) MMF Sirolimus Rituximab Other agents
Number of IST agents, median (range)* FAM Protocol, n (%)
23 21 (81) (88) * Systemic immunosuppressive agents used concomitantly or sequentially during the specified time period.
0.70
25 Page 25 of 31
Abbreviations: CSA: Cyclosporine; MMF: Mycophenolate Mofetil; TAC: Tacrolimus
26 Page 26 of 31
Table 3 Table 3: PFT Data for the ECP and Non-ECP-Treated Groups before and after Propensity-Score Matching Unmatched Cohort
Matched Cohort
ECP (n = 28)
No ECP (n = 46)
p
ECP (n = 26)
No ECP (n = 26)
p
PFT data before HCT FEV1PP FEV1/FVC ratio DLCOPP
86 (64-109) 0.7 (0.6-0.9) 84 (55-106)
96 (68-124) 0.7 (0.6-0.9) 82 (55-116)
0.05 0.76 0.91
87 (64-109) 0.7 (0.6-0.9) 84 (55-106)
91 (68-110) 0.7 (0.6-0.9) 81 (55-116)
0.72 0.49 1.0
PFT data at BOS diagnosis FEV1PP FEV1/FVC ratio DLCOPP
56 (23-74) 0.6 (0.3-0.7) 63 (42-102)
63 (16-74) 0.6 (0.4-0.7) 66 (38-113)
0.22 0.40 0.24
56 (23-74) 0.7 (0.3-0.7) 63 (42-78)
54 (16-74) 0.6 (0.4-0.7) 67 (38-96)
0.87 0.37 0.18
PFT data at ECP/index date FEV1PP FEV1/FVC ratio DLCOPP
42 (20-79) 0.6 (0.3-0.8) 60 (43-84)
64 (14-94) 0.6 (0.3-0.9) 65 (20-113)
0.001 0.22 0.03
43 (23-79) 0.6 (0.3-0.8) 60 (46-84)
52 (14-94) 0.5 (0.3-0.9) 61 (20-96)
0.20 0.86 0.40
PFT data at last follow up FEV1PP FEV1/FVC ratio DLCOPP
40 (14-74) 0.5 (0.3-0.8) 53 (27-75)
54 (10-94) 0.6 (0.3-0.9) 62 (29-96)
0.007 0.17 0.05
43 (17-74) 0.5 (0.3-0.8) 48 (27-75)
46 (10-86) 0.5 (0.3-0.9) 62 (29-95)
0.43 0.74 0.08
-4.5 (-16 to
-3.1 (-15 to -
0.5) -0.3 (-7.5 to 1.3)
0.7)
-4.5 (-16 to 0.5)
-3.6 (-15 to -0.7)
Rate of decline in FEV1, percent predicted per month Before ECP/index date After ECP/index date
0.0 (-3.1 to 19)
1
0.83
1
0.33 -0.2 (-2.3 to 1.3)
-0.5 (-2.5 to 3)
All data presented as median (range) 1
Wilcoxon signed-rank test for comparison of paired data (before and after ECP/index date) between ECP and nonECP groups
27 Page 27 of 31
Table 4 Table 4: Clinical Outcomes and Adverse events in the Matched Cohort ECP (n = 26)
No ECP (n = 26)
p
34 (12-126)
44 (1-135)
0.63
Home oxygen at last follow up, n (%)
6 (23)
4 (15)
0.72
Non-pulmonary chronic GVHD at last follow-up None Mild Moderate Severe
5 (27) 7 (27) 7 (27) 7 (27)
4 (31) 8 (31) 5 (19) 9 (35)
0.84
22 (85)
20 (77)
0.72
13 (50) 7 (27) 6 (23) 10 (38) 8 (30)
9 (34) 3 (12) 14 (54) 8 (30) 4 (15)
0.06
Karnofsky Performance Score (KPS), mean ± SD At BOS diagnosis At last outpatient follow up
78 ± 11 76 ± 14
82 ± 6 73 ± 16
0.061
Respiratory infections after BOS diagnosis Bacterial2 Viral3 Fungal4
19 (73) 15 (58) 9 (35) 7 (27)
16 (65) 13 (50) 7 (27) 4 (15)
0.55 0.78 0.76 0.49
10 (55) [n = 18]
-
-
[n = 6] 4 (75) 0 (0) 1 (7.5) 1 (7.5)
[n = 13] 9 (69) 1 (8) 1 (8) 2 (15)
Follow-up months after BOS diagnosis, median (range)
Immunosuppression at last follow up Corticosteroids None Prednisone 1−5 mg daily Prednisone >5 mg daily Calcineurin inhibitors Other immunosuppressive agents
Central venous catheter complications5 Cause of death, n (%) Respiratory infection Other infections Disease relapse Other / unknown
0.77 0.32
1.0
1
Wilcoxon signed-rank test for comparison of paired data (at BOS diagnosis and at last outpatient follow-up) between the ECP and non-ECP groups 2
Clinical or microbiological diagnosis
3
Microbiological diagnosis via Influenza/RSV PCR (2005-2015) and/or Film Array Respiratory Panel (2010-2015)
4
Microbiological diagnosis
5
Central venous catheter malfunction/malposition (n = 7), infection (n = 4), and thrombosis (n = 1).
28 Page 28 of 31
Table 5
Table 5: Factors Associated with Overall Survival in the Matched Cohort Univariable Analysis HR (95% CI) p 1.9 (0.2-19) 0.54 0.6 (0.2-1.5) 0.29 0.3 (0.1-0.8) 0.02
Age at HCT2 MAC (vs. RIC) MRD (vs. MUD)3 ABO Compatibility Compatible Major/Bidirectional mismatch Minor CMV+ (vs. CMV-) aGVHD: overall grade II-IV (vs. 0-I) Non-pulmonary cGVHD: mod-severe (vs. mild-none) FEV1PP at ECP/index date
2
DLCOPP at ECP/index date2 Rate of decline in FEV1PP before ECP/index date ECP (vs. no ECP)
2
Multivariable Analysis1 HR (95% CI) p 0.1 (0.03-0.5) 0.002
Ref
Ref
-
-
1.0 (0.3-3.0) 2.0 (0.3-7.4)
0.89 0.31
-
-
2.8 (0.7-18) 0.5 (0.2-1.5)
0.12 0.26
-
-
0.7 (0.2-1.9) 0.9 (0.9-1.0)
0.4 0.40
-
-
0.9 (0.9-1.0)
0.15
-
-
0.9 (0.8-1.0)
0.24
0.3 (0.1-0.8)
0.01
0.7 (0.6-0.8) 0.1 (0.01-0.3)
0.001 0.003
1
All factors tested in the univariable analysis were entered into the multivariable model, and were retained in the final model using backward selection method with a cutoff significance level of p < 0.05 2
HR per unit change in regressor, as a continuous variable
3
HCT from cord or haploidentical donors were excluded due to small numbers (n = 6).
Abbreviations MAC: myeloablatrive conditioning; MRD: matched related donor; MUD: matched unrelated donor; RIC: reduced intensity conditioning
29 Page 29 of 31
Figure 1
30 Page 30 of 31
Figure 2
31 Page 31 of 31