Bone Marrow Transplantation (2015) 50, 1217–1223 © 2015 Macmillan Publishers Limited All rights reserved 0268-3369/15 www.nature.com/bmt
ORIGINAL ARTICLE
Increased incidence of chronic GvHD and CMV disease in patients with vitamin D deficiency before allogeneic stem cell transplantation L von Bahr1,2, O Blennow3, J Alm2, A Björklund1,4, K-J Malmberg1,4,5,6, D Mougiakakos7, A Le Blanc1, PJ Oefner8, M Labopin9, P Ljungman1 and K Le Blanc1,2 Vitamin D has emerged as a central player in the immune system, with its deficiency being implicated in the pathogenesis of several autoimmune diseases, including chronic GvHD. This is a retrospective cohort analysis of 166 patients, who underwent allogeneic hematopoietic stem cell transplantation (HSCT) at the Karolinska University Hospital, evaluating GvHD, graft failure, infectious complications and survival after HSCT in relation to pre-transplantation vitamin D levels. Most of the patients were deficient in vitamin D before HSCT (median 42 nmol/L). In multivariate analysis, vitamin D level before HSCT was identified as a significant independent risk factor for development of cGvHD. The increased incidence of cGvHD was not coupled to better disease-free survival; instead there was a trend towards lower overall survival in the vitamin D-deficient patients. In addition, we found a significant correlation between vitamin D deficiency and incidence of CMV disease, with no case of CMV disease occurring in patients with sufficient levels of vitamin D before HSCT. Our results support a role of vitamin D in immune tolerance following HSCT. These findings could be highly relevant for the care of HSCT patients, and prospective, randomized studies on the effect of vitamin D supplementation are therefore needed. Bone Marrow Transplantation (2015) 50, 1217–1223; doi:10.1038/bmt.2015.123; published online 1 June 2015
INTRODUCTION Allogeneic hematopoietic stem cell transplantation (HSCT) is an important and potentially curative treatment of haematological malignancies, but GvHD and infections affect outcomes. Vitamin D has emerged as a central player in the immune system,1 affecting T and B cells, macrophages and dendritic cells (DCs).2–4 A vitamin D-enriched milieu maintains an immature DC phenotype associated with decreased ability to stimulate alloreactive T cells in MLCs.5 Vitamin D deficiency has been associated with an increased incidence of chronic GvHD,6 and supplementation with vitamin D to induce a tolerogenic DC population has been suggested for prevention of GvHD.7 One mechanism for this maintenance of an immature DC population has been shown to be the upregulation of indoleamine 2,3-dioxygenase (IDO),8 an enzyme converting tryptophan to kynurenine and a central feature in tolerizing DCs.9 Contrary to the suppressive effects on the immune system, vitamin D also has a protective effect against infections. This was first shown in tuberculosis, where the traditional treatments with sunlight and cod liver oil, rich in vitamin D, are viewed in a new light after the discovery of antimicrobial peptides induced by vitamin D. The two antimicrobial peptides under the influence of vitamin D are LL-37 (cathelicidin) and β-defensin 2,10,11 which
have activity against several bacteria, as well as viruses12 and fungi.13 We present here what we believe is currently the largest epidemiological study on adult HSCT patients examining clinical outcome with regards to vitamin D deficiency before HSCT, including both infectious and non-infectious complications. PATIENTS AND METHODS Patients The study was conducted in accordance with the Helsinki Declaration upon approval by the regional ethics committee in Stockholm. Serum specimens were collected from 166 consecutive patients (412 years of age) undergoing HSCT between 2005 and 2011 at the Karolinska University Hospital, Huddinge, (Table 1). Serum samples from 20 healthy donors were used as controls. All data were taken from the patient's medical records. A total of 139 patients were eligible for evaluation of chronic GvHD (cGvHD) excluding patients with graft failure (n = 13) or with a survival after HSCT of o100 days (n = 14). For analysis of infectious complications, only patients with a follow-up at the Karolinska University Hospital were selected (n = 137). When calculating disease-free survival (DFS), only patients with haematological malignancies (n = 156) were considered. For IDO analysis, 24 patients with vitamin D levels above or below 50 nmol/L were selected; 14 with cGvHD and 10 without cGvHD (Table 2). They were selected based on the following criteria: clear presence or
1 Department of Medicine, Hematology Centre, Karolinska University Hospital, Stockholm, Sweden; 2Center for Hematology and Regenerative Medicine, Division of Clinical Immunology and Transfusion Medicine, Department of Laboratory Medicine, Karolinska Institutet, Stockholm, Sweden; 3Division of Infectious Diseases, Department of Medicine, Karolinska Institutet, Stockholm, Sweden; 4Department of Medicine, Center for Infectious Medicine, Karolinska Institutet, Stockholm, Sweden; 5Department of Cancer Immunology, Institute for Cancer Research, Oslo University Hospital, Oslo, Norway; 6The KG Jebsen Center for Cancer Immunotherapy, Institute of Clinical Medicine, University of Oslo, Oslo, Norway; 7Department of Internal Medicine 5, Hematology and Oncology, University of Erlangen-Nuremberg, Erlangen, Germany; 8Institute of Functional Genomics, University of Regensburg, Regensburg, Germany and 9Service d'Hématologie et Thérapie Cellulaire, Hôpital Saint Antoine, Paris, France. Correspondence: Dr L von Bahr, Department of Laboratory Medicine, Karolinska Institutet, Hematology Centre R51, 141 86 Stockholm, Sweden. E-mail: [email protected] Received 14 February 2015; revised 14 April 2015; accepted 18 April 2015; published online 1 June 2015
Vitamin D deficiency and complications after HSCT L von Bahr et al
1218 Table 1.
Patient characteristics o25 nmol/L deficient N = 19 (11%) N
Age o40 years ⩾ 40 years Median (range) Vitamin D supplementation 1st year post transplantation
%
9 10
25–49 nmol/L insufficient N = 88 (53%) N
%
⩾ 50 nmol/L sufficient N = 59 (36%) N
P-value
%
47% 53% 36 (14–65) 4 21%
20 68
23% 77% 53 (14–67) 6 7%
17 42
29% 71% 53 (12–68) 8 14%
Sex Female Male
4 15
21% 79%
41 47
47% 53%
27 32
46% 54%
Donora Donor age, median (range) Matched related Matched unrelated o8/8 HLA matchb
8 7 2
29 (3–50) 47% 41% 12%
21 36 25
34 (18–64) 26% 44% 30%
11 39 7
27 (10–63) 19% 68% 12%
0.004* 0.9
0.18
0.14 0.007*
Sex matchc F to M Other
4 142
22% 78%
11 762
13% 87%
3 56
5% 95%
Diagnosis AML ALL CML MDS MPD/MF Myeloma Lymphoma/CLL PNH Solid tumour Non-malignant
7 2 1 3 1 2 1 — — 2
37% 11% 5% 19% 5% 11% 5% — — 11%
36 7 2 11 6 1 23 — 1 1
41% 8% 2% 13% 7% 1% 26% — 1% 1%
19 4 4 11 2 1 10 1 2 5
32% 7% 7% 19% 3% 2% 17% 2% 3% 8%
Stem cell source PBSC Bone marrow Cord blood
13 4 2
68% 21% 11%
80 4 4
91% 5% 5%
52 7 —
88% 12% —
0.02*
Conditioning RIC MAC
7 12
37% 63%
54 34
61% 39%
37 22
63% 37%
0.1
Anti-T cell No ATG Alemtuzumab Cyclophosphamide
7 12 — —
39% 61% — —
21 60 7 —
24% 68% 8% —
8 43 6 2
14% 73% 10% 3%
0.2
0.03*
Abbreviations: ATG = anti-thymocyte globulin (rabbit); F = female; HSCT = hematopoietic stem cell transplantation; M = male; MAC = myeloablative conditioning; MDS = myelodysplastic syndrome; MF = myelofibrosis; MPD = myeloproliferative disorder; RIC = reduced intensity conditioning. aCord blood transplants excluded. bMainly C mismatch. No haplo-transplants. Three unknown. cOne unknown. P-values are calculated using Χ2-test for trend or independent-samples median test. Table of all patients included in the study, in cohorts by vitamin D level before transplantation. *P-valueo 0.05.
absence of classical chronic GvHD (grade moderate to severe), no donor lymphocyte infusions or stem cell booster and no photopheresis. Follow-up ranged between 36 and 107 months after HSCT, with a median follow-up of 71 months.
Definitions For cGvHD diagnosis and scoring, the NIH consensus criteria were used14 and both classic chronic GvHD and overlap syndrome were included, but not late-onset acute GvHD. Only cGvHD of moderate and severe grades was considered in the analysis. Graft failure was defined as the lack of engraftment, engraftment with recipient cells or later developing full (495%) recipient chimerism in the absence of relapse of the underlying disease. DFS was defined as survival Bone Marrow Transplantation (2015) 1217 – 1223
with no evidence of relapse or progression of malignant disease. Overall survival (OS) was defined as the time from HSCT to death, regardless of the cause. We evaluated both agent-specific and non-agent-specific infectious complications during the first year after HSCT, censoring the follow-up at relapse of the malignant disease. Agent-specific outcomes were CMV disease, EBV-associated post-transplantation lymphoproliferative disorder, influenza, invasive fungal disease and bacteraemia. Non-agent-specific outcomes were pneumonia and days on IV antibiotics. Bacteraemia and days on IV antibiotics were analysed separately for neutropenic and non-neutropenic episodes. CMV disease was defined according to Ljungman et al.15 Invasive fungal disease was defined according to De Pauw.16 Only probable and proven © 2015 Macmillan Publishers Limited
Vitamin D deficiency and complications after HSCT L von Bahr et al
1219 Table 2.
Patients included in IDO analysis Chronic GvHD
Total Age Median (range) Sex Donor Donor age median (range) Stem cell source Conditioning Anti-T cell Time to cGvHD grade moderate/severe median weeks (range) 25-OH-D3 before transplantation, median (range
No chronic GvHD
Low vitamin D ( o50 nmol/L)
High vitamin D (450 nmol/L)
Low vitamin D (o50 nmol/L)
High vitamin D (450 nmol/L)
7
7
5
5
34 (15–55) 6 M, 1 F 2 MUD, 5 SIB 36 (17–50) 7 PBSC 6 MAC, 1 RIC 2 ATG, 5 none 29 (17–94)
53 (37–65) 3 M, 4 F 3 MUD, 4 SIB 49 (20–63) 7 PBSC 3 MAC, 4 RIC 4 ATG, 3 none 37 (17–87)
47 (25–59) 3 M, 2 F 5 MUD 23 (19–25) 5 PBSC 4 MAC, 2 RIC 5 ATG —
46 (22–55) 2 M, 3 F 5 MUD 26 (18–36) 5 PBSC 3 MAC, 2 RIC 5 ATG —
20 (15–43)
59 (51–66)
30 (26–46)
69 (57–64)
25-OH-cholecalciferol, nmol/l
Abbreviations: ATG = anti-thymocyte globulin (rabbit), F = female; M = male; MAC = myeloablative conditioning; MUD = matched unrelated donor; RIC = reduced intensity conditioning. Table of 24 patients selected for IDO analysis.
125 100 75 50
Insufficiency level
25
Deficiency level
no
rs
52 H
ea
lth y
do
w
26 w
w
0
0
to current routine standards at the hospital laboratory. In brief, blood samples collected in serum tubes were left for 30–120 min at room temperature, or up to 12 h at +8 °C, before serum separation by centrifugation. Specimens were stored in liquid nitrogen until analysis. Serum vitamin D was measured as 25-OH-cholecalciferol and expressed as nmol/L, hereafter referred to as 25-OH-D3. Analyses were performed by the laboratory for clinical chemistry (Karolinska University Hospital, Solna) using a chemiluminescence method (CLIA) approved by SWEDAC. Vitamin D deficiency was defined as o25 nmol/L, insufficiency as 26–49 nmol/L and sufficiency as ⩾ 50 nmol/L, in accordance with the Nordic Guidelines.19 Serum levels of tryptophan metabolites were analysed before and at 1, 2, 3, 6, 12 and 24 months post HSCT by liquid chromatography-tandem mass spectrometry.20 IDO activity was estimated by calculating the serum kynurenine-to-tryptophan ratio. A median of six out of seven time points per patient (range 2–7) were available for analysis.
Figure 1. Alterations of 25-OH-D3 levels in HSCT patients. Serum levels of 25-OH-D3 (median with interquartile range) in HSCT patients before conditioning (w0), and 26 and 52 weeks, respectively, after transplantation, as well as in healthy donors.
infections were considered in the analysis. Diagnosis of pneumonia required either a combination of new pulmonary infiltrates on chest X-ray or CT scan with symptoms of respiratory infection such as cough, dyspnoea or fever, excluding idiopathic pulmonary syndrome or autopsyverified infectious pneumonia. Bacteraemia was defined as the first positive blood culture during a 10-day time period. Repeated positive blood cultures 410 days after the first were considered new episodes.
Prophylaxis and surveillance GvHD prophylaxis consisted of cyclosporin A in combination with methotrexate or prednisone, or tacrolimus in combination with sirolimus. Patients with an unrelated donor were given anti-T-cell therapy, most commonly ATG (see Table 1 for details). Patients with a related donor did not receive anti-T-cell therapy. Patients were routinely monitored for CMV reactivation with PCR and pre-emptively treated as described previously.17 Antifungal, antibacterial and Pneumocystis jiroveci prophylaxis was given as previously described.18
Serum analysis Vitamin D levels were measured in serum specimens collected before the start of conditioning therapy and in 85 of the patients also at 6 and/or 12 months after HSCT. Specimens were collected and handled according © 2015 Macmillan Publishers Limited
Statistical analysis This is a retrospective cohort study. Data were analysed as of date of the last data collection, 14 April 2014. Sample size was based on a power analysis using data from a pilot study (unpublished). Primary end point was chronic GvHD; secondary end points were OS, DFS, graft failure and incidence of infectious complications. Cumulative incidence functions were used to estimate GvHD, considering death and relapse to be competing events. Probabilities of DFS and OS were calculated using the Kaplan–Meier estimates. Univariate analyses were performed using Gray’s test for cumulative incidence functions and the log-rank test for DFS and OS. Associations of patient and graft characteristics with outcomes were evaluated in multivariate analysis, using Cox proportional hazards model for dichotomous variables or negative binomial regression analysis for outcomes with repeated events or continuous variables. For two-sample comparisons, the Wilcoxon rank-sum test or Fisher’s exact test was used. GraphPad Prism 6 (GraphPad software, San Diego, CA, USA), IBM SPSS version 21 (SPSS Inc./IBM, Armonk, NY, USA) and R 3.0.1 software were used.
RESULTS Serum levels of 25-OH-D3 in HSCT patients The median level of 25-OH-D3 for all 166 patients before transplantation was 42 nmol/L (range 10–118), below the deficiency level of 50 nmol/L (20 ng/mL)21 and significantly lower than in the healthy controls (median 66.5 nmol/L, range 21–104, P o0.001). Only 59 patients (36%) had sufficient levels of 25-OHD3 (⩾50 nmol/L) before HSCT, whereas the majority (N = 88, 53%) were 25-OH-D3 insufficient (26–49 nmol/L) and 19 (11%) were even deficient (o 25 nmol/L). The baseline characteristics of the Bone Marrow Transplantation (2015) 1217 – 1223
Vitamin D deficiency and complications after HSCT L von Bahr et al
1220
Survival and relapse The 2-year OS was 63% in patients with vitamin D deficiency (o 25 nmol/L), 69% in patients with insufficiency (26–49 nmol/L) and 76% in patients with sufficient levels (⩾50 nmol/L; P = 0.24; Figure 3a). Vitamin D level was significantly associated with OS when adjusted for age (P = 0.02; Table 3), and in a multivariate model including age, stem cell source, anti-T-cell treatment,
Graft failure Patients developing graft failure showed a trend towards lower pre-HSCT levels of 25-OH-D3 with a median level of 34 nmol/L (range 16–65), as compared with 43 nmol/L (range 10–118) in the other patients (P = 0.06). As 11 out of 13 cases of graft failure were in patients having received reduced intensity conditioning (RIC), and RIC has previously been shown to be a risk factor for graft failure,22 we chose this subgroup for a post hoc analysis (Figure 4).
a Overall survival (%)
Acute and chronic GvHD The cumulative incidence of acute GvHD grade II–IV was 42% in patients with vitamin D deficiency ( o25 nmol/L), 48% in patients with insufficiency (26–49 nmol/L), and 37% in patients with sufficient vitamin D levels (⩾50 nmol/L), showing no correlation between the 25-OH-D3 level at the time of transplantation and incidence of aGvHD. For chronic GvHD (NIH grade moderate or severe), however, the 2-year cumulative incidence was 56% in patients with vitamin D deficiency, compared with 31% in patients with insufficiency, and 21% in the vitamin D sufficient group (P = 0.01; Figure 2). This association between 25-OH-D3 levels and cGvHD was confirmed in a multivariate landmark analysis at 100 days with death and disease relapse as competing risks, including vitamin D status (deficient, insufficient, sufficient), age, stem cell source, anti-T-treatment, haematological malignancy and sex mismatch (female donor to male recipient) as covariates. This analysis identified vitamin D status before HSCT as a significant independent risk factor for development of moderate to severe cGvHD (P = 0.04). HLA mismatch (o 8/8) was not associated with cGvHD in this material, and including it in the analysis did not change the result. The effect of 25-OH-D3 levels appeared to be continuous at the lower intervals, with a threshold at 60 nmol/L. The relative risk of cGvHD at 2 years after HSCT, measured as risk ratio, was 2.66 (95% CI: 1.03–6.87) for 25-OH-D3 below 60 nmol/L as compared with above 60 nmol/L before HSCT.
sex mismatch and haematological malignancy, a lower 25-OH-D3 level was a significant risk factor for death (P = 0.03). The 2-year DFS showed no significant differences between the groups, with 59% DFS in both the vitamin D-deficient and insufficient groups, versus 67% in the vitamin D sufficient group (Figure 3b). There was no significant difference in relapse incidence between the vitamin D-deficient (19%) and sufficient patients (21%).
100
P = 0.24
50
ORIGINAL ARTICLE
Increased incidence of chronic GvHD and CMV disease in patients with vitamin D deficiency before allogeneic stem cell transplantation L von Bahr1,2, O Blennow3, J Alm2, A Björklund1,4, K-J Malmberg1,4,5,6, D Mougiakakos7, A Le Blanc1, PJ Oefner8, M Labopin9, P Ljungman1 and K Le Blanc1,2 Vitamin D has emerged as a central player in the immune system, with its deficiency being implicated in the pathogenesis of several autoimmune diseases, including chronic GvHD. This is a retrospective cohort analysis of 166 patients, who underwent allogeneic hematopoietic stem cell transplantation (HSCT) at the Karolinska University Hospital, evaluating GvHD, graft failure, infectious complications and survival after HSCT in relation to pre-transplantation vitamin D levels. Most of the patients were deficient in vitamin D before HSCT (median 42 nmol/L). In multivariate analysis, vitamin D level before HSCT was identified as a significant independent risk factor for development of cGvHD. The increased incidence of cGvHD was not coupled to better disease-free survival; instead there was a trend towards lower overall survival in the vitamin D-deficient patients. In addition, we found a significant correlation between vitamin D deficiency and incidence of CMV disease, with no case of CMV disease occurring in patients with sufficient levels of vitamin D before HSCT. Our results support a role of vitamin D in immune tolerance following HSCT. These findings could be highly relevant for the care of HSCT patients, and prospective, randomized studies on the effect of vitamin D supplementation are therefore needed. Bone Marrow Transplantation (2015) 50, 1217–1223; doi:10.1038/bmt.2015.123; published online 1 June 2015
INTRODUCTION Allogeneic hematopoietic stem cell transplantation (HSCT) is an important and potentially curative treatment of haematological malignancies, but GvHD and infections affect outcomes. Vitamin D has emerged as a central player in the immune system,1 affecting T and B cells, macrophages and dendritic cells (DCs).2–4 A vitamin D-enriched milieu maintains an immature DC phenotype associated with decreased ability to stimulate alloreactive T cells in MLCs.5 Vitamin D deficiency has been associated with an increased incidence of chronic GvHD,6 and supplementation with vitamin D to induce a tolerogenic DC population has been suggested for prevention of GvHD.7 One mechanism for this maintenance of an immature DC population has been shown to be the upregulation of indoleamine 2,3-dioxygenase (IDO),8 an enzyme converting tryptophan to kynurenine and a central feature in tolerizing DCs.9 Contrary to the suppressive effects on the immune system, vitamin D also has a protective effect against infections. This was first shown in tuberculosis, where the traditional treatments with sunlight and cod liver oil, rich in vitamin D, are viewed in a new light after the discovery of antimicrobial peptides induced by vitamin D. The two antimicrobial peptides under the influence of vitamin D are LL-37 (cathelicidin) and β-defensin 2,10,11 which
have activity against several bacteria, as well as viruses12 and fungi.13 We present here what we believe is currently the largest epidemiological study on adult HSCT patients examining clinical outcome with regards to vitamin D deficiency before HSCT, including both infectious and non-infectious complications. PATIENTS AND METHODS Patients The study was conducted in accordance with the Helsinki Declaration upon approval by the regional ethics committee in Stockholm. Serum specimens were collected from 166 consecutive patients (412 years of age) undergoing HSCT between 2005 and 2011 at the Karolinska University Hospital, Huddinge, (Table 1). Serum samples from 20 healthy donors were used as controls. All data were taken from the patient's medical records. A total of 139 patients were eligible for evaluation of chronic GvHD (cGvHD) excluding patients with graft failure (n = 13) or with a survival after HSCT of o100 days (n = 14). For analysis of infectious complications, only patients with a follow-up at the Karolinska University Hospital were selected (n = 137). When calculating disease-free survival (DFS), only patients with haematological malignancies (n = 156) were considered. For IDO analysis, 24 patients with vitamin D levels above or below 50 nmol/L were selected; 14 with cGvHD and 10 without cGvHD (Table 2). They were selected based on the following criteria: clear presence or
1 Department of Medicine, Hematology Centre, Karolinska University Hospital, Stockholm, Sweden; 2Center for Hematology and Regenerative Medicine, Division of Clinical Immunology and Transfusion Medicine, Department of Laboratory Medicine, Karolinska Institutet, Stockholm, Sweden; 3Division of Infectious Diseases, Department of Medicine, Karolinska Institutet, Stockholm, Sweden; 4Department of Medicine, Center for Infectious Medicine, Karolinska Institutet, Stockholm, Sweden; 5Department of Cancer Immunology, Institute for Cancer Research, Oslo University Hospital, Oslo, Norway; 6The KG Jebsen Center for Cancer Immunotherapy, Institute of Clinical Medicine, University of Oslo, Oslo, Norway; 7Department of Internal Medicine 5, Hematology and Oncology, University of Erlangen-Nuremberg, Erlangen, Germany; 8Institute of Functional Genomics, University of Regensburg, Regensburg, Germany and 9Service d'Hématologie et Thérapie Cellulaire, Hôpital Saint Antoine, Paris, France. Correspondence: Dr L von Bahr, Department of Laboratory Medicine, Karolinska Institutet, Hematology Centre R51, 141 86 Stockholm, Sweden. E-mail: [email protected] Received 14 February 2015; revised 14 April 2015; accepted 18 April 2015; published online 1 June 2015
Vitamin D deficiency and complications after HSCT L von Bahr et al
1218 Table 1.
Patient characteristics o25 nmol/L deficient N = 19 (11%) N
Age o40 years ⩾ 40 years Median (range) Vitamin D supplementation 1st year post transplantation
%
9 10
25–49 nmol/L insufficient N = 88 (53%) N
%
⩾ 50 nmol/L sufficient N = 59 (36%) N
P-value
%
47% 53% 36 (14–65) 4 21%
20 68
23% 77% 53 (14–67) 6 7%
17 42
29% 71% 53 (12–68) 8 14%
Sex Female Male
4 15
21% 79%
41 47
47% 53%
27 32
46% 54%
Donora Donor age, median (range) Matched related Matched unrelated o8/8 HLA matchb
8 7 2
29 (3–50) 47% 41% 12%
21 36 25
34 (18–64) 26% 44% 30%
11 39 7
27 (10–63) 19% 68% 12%
0.004* 0.9
0.18
0.14 0.007*
Sex matchc F to M Other
4 142
22% 78%
11 762
13% 87%
3 56
5% 95%
Diagnosis AML ALL CML MDS MPD/MF Myeloma Lymphoma/CLL PNH Solid tumour Non-malignant
7 2 1 3 1 2 1 — — 2
37% 11% 5% 19% 5% 11% 5% — — 11%
36 7 2 11 6 1 23 — 1 1
41% 8% 2% 13% 7% 1% 26% — 1% 1%
19 4 4 11 2 1 10 1 2 5
32% 7% 7% 19% 3% 2% 17% 2% 3% 8%
Stem cell source PBSC Bone marrow Cord blood
13 4 2
68% 21% 11%
80 4 4
91% 5% 5%
52 7 —
88% 12% —
0.02*
Conditioning RIC MAC
7 12
37% 63%
54 34
61% 39%
37 22
63% 37%
0.1
Anti-T cell No ATG Alemtuzumab Cyclophosphamide
7 12 — —
39% 61% — —
21 60 7 —
24% 68% 8% —
8 43 6 2
14% 73% 10% 3%
0.2
0.03*
Abbreviations: ATG = anti-thymocyte globulin (rabbit); F = female; HSCT = hematopoietic stem cell transplantation; M = male; MAC = myeloablative conditioning; MDS = myelodysplastic syndrome; MF = myelofibrosis; MPD = myeloproliferative disorder; RIC = reduced intensity conditioning. aCord blood transplants excluded. bMainly C mismatch. No haplo-transplants. Three unknown. cOne unknown. P-values are calculated using Χ2-test for trend or independent-samples median test. Table of all patients included in the study, in cohorts by vitamin D level before transplantation. *P-valueo 0.05.
absence of classical chronic GvHD (grade moderate to severe), no donor lymphocyte infusions or stem cell booster and no photopheresis. Follow-up ranged between 36 and 107 months after HSCT, with a median follow-up of 71 months.
Definitions For cGvHD diagnosis and scoring, the NIH consensus criteria were used14 and both classic chronic GvHD and overlap syndrome were included, but not late-onset acute GvHD. Only cGvHD of moderate and severe grades was considered in the analysis. Graft failure was defined as the lack of engraftment, engraftment with recipient cells or later developing full (495%) recipient chimerism in the absence of relapse of the underlying disease. DFS was defined as survival Bone Marrow Transplantation (2015) 1217 – 1223
with no evidence of relapse or progression of malignant disease. Overall survival (OS) was defined as the time from HSCT to death, regardless of the cause. We evaluated both agent-specific and non-agent-specific infectious complications during the first year after HSCT, censoring the follow-up at relapse of the malignant disease. Agent-specific outcomes were CMV disease, EBV-associated post-transplantation lymphoproliferative disorder, influenza, invasive fungal disease and bacteraemia. Non-agent-specific outcomes were pneumonia and days on IV antibiotics. Bacteraemia and days on IV antibiotics were analysed separately for neutropenic and non-neutropenic episodes. CMV disease was defined according to Ljungman et al.15 Invasive fungal disease was defined according to De Pauw.16 Only probable and proven © 2015 Macmillan Publishers Limited
Vitamin D deficiency and complications after HSCT L von Bahr et al
1219 Table 2.
Patients included in IDO analysis Chronic GvHD
Total Age Median (range) Sex Donor Donor age median (range) Stem cell source Conditioning Anti-T cell Time to cGvHD grade moderate/severe median weeks (range) 25-OH-D3 before transplantation, median (range
No chronic GvHD
Low vitamin D ( o50 nmol/L)
High vitamin D (450 nmol/L)
Low vitamin D (o50 nmol/L)
High vitamin D (450 nmol/L)
7
7
5
5
34 (15–55) 6 M, 1 F 2 MUD, 5 SIB 36 (17–50) 7 PBSC 6 MAC, 1 RIC 2 ATG, 5 none 29 (17–94)
53 (37–65) 3 M, 4 F 3 MUD, 4 SIB 49 (20–63) 7 PBSC 3 MAC, 4 RIC 4 ATG, 3 none 37 (17–87)
47 (25–59) 3 M, 2 F 5 MUD 23 (19–25) 5 PBSC 4 MAC, 2 RIC 5 ATG —
46 (22–55) 2 M, 3 F 5 MUD 26 (18–36) 5 PBSC 3 MAC, 2 RIC 5 ATG —
20 (15–43)
59 (51–66)
30 (26–46)
69 (57–64)
25-OH-cholecalciferol, nmol/l
Abbreviations: ATG = anti-thymocyte globulin (rabbit), F = female; M = male; MAC = myeloablative conditioning; MUD = matched unrelated donor; RIC = reduced intensity conditioning. Table of 24 patients selected for IDO analysis.
125 100 75 50
Insufficiency level
25
Deficiency level
no
rs
52 H
ea
lth y
do
w
26 w
w
0
0
to current routine standards at the hospital laboratory. In brief, blood samples collected in serum tubes were left for 30–120 min at room temperature, or up to 12 h at +8 °C, before serum separation by centrifugation. Specimens were stored in liquid nitrogen until analysis. Serum vitamin D was measured as 25-OH-cholecalciferol and expressed as nmol/L, hereafter referred to as 25-OH-D3. Analyses were performed by the laboratory for clinical chemistry (Karolinska University Hospital, Solna) using a chemiluminescence method (CLIA) approved by SWEDAC. Vitamin D deficiency was defined as o25 nmol/L, insufficiency as 26–49 nmol/L and sufficiency as ⩾ 50 nmol/L, in accordance with the Nordic Guidelines.19 Serum levels of tryptophan metabolites were analysed before and at 1, 2, 3, 6, 12 and 24 months post HSCT by liquid chromatography-tandem mass spectrometry.20 IDO activity was estimated by calculating the serum kynurenine-to-tryptophan ratio. A median of six out of seven time points per patient (range 2–7) were available for analysis.
Figure 1. Alterations of 25-OH-D3 levels in HSCT patients. Serum levels of 25-OH-D3 (median with interquartile range) in HSCT patients before conditioning (w0), and 26 and 52 weeks, respectively, after transplantation, as well as in healthy donors.
infections were considered in the analysis. Diagnosis of pneumonia required either a combination of new pulmonary infiltrates on chest X-ray or CT scan with symptoms of respiratory infection such as cough, dyspnoea or fever, excluding idiopathic pulmonary syndrome or autopsyverified infectious pneumonia. Bacteraemia was defined as the first positive blood culture during a 10-day time period. Repeated positive blood cultures 410 days after the first were considered new episodes.
Prophylaxis and surveillance GvHD prophylaxis consisted of cyclosporin A in combination with methotrexate or prednisone, or tacrolimus in combination with sirolimus. Patients with an unrelated donor were given anti-T-cell therapy, most commonly ATG (see Table 1 for details). Patients with a related donor did not receive anti-T-cell therapy. Patients were routinely monitored for CMV reactivation with PCR and pre-emptively treated as described previously.17 Antifungal, antibacterial and Pneumocystis jiroveci prophylaxis was given as previously described.18
Serum analysis Vitamin D levels were measured in serum specimens collected before the start of conditioning therapy and in 85 of the patients also at 6 and/or 12 months after HSCT. Specimens were collected and handled according © 2015 Macmillan Publishers Limited
Statistical analysis This is a retrospective cohort study. Data were analysed as of date of the last data collection, 14 April 2014. Sample size was based on a power analysis using data from a pilot study (unpublished). Primary end point was chronic GvHD; secondary end points were OS, DFS, graft failure and incidence of infectious complications. Cumulative incidence functions were used to estimate GvHD, considering death and relapse to be competing events. Probabilities of DFS and OS were calculated using the Kaplan–Meier estimates. Univariate analyses were performed using Gray’s test for cumulative incidence functions and the log-rank test for DFS and OS. Associations of patient and graft characteristics with outcomes were evaluated in multivariate analysis, using Cox proportional hazards model for dichotomous variables or negative binomial regression analysis for outcomes with repeated events or continuous variables. For two-sample comparisons, the Wilcoxon rank-sum test or Fisher’s exact test was used. GraphPad Prism 6 (GraphPad software, San Diego, CA, USA), IBM SPSS version 21 (SPSS Inc./IBM, Armonk, NY, USA) and R 3.0.1 software were used.
RESULTS Serum levels of 25-OH-D3 in HSCT patients The median level of 25-OH-D3 for all 166 patients before transplantation was 42 nmol/L (range 10–118), below the deficiency level of 50 nmol/L (20 ng/mL)21 and significantly lower than in the healthy controls (median 66.5 nmol/L, range 21–104, P o0.001). Only 59 patients (36%) had sufficient levels of 25-OHD3 (⩾50 nmol/L) before HSCT, whereas the majority (N = 88, 53%) were 25-OH-D3 insufficient (26–49 nmol/L) and 19 (11%) were even deficient (o 25 nmol/L). The baseline characteristics of the Bone Marrow Transplantation (2015) 1217 – 1223
Vitamin D deficiency and complications after HSCT L von Bahr et al
1220
Survival and relapse The 2-year OS was 63% in patients with vitamin D deficiency (o 25 nmol/L), 69% in patients with insufficiency (26–49 nmol/L) and 76% in patients with sufficient levels (⩾50 nmol/L; P = 0.24; Figure 3a). Vitamin D level was significantly associated with OS when adjusted for age (P = 0.02; Table 3), and in a multivariate model including age, stem cell source, anti-T-cell treatment,
Graft failure Patients developing graft failure showed a trend towards lower pre-HSCT levels of 25-OH-D3 with a median level of 34 nmol/L (range 16–65), as compared with 43 nmol/L (range 10–118) in the other patients (P = 0.06). As 11 out of 13 cases of graft failure were in patients having received reduced intensity conditioning (RIC), and RIC has previously been shown to be a risk factor for graft failure,22 we chose this subgroup for a post hoc analysis (Figure 4).
a Overall survival (%)
Acute and chronic GvHD The cumulative incidence of acute GvHD grade II–IV was 42% in patients with vitamin D deficiency ( o25 nmol/L), 48% in patients with insufficiency (26–49 nmol/L), and 37% in patients with sufficient vitamin D levels (⩾50 nmol/L), showing no correlation between the 25-OH-D3 level at the time of transplantation and incidence of aGvHD. For chronic GvHD (NIH grade moderate or severe), however, the 2-year cumulative incidence was 56% in patients with vitamin D deficiency, compared with 31% in patients with insufficiency, and 21% in the vitamin D sufficient group (P = 0.01; Figure 2). This association between 25-OH-D3 levels and cGvHD was confirmed in a multivariate landmark analysis at 100 days with death and disease relapse as competing risks, including vitamin D status (deficient, insufficient, sufficient), age, stem cell source, anti-T-treatment, haematological malignancy and sex mismatch (female donor to male recipient) as covariates. This analysis identified vitamin D status before HSCT as a significant independent risk factor for development of moderate to severe cGvHD (P = 0.04). HLA mismatch (o 8/8) was not associated with cGvHD in this material, and including it in the analysis did not change the result. The effect of 25-OH-D3 levels appeared to be continuous at the lower intervals, with a threshold at 60 nmol/L. The relative risk of cGvHD at 2 years after HSCT, measured as risk ratio, was 2.66 (95% CI: 1.03–6.87) for 25-OH-D3 below 60 nmol/L as compared with above 60 nmol/L before HSCT.
sex mismatch and haematological malignancy, a lower 25-OH-D3 level was a significant risk factor for death (P = 0.03). The 2-year DFS showed no significant differences between the groups, with 59% DFS in both the vitamin D-deficient and insufficient groups, versus 67% in the vitamin D sufficient group (Figure 3b). There was no significant difference in relapse incidence between the vitamin D-deficient (19%) and sufficient patients (21%).
100
P = 0.24
50
