Allogeneic hematopoietic stem cell transplant (allo-HSCT) can offer a long-term cure to patients with Hodgkin lymphoma (HL) and non-Hodgkin lymphoma (NHL)1. However, often human leukocyte antigen (HLA)-matched related or unrelated donors are unavailable, particularly for patients who come from ethnic minority backgrounds. In these instances, unrelated cord blood (UCB) or haploidentical (haplo) relatives may be considered donor options. Deciding between these donor options typically relies on consideration of the transplant center’s experience, donor availability, and cost. No comparative analyses between UCB and haplo donors in this setting have been conducted to date, but studies involving these donor types, with heterogenous low-intensity conditioning regimens and graft-versus-host disease (GvHD) prophylaxis, are available.
In order to compare the outcomes of patients with HL or NHL who underwent non-myeloablative haplo or UCB allo-HSCT, Giancarlo Fatobene, Vanderson Rocha, and colleagues conducted a retrospective analysis based on patient data obtained from various international stem cell transplant registries. The results of the analysis are summarized in this article and were published in the Journal of Clinical Oncology1.
Study design
- Data were reported from the Lymphoma Working Party (LWP) of the European Society for Blood and Marrow Transplantation (EBMT), Eurocord, and the Center for International Blood and Marrow Transplant Research (CIBMTR)
- Patients (N = 740) with HL (n = 283) and NHL (n = 457) who received their first allo-HSCT between 2009 and 2016 were included:
- Subtypes of NHL included diffuse large B-cell lymphoma (DLBCL), follicular lymphoma (FL), mantle cell lymphoma (MCL), or T-cell lymphoma (TCL)
- Key patient characteristics are shown in Table 1
- Patients underwent haplo-HSCT (n = 526) or UCB-HSCT (n = 214)
- For haplo-HSCT the stem cell source was either bone marrow (BM; 68%) or peripheral blood (PB; 32%)
- Conditioning: 2 Gy of total-body irradiation, cyclophosphamide, and fludarabine
- Cyclophosphamide:
- UCB: 50 mg/kg
- Haplo: 29 mg/kg
- Fludarabine:
- UCB: 200 mg/m2
- Haplo: 150 mg/m2
- Cyclophosphamide:
- GvHD prophylaxis:
- Haplo transplantation: calcineurin inhibitor and mycophenolate, plus post-transplant cyclophosphamide (PTCy; 100 mg/kg)
- UCB transplantation: calcineurin inhibitor and mycophenolate
- Primary endpoint: overall survival (OS)
Table 1. Key patient characteristics by type of donor type1
|
|
Haplo-BM |
Haplo-PB |
UCB |
|---|---|---|---|
|
n |
357 |
169 |
214 |
|
Median age, years (range) |
48 (18–75) |
48 (19–75) |
48 (18–73) |
|
Sex, % Male Female |
37 37 |
68 32 |
64 36 |
|
Performance score, % 90–100 ≤80 Not reported |
74 22 4 |
62 35 3 |
69 25 6 |
|
Disease subtype, % HL DLBCL FL MCL TCL |
44 30 10 8 9 |
39 20 9 12 20 |
29 22 19 8 22 |
|
Disease status, % CR PR Chemo-refractory Untreated |
48 40 10 2 |
46 34 17 3 |
53 35 7 5 |
|
Prior ASCT, % |
52 |
56 |
59 |
|
Median follow-up, months (range) |
43 (3–100) |
30 (3–87) |
48 (3–112) |
ASCT, autologous stem cell transplant; BM, bone marrow; chemo, chemotherapy; CR, complete remission; DLBCL, diffuse large B-cell lymphoma; FL, follicular lymphoma; haplo, haploidentical; HL, Hodgkin lymphoma; MCL, mantle cell lymphoma; PB, peripheral blood; PR, partial remission; TCL, T-cell lymphoma; UCB, unrelated cord blood
Hematopoietic recovery
- Median time to neutrophil recovery, days (UCB vs haplo-BM vs haplo-PB): 20 vs 19 vs 17
- Day 28 neutrophil recovery was lower with UCB vs haplo-BM and haplo-PB, p < 0.0001
- UCB: 69% (95% CI, 58–78)
- Haplo-BM: 85% (95% CI, 79–91)
- Haplo-PB: 92% (95% CI, 86–97)
- Day 100 platelet recovery: lower with UCB vs haplo-BM and haplo-PB, p < 0.0001
- UCB: 67% (95% CI, 57–77)
- Haplo-BM: 91% (95% CI, 85–96)
- Haplo-PB: 90% (95% CI: 83–96)
GvHD rates
- Rates of acute GvHD (aGvHD) and chronic GvHD (cGvHD) are shown in Table 2
- Day 100 Grade 2–4 aGvHD was significantly higher with UCB vs haplo-BM (p < 0.0001) but did not vary between UCB and haplo-PB (p = 0.17)
- Day 100 Grade 3–4 aGvHD was also significantly higher with UCB compared to haplo-BM or haplo-PB, p < 0.0001
- Six-month cGvHD rates were higher with UCB compared to haplo-BM (p = 0.04) but not haplo-PB (p = 0.91)
- Two-year severe cGvHD was higher with UCB and haplo-PB vs haplo-BM (p = 0.02)
- Four-year cGvHD: no difference with UCB compared to haplo-BM (p = 0.31) and haplo-PB (p = 0.42)
Table 2. GvHD rates by donor type1
|
|
UCB |
Haplo-BM |
Haplo-PB |
|---|---|---|---|
|
Day 100 Grade 2–4 aGvHD, % (95% CI) |
43 (36–50) |
20 (16–25) |
35 (28–43) |
|
Day 100 Grade 3–4 aGvHD, % (95% CI) |
18 (13–24) |
5 (3–8) |
6 (3–11) |
|
Median time to onset of cGvHD, months (IQR) |
5 (4–8) |
6 (4–11) |
6 (4–9) |
|
6-month incidence of cGvHD, % (95% CI) |
17 (12–23) |
11 (8–14) |
18 (12–24) |
|
2-year severe cGvHD, % (95% CI) |
14 (10–20) |
8 (5–12) |
16 (10–22) |
|
4-year cGvHD rate, % (95% CI) |
28 (22–34) |
24 (19–29) |
32 (25–40) |
aGvHD, acute graft-versus-host disease; BM, bone marrow; cGvHD, chronic graft-versus-host disease; GvHD, graft-versus-host disease; haplo, haploidentical; IQR, interquartile range; PB, peripheral blood; UCB, unrelated cord blood
Efficacy
OS was significantly lower with UCB compared to haplo-BM or haplo-PB (p = 0.008), which was also true of progression-free survival (PFS, p = 0.002). Additionally, the lower survival associated with UCB transplant was attributed to higher transplant-related mortality (TRM), as shown in Table 3 and further demonstrated in multivariable analysis in Table 4.
Table 3. Efficacy outcomes by donor type1
|
|
UCB |
Haplo-BM |
Haplo-PB |
|---|---|---|---|
|
4-year PFS probability, % (95% CI) |
36 (30–43) |
46 (41–52) |
52 (43–61) |
|
4-year OS probability, % (95% CI) |
49 (42–56) |
58 (53–64) |
59 (50–68) |
|
4-year cumulative incidence TRM, % (95% CI) |
33 (27–40) |
20 (16–24) |
18 (13–25) |
|
4-year cumulative incidence of relapse/progression, % (95% CI) |
31 (25–37) |
34 (29–39) |
30 (22–38) |
BM, bone marrow; CI, confidence interval; haplo, haploidentical; PB, peripheral blood; OS, overall survival; PFS, progression-free survival; UCB, unrelated cord blood
Multivariable analysis
As shown in Table 4.
- UCB transplant was associated with a higher risk of Grade 2–4 aGvHD and cGvHD compared to haplo-BM, but not haplo-PB
- TRM was higher with UCB compared to haplo-BM or haplo-PB
- Risk of relapse/progression did not differ between UCB and haplo-BM but there was a lower risk associated with haplo-PB compared with UCB
- OS and PFS were lower with UCB compared with haplo-BM and haplo-PB
Table 4. Multivariable analysis of outcomes by donor type1
|
Outcome |
HR |
95% CI |
p value |
Adjusted for |
|---|---|---|---|---|
|
Grade 2–4 aGvHD UCB vs haplo-BM UCB vs haplo-PB |
1.83 1.04 |
1.37–2.44 0.76–1.43 |
< 0.0001 0.81 |
Age |
|
Grade 3–4 aGvHD UCB vs haplo-BM UCB vs haplo-PB |
2.61 2.31 |
1.55–4.40 1.18–4.54 |
0.0003 0.02 |
Age |
|
cGvHD UCB vs haplo-BM UCB vs haplo-PB |
1.53 1.01 |
1.09–2.15 0.69–1.48 |
0.02 0.97 |
— |
|
TRM UCB vs haplo-BM UCB vs haplo-PB |
1.91 2.27 |
1.37–2.65 1.45–3.54 |
0.0001 0.0002 |
Age, performance score, and prior ASCT |
|
Relapse/progression UCB vs haplo-BM UCB vs haplo-PB |
1.11 1.52 |
0.80–1.52 1.03–2.26 |
0.54 0.04 |
Age, lymphoma subtype, and disease status |
|
PFS UCB vs haplo-BM UCB vs haplo-PB |
1.44 1.86 |
1.14–1.82 1.39–2.50 |
0.002 < 0.0001 |
CMV serostatus, lymphoma subtype, and disease status |
|
OS UCB vs haplo-BM UCB vs haplo-PB |
1.55 1.59 |
1.19–2.03 1.15–2.20 |
0.001 0.005 |
Age, sex, CMV serostatus, lymphoma subtype, and disease status |
aGvHD, acute graft-versus-host disease; BM, bone marrow; cGvHD, chronic graft-versus-host disease; CMV, cytomegalovirus; haplo, haploidentical; HR, hazard ratio; OS, overall survival; PB, peripheral blood; PFS, progression-free survival; TRM, transplant-related mortality; UCB, unrelated cord blood
Subgroup analysis
- Patients aged 60–75 had a higher risk of Grade 2–4 aGvHD vs patients aged 18–39 (HR, 1.40; 95% CI, 1.03–2.30). No difference was seen for aGvHD for patients age 40–59 years and those age 18–30 years
- Patients had a higher risk of TRM if they were aged 40–59 or 60–75, had a performance score ≤ 80 (vs 90–100), or had received prior ASCT
- Relapse risk was higher for patients with DLBCL or MCL (vs HL) and patients with PR or chemotherapy-refractory disease at time of transplant (vs CR)
- In relation to PFS, risks were higher for cytomegalovirus (CMV) seropositive recipients, patients with DLBCL and MCL (vs HL), and those with PR or chemotherapy-refractory disease at time of transplant (vs CR)
- For OS, there was a higher risk of death associated with age 60–75, CMV seropositive recipients, DLBCL and TCL subtypes, and those with PR or chemotherapy-refractory disease at time of transplant (vs CR)
Deaths
- UCB: 98/193 (51%)
- Haplo-BM: 115/313 (37%)
- Haplo-PB: 58/159 (37%)
- Causes of death
- Most common: recurrent disease (rates did not differ between groups)
- Second most common: infection (rates did not differ between groups)
- GvHD (UCB vs haplo-BM vs haplo-PB): 18% vs 6% vs 9%, p = 0.01
- Other causes: interstitial pneumonitis and organ failure
- Both Grade 3–4 aGvHD and cGvHD were associated with higher mortality following UCB transplant compared with haplo-BM and haplo-PB
Strengths and limitations of study
Strengths
- Large sample size of patients with lymphoma (> 700)
- Long-term follow-up
Weaknesses
- Retrospective — treatment choices were physician/institution preference
- Heterogenous conditioning and GvHD prophylaxis regimens
- Differences in transplantation period were not accounted for
- Inability to adjust for unknown factors, such as comorbidities
Conclusion
This analysis has shown UCB transplant leads to lower OS and PFS compared with haplo-HSCT, specifically in patients with lymphoma. The lower OS and PFS with UCB transplant were attributed to higher TRM and not relapse/progression. Additionally, outcome based on donor type was independent of other factors, such as disease type or status at transplant. Finally, there were no differences in PFS or OS comparing BM to PB as stem cell source for haplo-HSCT.
The authors recommend performing haplo-HSCT (BM or PB) for patients with HL and NHL when no HLA-matched donor can be found and conducting UCB only in cases where a haplo donor is not available. Future studies using UCB for transplantation should investigate strategies to lower the risk of GVHD, such as the use of PTCy.