Haploidentical transplantation versus unrelated donor hematopoietic stem cell transplantation

The gold standard donor for hematopoietic stem cell transplantation (HSCT) is a matched sibling donor (MSD), but for most, the only available option is HSCT from a human leukocyte antigen (HLA) matched unrelated donor (MUD), and increasingly from a related haploidentical donor. Haploidentical donors are defined as being at least 50% HLA-identical to the recipient and are usually first-degree relatives. The introduction of post-transplant cyclophosphamide (PTCy) to prevent graft-versus-host disease (GvHD) has greatly improved the outcome of haploidentical HSCT.¹

Leonardo Javier Arcuri from the Insitituto Nacional de Câncer, Centro de Transplante de Medula Óssea, Rio de Janeiro, Brazil, and colleagues performed a systematic review² to compare overall survival (OS), non-relapse mortality (NRM), GvHD (acute [aGvHD] and chronic [cGvHD]) and relapse rates in patients with hematological malignancies treated with either haploidentical HSCT with PTCy or standard unrelated donor (URD) HSCT. In their Biology of Blood and Marrow Transplantation publication (August 2019) they detail their searches of the PubMed and Cochrane databases for studies comparing haploidentical HSCT with PTCy against URD HSCT indexed between 1^st January 2008 and 1^st January 2019. The group screened 113 abstracts, before reviewing 46 full texts for eligibility. There were 20 observational studies remaining after eligibility checks, which totalled 1,783 haploidentical HSCT recipients and 6,077 URD HSCT recipients. One study was not included in the meta-analysis due to a lack of extractable data on standard deviations. Study characteristics and a summary of results are detailed in Tables 1 and 2 respectively.

Results

OS: The study found no difference in overall survival between the two different donor HSCT treatments (hazard ratio (HR) 0.98 (95% CI, 0.88–1.08) and risk difference (RD) for death of 4 percentage points (pp) (95% CI, -8 to +1 pp).

Relapse: There was also no difference for relapse risk which was 2 pp higher (95% CI, -2 to +6 pp) in the haploidentical donor HSCT group and the hazard ratio was 1.06 (95% CI, 0.95–1.19). This was also true when only patients with active or high-risk disease were analyzed.

NRM: In studies in which all haploidentical donor HSCTs used a PTCy approach, NRM was lower for haploidentical donor transplantation by 6 pp (95% CI, -10 to -3) and had a HR of 0.85 (95% CI, 0.72–1.00).

GvHD: In terms of aGvHD, both grade II-IV and grade III-IV were lower in the haploidentical HSCT groups (grade II-IV: RD, -12 pp [95% CI, -17 to -7 pp]; HR, 0.52 [95% CI, 0.47–0.82] and grade III-IV: RD, -9 pp [95% CI, -13 to -5 pp]; HR, 0.66 [95% CI, 0.48–0.92]).

The RD of cGvHD was 12 pp lower in the haploidentical donor HSCT group (95% CI, -20 to -4 pp). However, the team reported a high heterogeneity (Ι² = 86%) and found that this was associated with the use of peripheral blood stem cells (PBSCs) and the proportional difference between the two donor groups (31% of participants in the haploidentical donor HSCT group received PBSCs vs 85% in the URD HSCT group). Meta-regression analysis found an association of PBSC transplantation with cGvHD in the haploidentical donor HSCT groups (cGvHD was 3.7 pp lower in the haploidentical donor HSCT group for each 10-pp PBSC proportion difference [p < 0.001; R² = 70%]).

 

Discussion

The meta-analysis found that there is no difference in OS between the two donor types. The team also reported no evidence of higher relapse rates in the haploidentical donor HSCT group, and this finding was upheld even when analying only high-risk disease studies.

However, the incidence rates of all forms of GvHD and NRM are lower with haploidentical donor HSCT compared to URD HSCT. The team point out that this could be partially explained by the use of PTCy-based GvHD prophylaxis for haploidentical donor HSCT as well as the lower use of PBSCs in this group.

The finding that haploidentical donor HSCT is comparable with URD HSCT in terms of OS and relapse rate, while showing lower rates of NRM and GvHD, encouraged the authors to recommend performing haploidentical donor HSCT with PTCy in patients who are being treated in experienced centers, and cannot wait 2–3 months for identification of an URD. This would allow speeding up of the donor identification process, especially for patients with hematologic malignancy with a high risk of relapse.

Arcuri and colleagues highlight some limitations of their study, such as incomplete data extraction from included studies, all of the studies included were retrospective, outcomes of studies were sometimes reported as cumulative incidence with no confidence intervals, and HRs either reported with inadequate reference categories or not reported at all. They also note that they were unable to find any randomized controlled trials comparing these donor transplantations.

 

Table 1: Study Characteristics

Study	Multi center	Disease	Number			% PBSC		% MAC		GVHD Prophylaxis, %		Age, year		Active or High-risk Disease, %
AML, acute myelogenous leukemia; MAC, myeloablative conditioning;  MDS, myelodysplastic syndrome; MMURD, mismatched unrelated donor; NHL, non-Hodgkin lymphoma; TCD, T cell depletion
										Haplo	URD
Haplo	URD	% MMURD	Haplo	URD	Haplo	URD	PTCy	PTCy or in vivo TCD	Haplo	URD	Haplo	URD
Burroughs et al, 2008 (3)	Yes	Hodgkin Lymphoma	28	24	25	0	100	0	0	100	0	32	28	43	38
Di Stasi et al, 2014 (4)	No	AML/MDS	19	26	0	5	62	0	0	100	100	55	62	0	0
Di Stasi et al, 2014 (4)	No	AML/MDS	13	82	0	0	51	0	0	100	100	52	62	100	100
Raiola et al, 2014 (5)	No	Malignant	92	43	0	0	40	77	72	100	100	45	42	58	42
Raiola et al, 2014 (5)	No	Malignant	92	43	100	0	35	77	72	100	100	45	47	58	56
Ciurea et al, 2015 (6)	Yes	AML	88	737	0	13	89	0	0	100	39	78% >50	95%>50	16	22
Ciurea et al, 2015 (6)	Yes	AML	104	1245	0	18	81	100	100	100	23	42% >50	43%>50	34	25
Garciaz et al, 2015 (7)	Yes	NHL	26	28	12	50	100	0	0	100	100	53	61	43	29
Baker et al, 2016 (8)	No	Malignant	54	59	39	56	68	0		100	91	50.5	57	44	51
Bashey et al, 2016 (9)	No	Malignant	116	178	0	45	82	40	51	100	28	51	53	40	31
Blaise et al, 2016 (10)	No	Malignant	31	63	0	87	95	0	0	100	100	62	64	39	30
Gaballa et al, 2016 (11)	No	Malignant and non malignant	60	46	100	3	17	0	0	100	100	45	51	38	40
Kanate et al, 2016 (12)	Yes	Lymphoma	185	241	0	13	91	0	0	100	100	55	55	6	18
Kanate et al, 2016 (12)	Yes	Lymphoma	185	491	0	13	94	0	0	100	0	55	55	6	11
Rashidi et al, 2016 (13)	No	AML	52	88	0	100	100	44	44	100	0	54	63	42	41
How et al, 2017(14)	No	Refractory AML	24	43	19	100	98	67	79	100	24	54	55	100	100
Martinez et al, 2017 (15)	Yes	Hodgkin Lymphoma	98	273	0	39	88	10	31	100	74	31	32	15	16
McCurdy et al, 2017 (16)	No	Malignant	372	120	0	0	0	0	100	100	100	55	50	19	29
Bashey et al, 2018 (17)	No	Malignant	33	57	4	48	79	6	30	100	4	66	65	17	24
Brissot et al, 2018 (18)	Yes	AML, relapsed/ refractory	199	1111	0	53	94	54	42	100	78	52	52	100	100
Brissot et al, 2018 (18)	Yes	AML, relapsed/ refractory	199	383	100	53	92	54	38	100	88	52	52	100	100
Devillier et al, 2018 (19)	No	AML	33	30	10	94	97	0	17	100	97	64% >65	50%>65	24	7
Dulery et al, 2018 (20)	Yes	Refractory malignant	27	29	24	78	100	78	79	100	100	42	63	100	100
Lorentino et al, 2018 (21)	Yes	AML, adverse karyotype	48	433	0	62	81	53	49	100	76	49	53	100	100
Lorentino et al, 2018 (21)	Yes	AML, adverse karyotype	48	123	100	62	83	53	54	100	86	49	51	100	100
Pagliardini et al, 2018 (22)	No	Malignant	81	81	0	73	96	17	28	100	100	50	50	32	25

 

 

Table 2: Summary of pooled results

Outcome	Risk Difference (95% CI) in percentage points (pp)*	Number of published studies (number of actual comparisons)	Hazard Ratio (95% CI)	Number of published studies (number of actual comparisons)
aGvHD, acute graft-versus-host disease; cGvHD, chronic graft-versus-host disease; CI, confidence interval; NRM, non-relapse mortality; OS, overall survival * For Risk Difference, negative values favor haploidentical donor hematopoietic stem cell transplantation and positive values favour unrelated donor hematopoietic stem cell transplantation † High heterogeneity Statistically significant results in bold
OS	-4 (-8 to +1)	11 (15)	0.98 (0.88–1.08)	11 (15)
Relapse	+2 (-2 to +6)	12 (16)	1.06 (0.95–1.19)	10 (13)
NRM	-6 (-10 to -3)	10 (14)	0.85 (0.72–1.00)	10 (13)
aGVHD grade II-IV†	-12 (-17 to -7)	10 (14)	0.52 (0.47–0.82)	7 (9)
aGVHD grade III-IV†	-9 (-13 to -5)	10 (13)	0.48 (0.32–0.72)	6 (8)
cGVHD, all†	-12 (-20 to -4)	12 (16)	0.25 (0.17–0.38)	3 (4)
cGVHD, moderate†	-5 (-14 to +4)	4 (5)	0.55 (0.31–0.99)	5 (5)