These authors contributed equally to this study.
Haematopoietic stem cell transplantation in haemophagocytic lymphohistiocytosis
Version of Record online: 5 MAY 2005
British Journal of Haematology
Volume 129, Issue 5, pages 622–630, June 2005
How to Cite
Horne, A., Janka, G., Maarten Egeler, R., Gadner, H., Imashuku, S., Ladisch, S., Locatelli, F., Montgomery, S. M., Webb, D., Winiarski, J., Filipovich, A. H., Henter, J.-I. and the Histiocyte Society (2005), Haematopoietic stem cell transplantation in haemophagocytic lymphohistiocytosis. British Journal of Haematology, 129: 622–630. doi: 10.1111/j.1365-2141.2005.05501.x
- Issue online: 23 MAY 2005
- Version of Record online: 5 MAY 2005
- Received 28 December 2004; accepted for publication 14 March 2005
- stem cell transplantation;
- familial haemophagocytic lymphohistiocytosis;
Haemophagocytic lymphohistiocytosis (HLH) poses major therapeutic challenges, and the primary inherited form, familial haemophagocytic lymphohistiocytosis (FHL), is usually fatal. We evaluated, including Cox regression analysis, survival in 86 children (29 familial) that received HLH-94-therapy (etoposide, dexamethasone, ciclosporin) followed by allogeneic stem cell transplantation (SCT) between 1995 and 2000. The overall estimated 3-year-survival post-SCT was 64% [confidence interval (CI) = ±10%] (n = 86); 71 ± 18% in those patients with a matched related donor (MRD, n = 24), 70 ± 16% with a matched unrelated donor (MUD, n = 33), 50 ± 24% with a family haploidentical donor (haploidentical, n = 16), and 54 ± 27% with a mismatched unrelated donor (MMUD, n = 13). After adjustment for potential confounding factors, estimated odds ratios (OR) for mortality were 1·93 (CI =0·61–6·19) for MUD, 3·31 (1·02–10·76) for haploidentical, and 3·01 (0·91–9·97) for MMUD, compared with MRD. In children with active disease after 2-months of therapy (n = 43) the OR was 2·75 (1·26–5·99), compared with inactive disease (n = 43). In children with active disease at SCT (n = 37), the OR was 1·80 (0·80–4·06) compared with inactive disease (n = 49), after adjustment for disease activity at 2-months. Mortality was predominantly transplant-related. Most HLH patients survived SCT using MRD or MUD, and survival with partially mismatched donors was also acceptable. Patients that responded well to initial pretransplant-induction therapy fared best, but some persisting HLH activity should not automatically preclude performing SCT.
Haemophagocytic lymphohistiocytosis (HLH) poses major diagnostic and therapeutic challenges. The primary autosomal recessive form, familial haemophagocytic lymphohistiocytosis (FHL) with an incidence of around 1:50 000 live-born children, is fatal without adequate treatment (Janka, 1983; Henter et al, 1991a,b, 1998; Aricòet al, 1996; Janka et al, 1998). In addition, central nervous system (CNS) involvement may be severe and cause permanent CNS dysfunction (Henter & Elinder, 1992; Haddad et al, 1997; Henter & Nennesmo, 1997). Prompted by earlier dismal therapeutic results (Janka, 1983; Henter et al, 1991a; Aricòet al, 1996), in 1994 the Histiocyte Society developed a consensus study protocol: the HLH-94 protocol. HLH-94 comprises an induction/continuation phase, which aims to induce a remission in these often dramatically sick patients, followed by haematopoietic stem cell transplantation (SCT) from the best available donor, which aims to replace the immune system and thereby induce a definitive cure, in patients with familial, persistent and recurrent disease (Henter et al, 1997, 2002). Recent studies have revealed perforin and hMunc 13–4 gene defects in FHL-families, providing a strong rationale to correct the disease with a substitution of haematopoietic stem cells (Stepp et al, 1999; Clementi et al, 2001; Göransdotter Ericson et al, 2001; Feldmann et al, 2002, 2003; Suga et al, 2002).
We have previously reported results of HLH-94, the first prospective international therapeutic study on HLH, with an overall estimated 3-year probability of survival of 55% (±9%) and 51% (±20%) in the familial cases. We also showed that the major risk of fatality now occurs after SCT, as opposed to in the pre-SCT phase (Henter et al, 2002). The present paper analysed the outcome of SCT of 86 children with HLH in more detail, all of whom were initially treated with the HLH-94 protocol. We focused on parameters associated with post-transplant disease-free survival, including multivariate analysis of donor- and treatment-related measures.
Material and methods
Information concerning the patients was submitted at regular intervals by the treating physicians on follow-up forms. As of June 2002, the HLH-94 data-base in Stockholm contained 107 patients aged <15 years at onset of HLH-94 therapy, who had not received any previous cytotoxic therapy, and had their first SCT performed between 1 January 1995 and 31 December 2000. Of these patients, 87 were eligible for the present study since they either fulfilled the diagnostic criteria (Henter et al, 1991a) (Table I) or had familial disease (as indicated by an affected sibling), whereas 19 remaining patients did not fulfil either of these criteria and there was missing data for one patient. We report on the 86 individuals that had complete information on the covariates included in the multivariate analysis (listed in ‘Statistics’ below). The patient excluded because of missing information was a boy below 1 year of age at onset of disease, who was transplanted with a matched unrelated donor and who is still alive (more than 4 years after transplantation). When comparing the 86 children studied and the 21 that were not eligible, there were no differences in age, consanguinity or CNS involvement, although significantly more girls did not fulfil the eligibility criteria, and there was no statistically significant difference in the estimated 3-year probability of survival between the two groups. Sixty-four of the patients were included in a previous report on HLH-94 outcome in which 65 patients underwent SCT (Henter et al, 2002); in that report, the term ‘bone marrow transplant’ also includes cord and peripheral stem cell transplants. Patient data and outcome were analysed as of 1 December 2003.
|• Fever: duration ≥7 days|
|• Splenomegaly ≥3 cm below the costal margin|
|• Cytopenia affecting two or all three lineages in the peripheral blood with haemoglobin <9·0 g/dl, platelets <100 × 109/l, neutrophils <1·0 × 109/l|
|• Hypertriglyceridaemia and/or hypofibrinogenaemia (fasting triglycerides ≥2·0 mmol/l or ≥3 SD of the normal value of age, fibrinogen ≤1·5 g/l or ≤3 SD)|
|• Haemophagocytosis in bone marrow or spleen or lymph nodes. No evidence of malignancy.|
HLH disease status
In the HLH-94 protocol, inactive disease was defined as having no clinical signs of disease; i.e. no fever except if infection-induced, no hepatosplenomegaly, no clinical signs of active CNS disease, and no cytopenias (except if drug-induced). CNS disease was defined as having an abnormal neurological examination and, in addition, cerebrospinal fluid (CSF) pleocytosis and/or elevated CSF protein. In the nine cases with a discrepancy between the report on active/inactive disease at 2 months and the separate specific reports on the clinical and laboratory data included in the definition for active disease, the reporting physicians rechecked the data prior to the final analysis.
All 86 patients were initially treated according to the HLH-94 protocol (Henter et al, 1997, 2002), consisting of 8 weeks initial therapy with etoposide (VP-16) and dexamethasone (Fischer et al, 1985; Henter et al, 1986, 1993). In addition, intrathecal (IT) methotrexate (MTX) was administered to 11 of 24 patients with CNS disease at registration. Continuation therapy with alternating weekly pulses of VP-16 and dexamethasone together with daily ciclosporin A (CSA) (Stéphan et al, 1993; Imashuku et al, 1999a), followed for all but three patients who changed therapy after the initial treatment.
Donor typing, transplantation conditioning and GVHD prophylaxis
Human leucocyte antigen (HLA)-typing was performed using genomic techniques in 50 patients (58%) and serological techniques in 18 (21%); information regarding technique used was unavailable for 18 patients (21%). Both the conditioning regimen and graft-versus-host disease (GVHD) prophylaxis were determined by the treating transplant unit, although a suggestion was included in the HLH-94 protocol (Henter et al, 1997, 2002). The suggested preparative regimen consisted of oral busulphan at 4 mg/kg on days −9, −8, −7 and −6 (total dose 16 mg/kg), intravenous cyclophosphamide at 50 mg/kg on days −5, −4, −3 and −2 (total dose 200 mg/kg) and intravenous VP-16 at 300 mg/m2 on days −5, −4 and −3 (total dose 900 mg/m2). In practice, 80 patients (93%) were reported to have received treatment with busulphan, 78 (91%) were administered cyclophosphamide and 43 (50%) received VP-16; 41 of 86 (48%) received all three drugs. Additional immunosuppression was recommended for unrelated donor transplants, with intravenous antithymocyte globulin (ATG) on days −2, −1, +1 and +2. This treatment was reported to have been administered to 28 of 46 patients (61%) with SCT from unrelated donors.
The suggested GVHD prophylaxis included intravenous MTX 15 mg/m2 on day +1 and 10 mg/m2 on days +3, +5, and +11, in combination with CSA beginning on day −3. In practice, 75/85 patients (87%) were reported to have received treatment with CSA and 43 of 84 patients (50%) with MTX (all of whom also had CSA therapy). Seven of the 10 children not receiving CSA were children with family haploidentical stem cell donors (Table II).
|Matched related||Matched unrelated||Family haploidentical||Mismatched unrelated||Total|
|Number of evaluated patients||24||33||16||13||86|
|Males||14 (58)||21 (64)||13 (81)||8 (62)||56 (65)|
|Age <12 months at start*||18 (75)||19 (58)||12 (75)||10 (77)||59 (69)|
|Age 12–24 months at start*||4 (17)||8 (24)||2 (12)||2 (15)||16 (19)|
|Age >24 months at start*||2 (8)||6 (18)||2 (12)||1 (8)||11 (13)|
|Parental consanguinity||8/23 (35)||6/32 (19)||0 (0)||3 (23)||17/84 (20)|
|Familial disease||9 (38)||12/32 (38)||5 (31)||3 (23)||29/85 (34)|
|CNS-involvement at start*||8 (33)||8 (24)||6 (38)||3 (23)||25 (29)|
|Active disease 2 months after start*||15 (62)||15 (45)||8 (50)||5 (38)||43 (50)|
|≤180 days from start to SCT*||21 (88)||9 (27)||6 (38)||8 (62)||44 (51)|
|SCT 1995–96||10 (42)||11 (33)||2 (12)||3 (23)||26 (30)|
|SCT 1997–98||8 (33)||11 (33)||10 (62)||7 (54)||36 (42)|
|SCT 1999–2000||6 (25)||11 (33)||4 (25)||3 (23)||24 (28)|
|Active disease at time for SCT||14 (58)||10 (30)||8 (50)||5 (38)||37 (43)|
|Conditioning with VP-16||12 (50)||18 (55)||1 (6)||12 (92)||43 (50)|
|Conditioning with busulphan||24 (100)||27 (82)||16 (100)||13 (100)||80 (93)|
|Conditioning with cyclophosphamide||22 (92)||29 (88)||14 (88)||13 (100)||78 (91)|
|Conditioning with ATG||9/22 (41)||20 (61)||13/15 (87)||8 (62)||50/83 (60)|
|Conditioning with irradiation||1 (4)||6 (18)||0 (0)||1 (8)||8 (9)|
|GVHD prophylaxis with CSA||23 (96)||30/32 (94)||9 (56)||13 (100)||75/85 (88)|
|GVHD prophylaxis with MTX||13/23 (57)||21/32 (66)||3 (19)||6 (46)||43/84 (51)|
Differences in distribution were compared by using the chi-square test or, where frequencies were small, the two-tailed Fisher's exact test. Survival rates were analysed using the Kaplan–Meier life table method and univariate comparison of survival using the log rank test. Multivariate analysis using Cox proportional hazards regression was performed with time to death as the endpoint and using the maximum follow-up time available. The covariates used were: sex, age at start of treatment, CNS involvement at start of therapy, disease activity at 2 months after start of treatment, disease activity at SCT, time to SCT, and donor type; matched related donors (MRD), matched unrelated donors (MUD), family haploidentical donors (HAPLO), and mismatched unrelated donors (MMUD).
All analyses were carried out using the Statistical Package for the Social Sciences (spss, version 11·5; Chicago, IL, USA). P-values <0·05 were considered significant. The study was approved by the Histiocyte Society and the Ethics Committee of the Karolinska Institute.
Eighty-six patients from 18 countries were eligible for inclusion in the analyses. The characteristics and donor groups of these patients are detailed in Table II. Family history of the disease was present in 29 of 85 patients (34%) and parental consanguinity in 17 of 84 (20%). The diagnostic criteria were completely fulfilled by 76 patients (88%). Of those who did not fulfil all criteria but were evaluated, all 10 had a positive family history. CNS involvement was found in 25 patients (29%) on disease diagnosis. Neurological alterations were reported in 12 of 83 patients (14%) at 2 months after start of treatment and in 13 of 80 (16%) at the time of SCT.
The median age at SCT was 13 months (range 4 months to 11 years) and the median time from start of therapy to SCT was 6 months (range 2 months to 10 years); being 8, 19, 13 and 11 months in the MRD, MUD, HAPLO, and MMUD groups, respectively. MRD were utilized in 24 transplantations (21 from siblings), MUD in 33 (including two umbilical cord transplants), HAPLO in 16, and MMUD in 13 (including five umbilical cord transplants).
Engraftment, graft failure, second transplantation and relapse
Information on engraftment was available for 83 patients, 75 of whom (90%) were reported to have achieved engraftment (for details on donor types see Table III). All three patients for whom information was missing died during the first 100 days after SCT. Data on the number of cells provided and the timing of engraftment is not available. Of the eight children who never engrafted, seven had active and one inactive disease at 2 months (P = 0·029, Fisher's exact test, two-sided) whereas at the time of SCT, five had active and three had inactive disease (not significant). Of the eight children who did not achieve engraftment (none of whom had a cord transplant), two received a second transplant; one is alive at the last follow-up (32 months after the second transplant). Among the patients who engrafted, secondary graft failure was reported in three, all within the first 100 days (all died, one after a second transplant).
|Matched related||Matched unrelated||Family haploidentical||Mismatched unrelated||Total|
|Number of evaluated patients||24||33||16||13||86|
|Engraftment||21/23 (91)||28/31 (90)||13/16 (81)||13 (100)||75/83 (90)|
|Acute GVHD||4/22 (18)||9/30 (30)||5/14 (36)||7/12 (58)||25/78 (32)|
|Disease relapse despite sustained engraftment*||0/20 (0)||1/27 (4)||0/12 (0)||1/12 (8)||2/71 (3)|
|Second transplantation||0/22 (0)||2/31 (6)||1/16 (6)||0/13 (0)||3/82 (4)|
|Alive 100 days after SCT||17 (71)||27 (82)||9 (56)||7 (54)||60 (70)|
|Alive 1 year after SCT||17 71)||24 (73)||9 (56)||7 (54)||57 (66)|
|Alive at last follow-up||17 (71)||23 (70)||8 (50)||7 (54)||55 (64)|
In total, there were seven patients with recurrent disease, four of whom had no engraftment (two MUD and two HAPLO), two of these children received a second transplant (see above). Two individuals had active disease at SCT, and two had inactive disease.
Among the engrafted patients without known graft failure, three (with MRD, MUD and MMUD, respectively) were reported to have suffered recurrent episodes of HLH. One of these died on day +160 (no information is available on whether graft failure occurred). The other two received additional HLH-therapy post-SCT and thereafter stabilized; they are now alive off therapy with inactive disease. They both had full donor haematopoiesis at the time of relapse, both received VP-16, and one had an EBV-induced post-SCT-HLH. The other patient, who experienced two bouts of HLH activity following cord transplant, became severely ill and needed artificial respiration for 58 days and VP-16 was administered for nine consecutive weeks (until day +144) (Schwinger et al, 1988).
Acute GVHD (grade 2–4) was reported in 25 of 78 patients (32%), with 18% in the MRD group as compared with 30%, 36% and 58% in the MUD, HAPLO and MMUD groups, respectively (Table III). At 1 year after transplant, 57 patients were alive. Among these, chronic GVHD was reported in four of 44 patients (9%); one of 14 in the MRD group, as compared with one of 16 and two of six in the MUD and HAPLO groups, respectively.
Of the seven children with recurrent disease, two had acute GVHD. Both had an initial engraftment, and one had a sustained engraftment and was alive at follow-up, and for the other child, who died, no information was available on whether graft failure occurred. None of the three children with recurrent disease that were alive after 1 year had chronic GVHD reported at that time (missing data for one child).
Overall, the 3-year probability of survival (with 95% confidence interval) was 64% (±10%) as shown in Fig 1A, and in the 29 familial cases 66% (±17%). At the time of analysis 55 patients were alive, with a median follow-up period of 4·1 years post-SCT (range 1·1–7·2 years), all reported as being without active disease. With regard to their quality of life, we used a measure of ‘ability to play’ as an indicator of quality of life; 44 children had a normal ability to play, seven had a mild to moderate inhibition to play and two had a moderate to severe inhibition to play, according to the last report (missing data in two cases) (Lansky et al, 1987). Of the 25 patients with CNS involvement at diagnosis, 15 (60%) were long-term survivors (nine died within the first year after transplant, and one died during the second post-transplant year).
Of the 55 patients alive, data on chimaerism at 1 year or more after SCT was available in 43 patients. Eight of these were reported to have mixed chimaerism, all with inactive disease, at a median time of 3 years after SCT.
Causes of death
At the time of analysis 31 patients had died. Of these, 26 were reported to have had transplant-related mortality (TRM), two died after relapse of HLH (one had a graft failure, and the other patient is the one without information on whether graft failure occurred), one died of secondary acute myeloid leukaemia (AML) as previously reported (Henter et al, 2002), one died of a respiratory disease of unknown cause and one died during surgery (unrelated to HLH). The two children who died following relapse of HLH both had active disease at the time of SCT.
A total of 23 deaths occurred within 100 days after SCT. Causes of death were GVHD (n = 3); cytomegalovirus (CMV; n = 3); pneumonia, bronchiolitis obliterans organizing pneumonia, or respiratory failure of unknown origin (n = 8); veno-occlusive disease (VOD) (n = 4, including one with additional CMV and one with additional septicemia); septicemia (n = 1); interstitial pneumonia with CMV and adenovirus (n = 1). The further three cases were reported as SCT complications, but were not more closely defined. Thus, at least half of the early deaths were attributed to treatment-related complications of the lung and the liver.
Of the eight deaths that occurred more than 100 days after SCT, two occurred after more than 1-year (one following graft failure, relapse and subsequent AML at day SCT+ 450, and the other at day SCT+ 550 because of respiratory disease). The other six deaths were caused by relapse (n = 2); infections (n = 1); post-transplant lymphoproliferative disease (n = 1); GVHD and aspergillus infection (n = 1) and surgical haemorrhage (n = 1).
Univariate analyses of survival are shown in Table IV. For the donor groups, survival was 71% (±18%) for MRD (n = 24, three with non-sibling donors), 70% (±16%) for MUD (n = 33), 50% (±24%) for HAPLO (n = 16) and 54% (±27%) for MMUD (n = 13), at 3 years after SCT (Fig 1B). Multivariate analysis was performed and the results are shown in Table V. There was no statistically significant difference in survival between MRD and MUD. With regard to unadjusted OR there was no significant difference between HAPLO and MMUD. However, the adjusted OR for HAPLO compared with MRD was 3·31 (1·02–10·76) whereas the OR for MMUD compared with MRD was 3·01 (0·91–9·97).
|Variable||n||Probability (95% CI)||P-value|
|All evaluated patients||86||64 (54–74)|
|Age group at start of treatment|
|<12 months||59||64 (52–76)||†|
|12–24 months||16||63 (39–87)||0·912|
|>24 months||11||64 (36–92)||0·888|
|CNS-involvement at registration|
|Disease activity 2 months after start of treatment|
|Time from start of treatment to SCT|
|≤180 days||44||59 (44–74)||0·362|
|>180 days||42||69 (55–83)|
|Disease activity at time for SCT|
|Year of SCT|
|Matched related||24||71 (53–89)||†|
|Matched unrelated||33||70 (54–86)||0·972|
|Family haploidentical||16||50 (26–74)||0·202|
|Mismatched unrelated||13||54 (27–81)||0·323|
|OR of death||95% CI||P-value||OR of death||95% CI||P-value|
|Age <12 months at start†||1||1|
|Age 12–24 months at start†||1·04||0·42–2·59||0·925||0·68||0·25–1·89||0·464|
|Age >24 months at start†||1·08||0·37–3·16||0·882||1·26||0·41–3·91||0·685|
|CNS at start†||1·13||0·53–2·40||0·747||0·99||0·44–2·22||0·985|
|Active disease 2 months after start†||2·62||1·23–5·57||0·012||2·75||1·26–5·99||0·011|
|Active disease at SCT||1·93||0·95–3·91||0·070||1·80||0·80–4·06||0·154|
|>180 days from start to SCT†||0·72||0·35–1·47||0·365||0·57||0·24–1·37||0·209|
|Matched related donor||1||1|
|Matched unrelated donor||1·02||0·39–2·68||0·420||1·93||0·61–6·19||0·265|
|Family haploidentical donor||1·93||0·70–5·32||0·205||3·31||1·02–10·76||0·047|
|Mismatched unrelated donor||1·75||0·59–5·22||0·314||3·01||0·91–9·97||0·071|
Disease activity at 2 months after start of HLH-therapy
Our data indicated that children with active disease at 2 months after start of HLH-94 treatment had a significantly worse outcome after SCT (51 ± 15%) than those with inactive disease (77 ± 12%) (P = 0·009) (Table IV). This increased risk of mortality post-SCT for these patients remained statistically significant after adjustment for potentially confounding factors (OR = 2·75, 1·26–5·99, P = 0·011) (Table V).
Disease activity at time for transplantation
Active disease at the time of SCT was associated with a decreased survival indicated by an unadjusted odds ratio of 1·93 (0·95–3·91, P = 0·070) from Cox regression analysis reported in Table V. Adjustment for the potential confounding factors reduces these odds to 1·80 (0·80–4·06, P = 0·154).
A number of conclusions with regard to SCT in HLH can be drawn from the HLH-94 study, the first prospective international study on HLH: (1) The probability of surviving 3 years after SCT is satisfactory (64%), in view of the mixed clinical and geographical backgrounds of the patients. (2) Survival for MUD transplant recipients is not significantly lower than that for MRD transplant recipients, and although the use of HAPLO and MMUD gives less favourable survival, the outcome is still acceptable, supporting the use of alternative donors where matched donors are unavailable. (3) The survival curve reaches a plateau 2-years after SCT, consistent with persistent control of underlying HLH. (4) Some degree of disease activity at time for transplantation should not automatically preclude performing SCT, but this topic deserves closer scrutiny in future studies. (5) Persistent disease activity at 2 months after start of HLH treatment appears to suggest a worse long-term prognosis.
Fischer et al (1986) were the first to show that a cure for FHL could be achieved by SCT, as later confirmed by others (Todo et al, 1990; Blanche et al, 1991; Nespoli et al, 1991; Bolme et al, 1995; Baker et al, 1997; Jabado et al, 1997; Dürken et al, 1999; Imashuku et al, 1999b). Initially, only matched donors were used but, since the outcome of FHL is invariably fatal without a successful SCT, HLA non-identical donors are now used more often and represent a third of the donors in this analysis. We have demonstrated that the use of MUD (6/6) provide results comparable with those achieved with MRD. Furthermore, 15 of 29 (52%) patients were alive at last follow-up following HLA non-identical transplantation, with a median follow-up time of 50 months in the survivors. These data support SCT with the use of alternative donors when matched donors are unavailable. We suggest these SCT are performed in experienced centres.
In FHL, two steps are essential for survival and cure; first, an effective initial and continuation therapy, and second, a successful SCT. HLH-94 is highly successful in achieving symptomatic remission, allowing most patients to be admitted to SCT (with a pre-SCT fatality rate of around 20%) (Henter et al, 2002). The overall estimated 3-year probability of survival has increased to around 50% when the protocol has been followed (Henter et al, 2002). This represents a remarkable improvement in disease control, particularly when compared with data from the early 1980s, when the cure rate in FHL was 0% and the median survival time after diagnosis was only 1–2 months (Janka, 1983). Moreover, the importance of SCT in the cure of HLH is highlighted by our previous report of the HLH-94 data, in which none of the patients with verified familial disease survived without having a SCT (Henter et al, 2002).
Opinions vary as to what degree of disease activity immediately prior to transplantation compromises long-term outcome. In this study, there was a tendency towards better survival in patients with inactive disease at SCT, but this association failed to achieve statistical significance in univariate analysis, and adjustment for potential confounding factors further eroded the risk associated with active disease at SCT. Our results suggest that some degree of disease activity, e.g. residual (but reduced) splenomegaly, at the time for transplantation should not automatically preclude performing SCT, which is worthy of consideration, particularly for patients in whom a state of inactive disease is difficult to achieve. This topic deserves closer scrutiny in future studies. Apart from donor type, the only independent statistically significant association with improved survival was inactive disease after 2 months of HLH-94 therapy. Whether or not this association is because of different underlying molecular abnormalities remains to be elucidated. It could be speculated that poor responders to chemotherapy may respond poorly to conditioning treatment and may only suffer toxicity and not benefits. The analyses on disease activity should, though, be interpreted with caution. To validate our findings, the answers to the question on active or inactive disease at 2 months have been compared with answers on specific clinical and laboratory data after 2 months of treatment that were reported separately. In the few cases with inconsistencies, the data have been rechecked by the reporting centres and amended according to the definitions set out in the study. One potential limitation with our study is the possibility of time censoring, since the time to SCT was different in the various groups (longest in MUD patients). However, in the HLH-94 study most of the deaths prior to SCT occurred during the first 3 months (Henter et al, 2002), suggesting that the risk of time censoring in the current study may not be that great.
A substantial proportion of early deaths post-SCT (<100 days) were attributed to regimen-related toxicity: non-infectious pulmonary toxicity and VOD. Post-transplant complications associated with fevers, capillary leak, vascular instability, respiratory distress and/or hepatic dysfunction, sometimes referred to as the ‘hypercytokinaemia syndrome’, occur more frequently in HLH patients than in SCT patients with other diagnoses (unpublished observations). This constellation of symptoms often occurs around the time of engraftment or with the onset of GVHD, and has been reversed with the use of very high doses of corticosteroids and/or anti-tumour necrosis factor agents. The hypercytokinaemia syndrome is usually observed in the context of complete or predominant donor type engraftment and differs from the recurrent HLH experienced later after SCT, particularly in patients who have waning engraftment or rejection. We have no data on how many of the patients reported here developed this syndrome, since we did not specifically ask for this information. Early recognition and aggressive management of the hypercytokinaemia syndrome may improve the overall success of SCT. Whether this syndrome is associated with the underlying biological deficiency in HLH remains to be elucidated.
This study confirms that successful transplantation will induce not only prolonged survival but also cure in HLH. The vast majority of deaths occurred within the first year after SCT, two during the second year, and then the survival curve was flat, indicating a limited likelihood of late relapse (Fig 1A). Importantly, HLH may develop in the early post-SCT period despite apparent earlier engraftment, as reported in a few of our patients. There are at least two possible explanations in addition to graft rejection. One possibility is delayed and/or insufficient lymphocyte recovery, resulting in delayed normalization of NK cell function and relapse despite complete three-lineage engraftment (Schwinger et al, 1988). Another is that the immunocompromised child experiences an episode of secondary HLH after SCT, such as in connection with a viral infection, which may be treatable. In both situations, increasing immunosuppression with CSA and corticosteroids and, in addition, administration of VP-16 may result in ultimate remission.
To conclude, we report the largest collection of SCT data so far presented on this rare disease, the result of a broad international collaboration. Further, the study was prospective and, to our knowledge, reported the only multivariate analysis of survival following SCT in patients with HLH.
We would like to acknowledge all collaborating colleagues and families. Supported by grants provided by the Children's Cancer Foundation of Sweden, the Cancer and Allergy Foundation of Sweden, the Swedish Research Council (no. B0661), the Tobias Foundation, the Ronald McDonald Foundation, the Märta and Gunnar V. Philipson Foundation, and the Histiocytosis Association of America.
- 1996) Hemophagocytic lymphohistiocytosis. Report of 122 children from the International Registry. FHL Study Group of the Histiocyte Society, Leukemia, 10, 197–203. , , , , , , & (
- 1997) Successful correction of hemophagocytic lymphohistiocytosis with related or unrelated bone marrow transplantation. Blood, 89, 3857–3863. , , , , , , & (
- 1991) Treatment of hemophagocytic lymphohistiocytosis with chemotherapy and bone marrow transplantation: a single-center study of 22 cases. Blood, 78, 51–54. , , , , & (
- 1995) Allogeneic bone marrow transplantation for hemophagocytic lymphohistiocytosis in Sweden. Bone Marrow Transplantation, 15, 331–335. , , , , , & (
- 2001) Six novel mutations in the PRF1 gene in children with haemophagocytic lymphohistiocytosis. Journal of Medical Genetics, 38, 643–646. , , , , , , , , , , & (
- 1999) Improved outcome in haemophagocytic lymphohistiocytosis after bone marrow transplantation from related and unrelated donors: a single-centre experience of 12 patients. British Journal of Haematology, 106, 1052–1058. , , , , , , , , , , & (
- 2002) Functional consequences of perforin gene mutations in 22 patients with familial haemophagocytic lymphohistiocytosis. British Journal of Haematology, 117, 965–972. , , , , , , , , , , & (
- 2003) Munc13–4 is essential for cytolytic granules fusion and is mutated in a form of familial hemophagocytic lymphohistiocytosis (FHL3). Cell, 115, 461–473. , , , , , , , , , , , , , , & (
- 1985) Treatment of four patients with erythrophagocytic lymphohistiocytosis by a combination of epipodophyllotoxin, steroids, intrathecal methotrexate, and cranial irradiation. Pediatrics, 76, 263–268. , , , , & (
- 1986) Allogeneic bone marrow transplantation for erythrophagocytic lymphohistiocytosis. Journal of Pediatrics, 108, 267–270. , , , , , , , & (
- 2001) Spectrum of perforin gene mutations in familial hemophagocytic lymphohistiocytosis. American Journal of Human Genetics, 68, 590–597. , , , , , , , , , , , , , , , , & (
- 1997) Frequency and severity of central nervous system lesions in hemophagocytic lymphohistiocytosis. Blood, 89, 794–800. , , , , & (
- 1992) Cerebromeningeal haemophagocytic lymphohistiocytosis. Lancet, 339, 104–107. & (
- 1997) Neuropathologic findings and neurologic symptoms in twenty-three children with hemophagocytic lymphohistiocytosis. Journal of Pediatrics, 130, 358–365. & (
- 1986) Successful induction with chemotherapy including teniposide in familial erythrophagocytic lymphohistiocytosis. Lancet, 2, 1402. , , & (
- 1991a) Diagnostic guidelines for hemophagocytic lymphohistiocytosis. Seminars in Oncology, 18, 29–33. , & (
- 1991b) Incidence in Sweden and clinical features of familial hemophagocytic lymphohistiocytosis. Acta Paediatrica Scandinavica, 80, 428–435. , , & (
- 1993) Familial hemophagocytic lymphohistiocytosis and viral infections. Acta Paediatrica Scandinavica, 82, 369–372. , , & (
- 1997) HLH-94: a treatment protocol for hemophagocytic lymphohistiocytosis. HLH study Group of the Histiocyte Society. Medical and Pediatric Oncology, 28, 342–347. , , , , , , , , , , & (
- 1998) Familial hemophagocytic lymphohistiocytosis. Primary hemophagocytic lymphohistiocytosis. Hematology/Oncology Clinics of North America, 12, 417–433. , , , & (
- 2002) Treatment of hemophagocytic lymphohistiocytosis with HLH-94 immunochemotherapy and bone marrow transplantation. Blood, 100, 2367–2373. , , , , , , , , , , & (
- 1999a) Effective control of Epstein-Barr virus-related hemophagocytic lymphohistiocytosis with immunochemotherapy. Blood, 93, 1869–1874. , , , , , , , , , , , , , , , , , , , , , , & (
- 1999b) Allogeneic hematopoietic stem cell transplantation for patients with hemophagocytic syndrome (HPS) in Japan. Bone Marrow Transplantation, 23, 569–572. , , , , , , , , , , , , , , , , , , & (
- 1997) Treatment of familial hemophagocytic lymphohistiocytosis with bone marrow transplantation from HLA genetically nonidentical donors. Blood, 90, 4743–4748. , , , , , , , , & (
- 1983) Familial hemophagocytic lymphohistiocytosis. European Journal of Pediatrics, 140, 221–230. (
- 1998) Infection- and malignancy-associated hemophagocytic syndromes. Secondary hemophagocytic lymphohistiocytosis. Hematology/Oncology Clinics of North America, 12, 435–444. , , , & (
- 1987) The measurement of performance in childhood cancer patients. Cancer, 60, 1651–1656. , , , & (
- 1991) Familial haemophagocytic lymphohistiocytosis treated with allogeneic bone marrow transplantation. Bone Marrow Transplantation, 7, 139–142. , , , , , , , , & (
- 1988) Unrelated 5/6-locus matched umbilical cord blood transplantation in a 23-month-old child with hemophagocytic lymphohistiocytosis Unrelated 5/6-locus matched umbilical cord blood transplantation in a 23-month-old child with hemophagocytic lymphohistiocytosis. Bone Marrow Transplantation, 22, 393–396. , , , , , , & (
- 1993) Treatment of familial hemophagocytic lymphohistiocytosis with antithymocyte globulins, steroids, and cyclosporin A. Blood, 82, 2319–2323. , , , , & (
- 1999) Perforin gene defects in familial hemophagocytic lymphohistiocytosis. Science, 286, 1957–1959. , , , , , , , , , & (
- 2002) Perforin defects of primary haemophagocytic lymphohistiocytosis in Japan. British Journal of Haematology, 116, 346–349. , , , , , , & (
- 1990) Allogeneic bone marrow transplantation for familial erythrophagocytic lymphohistiocytosis with high dose VP16-containing conditioning regimen. Leukemia and Lymphoma, 1, 361–364. , , , , , , , , & (