Seidel, Markus G. MD*; Minkov, Milen MD*; Witt, Volker MD*; Matthes-Martin, Susanne MD*; Pötschger, Ulrike MSc*; Worel, Nina MD†; Leitner, Gerda MD†; Stary, Jan MD‡; Gadner, Helmut MD*; Peters, Christina MD*
Granulocyte transfusion (GT) therapy as means to shorten neutropenia and combat tissue pathogens in neutropenic patients has recently drawn new attention owing to the availability of new granulocyte mobilization and apheresis methods (reviewed in Refs. 1,2). For pediatric patients, GT with sufficient cell numbers can be harvested from both granulocyte colony-stimulating factor (GCSF) and prednisolone-primed donors, whereas for adult patients only GCSF-stimulated GTs are sufficient.3–9 Being practiced in many centers based on empirical evidence and results of earlier controlled studies,10,11 randomized trials are still needed to definitively prove the efficacy of GT and particularly, whether sustained increment of the neutrophil counts results in improved survival.12
In this investigation, we analyzed the impact of granulocyte dose (per kilogram recipient's body weight), dose intensity of GT (cumulative dose within the first 5 d of GT treatment), presence of sepsis-related organ dysfunction on side effects, ANC increment, infection elimination, and survival. As a prospective clinical study, this analysis aimed to define response parameters and risk factors of adverse reactions of GT treatment and to identify a target patient population most likely to benefit from this still rarely used and controversially discussed therapeutic option.
PATIENTS AND METHODS
Patients: Between 1995 and 2005, 59 immunocompromised neutropenic children and young adults with severe infections were included in this trial. Data collected from the first 30 trial patients were published in part previously.7 After the initial publication, another 29 patients were included in the prospective study, thus forming a total cohort of 59 patients, treated with a total of 778 GTs for 70 infection episodes. The median age of the total cohort was 7.0 years (range: 0.1 to 28.0). To facilitate detection of dose-dependent effects the population was divided into 2 groups, namely, patients <18 years and ≥18 years, reflecting a division into individuals below and above 50 kg bodyweight (Table 1). Fifteen patients had various degrees of treatment-related comorbidity with most frequent diagnoses such as toxic cardiomyopathy, hepatopathy, and neurologic and endocrinologic disorders. Sepsis-related organ dysfunction was separately analyzed (Table 2). The study was approved by the local ethics committee and conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. Informed consent was obtained from all patients and all related and unrelated donors.
Inclusion criteria: (1) diagnosis of a hematologic disorder or malignancy (Table 1); (2) invasive bacterial or fungal infection (Table 2) with insufficient response to standard antibacterial and antifungal therapy; (3) ANC <0.2×109/L (=200/μL) after conventional chemotherapy (n=14), conditioning for autologous (n=4) or allogeneic (n=40) stem cell transplantation, or owing to the underlying disease itself (hemophagocytic lymphohistiocytosis, n=1) despite GCSF treatment. In the presence of a fulminant infection and expected prolonged severe aplasia (eg, after high-dose chemotherapy or conditioning for allogeneic stem cell transplantation) GTs were commenced even if the ANC were still >1000/μL with falling tendency. (4) Expected time to hematopoietic regeneration or engraftment of >5 days.
Exclusion Criteria: (1) adult respiratory distress syndrome or (2) participation in an investigational drug study.
End points: (1) survival on day 28 after first GT as surrogate parameter for effective anti-infective treatment (see below); (2) stable neutrophil reconstitution (ANC >1000/μL). Secondary end points were the granulocyte-dose-dependency of neutrophil increment (related to body weight and administration frequency), infection elimination (as determined by clinical and serologic inflammation parameters, microbiologic and radiologic findings), and adverse effects. Four-week survival after initiation of a treatment supplement for an infectious episode, although being an unspecific parameter of the clinical course, was chosen as primary end point because the course of other infection-related parameters such as fever, C-reactive protein, arterial hypotension, or other clinical signs proved too variable for statistical analyses.
Donors: exclusion of viral infections (human immunodeficiency virus 1/2, hepatitis C virus, cytomegalovirus, hepatitis B surface antigen) and stimulation with 5 μg/kg/d GCSF (filgrastim, Neupogen, Amgen, or lenograstim, Granocyte, Chugai; n=361 GTs), or prednisolone (50 mg; n=417 GTs) 6 to 10 hours preapheresis were performed as published.7,13 Leukapheresis and irradiation of GTs were carried out as described previously.7,13 GTs were applied within 6 hours of apheresis after premedication of the recipient with antihistamines, and in rare cases with prednisolone (1 mg/kg intravenously). GTs were scheduled daily at a minimum recommended cell content per concentrate of 3×108 neutrophils/kg bodyweight.7 GCSF treatment of the recipients (5 μg/kg/d intravenously over 1 h) was continued during GT therapy. Administration of GT was intended to be continued until (1) hematopoietic regeneration (ie, ANC >1000/μL), (2) recovery from infection (defined by stable improvement of laboratory and clinical parameters), (3) presence of adverse effects in donor or recipient, or (4) unavailability (Table 2).
Statistics: the comparison between different subgroups was performed with the log-rank test. A P value of <0.05 was considered significant. Survival probability was calculated by the Kaplan-Meier method.
The differences of absolute leukocyte and neutrophil content between GCSF and prednisolone-stimulated donor-derived granulocyte concentrates, as well as the respective quantities per recipient's body weight, were statistically significant in favor of GCSF-elicited granulocyte concentrates (not shown) in line with published experience.7
Seventy treatment courses with a total of 778 GTs were performed in 59 patients (1 course in 51, 2 or more courses in 8 patients). The GT for each recipient were derived from a median of 4 (range: 1 to 51) different related or unrelated donors. A median of 8 GTs (range: 1 to 65) was given per treatment course, 5.5 (range: 1 to 14) of them were administered within the first 5 days (Table 3).
Total Applied Granulocyte Dose and Dose Intensity
The median total granulocyte dose per treatment course was 11×109 neutrophils/kg (range: 1 to 91) and the median dose intensity (dose with the first 5 d) was 6.07×109 neutrophils/kg (Table 3). Owing to the higher body weight, the patient group ≥18 years (all >50 kg) received fewer neutrophils per kilogram as compared with younger individuals (median 2.67×109/kg vs. 7.21×109/kg; Table 3), but still more than the minimum recommended cell dose.
While on day 0 of GT therapy the median ANC was 0/μL in most children and adults, in some individuals (n=8) GT treatment was initiated because of anticipated prolonged neutropenia after myeloablative chemotherapy before reaching the nadir (ANC still >1000/μL with falling tendency). Those patients were excluded from the calculation of ANC increments. Hence, the day 5 ANC increment in eligible patients was 800/μL (≥18 y) and 1570/μL (<18 y). The ANC increment correlated to the dose intensity with borderline significance (Spearman correlation coefficients 0.316; P=0.04; not shown). However, we could not detect a difference with respect to probability and time to reach ANC >1000/μL between children (ie, patients with lower body weight, high-dose intensity, and high ANC increments) and adults (lower dose intensity and lower ANC increments; Fig. 1).
Infection Control and Survival
Episodes of GT treatment were terminated because of (1) infection control (n=40), (2) neutrophil regeneration/engraftment (n=10), (3) no detected benefit (n=8), and (4) death during GT-treated infectious episode (n=6; Table 2). The 28-day survival probability of the total patient cohort was 0.72±0.06 and the 100-day survival was 0.52±0.07. To clarify whether the effect of GT was dose-dependent, we analyzed the clinical course (28-d and 100-d survival as marker of infection control) of subgroups divided according to either (1) the 5-day cumulative dose of given neutrophils per kilogram body weight or (2) the detected ANC increment. There was no significant difference in the survival between patients who received a higher or lower than median 5-day cumulative neutrophil dose [28-d and 100-d survival pSUhigh-dose=0.66±0.09 and 0.52±0.11 (n=25) vs. pSUlow-dose=0.70±0.1 and 0.59±0.11 (n=20), P=not significant; not shown]. Likewise, the achieved ANC increment at day 5 of a GT treatment episode did not affect the probability to survive the infectious episode in our cohort as determined by the reasons of death (Table 2), and the 28-day and 100-day survival [pSUhigh-increment=0.66±0.1 and 0.46±0.1 (n=22) vs. pSUlow-increment=0.57±0.13 and 0.48±0.14 (n=14), P=not significant; calculated for patients with ANC <1000/μL on day 0]. The GT-treated infection was the main reason of death during the first 28 days after initiation of GT treatment (n=8, Fig. 2 and Table 2). We hypothesized, that those 8 patients who died from the infection despite granulocyte support might have received insufficient GT with lower neutrophil content than those with better infection control, but this was not the case (median 18.6×109/kg 5-day cumulative neutrophil dose, as compared with overall median of 6.07×109/kg; Table 3).
Neither the body weight of the recipients nor the type of donor stimulation (prednisolone vs. GCSF) played a role as independent markers for infection control and survival. The relative neutrophil doses given were higher in low-weight individuals and GCSF-elicited GT than in patients >18 years (all >50 kg) or after prednisolone-elicited GT in accordance with previous observations7 (Table 3). Sepsis-related cardio-respiratory distress and/or organ dysfunction (apart from preexisting comorbidity) at time of initiation of GT treatment was associated with an adverse outcome as compared with no organ involvement in both children and adults (pSU(<18) 0.65±0.1 vs. 0.89±0.1 and pSU(≥18) 0.58±0.14 vs. 1), but no alteration of the response to GT with respect to ANC increment or adverse effects was observed in patients with cardio-respiratory, renal, or multiorgan failure (not shown). The risk of death after GT administration for a bacterial infection was significantly lower than in fungal infections [28-d and 100-d survival pSUbacterial=0.89±0.06 and 0.65±0.09 (n=29) vs. pSUfungal=0.51±0.12 and 0.40±0.11 (n=20); P=0.039; Fig. 3]; these differences were independent of the patients' ages. Most of the deaths in patients receiving GT for fungal infections occurred during the first 2 weeks of GT treatment, whereas the deaths of patients with bacterial infections were more evenly distributed within the shown period until day 100 (Fig. 3).
GT administration was safe and accompanied by little acute toxicity. Fever (≤14%, mostly ≤ World Health Organization grade I to III) and chills (3%) were the most frequently observed adverse events of all transfused GT. Skin rash (grade I), airway obstruction (grades I to III) occurred in 1% to 3%, respectively; transfusion-associated acute lung injury was not observed. Furthermore, we did not detect an increased incidence or severity of graft-versus-host disease (GvHD) in allografted recipients of GT (n=40) as compared with hematopoietic stem cell transplantation (HSCT) recipients without GT treatment; neither in episodes with GT during aplasia owing to conditioning for HSCT (GT after HSCT, n=28, 2 patients with grade IV, 3 patients grade III, 8 with grade I to II acute GvHD), nor in patients who received HSCT after one or more courses of GT (n=15; 1 grade IV acute GvHD, 5 patients with grade I to II acute GvHD). There was no difference between GCSF-primed and prednisolone-primed GT with respect to adverse effects in the recipients. As could be expected from the donor selection, no cytomegalovirus infection owing to GT was detected.
This prospective clinical study of 778 GTs in 70 episodes of neutropenic fever owing to refractory infections in 59 pediatric and young adult patients confirms accumulating evidence that (1) GT may help to combat severe infections in neutropenia, (2) GT treatment is generally safe, and (3) GT may shorten treatment-related neutropenia in children4–7,9 and adults (reviewed in Refs. 1,2,12) with severe bacterial and fungal infections.
The overall incidence of adverse events was low with fever and chills in less than every sixth patient being the most frequent adverse reactions, and no severe complication or consequences on the following course was observed. Hence, no donor-related or patient-related risk factors for adverse reactions of GT could be detected. Some studies suggest increased risk of adverse reactions and reduced efficacy of mismatched concentrates, and recommend human leukocyte antigen-match or ABO-match for GT,14–16 at least for prophylactic granulocyte substitution in subjects with a history of invasive infections undergoing ablative chemotherapy.17 Owing to the tight schedule of therapeutic GT administration in our study, we had to apply GT from more than 1 donor per treatment course in most of the patients. Using a single human leukocyte antigen-matched or ABO-matched donor would carry the disadvantage of delayed treatment or lower granulocyte dosage. Although this study did not address the influence of the GT match on the treatment efficacy, our data do not support its influence on the tolerability of GT.
The most intriguing question besides the lack of valid randomized, controlled trials for the efficacy of GT is whether there is a detectable dose-effect of the transfused neutrophils on infection control and survival. Few recent studies compared the efficacy of GCSF-versus corticoid-elicited GT with or without control groups.7,8,18 It has been shown that infused granulocytes remain in the body, are primarily attracted to inflammated tissues, increase the ANC and shorten neutropenia.7,8,10,11,14,19,20 However, published studies used different administration schedules and neutrophil doses, and concerns about the ideal read-out system for GT efficacy have been claimed.8,21 Neutrophil tissue infiltration precedes ANC increments during neutrophil regeneration after neutropenia, thus, oral mucosal neutrophil counts might be superior to blood counts to assess GT-induced neutrophil increments.21 Our results show a correlation of ANC increment to the administered neutrophil dose, but no difference in infection control and overall outcome depending on the GT dose intensity or ANC increment was detected. Potentially, this might be owing to the relatively high dose of neutrophils transfused to all patients included in this study, as even young adults, but more strikingly children, received concentrates containing at least the 6 to 10-fold recommended minimum amount of neutrophils according to a recent Cochrane meta-analysis12 and the American Society of Blood Banks of 1.4×108/kg, which substantially exceeds the number most frequently available in published adult studies. Second, the administration schedule in our study was very dense, yielding a median of 5 to 6 GTs within the first 5 days of a treatment episode. This high relative dose and dose intensity might contribute to an overall similarly positive effect on both (1) the used read-out marker (ANC increment) and (2) infection elimination/survival of the 2 cohorts of patients compared in this study. The time point when 70% of patients reached 1000/μL ANC was substantially earlier in this pediatric study than in a simultaneously analyzed cohort of adult patients randomized for GT treatment in similar situations for another study, who received less intensive or no GT treatment (10 d vs. 21 to 28 d22). Patients in this study received a median number of neutrophils of 13×108/kg/concentrate (<18 y)–8×108/kg/concentrate (≥18 y) on daily basis, whereas the compared adult patients received a median of 6×108/kg/concentrate on a 3-week schedule (median age 47 y22), suggesting that the critical minimum neutrophil content of 1 GT might be higher than the World Health Organization-recommended dose of 1.4 to 3×108/kg, and that the tight schedule might also be necessary.
There are several factors complicating data interpretation. First, it is likely that patients at highest risk or those who responded least to GT as monitored by daily blood counts were treated even more intensively than others, who responded well in both ANC and clinically, suggesting that an inverse relation of administered cell dose and response might be possible in some individuals. Second, it is inherent in the treatment method that the time point of initiation of GT administration varies substantially between patients with regard to the preceding duration of neutropenia and their disease state. Thus, a comparison of the treatment success between individuals is difficult. Although these data cannot show a dose-effect of transfused neutrophils on infection control and survival, and this study lacks a control cohort who received markedly less or no GT, our data corroborate the notion that efficient neutrophil substitution regimens require a rapid availability of GCSF-stimulated donors and centers for transfusion medicine on a 24-hour basis. We did not perform a statistical comparison with a matched historical cohort because (1) practically all patients at our institution after 1995 with prolonged neutropenia and an invasive infection were included in this study and received GT and (2) the available anti-infective treatment before 1995 differed substantially from the study time frame (availability of linezolid, carbopenems, new antifungal agents like caspofungin, voriconazole, and posaconazole), thus causing a bias. Likewise, we did not provide a comparison with published results, because it was not feasible to identify a matched cohort of consecutive pediatric patients, selected according to the inclusion criteria of this study.
Overall survival in the study cohort was 72% on day 28 and 52% on day 100 after the first administration of GT, reflecting the severely life-threatening situation of these patients owing to underlying disease, comorbidity, and infection in prolonged neutropenia. Unfortunately, it was impossible to identify a subgroup of diagnoses associated with an outstanding benefit from GT therapy in our study. Sepsis-related organ dysfunction coincided with many cases of life-threatening invasive fungal infections and was independently shown to adversely affect the outcome. Individual data reveal that these patients were given GT in quickly deteriorating clinical conditions as ultima ratio. Of note, the main difference of the risk of death between bacterial and fungal infections lies within the first 14 days after GT administration (Fig. 3). In contrast to bacterial infections treated with GT, a proportion of fungal infections were rapidly lethal, suggesting that GTs were insufficient or started too late to affect the outcome of this subgroup. The constant decrement and missing early drop of the survival curve of GT-treated patients with bacterial infections might suggest that infections were eliminated better in this cohort than in those with fungal infections. However, varying underlying conditions between these cohorts might prohibit direct comparison (eg, proportion of patients with prolonged preceding phases of neutropenia, history of invasive mycosis, comorbidity, toxicity). In a very high-risk subgroup of patients (ie, with history of proven infiltrative mycosis and preceding periods of long-term neutropenia), it might be necessary to initiate administration of GT already as secondary prophylaxis.
In summary, these results provide another piece of evidence for the safety and feasibility of GT in children and adults. According to our data, the effect of rapidly and daily available GT over at least 5 days containing a minimum 3×108/kg neutrophils per concentrate to generate a stable ANC increment, shorten neutropenia, and support the control of infections in neutropenic patients with high-risk infections cannot be increased by even higher relative doses of neutrophils. However, very early initiation of GT, or even prophylactic treatment might be required to yield a detectable benefit for infection control in life-threatening invasive fungal infections. Despite expectable obstacles with randomization in life-threatening infectious complications during neutropenia, controlled phase 3 studies are required to confirm the empirically acknowledged benefits of prophylactic and therapeutic GT treatment.
The authors thank Dr S. Karlhuber for excellent documentary assistance and Amgen for travel grants.
1. Price TH. Granulocyte transfusion: current status. Semin Hematol. 2007;44:15–23.
2. Schiffer CA. Granulocyte transfusion therapy 2006: the comeback kid? Med Mycol. 2006;44(suppl):383–386.
3. Bensinger WI, Price TH, Dale DC, et al. The effects of daily recombinant human granulocyte colony-stimulating factor administration on normal granulocyte donors undergoing leukapheresis. Blood. 1993;81:1883–1888.
4. Grigg A, Vecchi L, Bardy P, et al. G-CSF stimulated donor granulocyte collections for prophylaxis and therapy of neutropenic sepsis. Aust N Z J Med. 1996;26:813–818.
5. Grigull L, Pulver N, Goudeva L, et al. G-CSF mobilised granulocyte transfusions in 32 paediatric patients with neutropenic sepsis. Support Care Cancer. 2006;14:910–916.
6. Hester JP, Dignani MC, Anaissie EJ, et al. Collection and transfusion of granulocyte concentrates from donors primed with granulocyte stimulating factor and response of myelosuppressed patients with established infection. J Clin Apher. 1995;10:188–193.
7. Peters C, Minkov M, Matthes-Martin S, et al. Leucocyte transfusions from rhG-CSF or prednisolone stimulated donors for treatment of severe infections in immunocompromised neutropenic patients. Br J Haematol. 1999;106:689–696.
8. Price TH, Bowden RA, Boeckh M, et al. Phase I/II trial of neutrophil transfusions from donors stimulated with G-CSF and dexamethasone for treatment of patients with infections in hematopoietic stem cell transplantation. Blood. 2000;95:3302–3309.
9. Sachs UJ, Reiter A, Walter T, et al. Safety and efficacy of therapeutic early onset granulocyte transfusions in pediatric patients with neutropenia and severe infections. Transfusion. 2006;46:1909–1914.
10. Alavi JB, Root RK, Djerassi I, et al. A randomized clinical trial of granulocyte transfusions for infection in acute leukemia. N Engl J Med. 1977;296:706–711.
11. Vogler WR, Winton EF. A controlled study of the efficacy of granulocyte transfusions in patients with neutropenia. Am J Med. 1977;63:548–555.
12. Stanworth SJ, Massey E, Hyde C, et al. Granulocyte Transfusions for Treating Infections in Patients With Neutropenia or Neutrophil Dysfunction. Cochrane Database of Systematic Reviews. 2005, CD005339. DOI: 10.1002/14651858. CD005339.
13. Worel N, Kurz M, Peters C, et al. Serial granulocytapheresis under daily administration of rHuG-CSF: effects on peripheral blood counts, collection efficiency, and yield. Transfusion. 2001;41:390–395.
14. Adkins DR, Goodnough LT, Shenoy S, et al. Effect of leukocyte compatibility on neutrophil increment after transfusion of granulocyte colony-stimulating factor-mobilized prophylactic granulocyte transfusions and on clinical outcomes after stem cell transplantation. Blood. 2000;95:3605–3612.
15. Dutcher JP, Schiffer CA, Johnston GS, et al. Alloimmunization prevents the migration of transfused indium-111-labeled granulocytes to sites of infection. Blood. 1983;62:354–360.
16. McCullough J, Weiblen BJ, Clay ME, et al. Effect of leukocyte antibodies on the fate in vivo of indium-111-labeled granulocytes. Blood. 1981;58:164–170.
17. Vamvakas EC, Pineda AA. Determinants of the efficacy of prophylactic granulocyte transfusions: a meta-analysis. J Clin Apher. 1997;12:74–81.
18. Hubel K, Carter RA, Liles WC, et al. Granulocyte transfusion therapy for infections in candidates and recipients of HPC transplantation: a comparative analysis of feasibility and outcome for community donors versus related donors. Transfusion. 2002;42:1414–1421.
19. Adkins D, Goodgold H, Hendershott L, et al. Indium-labeled white blood cells apheresed from donors receiving G-CSF localize to sites of inflammation when infused into allogeneic bone marrow transplant recipients. Bone Marrow Transplant. 1997;19:809–812.
20. Freireich EJ, Levin RH, Whang J, et al. The function and fate of transfused leukocytes from donors with chronic myelocytic leukemia in leukopenic recipients. Ann N Y Acad Sci. 1964;113:1081–1089.
21. Akpek G, Knight RD, Wright DG. Use of oral mucosal neutrophil counts to detect the onset and resolution of profound neutropenia following high-dose myelosuppressive chemotherapy. Am J Hematol. 2003;72:13–19.
22. Seidel MG, Peters C, Wacker A, et al. Randomized phase III study of granulocyte transfusions in neutropenic patients. Bone Marrow Transplant. 2008;42:679–684.
© 2009 Lippincott Williams & Wilkins, Inc.