Chronic granulomatous disease is caused by mutations leading to defects in individual subunits of the phagocyte NADPH-oxidase (gp91phox in X-linked-; p22-, p47-, p67-, p40phox, and EROS in autosomal recessive-CGD) (1–4). The NADPH-oxidase-myeloperoxidase system generates microbicidal oxidants required for host defense and control of inflammation. CGD affects ~1:200,000–250,000 live-births (5–7) and X-linked-CGD accounts for approximately two-thirds of patients. P47phox-mutations are the most common AR-defects. Rarely, female carriers of X-CGD with random X-lyonization of <15% of circulating NADPH-oxidase-producing neutrophils present with CGD-symptoms (7–9). Symptoms comprise of invasive infections and chronic autoinflammatory diseases leading to frequent medical interventions, impaired quality-of-life, and increased morbidity/mortality (10–13). The majority of patients are diagnosed in childhood, while some develop symptoms in adulthood (7, 14, 15). Due to residual NADPH-oxidase activity, patients with AR-p47phox-mutations survive longer than X-CGD-patients (survival >40 years: >80 vs. 55%) (16). The clinical course may be unpredictable even in individuals of families with identical CGD-mutations (17). Short stature, osteoporosis, organ failure, and amyloidosis are long-term complications (18). There is still paucity of data on quality-of-life and emotional health in patients with CGD (11, 12, 14, 19, 20). Today, 90% of children with CGD are reaching adulthood and the transition into adult care is challenging (21, 22).

The infections typically affect lungs, lymph nodes, skin, liver, perianal region, gingiva and bone and are mainly caused by Staphylococcus aureus, Burkholderia cepacia, Nocardia, Serratia marcescens, and Aspergillus species. Klebsiella pneumoniae, Salmonella (7, 23), Mycobacteria (21, 24), Actinomyces, Granulibacter bethesthensis (25–27). Infections caused by Chromobacterium violaceum and B. pseudomallei (28–30) are less frequently encountered. The use of life-long antibacterial prophylaxis with trimethoprim-sulfamethoxazole is recommended. Pulmonary Aspergillus-infections are the leading cause of mortality (31). Anti-fungal prophylaxis, mainly with itraconazole (32, 33), can reduce the incidence of fungal infections, but the emergence of azole-resistant aspergillus species and dematiaceous molds is becoming a clinical challenge (34).

Absent or reduced NADPH-oxidase activity in monocytes/macrophages causes impairment of efferocytosis and autophagy (35, 36). Ineffective apoptotic cell clearance increases the risks of developing autoinflammation (37). Progressive granulomatous lung disease (PGLD), Crohn-like enterocolitis (38, 39) and obstructive genitourinary inflammation (40, 41) are relevant autoinflammatory syndromes and their risk increase steadily during life (14). Initial gastrointestinal involvement without infections has often been misdiagnosed as Crohn‘s disease (38, 39). Treatment of autoinflammation includes steroids (42) and more recently IL1- or TNF-alpha inhibitors to replace steroids, however, all of these drugs bear the risk of increasing the risks for invasive infections (43). Thalidomide (44–46), vedolizumab, ustekinumab (47, 48), as well as pioglitazone (36) can be beneficial to reduce autoinflammation in CGD and regular IFN-gamma injections decrease the incidence of bacterial infections with no impact on the incidence of colitis (21, 49–55).

Before HSCT, screening for infections is mandatory in biopsies of infectious lesions and in bronchoalveolar lavage specimens (14, 56). Steroids added to antimicrobials can accelerate the regression of infectious lesions (42, 57, 58) and can help to avoid extensive surgery (59). Granulocyte-transfusions should be strictly indicated to prevent CMV-transmission and sensitization to blood cell antigens (60–62). The McLeod-blood group should be evaluated in X-CGD-patients to minimize the sensitization against Kx-positive red cell transfusions (63–65).

In Europe, the first major survey of the Inborn Errors Working Party of the EBMT reported on 27 patients with CGD who had been transplanted between 1985 and 2000. At HSCT, nine of 27 patients had intractable invasive infections and received antibiotics as well as granulocyte-transfusions (seven of nine). Eighteen of 27 patients were free of infection at HSCT. Seven of the 18 patients without overt infection had signs of active ongoing autoinflammation including enterocolitis and PGLD. Twenty-five of the 27 patients received MSD-transplants (five heterozygous carriers). Two patients with no overt infection or autoinflammation received a MUD-transplant. Conditioning-regimens were mainly myeloablative with full-dose busulfan/cyclophosphamide and mainly without serotherapy (67). Recovery from refractory infection, remission of inflammatory organ dysfunction and catch-up growth were observed (67). Patients without overt autoinflammation/infections had an OS of 100%, whereas patients with ongoing infections at transplant had a TRM of 44% (four of nine) (67). The OS/EFS was 85/81%, respectively. The GF and chronic GVHD rate were 7 and 11%, respectively. The majority of surviving patients had >95% circulating myeloid cells of donor origin. This important paper showed that myeloablative HSCT based on busulfan/cyclophosphamide and no in vivo T-cell depletion was overall efficient in sibling transplants but induced exuberant inflammation in patients suffering from ongoing infections at transplant. The same was observed in a transplantation model in non-infected CGD mice after myeloablative allogeneic HSCT resulting in marked infiltration of the lungs with inflammatory cells, in contrast to normal mice (81). Cultured monocytes from the CGD-mice produced 3-fold TNF-alpha (81), explaining the higher incidence of severe GvHD in patients with pre-existing overt infections treated with HSCT without serotherapy. Myeloablative regimens containing cyclophosphamide were greatly abandoned in Europe after this experience. The authors at that time concluded that all infectious/inflammatory foci had to be detected and treated before HSCT and that HSCT should be mainly restricted to children with MSD/MFD (67).

Almost simultaneously to the above mentioned European experience, the NIH in the USA used for the first time a reduced intensity conditioning (RIC) comprising of non-myeloablative fludarabine/cyclophosphamide followed by in-vitro T-cell depleted grafts. This approach resulted in clearly increased GF-rates (20%), even with the use of matched family/sibling donors (66). Donor-lymphocyte infusions were necessary to prevent falling DC but unfortunately induced severe acute GVHD and resulted in a transplant-related mortality rate of 30% (66, 82). RIC-regimens including melphalan and fludarabine were associated with similarly high GF-rates (30%) (75).

RIC-regimens based on reduced or targeted busulfan, fludarabine and serotherapy were more successful and achieved sufficient myeloablation and clearly lower rates of GF and chronic GVHD (38, 63, 83–85). These busulfan-fludarabine-based RIC-regimens were first used in adult high-risk CGD-patients suffering from invasive Aspergillus-infections and/or enterocolitis using MSD/MFD- or MUD transplants. The OS/EFS rates were 100% in these small initial series (38, 84). Administration of anti-T-cell/thymocyte globulins as well as of a humanized monoclonal anti-CD52 antibody (Campath IH; alemtuzumab) were shown to deplete successfully T-cells and allo-stimulatory dendritic cells (86) of recipient origin. The importance of using serotherapy for in-vivo T-cell depletion to reduce both GF and chronic GVHD after HSCT for CGD became obvious. Viral reactivations after serotherapy were fortunately rare or well manageable rendering clinical HSCT outcomes with MUD-donors vastly similar to MSD/MFD-donors (68, 69, 71).

Busulfan-based RIC-conditioning was further refined by investigating the interindividually variable busulfan clearance and exposure in patients (87, 88). Therapeutic drug monitoring (TDM) helped optimize both safety and efficacy of busulfan-administration. The assessment of the cumulative AUC (cAUC) turned out as an appropriate tool to measure the total busulfan-exposure and -toxicity (87, 89). A 10-year prospective study on 56 pediatric/adult CGD-patients (2/3 high-risk patients) treated with submyeloablative busulfan (half-dose or cAUC 45–65 mg/L × h) yielded, indeed, excellent results. Busulfan-dose adjustments (90) were necessary in 14/44 patients (32%) (63). Immunoablation was achieved with fludarabine and serotherapy including rabbit ATG or alemtuzumab. After a follow-up time of 21 months, the OS/EFS-rates were 93 and 89%, respectively. However, GF could not be abolished and occurred in 5% of patients. The cumulative incidences of grade III–IV acute GVHD and chronic GHVD were low with 4 and 7%, respectively. Stable ≥90% myeloid DC was documented in 93% surviving patients leading to resolution of infectious and inflammatory lesions. Equivalent outcomes were observed between MFD/MSD and MUD rendering matched unrelated donors a good donor choice in the absence of matched sibling donors. Outcomes were not different between 9/10-HLA- (n = 10) and 10/10-matched MUD (n = 25), but the numbers were low. Two fatherhoods were documented after successful HSCT. To further reduce the risk of graft failure with this RIC-regimen, some investigators have narrowed the submyeloablative target of the cumulative AUC of busulfan to 55–65 mg/L × h (83) and have started using busulfan starting doses based on a new body weight-dependent busulfan dosing nomogram (91).

Morillo-Gutierrez et al. (74) showed in a large retrospective European study of the EBMT on 70 CGD-children that HSCT after treosulfan-based conditioning was well tolerated and achieved OS/EFS-rates of 91.4/81.4%, respectively. Treosulfan, an alkylating drug with both myeloablative and immunosuppressive effects, exhibited an overall low acute toxicity in CGD transplants. If used as a single alkylator, treosulfan may be less gonadotoxic than other alkylators, however, there is no study yet available convincingly proving this assumption (92–94). Excellent myeloid DC (>95%) was documented in 80% of surviving patients. With this paper, treosulfan-based RTC was shown to be an alternative conditioning to targeted busulfan-based-RIC, although it remained unclear which treosulfan systemic exposure was more likely to be myeloablative or submyeloablative. Graft failure remained a problem occurring in 12% of the patients (74). Some centers have therefore started to add thiotepa to treosulfan to further reduce the risk of GF (80), probably at the expense of augmented gonadotoxicity (94).

The experience with unrelated 4/6–6/6-HLA-compatible cord blood transplants (CBT) in CGD is scarce, but there are a few examples of successful transplants using cord blood in patients lacking MSD or MUD (72, 95, 96). Due to low HSC-numbers in CB, CBT is usually restricted to patients with low body weight (<20 kg) and viral reactivations may be of concern. CBT usually requires myeloablative and therefore more gonadotoxic conditioning, e.g., busulfan (cAUC 80–100 mg/L × h) or treosulfan/thiotepa, to achieve sufficient myeloid engraftment.

For this review, we have analyzed the results of the two above mentioned major European studies together with 15 other relevant international papers published between 2010 and 2019. We have summarized the results of 386 CGD-patients receiving transplants from mainly MSD/MFD- and MUD-donors in Table 1. The median overall incidences of OS, EFS, graft failure, chronic GVHD and de-novo autoimmune disease in these 17 papers were 92, 81, 14, 9, and 15%, respectively. The most important secondary problems were graft failure including patients with slowly falling myeloid DC <10% (DHR/NBT-tests <10%), de-novo autoimmunity and chronic GVHD. Graft failure or low donor myeloid DC was associated with reappearance of CGD associated symptoms, and chronic GVHD clearly impacted negatively on quality-of-life and life expectancy (Table 1).

Hoenig et al. demonstrated for the first time that haploidentical HSCT was curative CGD (97). They used myeloablative conditioning (full-dosed busulfan, thiotepa and alemtuzumab) and in-vitro selected peripheral HSCs and achieved full donor donor cell engraftment and complete resolution of pulmonary aspergillosis. More recently, haploidentical TCR alpha-beta -/CD19-depleted grafts were shown to successfully achieve myeloid donor cell engraftment without inducing relevant GVHD (103–105). The advantage of these in-vitro T-cell depletion techniques is that chronic GVHD is rare (14, 80, 106). In-vivo T-cell depletion strategies in haploidentical transplants include the use post-transplant cyclophosphamide (PT/CY) (50 mg/kg/day), administered on day+3 and day +4 (107). PT/CY is non-toxic to donor HSCs, but efficiently eliminates activated alloreactive donor-derived CD3+T-cells while sparing resting CD3+T-cells with potential anti-infective properties. The first successful haplo HSCT with PT/CY in CGD was reported in the USA after the administration of targeted busulfan (cAUC 40 mg/L × h), fludarabine, cyclophosphamide and 2 Gy TBI (98). However, in a very recent follow-up paper by Parta et al. on seven patients with CGD a rather high rate of severe GVHD was observed leading to death in two patients (OS and EFS 71%, respectively). The estimated total cumulative of busulfan ranged from 30–52 mg/L × h (2,461–4,250 min × micromol/L × 3 days). They used a protocol with mainly PBSC grafts and sirolimus for GVHD-prophylaxis (101). Patients' age ranged between 14 and 26 years and comprised of mainly adults. Severe grade III acute GVHD were observed in three patients with enterocolitis.

Another currently investigated RTC-protocol is currently used in our institution. It comprises of up-front rabbit ATG (30–40 mg/kg), fludarabine (180 mg/sqm) and targeted busulfan (cAUC 65–75 mg/L × h) followed by haplo-HSCT with PT/CY and GVHD prophylaxis with CSA and MMF (starting at day +5). We believe that both haploidentical HSCT with PT/CY and with antibodies containing magnetic beads are promising alternatives in high-risk patients with CGD when HLA-matched related or unrelated donors are unavailable (105). To further explore the rates of graft failure and cGVHD after haploidentical HSCT comparative studies of both techniques are urgently needed (Table 2).

Autologous gene-therapy (GT) of HSCs leads to partially functional correction of defective phagocytes and is a potentially curative treatment approach in CGD. Graft failure may occur after GT, but the risk of GVHD is zero (102, 108, 109). While early studies with unconditioned transfusions of retrovirally transduced HSCs were unsuccessful (110), autologous infusion of HSCs transduced with a gamma-retroviral vector after busulfan-based myeloablative conditioning helped to successfully engraft 4 CGD-patients (2 adults, 2 children) (109). Approximately 15% of gp91phox-expressing neutrophils had been detectable within the first 5 months after GT leading to resolution of life-threatening invasive fungal infections. Unfortunately, methylation with downregulation of the transduced gene and clonal expansion of transduced myeloid cells due to random viral integrations were observed, leading to activation of endogenous oncogenes and development of MDS with or without monosomy 7. Both children treated with GT survived after subsequent allogeneic HSCT (65), while 2 adult patients died due to secondary MDS and AML, respectively. Recently, nine X-CGD-patients (age 2–27 years) received GT using a self-inactivating lentiviral vector designed to limit the risk of mutagenesis (102). Patients were pretreated with myeloablative busulfan exposures (cAUC 70–75 mg/L × h). Two patients died within 3 months from GT due to severe pulmonary disease and hemorrhage. At 12 months, 6/7 surviving patients demonstrated persistence of sufficiently NADPH-oxidase-expressing neutrophils (16–46%) and stable vector copy numbers. One patient had graft failure with a decline <5% enzyme-producing neutrophils. There was no evidence of clonal dysregulation or transgene silencing. Surviving patients did not develop new CGD-related infections, and six have been able to discontinue antibiotic prophylaxis (OS/EFS >12 mo.: 78/66%, respectively) (Table 2).

Traditionally, indications for HSCT in CGD had been the following: (1) > 1 invasive life-threatening infections, (2) non-tolerability of prophylactic drugs, (3) non-compliance, (4) severe autoinflammation, or (5) unavailability of a CGD-experienced physician (14, 21, 22, 111). Due to the above mentioned favorable results, there is nowadays agreement that HLA-matched HSCT is indicated in any CGD-patient with absent NADPH-oxidase enzyme activity (16). Small children with CGD may clearly benefit from 5/6- or 6/6-HLA-matched CBT in experienced centers. Less than 10/10-HLA-MUD should probably not be offered to asymptomatic CGD-patients since the rates of graft failure and chronic GVHD are higher than in completely matched transplants. The indication for HSCT in adults should be carefully assessed by the treating physician, although the results in recent years have been encouraging (38, 63, 67, 78, 84, 112). We believe that haploidentical transplants and GT in X-CGD should only be offered to high-risk CGD-patients suffering from severe infectious and/or autoinflammatory complications with no other treatment alternatives. Ideally, high-risk CGD-patients without matched donors should be prospectively investigated in trials comparing GT vs. haploidentical HSCT.

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The handling editor AL declared a collaboration with the authors RC, TG.