The study included 18 C1q deficient patients from 11 referral centres, of whom 14 were previously unreported. In addition, two of four previously reported cases (P15, P16, P17, P18) were updated (Data summarized in Table 1).

Eleven (61%) patients were female. The median age at disease onset was 2.5 years (range, 0.5 months – 9 years). The C1q genetic defect was determined in 17 patients, with mutations in C1QA in 11, C1QB in 4 and C1QC in 2 patients. The most frequent variant was Gln208X in the C1QA gene, present in 6 patients. P14 and P15 were siblings with the same homozygous mutation (c.187 + 1G > T).

All patients demonstrated an autoimmune/autoinflammatory phenotype with a broad spectrum of clinical manifestations: mucocutaneous involvement was reported in all 18 patients in combination with cytopenia in 7 cases (39%), neurologic involvement in 5 cases (28%) and glomerulonephritis in 2 cases (11%). Three patients (17%) had lymphoproliferation-associated disorders, such as lymphadenitis and splenomegaly, and one patient (P6) developed mucosa-associated lymphoid tissue (MALT) lymphoma in the context of Sjogren syndrome. Eight (44%) patients had exhibited symptoms including recurrent fever, arthralgia, and weight loss.

All these disease manifestations resulted in a significant disease burden, that required use of steroids and/or various immunosuppressive treatments leading to important side effects such as osteonecrosis, hypertension, and growth retardation. In 8 patients (44%) FFP infusions were given in conjunction with immunosuppressive drugs. In 6 cases (33%) severe and/or recurrent infections were reported and 7 patients (39%) were receiving antibiotic prophylaxis. Of note, only three patients had history of severe infections with one case of S. pneumoniae sepsis (P3) and two cases of meningitis (P16, P18).

We reviewed 89 patients with genetically confirmed C1Q deficiency (including 14 previously unreported cases from our cohort). To the cohort of 77 patients C1Q analysed by Triaille and colleagues [23], (which already included P2, P3, P15, P17, P18), we have added 12 genetically confirmed C1q deficient patients from our cohort (P1, P4, P5, P6, P7, P8, P9, P10, P11, P12, P13, P14).

Variants were seen in C1QA, CQ1B, and CQ1C in 56%, 12%, and 32% of the 89 patients, respectively.

The most frequent mutations were Gln208X in C1QA reported in 31 cases (35%), Arg69X in C1QC reported in 8 cases (9%) and Gly34Arg in C1QC in 8 cases (9%).

In 69 patients, mutation data were available, specific autoantibodies were tested and main clinical manifestations were recorded. In this subgroup of patients, we investigated if specific genotypes and/or autoantibody subset were associated with an autoimmune-driven CNS or renal involvement.

As in our 18 transplant patients, in this larger cohort the mucocutaneous involvement was the most common manifestation, reported in 62 (90%), and a significant percentage of patients 21 (30%) had neurologic involvement. On the other hand, patients with renal involvement and severe infections were more frequent in this cohort, respectively 16 (23%) and 21 (30%).

We found no association between the three most frequent gene variants (Gln208X, Arg69X, Gly34Arg) and different clinical manifestations. Anti-nuclear antibody (ANA) titres were positive in 65 (94%) of patients, with anti-Ro specificity in 37/69 (54%), anti-Sm in 32/65 (49%), anti-RNP in 22/65 (34%) and anti-DNA in 13/65 (20%).

Analysing different autoantibody specificities, we found that anti-Ro associated with CNS involvement (OR 4.11; IC95% 1.30–13.10) and anti-RNP and anti-DNA with renal involvement (respectively OR 5.69; IC95% 1.72–18.9 and OR 6.09; IC95% 1.66–22.40) (Fig. 1).

Possible markers of disease severity. Data are summarised as the percentage of specific mutations and autoantibodies within four different clinical groups defined according to the presence or absence of CNS and renal involvement: No-CNS, CNS, No-Renal and Renal

Two patients (P3, P13) had two HSCTs, thus 20 HSCTs were performed in total (Table 2). In 17/18 patients the indication for HSCT was the persistence of symptoms despite ongoing treatments. P6 underwent HSCT because of high-risk disease with early development of MALT lymphoma on a background of Sjogren syndrome, quiescent at the time of HSCT (Fig. 2summarises the main baseline features and the overall HSCT outcome).

Clinical features and HSCT outcome. Gray squares represent the presence of a clinical feature/phenotype. Green squares indicate that patients survive after HSCT. Yellow squares indicate that patients had a graft failure. Red squares indicate that patients died after HSCT. For patients P3 and P13 that had two HSCT the outcome of both transplants is indicated. * Recurrent fever, arthralgia and weight loss

The median age at HSCT was 10 years (range 0.9—19 years) with a median time between symptom onset and HSCT of 7.2 years (range 0.8 – 14 years). Different donors were used: 5 (25%) matched sibling donors (MSD), 3 (15%) matched related donors (MRD), 7 (35%) matched unrelated donors (MUD), 1 (5%) mismatched related donor (MMRD) and 4 (20%) mismatched unrelated donors (MMUD). The stem cell source was bone marrow in 10 cases (50%) and peripheral blood stem cells (PBSC) in the remaining cases. Different conditioning regimens were used, both myeloablative and reduced toxicity, based on Treosulfan in 10, Melphalan in 7 and Busulfan in 3 cases.

All patients achieved neutrophil engraftment after a median of 15 days (range 9 – 36 days). The OS in the whole group was 71% (95%CI 44–87%) at 2 years and the EFS was 59% (95%CI 32–78%) at 2 years (Fig. 3 ).

Overall survival (OS) and Event Free Survival (EFS) of the whole cohort. The overall survival at two years was 71% (95%CI 44–87%). For patients who had two HSCT, 2nd HSCT was considered as baseline. The event free survival at two years was 59% (95%CI 32–78%). Event: aGvHD ≥ grade III; disease recurrence due to loss of chimerism; death

Seven patients (39%) developed aGvHD of at least overall grade II, with involvement of skin in three (P1, P6, P13.2), gut in three (P7, P9 and P17) and both skin and gut in two patients (P10, P13.1). Of note, P13 developed aGvHD both after the first and the second HSCT. Only one patient (P6) reported chronic GvHD of the lung (bronchiolitis obliterans).

In 11 patients (61%), HSCT led to resolution of autoimmune features allowing for discontinuation of immunosuppressive treatment (median follow-up time since HSCT 33 months, range 3–84 months). In ten, there was a documented normalization of the function of the classical complement pathway (CH50) and/or of C1q level after HSCT (P10 had no available CH50 and C1q level post-HSCT) (Supplementary Table 1).

Ten patients demonstrated full donor chimerism at the time of last follow up and one (P18) had mixed monocyte chimerism (45%) at 24 months maintaining good CH50 value and disease remission. In this group, 3 patients (P1, P6, P10) developed aGvHD of at least grade II. Two patients (P16 and P18) had Epstein-Barr Virus (EBV) reactivation with consequent development of post-transplant lymphoproliferative disorder (PTLD) in P18, both treated successfully with Rituximab.

After initial engraftment, three patients (P3, P4, P13), experienced secondary graft loss with a recurrence of autoimmunity. Two of them received a reduced intensity conditioning regimen, based on Treosulfan and Fludarabine in P3 and P4 and one received a reduced toxicity conditioning based on Busulfan (total dose received 177 mg/kg; target AUC 60–70 mg*h/L) and Fludarabine in P13.

P4 reached a chimerism of less than 20% 14 months after HSCT with a simultaneous drop of the C1q and CH50 levels and consequent recurrence of malar rash and oral ulcers that required further treatment with hydroxychloroquine and FFP infusions.

The first HSCT of P13 was complicated by grade III acute GVHD involving the skin and gut and by a severe autoimmune pancytopenia requiring treatment with steroids and immunosuppressors. After 146 days, she had secondary graft failure. Due to persistence of pancytopenia (considered as a possible manifestation of the underlying disorder), she underwent a second HSCT 9 months later achieving normalization of complement activity with initial improvement of pancytopenia. However, one month after the second procedure, she developed autoimmune haemolytic anaemia (likely transplant-related considering the persistence of 99% chimerism) that still requires treatment with steroids, cyclosporine, and Rituximab.

P3 had 7% chimerism 3 months after mismatched carrier related-donor HSCT. The nucleated cell dose in the graft was lower than desired (1.9 × 10^8/kg vs 3.0 × 10^8/kg as centre target dose). Despite initial normalization of classical complement function and disease control, 27 months after the HSCT she relapsed with CNS vasculitis (at that time the chimerism was 0%), requiring treatment with high dose of steroids, mycophenolate mofetil (MMF) and Rituximab. Considering the severity of the disorder, a second HSCT was attempted 4 years after the initial transplant. Despite establishing neutrophil engraftment, she developed progressive and irreversible respiratory failure secondary to aspergillus pneumonia and died 32 days after HSCT. Four other patients (P7, P9, P12 and P17) died after establishing neutrophil engraftment: a 13-year-old girl (P7), 3 months after HSCT, with multiorgan failure (MOF) secondary to transplant-associated thrombotic microangiopathy (TA-TMA), gastrointestinal GVHD (grade IV) and acute respiratory distress syndrome due to Methicillin-Resistant Staphylococcus Aureus (MRSA) pneumonia; a 16 year old girl (P9) with encephalopathy due to idiopathic hyperammonaemia after acute gastrointestinal GVHD (grade III); a 1 year old boy (P12) with respiratory failure secondary to cytomegalovirus (CMV) pneumonia; and a 9-year-old boy (P17), 4 months after HSCT, with MOF due to gastrointestinal acute GVHD (grade II) occurring after lymphocyte infusion for EBV-PTLD [20]. Of note, P3, P7 and P17 had a severe underlying disorder with neurologic involvement. Additionally, P7 had glomerulonephritis (grade IV) with active proteinuria and pulmonary hypertension at the time of HSCT.

At the time of HSCT, P9 was 16 years old and exhibited severe cutaneous and musculoskeletal involvement causing a very low performance status (Lansky score 30). After HSCT, she developed mood disorders with fluctuation in the level of consciousness secondary to idiopathic hyperammonaemia. At that time the chimerism was 95% (C1q and CH50 level not available). Due to the subsequent rapid deterioration of the neurological picture to death, cerebral magnetic resonance imaging (MRI) was not performed, and the cause of the encephalopathy remained undetermined. Underlying disease-related CNS involvement cannot be excluded given the absence of pre-transplant brain imaging.

Even though HSCT was performed at an early age before the development of organ damage, P12 died of CMV pneumonia. CMV serostatus was positive in the recipient and negative in the donor, and a CD34 + selected graft was used.

Overall, 5 patients (28%) had a baseline neurologic and/or renal involvement, both clinical markers of disease severity. As summarized in Fig. 4, the OS at 2 years in this subgroup was lower in comparison with the OS in the subgroup of patients without these complications (40% vs 84%; p = 0.034). We did not find any significant difference in the EFS between the two groups (60% vs 59%; p = 0.596)

Comparison of overall survival and of event free survival between patients with severe and no-severe baseline disease. The presence of neurological and/or renal involvement were considered as markers of severe disease. Patients in severe group had worst OS (40% vs 84%; p = 0.034), while there was no difference in EFS at two years (60% vs 59%; p = 0.596) between the two groups. In the overall survival analys for the patients who had two HSCT, 2nd HSCT was considered as baseline