Absence of identifiable granulopoiesis at diagnosis of childhood acute lymphoblastic leukemia conveys an extremely high risk of early blood culture‐positive infection

Infectious complications remain the most common cause of treatment-related mortality (TRM) among children with acute lymphoblastic leukemia (ALL).1, 2 The infection-related TRM rate in childhood ALL is 2%–4% in high-income countries, and susceptibility to infection peaks during induction chemotherapy.1-4 Standard induction chemotherapy will likely remain largely unchanged for the foreseeable future, although immunotherapies such as blinatumomab are expected to markedly alter the therapeutic landscape and toxicity profile of pediatric precursor B-ALL. This underscores the continued importance of reducing induction-related mortality in the era of immunotherapy.5 While patients with T-ALL are generally considered to have a high infectious burden, we observed that children with precursor B-ALL experienced more infectious complications during induction therapy. We therefore hypothesized that a higher degree of leukemic blast infiltration in the bone marrow (BM) in precursor B-ALL patients lead to greater suppression of normal granulopoiesis, resulting in an increased risk of infection during induction chemotherapy until BM recovery.6 Among children with ALL, 50%–70% of episodes of fever of unknown origin (FUO) are ultimately confirmed as bacterial infections.1, 7, 8 The appropriate duration of empirical antibiotic treatment (EAT) for FUO and the criteria for its discontinuation remain subjects of ongoing debate. In 2023, the International Pediatric Fever and Neutropenia Panel published an updated clinical practice guideline (CPG) on EAT cessation, recommending continuation of EAT in high-risk patients until BM recovery.9 In contrast, the European Conference on Infections in Leukaemia (ECIL) guidelines provide a level B recommendation to discontinue EAT in high-risk patients, including those undergoing induction therapy for ALL, if the patient has been afebrile for more than 48 h and antimicrobial therapy has lasted at least 72 h, irrespective of neutrophil count.7, 10 In this retrospective, single-center study, we sought to identify factors predisposing to blood culture–positive infections during induction chemotherapy for childhood ALL. All 73 patients with ALL who were under 16 years at diagnosis and were treated according to the NOPHO ALL-2008 protocol between 2008 and 2020 at Oulu University Hospital were included. Four patients with Down syndrome and one patient who died before initiation of chemotherapy were excluded from the analysis. All patients received a 4-week, three-drug induction regimen consisting of vincristine, daunorubicin, and either prednisolone or dexamethasone. Data were collected from the NOPHO ALL-2008 registry and the patients' medical records. Data on infection onset, neutrophil counts, blood culture pathogens, and EAT discontinuation were extracted manually (details in Supporting Information S1: Appendix 1). BM morphology at diagnosis was reviewed to determine whether any normal granulopoiesis could be identified on the slides. Statistical analyses were performed using IBM SPSS Statistics, version 29. Demographic data were compared between groups using Fisher's exact test, the chi-square test, or the independent samples t-test, as appropriate. Multivariable logistic regression was used to assess the effects of total absence of granulopoiesis (TAG) at diagnosis, neutrophil count, glucocorticoid type, and continuous antibiotic therapy until BM recovery on blood culture–positive infections. Ethical approval was obtained from the Regional Ethics Committee of the Northern Ostrobothnia Hospital District, Finland. Informed consent was obtained from all patients or their legal guardians prior to participation. Twenty patients (27%) developed a blood culture–positive infection during induction therapy, which is comparable to the rate previously described with the same treatment protocol.11 A total of 61 of 73 patients (84%) received intravenous antibiotics during induction due to clinical indications (Supporting Information S1: Table 1). No patient received primary antibiotic prophylaxis. Three patients with blood culture–positive infections required admission to the intensive care unit, but no deaths occurred during induction in this cohort. Eleven patients presented with TAG at diagnosis, while the remaining 62 patients had identifiable granulopoiesis (Table 1). No significant differences in age, sex, or final risk stratification were observed between groups, but patients with TAG had significantly lower hemoglobin levels and a higher degree of blast infiltration in the BM. Patients with precursor B-ALL indeed had a higher mean percentage of blasts in the BM than patients with T-ALL (92% vs. 71%, P < 0.001). All patients with TAG had precursor B-ALL (P < 0.05) and were treated according to the non–high-risk induction arm with prednisolone. The proportion of patients with an undetectable neutrophil count at diagnosis was significantly higher in the TAG group (P < 0.05), but no difference was observed between groups at day 15. A substantial proportion (64%) of TAG patients developed a blood culture–positive infection. Two of these had a positive culture already at diagnosis, with Pseudomonas aeruginosa and Staphylococcus aureus. Microbiological findings from the 20 positive blood cultures are shown in Figure 1. Distribution of the pathogens were similar between groups. Seventeen of the 20 infections (85%) occurred before Day 18 of induction therapy. During the first 2 weeks of induction, TAG at diagnosis was the strongest predictor of blood culture–positive infection (6/11 vs. 7/62; RR 4.8, 95% CI 2.00–11.67, P < 0.001). Between Days 16–29, there were more infections in patients with absolute neutrophil count (ANC) < 0.1 × 10⁹/L on Day 15 (6/35 vs. 1/38; RR 6.5, 95% CI 0.82–51.45, P = 0.075), but the difference was not statistically significant. Two of the 20 blood culture–positive infections led to temporary discontinuation or significant reduction of induction chemotherapy; both patients remain in first complete remission following completion of protocol treatment. All patients with TAG received antibiotic treatment during induction, six of whom were started on antibiotics (all with Gram-negative coverage) already at ALL diagnosis. In four of these, antibiotics were discontinued before evidence of BM recovery; three subsequently developed blood culture–positive infections, and one developed blood culture–negative fever. The two patients with TAG who received antibiotics until BM recovery did not experience recurrent fever during induction treatment. Overall, 17 of 61 patients who received antibiotics for FUO during induction had EAT discontinued before BM recovery, of whom twelve (71%) were readmitted with fever, and seven (41%) had blood culture–positive infections. The mean latency was 5.5 (SD 5.1, range 1–14) days from EAT discontinuation to blood culture–positive infection, and 3.0 (SD 2.3, range 1–5) days from EAT discontinuation to blood culture–negative fever. For the multivariable logistic regression analysis, a forward stepwise approach was used to find the optimal model. The final multivariable model included TAG, uninterrupted antimicrobial therapy until BM recovery (ANC ≥ 0.1 × 10⁹/L), and glucocorticoid type (dexamethasone/prednisolone), but not ANC. In the final model, TAG (adjusted odds ratio [aOR] 10.6, 95% CI 2.1–55.5, P < 0.01) and uninterrupted antimicrobial therapy until BM recovery (aOR 0.22, 95% CI 0.05–0.91, P < 0.05) were significantly associated with blood culture–positive infection risk, whereas dexamethasone was not (aOR 0.44, 95% CI 0.11–1.79, P = 0.25). Here, we describe that TAG in BM at the time of childhood ALL diagnosis is a major risk factor for blood culture–positive infection during induction. All patients with TAG in our study had precursor B-ALL and were finally stratified into standard- or intermediate-risk groups. This finding highlights the need for optimized supportive care in this subgroup, which otherwise represents a favorable prognostic population, in whom minimizing TRM is crucial. We have previously described a patient included in the present cohort with TAG at B-ALL diagnosis and an extremely problematic disseminated P. aeruginosa-infection following early discontinuation of EAT, serving as an example of the importance of uninterrupted antibiotic therapy in such patients.12 Currently, there is no consensus on when to discontinue antibiotics in high-risk patients. The CPG by Lehrnbecher et al. published in 2021 recommends continuation of EAT until BM recovery, while ECIL guidelines allow discontinuation in high-risk patients, including those receiving induction therapy for ALL.9, 10 Trials by Clement et al. and Klaassen et al. have reported higher readmission rates due to infection when antibiotics were discontinued before neutrophil recovery.2, 13 There is also clinical data indicating that ANC < 0.1 × 10⁹/L at the time of EAT discontinuation and ongoing induction therapy are risk factors for reinfection.14 Based on the present cohort, discontinuing EAT during ALL induction before neutrophil recovery is not safe. In our institutional practice, tailored ambulatory infusion pumps with weight-based dosing are used for in-home intravenous antibiotic administration, facilitating safer early discharge.15 The Infectious Diseases Society of America and a multinational CPG panel do not routinely recommend antibiotic prophylaxis for patients undergoing ALL induction, emphasizing the need to define subgroups who might benefit from prophylaxis.3, 16 Given the 100% infection rate and 64% blood culture positivity observed among patients with TAG at diagnosis, antibiotic prophylaxis during induction may be justified in this subgroup. Such an approach is unlikely to significantly increase antimicrobial resistance, as these patients appear to require antibiotics regardless. Limitations regarding the present study arise from possible interobserver variability in BM smear evaluation, although done by experienced hematopathologists. Validation of TAG as a risk factor for early sepsis should also be pursued in a larger cohort. Our key finding is that TAG at ALL diagnosis, enriched in patients with favorable risk B-ALL, renders patients severely immunocompromised already before chemotherapy and represents an exceptionally strong risk factor for infection during induction. We would like to thank all the families who participated in this study. Open access publishing facilitated by Oulun yliopisto, as part of the Wiley - FinELib agreement. Elli-Maija Ukonmaanaho: Conceptualization; methodology; resources; writing—original draft; visualization; writing—review and editing; data curation; formal analysis. Riitta Niinimäki: Conceptualization; funding acquisition; resources; writing—review and editing. Kimmo Halt: Resources; data curation. Henri Aarnivala: Conceptualization; investigation; methodology; resources; supervision; visualization; writing—original draft; writing—review and editing; data curation; formal analysis. The authors declare no conflict of interest. The study was performed in accordance with the Declaration of Helsinki and approved by the ethics committee of the Northern Ostrobothnia Hospital District. No external funding was received for this study. The data that support the findings of this study are available from the corresponding author upon reasonable request. Anonymized data that support the findings of this study are available from the corresponding author upon reasonable request. Supporting information Additional supporting information can be found in the online version of this article. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.