Predictors and Outcomes of Viridans Group Streptococcal Infections in Pediatric Acute Myeloid Leukemia

From the Canadian Infections in AML Research Group

Abstract

Background: 

Viridans group streptococci (VGS) cause significant morbidity in children treated for acute myeloid leukemia (AML). Our goals were to determine the occurrence and impact of these infections in children treated for AML and to understand the factors that increase the risk of VGS infections and viridans streptococcal shock syndrome (VSSS) in this population.

Methods: 

We conducted a retrospective, population-based cohort study that included children ≤18 years of age with de novo AML treated at 15 Canadian centers. We evaluated factors related to VGS infection and VSSS.

Results: 

Among 341 children with AML, VGS occurred in 78 (22.9%) children over the entire course of therapy and 16 had recurrent episodes. VGS infection occurred in 97 of 1277 courses of chemotherapy (7.6%). VSSS occurred in 19.6% of these episodes and included 11 patients who required intensive care services with 2 VGS infections resulting in death. In multiple regression analysis, factors independently related to VGS included treatment on a Medical Research Council-based protocol (odds ratio (OR) 2.87, 95% confidence interval (CI) 1.53–5.39; P = 0.001), cytarabine dose per gram/m² (OR 1.04, 95% CI 1.01–1.07; P = 0.002) and prolonged neutropenia (OR 1.58, 95% CI: 0.97–2.56; P = 0.06). None of the evaluated factors were predictive of VSSS.

Conclusions: 

VGS infections occur in 7.6% of chemotherapy courses and remain an important cause of morbidity and even mortality in children being treated for AML. Interventions to reduce VGS need to be identified.

Outcomes of children treated for acute myeloid leukemia (AML) have improved over the last few decades due to improved risk stratification, hematopoietic stem cell transplantation (SCT) for specific subgroups of patients and enhancements in supportive care.^1–5 The latter remains important because of the risk of morbidity and mortality from infections throughout the treatment of AML.^6–10 Although fungal infections continue to be a major contributor to mortality and are the most difficult to treat in patients with AML,¹¹ bacterial infections including those with viridans group streptococci (VGS) also impact on outcomes of AML treatment.^12–17

VGS is a normal commensal organism in the human oral cavity, the gastrointestinal tract and female genitalia.¹⁸ Pathogenicity may capitalize on the absence of neutrophils as well as oral ulceration caused by chemotherapy for AML.^19–21 Translocation of these organisms into the bloodstream may result in septicemia, add to morbidity and mortality, and may decrease overall outcomes in such patients. The widespread use of empiric and prophylactic antibiotics with restricted Gram-negative activity may have contributed to the increasing frequency of VGS septicemia and pneumonia.²²

Viridans streptococcal shock syndrome (VSSS), a toxic shock-like syndrome, is another manifestation of VGS infection and occurs when bacteremic children have concurrent hypotension and acute respiratory distress syndrome. The incidence of this manifestation in cancer and SCT recipients may be as high as 18% and factors relating to occurrence include peak temperature before start of antibiotic therapy as well as inpatient status.²³ Even with best treatments, mortality rates may be as high as 40–100%.^24,25

While VGS and VSSS are well recognized in pediatric AML, most reports are derived from cooperative group trials or single-center studies. A population-based approach may be a better way to evaluate VGS to minimize the risk of selection bias and to maximize generalizability. Thus, we conducted a retrospective population-based study within a cohort of children treated for AML at participating institutions in Canada. The primary objective was to describe factors associated with VGS in newly diagnosed children with AML. The secondary objective was to describe factors associated with VSSS among children with AML who develop VGS infection.

MATERIALS AND METHODS

This article of VGS and VSSS is a subanalysis of a larger study in which we included patients diagnosed with AML in each Canadian province except Saskatchewan.²⁶ Research ethics board approval was obtained from each participating site.

Study Sample

To examine factors associated with VGS, we included children and adolescents with de novo AML diagnosed between January 1, 1995, and December 31, 2004, who were age ≤ 18 years at diagnosis. We included children with Down syndrome. Exclusion criteria were diagnosis of acute promyelocytic leukemia, secondary AML, previous diagnosis of immunodeficiency and children with Down syndrome receiving only low-dose cytarabine. To examine factors associated with VSSS, we limited the sample to children with VGS isolated from a normally sterile site.

The period at risk for infection extended from initiation of AML treatment until hematopoietic recovery from the last cycle of chemotherapy, conditioning for SCT, relapse, persistent disease or death (whichever occurred first). Data abstraction was conducted by trained clinical research associates who travelled to each site to review the patient charts.

Outcome Measures

For the primary objective, occurrence of VGS sterile site infection was noted over the entire treatment period (excluding SCT) and for each course of chemotherapy. For the secondary objective, the occurrence of VSSS was delineated by the presence of sepsis associated with VGS infection. Sepsis was defined as systemic inflammatory response syndrome in the presence of suspected or proven infection and organ dysfunction according to international consensus guidelines.^27,28 Persistence of infection was defined as a positive culture for VGS from a sterile site ≥72 hours after initiation of appropriate antimicrobial therapy. Recurrence of infection was defined as the reappearance of VGS in the original or another sterile site after discontinuation of organism-specific antimicrobial therapy.

Potential Predictors

The following factors were evaluated to determine their potential association with VGS or VSSS: (1) child characteristics at diagnosis (age, Down syndrome and obese versus nonobese); (2) treatment characteristics [Pediatric Oncology Group (POG), Children’s Cancer Group or United Kingdom Medical Research Council (MRC) protocols, registration on AML trial, diagnosis before January 1, 2000, and cumulative dose of cytarabine in grams/m²] and (3) course characteristics [neutropenia at the start of the course, neutropenia >15 days [threshold chosen a priori), and days systemic corticosteroids were administered for any reason]. Additional factors evaluated for VSSS were maximum temperature, presence of mucositis and receipt of H₂ receptor antagonists or proton pump inhibitor therapy at onset of VGS infection.

Obesity was defined as body mass index percentile ≥95% for age and gender according to the Centers for Disease Control and Prevention for those at least 2 years of age.²⁰ The Children’s Oncology Group protocol AAML03P1³ was classified as MRC-based.

Statistics

To determine factors associated with VGS, analyses were conducted at the course level (in other words, to compare courses with and without VGS) with repeated logistic regression analysis using generalized estimating equations. This approach was used because each child could receive multiple courses of chemotherapy. To identify factors associated with VSSS, analysis was conducted at the infection level also using repeated measures logistic regression with generalized estimating equations. For both analyses, odds ratios (ORs) with 95% confidence intervals (CIs) were used to express the association, and variables significant in univariate analysis were included in a multivariable model. All tests of significance were 2-sided, and statistical significance was defined as P < 0.05. Statistical analysis was performed using the SAS statistical program (SAS-PC, version 9.3; SAS Institute Inc., Cary, NC).

RESULTS

In the overall AML study, there were 341 patients included; their demographics were previously described but in brief, 168 (49.3%) were male and the median age was 7.1 (interquartile range 2–13.5) years.²⁶ Over the course of therapy, VGS occurred in 78 (22.9%) children. There were 13 children with 2 episodes of VGS across 2 different courses and 3 children with VGS across 3 different treatment courses. No child had VGS in 4 or more treatment courses, and no child had recurrence within the same chemotherapy course. There were 1277 courses of chemotherapy administered and VGS occurred in 97 (7.6%). Table 1 illustrates the characteristics of the courses in which VGS occurred.

TABLE 1: Course Characteristics of Children With Viridans Group Streptococcal Infections (N = 97)

Table 2 illustrates the infection characteristics. Almost all children were neutropenic and in this study, all sterile site VGS infections were isolated from the bloodstream. Persistence was rare. In terms of treatment, patients received a median of 19.0 days of intravenous antibiotics [interquartile range (IQR) 15.0–26.0] and a median of 11.5 days of vancomycin [IQR 5.0–17.5]. Nineteen (19.6%) infections were associated with sepsis and thus were classified as VSSS. Eleven children were admitted to the intensive care unit and 2 children died of VSSS.

TABLE 2: Infection Characteristics of Children With Viridans Group Streptococcal Infections (N = 97)

Table, Supplemental Digital Content 1, https://links.lww.com/INF/B705, illustrates the factors associated with VGS. MRC relative to POG protocol, more recent diagnosis, higher cumulative dose of cytarabine and >15 days of neutropenia were all associated with VGS infection. In multiple regression analysis, factors independently associated with VGS were MRC protocol (OR 2.87, 95% CI: 1.53–5.39; P = 0.001), cytarabine dose per g/m² (OR 1.04, 95% CI: 1.01–1.07; P = 0.002) and prolonged neutropenia (OR 1.58, 95% CI: 0.97–2.56; P = 0.06). Importantly, era was no longer significantly associated with VGS infection. Table, Supplemental Digital Content 2, https://links.lww.com/INF/B706, illustrates that none of the examined parameters were predictive of VSSS.

DISCUSSION

Our study demonstrates that 23% of children with AML experienced at least 1 VGS infection and that 8% of courses were complicated by VGS infection during treatment. One in 5 children with VGS had VSSS. Factors associated with VGS infection were longer duration of neutropenia, higher dose of cytarabine and treatment with MRC-based regimens. Two patients died from VSSS in this study.

Unlike the study by Brunet et al²⁹ showing that VGS occurred in 3/78 (3.8%) of pediatric AML with no VSSS deaths, our study is congruent with other studies of pediatric AML that have documented higher rates of VGS infection and sepsis.^{6,12,17,30,31} The factors that we identified as being predictive of VGS infection are consistent with other studies involving leukemia and SCT recipients and include prolonged neutropenia, higher doses of cytarabine and overall intensity of therapy.^{7,19,23,32–35} Younger age was a significant risk factor for VGS infection in a previous study¹² although not in ours. In addition, our results are not concordant with a study by Gassas et al that identified peak temperature before initiation of antibiotic therapy as well as inpatient status as risk factors for VSSS,²³ although we did not evaluate inpatient status in our study. The difference between the 2 studies may be related to the larger sample size of our study or differences in VSSS risk factors between institutions.

Interestingly, patients treated on MRC-based treatment protocols experienced significantly more VGS infections compared with POG-based protocols. While all strategies are intense, they have different approaches to treatment intensification.³⁶ How these different approaches may lead to differences in VGS risk is not well-understood but would provide mechanistic insight into why some patients develop specific infections.

Optimal strategies to prevent VGS and VSSS have yet to be adequately defined. Brunet et al²⁹ credit oral cavity and gut decontamination, vancomycin mouth rinses and empiric systemic vancomycin as the reasons for the low level of VGS infection in their group of pediatric AML patients. In another series, Kurt et al³⁷ also demonstrated a low risk of VGS with prophylactic intravenous vancomycin in children with AML. Use of oral antibiotics in their series was not beneficial.³⁷ These studies highlight the potential role of prophylactic antibiotics including prophylactic vancomycin as an effective approach to reduction of VGS[31] although the trade-offs in terms of toxicities and emergence of antibiotic resistance is not yet fully known.^38–40 Other manoeuvres which may reduce VGS that warrant future investigation are interventions that reduce oral mucositis and central line-associated bloodstream infections.

The strengths of our study include the large number of children with AML, the multicenter nature of the project and the careful collection of both risk factors and outcomes of VGS by trained data abstractors. The population-based aspect of the study ensures that all children with AML were included and avoided selection bias that may occur in cooperative group trials. These factors enhanced the generalizability of our finding and allowed us to evaluate features such as type of protocol therapy. The retrospective nature of the analysis does not create important bias in terms of the endpoints (VGS and VSSS) as they are objective and should be well-documented in the patient chart. However, our results are limited by the varied standards of supportive care that were locally determined and specific to each institution.

In conclusion, VGS infections continue to be an important cause of morbidity and even mortality in patients being treated for AML. Prolonged neutropenia, higher doses of cytarabine and MRC-based chemotherapy are the major determinants of these infections. Effective approaches to reduce VGS and VSSS are needed.

ACKNOWLEDGMENTS

Canadian infections in AML research group: David Dix (PI) and Nita Takeuchi (CRA) from British Columbia Children’s Hospital; Kent Stobart (PI), Brenda Ennis (CRA) and Linda Churcher (CRA) from Stollery Children’s Hospital; Victor Lewis (PI), Janice Hamilton (CRA) and Karen Mazil (CRA) from Alberta Children’s Hospital; Sonia Cellot (PI), Dominique Lafreniere (CRA) and Catherine Desjean (CRA) from Hospital Sainte-Justine; Victoria Price (PI), Tina Bocking (CRA), Lynn Russell (CRA) and Emily Murray (CRA) from IWK Health Centre; Lynette Bowes (PI) and Gale Roberts (CRA) from Janeway Child Health Centre; Carol Portwine (PI) and Sabrina Siciliano (CRA) from McMaster Children’s Hospital at Hamilton Health Sciences; Joseph Beyene (Collaborator) from McMaster University; Mariana Silva (PI) from Kingston General Hospital; Rochelle Yanofsky (PI), Rebekah Hiebert (CRA) and Krista Mueller (CRA) from CancerCare Manitoba; Shayna Zelcer (PI), Martha Rolland (CRA) and Julie Nichols (CRA) from London Health Sciences; Donna Johnston (PI) and Elaine Dollard (CRA) from Children’s Hospital of Eastern Ontario; David Mitchell (PI), Martine Nagy (CRA) and Margaret Hin Chan (CRA) from Montreal Children’s Hospital; Bruno Michon (PI), Josee Legris (CRA) and Marie-Christine Gagnon (CRA) from Hospitalier Universitaire de Quebec; Josee Brossard (PI) and Lise Bilodeau (CRA) from Centre Hospitalier Universitaire de Sherbrooke; Lillian Sung (PI), Biljana Gillmeister (CRA), Marie-Chantal Ethier (CRA), Renee Freeman (Collaborator), Jeffrey Traubici (Collaborator), and Upton Allen (Collaborator) from The Hospital for Sick Children.