Study of Complications of Varicella-Zoster Virus Infection in Hospitalized Children at a Reference Hospital for Infectious Disease Treatment
OBJECTIVES: Varicella is a disease with potentially severe complications. We aimed to investigate characteristics of hospitalized children with varicella in Brazil in the prevaccine period and to identify predictors for requiring intensive care treatment.
METHODS: A prospective cohort study was conducted from May 2011 to April 2014. Patients up to 13 years of age with varicella diagnosis were included. Information was collected through interview and review of medical records. Logistic regression analysis was performed.
RESULTS: A total of 669 patients were admitted. The median age of subjects was 2.7 years (range 0–14 years) with a predominance of boys (56.6%). The main causes of hospitalization were bacterial complications (77.7%), viral complications (11.4%), and at-risk patients (10.9%). Main bacterial complications were skin infection and pneumonia. Main viral complications were herpes zoster, cerebellitis, and encephalitis. Most at-risk patients used corticosteroids or had a diagnosis of leukemia. At-risk patients were hospitalized earlier (P < .01) and remained hospitalized for longer periods (P = .03). A total of 44 patients (6.6%) were admitted to the ICU, and 5 (0.8%) died of septic shock. Thrombocytopenia was associated with more severe illness in patients with bacterial infections (P = .001). The long-time interval between onset of infection and admission was associated with the need for intensive care in all groups (P = .007).
CONCLUSIONS: Secondary bacterial infection is the main cause of hospitalization, and thrombocytopenia in these patients leads to worse outcomes. Difficulties of access to the health system and delay in medical care are determining factors of greater severity in this population.
Varicella-zoster virus (VZV) is the etiological agent of varicella or chickenpox, an infectious disease of childhood characterized by fever and exanthema. Infection occurs through direct contact with patient’s lesions or by viral transmission through respiratory tract; in most cases the diagnosis is based on history of exposure and description of typical vesicular rash.1,2 VZV establishes latency in neural tissue after primary infection and can be reactivated form dorsal root ganglia resulting in herpes zoster.1,2 High population density and crowding are indicated as main factors for viral transmission.3–5
Despite the public perception of varicella infection as a harmless childhood disease, it can have potentially severe complications that require hospitalization. The most common complication of varicella is bacterial superinfection of skin lesions caused most often by Staphylococcus aureus and group A Streptococcus.5 In some cases, VZV can also affect the central nervous system (CNS), resulting in encephalitis, meningitis, or cerebellitis. Hematologic, pulmonary, gastrointestinal, and osteoarticular complications are also described.5 Some patients, such as adolescents, adults, pregnant women, oncological patients, and newborns, may present symptoms of the disease more severely and have a higher incidence of complications.1,3,6,7
Hospitalization rates due to chickenpox complications are considerably higher among children in developing countries.8–10 From 2008 to 2012, 45 495 hospitalizations for varicella were recorded in the Brazilian Hospital Inpatient System, with an annual average of 9099 cases.4 Varicella hospitalizations generate substantial cost for the health care system, either through the treatment of complications or as indirect costs as a result of parent workdays lost while caring for their sick child.6,11,12
In Brazil, universal vaccination against varicella began in September 2013 through the National Immunization Program. A single dose is administered at the age of 15 months.4 Universal vaccination is likely to change the morbidity of varicella in Brazil, but baseline data are required to evaluate the impact of the vaccination program. In this study, we attempt to investigate clinical and epidemiologic characteristics of hospitalized children with VZV infection in Brazil in the prevaccine period. As a secondary outcome, we aim to identify clinical and laboratory predictors for requiring intensive care treatment during hospitalization.
This was a prospective cohort study of children hospitalized for VZV infection at our 100-bed, freestanding children’s hospital in Belo Horizonte, Brazil, between May 2011 and April 2014. The hospital offers free public health care service for patients between 1 month and 13 years of age and is a reference for the treatment of pediatric infectious disease. The diagnosis of VZV was made according to clinical and epidemiologic data obtained by a physician at patient admission. We included patients ≤13 years of age, admitted to the infectious disease pediatric ward during the study period with any stage of varicella or herpes zoster, or admitted with a VZV-related event occurring within 3 weeks after the onset of VZV symptoms. An informed consent form was signed by legal guardians of participants under 7 years of age and an assent form was obtained for participants over age 7 years. Those who refused to provide informed consent were excluded from the study.
For each included inpatient who met the study inclusion criteria, a standardized questionnaire was prospectively completed through interviews, and information was collected from hospital charts. Relevant data included age, sex, the presence of underlying chronic disease, previous contact with chickenpox, duration of varicella symptoms, vaccination status, reason for hospitalization, treatment, patient outcome, and laboratory test results such as platelet counts, hemoglobin level, and blood culture. Low platelet count was defined as ≤150 000 platelets per µL. Anemia was defined as hemoglobin level less than the fifth percentile for age.
Patients were stratified into 3 groups according to reason for hospitalization: (1) hospitalization with secondary bacterial complications, (2) hospitalization with complications related to VZV infection, and (3) hospitalization of at-risk groups. Complications associated with VZV infection were defined as herpes zoster or events occurring within 3 weeks after varicella that might be related to viral infection.13 Herpes zoster was included because it is a recurrent manifestation of viral infection. At-risk patients were defined as those with congenital or acquired immunodeficiency, neoplasia, use of oral or inhaled corticosteroid for at least 15 days, and use of immunosuppressive agents in the 30 days before varicella onset.
Admission to the ICU during hospitalization occurred because of the need for advanced airway support, presence of hemodynamic instability, or CNS depression. Variables associated with ICU admission were analyzed for the entire study group and also the subgroup of patients hospitalized because of secondary bacterial complications. Because of the small number of patients in the other 2 subgroups, a comparative analysis was not performed. Variables considered for analysis for the entire group were age, sex, household contact, symptom duration before hospitalization, and reason for hospital admission. Variables considered for analysis in patients hospitalized because of secondary bacterial complications were age, sex, household contact, laboratory findings, and site of bacterial infection.
We used SPSS version 12.0 (SPSS Inc, Chicago, IL) to perform the analysis. For descriptive analysis, we determined frequency, percentage, and measures of central tendency. Demographic and clinical characteristics were compared between groups according to reason for hospitalization by using χ2 or Fisher’s exact test for categorical variables and Student’s t test, analysis of variance, or Kruskal-Wallis tests for continuous variables. Patients belonging to >1 group were not considered in this analysis. Logistic regression analysis was performed to identify independent risk factors for ICU admission. Univariate analysis was performed by estimating odds ratios (ORs) and 95% confidence intervals (CIs) for the entire population and the group of patients hospitalized because of secondary bacterial complications. For multivariate regression analysis, a preliminary selection of variables was made, including those with P < .20. To construct the final model, the level of significance was P < .05.
This study was approved by the Research Ethics Committee of the Hospital Foundation of the State of Minas Gerais.
Over the study period, 688 patients with VZV infection and VZV-associated diseases were hospitalized; 3 were >13 years old and 16 refused to participate. In all, 669 patients were included. A seasonal pattern was observed, with hospitalizations following a normal distribution with a peak in October and gradual decline to a nadir in March. This seasonality was evident in each year of the study period, with the highest admission rates occurring during August to December.
Table 1 reveals the main characteristics of hospitalized patients. Median participant age was 2.7 years (interquartile range [IQR] 25%–75%: 1.4–4.6 years), with 56.6% boys. Median time from varicella onset to hospitalization was 4 days (IQR 25%–75%: 3–6 days), and median duration of hospitalization was 5 days (IQR 25%–75%: 3–7 days). Most patients were previously healthy; asthma was the most frequent underlying chronic disease, followed by epilepsy and atopic dermatitis. Most patients reported previous contact with individuals with chickenpox, mostly at home (51.4%), schools, or day care centers (24.6%). The main reason for hospitalization was occurrence of bacterial complications (77.7%), followed by complications associated with VZV (11.4%) and hospitalization of at-risk patients (10.9%). A total of 44 patients (6.6%) were hospitalized for >1 reason, and 44 (6.6%) were admitted to the ICU. Thirty-eight (74.5%) had bacterial complications, 4 (9.1%) had complications associated with VZV infection, and 2 (4.5%) were at-risk patients. Most patients recovered from the disease and its complications. The mortality rate was 0.7% (5 patients). Death occurred in a median 6 days (range 1–9 days) after hospitalization; septic shock was the cause of death in all cases.
Among 554 patients hospitalized with bacterial complications, skin and subcutaneous tissue infections occurred in 442 patients (79.7%), followed by pneumonia in 51 patients (9.2%). Other bacterial infections included sepsis, conjunctivitis, pyelonephritis, and septic arthritis. The most common skin infections were cellulitis and impetigo. Confirmed bacterial etiology was established in 11 cases (2%), with 5 cases of S aureus infections, 1 of Streptococcus pneumoniae, and 5 of Streptococcus viridans. Among 81 patients with complications associated with VZV infection, herpes zoster was the most frequent (27.1%), followed by cerebellitis (13.5%), encephalitis (9.9%), febrile seizure (8.6%), and pneumonitis (8.6%). Cerebrospinal fluid of patients with encephalitis showed mild pleocytosis (range 8–28 cells), proteins levels 18 to 69 mg/dL, and glucose levels 45 to 121 mg/dL. Screening of cerebrospinal fluid for VZV by polymerase chain reaction was not performed. Other serious neurologic complications, such as acute disseminated encephalomyelitis, Guillain-Barré Syndrome, and meningitis, were not observed. At-risk patients received oral (20.5%) or inhaled corticosteroids (12.8%) and had a diagnosis of leukemia (19.2%), primary immunodeficiency (9%), solid tumors (7.7%), HIV infection (7.7%), and lymphoma (6.4%). A total of 41 children (52.6%) in this group reported previous contact with patients with varicella, mostly at home.
A total of 557 patients (83.2%) received antibiotics for a median 11 days (range 5–32 days). The most common treatments were oxacillin or cephalexin (87.3%). Clindamycin, sulfamethoxazole and/or trimethoprim, and vancomycin were used in 3.2%, 2%, and 1% of patients, respectively. A total of 124 patients (18.5%) received intravenous acyclovir for an average of 6 days (range 4–10 days) before switching to the oral form.
A comparative analysis of the 3 subgroups revealed that patients hospitalized with bacterial complications were younger than those hospitalized with complications related to VZV or at-risk patients. It was also observed that at-risk patients were hospitalized earlier than those other groups and remained hospitalized for longer periods (Table 2).
Univariate analysis of the predictive factors of ICU admission for patients with bacterial complications revealed that thrombocytopenia, anemia, and absence of skin infection were all related to higher risk (Table 3). After adjusting for other variables in multivariate analysis, only thrombocytopenia remained significant (Table 4). Univariate analysis of the entire study population revealed that a longer time interval between onset of infection and hospital admission was associated with the need for intensive care treatment (Table 5). No other variables were included in the final analysis.
Varicella causes high rates of hospitalization and deaths worldwide. Hospitalization rates differ by country and are related to sociodemographic differences and different health policies.5 In this study, we included the highest number of hospitalized patients with VZV infection in Brazil, where the disease is responsible for 34 hospitalizations per day.14
Complications leading to hospitalizations peaked in spring, paralleling the highest incidence of VZV infection in Brazil. Studies reveal that most hospitalizations occur from ages 1 to 4 years.4,5,14–18 Although there is great concern about the involvement of patients with underlying chronic disease, most hospitalized children with varicella were previously healthy, probably because of the high transmissibility of VZV.1,7,19 The importance of home and schools as environments of transmission has been highlighted in most studies.20 Patients with home contact constitute a group at risk for unfavorable outcomes and complication.
Secondary bacterial infection mostly affects infants and toddlers with varicella disease, with skin and soft tissue infections the most common complication.15,20,21 Bacterial infections and septic shock also result in high death rates in this population.11,16,20 Use of nonsteroidal anti-inflammatory drugs, recurrence of fever, and age <1 year are predisposing factors for bacterial complications.7,15,21
Thrombocytopenia was associated with the highest risk of ICU admission in patients with VZV, complicated by bacterial infection. In general, infection can cause thrombocytopenia because of interference with platelet production and consumption; in cases of bacterial infection, thrombocytopenia may also occur because of septic conditions.22,23 Previous studies reveal that Escherichia coli and S aureus toxins stimulate disturbances in platelet membranes and act directly in a cytotoxic manner. When thrombocytopenia is present, these infections are associated with worse prognosis, including increased mortality rates.24
It is plausible that thrombocytopenia reflects the severity of underlying infectious processes and critical illness. Patients with thrombocytopenia tend to be sicker and more likely to develop major bleeding disorders.23 Previous studies demonstrate that thrombocytopenia is associated with acute renal failure and the need for vasopressors and blood products, which may justify the imminent need for intensive care and higher risk of unfavorable outcomes.22–25 No researchers have evaluated thrombocytopenia as a predictor of poor outcome in patients with VZV infection; however, it is recognized as an independent risk factor for longer ICU length of stay in patients with bacterial infections.23
Among VZV complications, CNS involvement has been well described with incidences of 10% to 20.4%.5,18 The spectrum of neurologic viral complications in our cohort was similar to that of previous studies in which febrile seizures and cerebellitis were the most common manifestations.18,26,27 Complete neurologic recovery has been observed in patients with acute cerebella ataxia, isolated seizures, and meningitis. In contrast, acute disseminated encephalomyelitis is related to long-term neurologic sequelae.27,28
Chickenpox is highly lethal when it affects groups at risk for severe disease.29 In this study, children belonging to the at-risk group were admitted earlier in the course of VZV infection and were hospitalized for a longer period, demonstrating the necessity for prompt treatment. Young age, severe neutropenia, and underweight are the main risk factors for development of varicella-related complications in this group.29
Treatment of VZV infections is usually effective if timely care is sought. VZV and its complications were responsible for a significant number of ICU admissions in our study, and delayed treatment seeking was identified as a determinant of complications. Caregivers’ treatment-seeking behavior is instrumental in reducing child deaths in low- and middle-income countries.30 A delay of >7 days, after the onset of symptoms, was observed as a risk factor for the development of varicella-related complications among pediatric oncology patients infected with VZV in Pakistan.29 Peterson et al31 observed that children with varicella who have delayed contact with health services have a greater chance of developing invasive group A streptococcal disease. Despite cost being cited in most studies as a reason not to seek care, several studies revealed low uptake of government services despite being free at the point of access.30 In our study, health care was offered free of charge. The challenge, therefore, is to implement ongoing programs to educate caregivers and increase the ability to recognize when to seek appropriate care.
One study limitation is that sample was not probabilistic. However, we believe this did not cause bias because the hospital is a reference unit for VZV treatment and provides a good picture of the universe of patients studied. Furthermore, serological tests to confirm the diagnosis of chickenpox were unavailable during the study period. Also, better diagnostic methods are needed for diagnosing varicella in the CNS. Many challenges remain in developing nations that cannot afford routine use of real-time polymerase chain reaction when CNS disease is suspected. Our results could be influenced by the absence of specific diagnosis tests that would aid in the differential diagnosis of CNS viral infections.
Universal vaccination in Brazil has replaced vaccination of at-risk groups; this has proven to be of great importance as evidenced by the high prevalence of healthy children in this study. In countries that have introduced single-dose treatment of varicella vaccine, studies reveal high effectiveness in preventing hospitalizations.2,4,32,33 Likewise, it is expected that the introduction of single-dose VZV vaccine in Brazil will reduce hospitalizations and disease complications. Monitoring the outcomes and complications of VZV infection is an important component of assessing the impact of childhood infection, as well as that of the Brazilian vaccination program.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Gershon AA
- ↵American Academy of Pediatrics. Varicella-zoster infections. In: Kimberlin DW, Brady MT, Jackson MA, Long SS, eds. Red Book: 2015 Report of the Committee on Infectious Disease. 30th ed. Elk Grove Village, IL: American Academy of Pediatrics; 2015:856–859
- ↵Brasil Ministério da Saúde; Secretaria de Vigilância em Saúde; Departamento de Vigilância Epidemiológica Coordenação geral do Programa Nacional de Imunizações. Informe Técnico de Introdução da vacina tetra viral. Vacina sarampo, caxumba, rubéola e varicela. Brasília, Brazil: Ministério da Saúde; 2013
- Bozzola E,
- Bozzola M
- Ziebold C,
- von Kries R,
- Lang R,
- Weigl J,
- Schmitt HJ
- Marin M,
- Zhang JX,
- Seward JF
- Capanema FD,
- Resende MB,
- Luz FF,
- et al
- Maia C,
- Fonseca J,
- Carvalho I,
- Santos H,
- Moreira D
- Wen SC,
- Miles F,
- McSharry B,
- Wilson E
- Cameron JC,
- Allan G,
- Johnston F,
- Finn A,
- Heath PT,
- Booy R
- Blumental S,
- Sabbe M,
- Lepage P
- Carvalho I,
- Caldeira T,
- Santos F
- Dubos F,
- Grandbastien B,
- Hue V,
- Martinot A
- Dunaiceva J,
- Sabelnikovs O
- Bozzola E,
- Bozzola M,
- Tozzi AE,
- et al
- Alam MM,
- Qamar FN,
- Khan ZW,
- Kumar V,
- Mushtaq N,
- Fadoo Z
- Gowin E,
- Wysocki J,
- Michalak M
- ↵Brasil Ministério da Saúde; Secretaria de Vigilância em Saúde; Departamento de vigilância de Doenças Transmissíveis. Manual dos Centros de Referência para Imunobiológicos Especiais/Ministério da Saúde, Secretaria de Vigilância em Saúde, Departamento de Vigilância das doenças Transmissíveis. 4th ed. Brasília, Brazil: Ministério da Saúde; 2014:162
- Copyright © 2018 by the American Academy of Pediatrics