Potentially serious bacterial infections (PSBIs) in febrile infants aged less than 3 months are more frequent, carry a poorer prognosis and have a different aetiology compared to older children. Streptococcus agalactiae (S. agalactiae, historically the leading causative agent of bacteraemia and meningitis in infants <3 months), gram-negative rods, especially Escherichia coli (E. coli), Listeria monocytogenes (L. monocytogenes) and Enterococcus species are the agents typically involved in PSBIs in this age group. With the exception of E. coli, which continues to be the aetiological agent most frequently involved in urinary tract infections throughout childhood, the prevalence of infection by these microorganisms is low in older children.1–7 All of the above warrants the performance of diagnostic tests and hospital admission with antibiotic therapy (ampicillin and an aminoglycoside or a third-generation cephalosporin to cover the bacteria detailed above) in the management of all febrile infants aged less than 3 months with risk factors.6–9

A study of febrile infants aged less than 3 months that received care in a Spanish emergency department over a period of 5 years described changes in the aetiology of bacteraemia; E. coli was the most prevalent bacterium and there were no cases of infection by L. monocytogenes.10 Recent studies in the United States have also shown that the incidence of bacteraemia and meningitis caused by S. agalactiae is decreasing, so that E. coli is currently the most prevalent pathogen not only in urinary tract infections, but also in cases of bacteraemia and meningitis in infants aged less than 3 months.11–14

The main objective of this study was to describe the microbiology and outcomes of infants aged less than 91 days with fever without source (FWS) that received a diagnosis of PSBI in Spain.

We performed a subanalysis in the framework of a prospective, multicentric study conducted between October 1, 2011 and September 30, 2013 in 19 Spanish Children's hospitals in 8 autonomous communities (Table 1) members of the Working Group on the Febrile Infant of the Spanish Pediatric Emergency Research Group of the Spanish Society of Pediatric Emergencies (RISeuP/SPERG.)

To be included in the study, infants aged less than 91 days had to meet the following criteria: seeking care for FWS, availability of the results of complete blood count, C-reactive protein test, blood culture, urine analysis (with urine test strip), and urine culture of a sample obtained by catheter, and signed informed consent by the parents or legal guardians.

We excluded patients whose temperature had been estimated at home without a thermometer and who were afebrile in the emergency department.

The study did not change the approach to the management of patients, which conformed to the established protocols of each hospital.

We collected the following data for each patient: age, sex, date of service, personal history, time elapsed from the detection of fever to receiving care at the emergency department, temperature at home and at the emergency department, general health status, results of diagnostic tests, final diagnosis and patient destination after discharge. The follow-up protocol included a telephone call within 30 days from the emergency department visit to gather information about the outcome of patients that were not admitted to hospital.

The principal investigator in each hospital registered patient data using a form created for the purpose in Google Drive®. The registers of all hospitals were consolidated into a single database (Microsoft Excel®) to which only principal investigators were given access. No personally identifiable data were entered in the database for any of the patients.

The study was approved by the Ethics Committee of the Hospital Infantil Universitario Niño Jesús de Madrid and, when required, by the institutional ethics committees of participating hospitals.

Over the 2-year study period, there were 1612212 visits to the 19 participating emergency departments, of which 4008 (0.25%) corresponded to infants aged less than 91 days with FWS. Finally, after applying the exclusion criteria, 3401 children were included in the study (Fig. 1) with the following age distribution: ≤28 days, 905 (26.6%), 29–59 days, 1404 (41.3%); 60–90 days, 1092 (32.1%).

Figure 1.

Flow chart of the patients included in the study.

(0.08MB).

Table 2 summarises the characteristics of the patients.

Among all the cultures done in the 3401 infants aged less than 91 days with FWS, 896 were positive: 766 urine cultures (85.5%), 100 blood cultures (11.2%), 18 CSF cultures (2%), 10 stool cultures (Salmonella in 6; C. jejuni in 4) and 2 umbilical cord cultures (E. coli in 2). A diagnosis of PSBI was made in 784 infants (23%), 107 (3.1%) of them had an IBI.

Bacteria were isolated from more than one type of culture in 64 patients (Table 3). A single bacterium was isolated in 44 of the children with positive blood and urine cultures (E. coli in 39 infants; and Klebsiella pneumoniae, Enterobacter cloacae, Enterococcus faecalis, S. agalactiae and S. aureus in one patient each), in all of the infants with positive blood and CSF cultures (E. coli 4; S. agalactiae 3; Streptococcus pneumoniae [S. pneumoniae] 2; Morganella morganii 1; Pasteurella multocida 1), in 1 infant with urine and CSF cultures positive for E. coli and in 1 infant with stool and blood cultures positive for C. jejuni.

The proportion of patients with 2 different positive cultures was greater in newborns (30/241; 12.4%) than in infants aged 29–59 days (20/297; 6.7%) or 60–90 days (14/294; 4.8%); differences that were statistically significant (p=0.003).

E. coli was the bacterium isolated most frequently in urine, blood and CSF cultures (Table 4).

In the subgroup of newborns (905 patients aged ≤28 days of life), in which there were 268 (29.6%) positive cultures (213 urine, 45 blood and 10 CSF cultures), the distribution of bacterial species was similar to the one found in the general population:

• –

  Urine cultures: E. coli (171/213; 80.3%), Klebsiella spp (15/213; 7%) and Enterococcus spp (14/213; 6.6%). Two urine cultures were positive for S. agalactiae in 2 newborns aged more than 7 days, one of who had bacteraemia.

• –

  Blood cultures: E. coli (25/45; 55.5%) and S. agalactiae (11/45; 24.4%; three of these patients were aged less than 8 days).

• –

  CSF cultures: E. coli (6/10) and S. agalactiae (2/10; associated with bacteraemia in both patients).

Two cultures tested positive for L. monocytogenes: the blood culture of a male newborn that presented with a fever lasting 17h, feeding refusal, irritability and poor general appearance; and the CSF culture of a well appearing boy aged 36 days with fever lasting 12h.

There were 7 cases of IBI caused by S. pneumoniae (4 with positive blood culture, 1 with positive CSF culture, 2 with positive blood and CSF cultures), of which 2 corresponded to infants aged 29–59 days and 5 to infants aged more than 59 days.

Three patients received a diagnosis of herpes simplex infection. One was a well appearing girl aged 8 days that presented with fever lasting 5h with no associated symptoms, treated with acyclovir following the detection of herpes simplex type 1 by PCR from a CSF sample (traumatic lumbar puncture), who had a favourable outcome free from sequelae. Another corresponded to a previously healthy girl aged 7 days that presented with a fever lasting 3h, irritability and well general appearence and who got worse during the hospital stay and died 3 days after. The autopsy revealed an infection by herpes simplex virus type 1. The results of the lumbar puncture analysis had been normal. Another was a boy aged 61 days that presented with a fever lasting 12h, irritability and well general appearence and was admitted for antibiotherapy with a diagnosis of meningitis (77 white blood cells/mm3 in CSF); herpes simplex was detected by PCR and the patient was treated with acyclovir, in spite of which he developed severe neurologic sequelae (epilepsy, psychomotor impairment).

Four patients died, all of them had been admitted for intravenous antibiotherapy in their first visit. A girl aged 77 days, previously healthy, with fever lasting 12h, poor general appearence and a consolidation in the evinced by chest radiography. She was diagnosed with sepsis and died due to multiple organ failure. Bacterial culture results were negative. A girl aged 77 days with hypotonia on physical examination that presented with fever lasting 3h and in poor general appearance. She was admitted with respiratory distress and received a diagnosis of bronchiolitis. Bacterial culture results were negative. A previously healthy boy aged 32 days presenting with fever lasting 3h, poor general appearance and a purpuric rash. Neisseria meningitidis group B was isolated from the blood culture, and E. coli from the urine culture. Also, the girl aged 7 days mentioned above with a herpes simplex type 1 infection identified during the autopsy.

Another 7 patients had complications or serious sequelae: 3 patients aged 20, 58 and 71 days with bacterial meningitis caused by E. coli, S. agalactiae and S. pneumoniae, respectively, developed complications and severe and permanent neurologic sequelae; one boy aged 13 days with meningitis caused by enterovirus complicated with myocarditis; the boy aged 61 days with meningitis caused by herpes simplex described above; one boy aged 38 days with sepsis due to S. agalactiae associated with arthritis and left-shoulder myositis requiring surgical drainage; and one boy aged 66 days with left coronary artery ectasia secondary to Kawasaki disease.

In our study, E. coli was the bacterium involved most frequently in invasive and noninvasive PSBIs in febrile infants aged less than 3 months, which confirms that the epidemiology of PSBIs in this age group is changing; we also observed that the proportion of patients with urinary tract infections (UTIs) had increased considerably compared to the proportion of patients with bacteraemia or meningitis, a shift that has also been described in the United States.13,14 Another salient finding was that nearly 50% of positive blood cultures were associated with a UTI. These percentages are slightly above those reported by Gómez et al, who analysed 1018 blood cultures ordered for 1125 infants aged less than 3 months presenting with FWS in the emergency department of a tertiary care hospital in the Basque Country (2003–2008), and found that 8 (34.8%) of the 23 diagnosed cases of bacteraemia were associated with a UTI.10 Overall, these findings emphasise the importance of urine testing and urine culture in the assessment of infants aged less than 3 months presenting with FWS.

As is the case in other countries, E. coli has become the bacterium most frequently isolated from urine, blood and CSF cultures in infants aged less than 3 months presenting with FWS10–14 and, also in newborns. All of this may be a reflection of some of the preventive measures implemented in recent years that have led to decreases in the incidence of infections by S. agalactiae, L. monocytogenes and S. pneumoniae. Programmes for S. agalactiae screening during pregnancy with antibiotic treatment of positive cases at the time of delivery have been associated with a decreased incidence of early-onset neonatal sepsis (≤7 days of life)15–17; in our series, all but 3 of the cultures positive for S. agalactiae were in infants aged more than 7 days. L. monocytogenes is a gram-positive bacillus that can be vertically transmitted during pregnancy or acquired through the consumption of contaminated food.18,19 Our study seems to suggest that the educational campaign recommending pregnant women to avoid risky foods has been effective. Although the incidence of listeriosis in Spain has increased in the last decade (in 2012, Spain had the second largest number of reported cases of infection by L. monocytogenes in the European Union),19 in our series we only found 2 IBIs caused by this bacterium. The introduction of the pneumococcal conjugate has reduced the incidence of invasive disease in vaccinated as well as unvaccinated children,20–25 which may reflect, as Poehling et al. suggested,26 the benefits of herd immunity for infants that had yet to receive any doses due to their age. We should also mention that none of the 7 pneumococcal infectious were diagnosed in newborns.

The main goal of paediatricians that manage infants aged less than 3 months presenting with FWS is the early diagnosis and treatment of PSBIs and especially IBIs (bacteraemia and meningitis). We should highlight that in our study, 6 of the 11 patients that died or developed severe sequelae had a meningeal infection, but in 3 of them the aetiology was viral (2 herpes simplex and one enterovirus). It is important to remember that early treatment of herpes infections is a significant prognostic factor and that, while infrequent, they must be included in the differential diagnosis on account of the considerable associated morbidity. Some care guidelines for the management of FWS in infants aged less than 3 months already recommend considering this infection, especially in newborns, if the patient presents with cutaneous vesicles, seizures, elevated transaminases or critical illness.27

Last of all, this case series with patients aged less than 3 months presenting with FWS is fairly representative of the management of these patients in Spanish emergency departments. Despite the current protocols for the management of FWS,8,9 lumbar puncture and antibiotic treatment were only performed in 6 out of 10 newborns. It seems that clinicians prefer to adopt a watchful waiting approach, and nearly 90% of newborns with fever are admitted to hospital. These findings are similar to those recently published by Jain et al,28 who found that a full sepsis evaluation (blood, urine and CSF tests and culture) was performed in 73% of the 2253 febrile newborns that visited 36 emergency departments in the United States in 2010. As one would expect, the approach was even less aggressive in infants aged more than 28 days.

It is important that further studies on this subject be conducted in Spain in order to detect epidemiological changes, especially with the purpose of determining which is the most appropriate empirical antibiotic treatment for infants aged less than 3 months presenting with FWS that are not considered low-risk. Another interesting finding was that S. aureus was the third most frequent organism isolated from blood cultures, following E. coli and S. agalactiae. For the time being, there is no question that empirical treatment should include an aminoglycoside or a third-generation cephalosporin to cover gram-negative bacilli. Ampicillin should also be part of the initial treatment because L. monocytogenes, while rare, has not been eradicated, and also there are also still UTIs as well as cases of bacteraemia caused by Enterococcus spp, although we must be particularly mindful of changes in listeriosis and IBI caused by S. aureus.

There are two limitations to this study. The first is that it only included febrile infants, and did not take into account all infants with a positive culture, including those without fever. It is well known that infants aged less than 3 months may have PSBIs in the absence of fever, and in some series the proportion of febrile patients of this age with bacteraemia has reached up to 7%,10 but the original study was focused on patients with fever, not on positive cultures. Nevertheless, our study contributes relevant data on a well-defined population and, in fact, its findings are similar to those of studies that included all positive cultures in febrile and afebrile patients.11,13 The second limitation is that we did not gather data on antibiotic susceptibility, when knowledge of antibiotic resistance profiles in bacteria is essential to choose the most appropriate treatment.

To conclude, we would like to highlight that at present E. coli is the bacterium isolated most frequently in urine, blood and CSF cultures in infants aged less than 91 days with FWS in Spain, even if we only take newborns into account. In our country, the prevalence of infection by Listeria monocytogenes is very low, which is consistent with the reports of other case series. Although the prevalence of herpes simplex infection is low, we need to take into account the considerable morbidity and mortality associated to it in these patients.

The authors have no conflict of interests to declare.