Background
Acinetobacter species are Gram-negative coccobacilli many of which are found in soil and fresh water throughout our natural habitats [
1]. However, certain
Acinetobacter species are frequently isolated increasingly from healthcare facilities and are concomitantly the source of many nosocomial infections [
1]. Specifically,
A. baumannii,
A. pittii, and
A. nosocomialis of the
Acinetobacter calcoaceticus-baumannii (
Acb) complex have become the most medically relevant members of the genus as they are most frequently isolated from health care facilities as well as human tissues. Patients with impaired host defenses in intensive care unit (ICU) settings appear to be an at risk group of acquiring
Acinetobacter infections [
2]. Despite this knowledge, the infection source in outbreaks often cannot be determined, leaving recommendations to prevent future outbreaks limited [
3]. Additionally, in recent years, some
Acinetobacter strains have demonstrated a propensity to acquire resistance to multiple classes of antibiotics, rendering treatment of such hospitalized patients extremely difficult [
4]. Furthermore, in pediatric patients,
A. baumannii is thought to be the most prevalent organism, which is associated with bacteremia, ventilator-associated pneumonia, bronchopulmonary dysplasia, meningitis, and neutropenia [
3,
5,
6].
Currently, our knowledge of the clinical impact of
Acinetobacter infections in the pediatric patient population in developing countries and countries with suboptimal infection control resources have been well documented [
2,
3,
7‐
11]. However, invasive
Acinetobacter infections associated with large pediatric, academic institutes in developed countries are still poorly defined due to unclear risk factors. Recent studies documented risk factors for invasive infections as catheter insertions, prolonged use of antibiotics, as well as underlying chronic diseases [
5,
12]. It is not clear if these risk factors are universal in all hospital settings, including locations outside of intensive care units. Additionally, improvements in bacterial sequencing, in particular the
Acinetobacter rpoB gene
, have allowed us to more accurately identify clinically important
Acinetobacter species [
13‐
15].
Therefore, the purpose of this study was to retrospectively review cases of invasive Acinetobacter infections occurring within an academic, pediatric setting in a developed country from a clinical perspective to define additional risk factors. Furthermore Acinetobacter strains were sequenced typed and preliminary characterized for motility and biofilm formation/maintenance. We hypothesized that Acinetobacter infections would be restricted to ICU settings and predominantly comprised of strains of A. baumannii.
Methods
Study design
The study was an observational review of patients with invasive
Acinetobacter infections correlated with comparative sequencing of hyper-variable regions of the
rpoB gene of
Acinetobacter isolates. It was reviewed and approved by the Institutional Review Board of Nationwide Children’s Hospital (IRB14-00145). As aggregate patient data was used, individual consent was not obtained. The review population consisted of pediatric patients at a single, large, academic, pediatric institution identified during the years of 2009–2013. Inclusion criteria included pediatric patients less than 21 years old with at least one positive blood, bone, endotracheal, peritoneal, or cerebrospinal fluid culture result positive for
Acinetobacter species [all identified as
A. baumannii or
A. baumannii complex by the clinical laboratory using Vitek 2 (bioMerieux, Durham NC) and other phenotypic methods as needed]
. Antimicrobial susceptibility testing was performed on all isolates via the Vitek 2 using established breakpoints [
16]. Multi-drug resistant isolates (MDR) was defined as isolates non-susceptible to ≥ 1 antimicrobial agent in ≥ 3 antimicrobial categories as defined by the joint initiative of the European Centre for Disease Prevention and Control (ECDC) and the Centers for Disease Control and Prevention (CDC) [
17]. Study samples represented a convenience sample of existing isolates available in the clinical laboratory
. Patients with cystic fibrosis respiratory isolates of
Acinetobacter were excluded from the study. Isolates were frozen and stored at −80 °C.
A comprehensive review of clinical and demographic patient information was performed surrounding positive culture identification. Patient demographics included gender, age, race, county of residence, and insurance type. Information related to determining the source of the infection including the patient’s location within the hospital, source of the culture, and susceptibilities were recorded. A 48 h time period after cultures were obtained was chosen as an initial analysis point based on standard hospital time-based culture practices needed for determining culture results. At this time point antibiotic selection by the practitioner might change based on available antimicrobial susceptibilities. Underlying diagnoses, co-morbidities, secondhand smoke exposure, and vital signs were recorded. Laboratory data included complete blood count with differential, creatinine, liver function tests, as well as the next 5 culture results for each patient. C-reactive protein and erythrocyte sedimentation rates were available on less than 20 % of the cohort and therefore not included. Information regarding the patient’s hospital stay included length of stay, febrile status, oxygen use, ventilator use, imaging and invasive procedures, and mortality. Antibiotic treatment including antibiotics given and length taken by the patient along with antibiotic usage in the past 30 days were noted. Home medication including the use of chronic immunosuppressants, antibiotics, and gastric acid suppressants were all recorded.
Comparative sequence analysis of the rpoB gene from clinical Acinetobacter isolates
Acinetobacter rpoB hyper-variable region sequencing was performed on
Acinetobacter isolates according to a previously established protocol [
18] with the following modifications. Genomic DNA (gDNA) was prepared from
Acinetobacter isolates utilizing the Gentra Puregene Yeast/Bact. Kit B (Qiagen) according to the manufacturer’s protocol. One hundred nanograms of gDNA from each isolate were used in a PCR with one of two sets of primers. The first set of primers, spanning zone 1, were Ac696F (5′-TAYCGYAAAGAYTTGAAAGAAG-3′) and Ac1093R (5′-CMACACCYTTGTTMCCRTGA-3′) and the second set of primers, spanning zone 2, were Ac1055F (5′-GTGATAARATGGCBGGTCGT-3′) and Ac1598R (5′-CGBGCRTGCATYTTGTCRT-3′).
DreamTaq DNA polymerase (Thermo Scientific) was used according to manufacturer’s protocol with the following thermocycling conditions: 1 cycle at 95 °C for 2 min; 30 cycles at 95 °C for 30 s, 55 °C for 30 s, and 72 °C for 1 min; 1 cycle at 72 °C for 10 min. PCR products were purified using the QiaQuick PCR purification kit (Qiagen) according to the manufacturer’s protocol. Clean PCR products were verified by electrophoresis and sent off for sequencing by Eurofins MWG Operon. Raw sequence files were trimmed and edited using MegAlign and EditSeq software applications (DNASTAR). Trimmed sequence files for the two hyper-variable regions of a single isolate were combined into a single FASTA file for phylogenetic analysis. The combined FASTA files for each isolate, as well the corresponding sequences from the reference Acinetobacter strains used for phylogenetic analysis, were aligned initially by the Clustal W method utilizing UPGMB for the cluster method and the Kmer4-6 method for the distance measure. Later parameters utilized UPGMB for the cluster method and the Kimora % identity method for distance measure. Sequence distances were computed with the metric uncorrected pairwise distance and assembled into a phylogenetic tree using the MegAlign Pro software (DNASTAR).
Motility assays
Twitching motility was investigated per previously described protocols with the following modifications [
19]. Twitching plates contained 10 g tryptone/L and 10 g agarose/L. Briefly, bacterial strains were grown overnight and a single colony was used to inoculate each twitching plate to the bottom of the petri dish with a sterile wooden applicator stick. Twitching plates were incubated at 37 °C in a humidified incubator for 16 h. To visualize zones of twitching motility, the agarose was removed, the adherent bacterial population was washed with phosphate buffered saline, and stained with 0.1 % crystal in water. Bacteria positively exhibiting twitching motility were defined by demonstrating a zone of motility of >10 mm around the site of inoculation. Surface-associated motility was simultaneously assessed per our previously published methodologies [
19]. Bacteria positively exhibiting surface-associated motility were defined as strains with a halo growth zone of >20 mm. Assays were performed in triplicate.
Crystal violet retention assay
In order to assess the ability of each isolate to form biofilms, the crystal violet retention assay was performed as described previously [
20] with minor modifications. Mueller Hinton (MH) broth was inoculated with one bacterial colony and incubated overnight at 37 °C for approximately 16–18 h. Cultures were subsequently diluted 1:100 in fresh MH broth and 100 μl of diluted culture was added to each well in 96 well microtiter plate and incubated overnight at 37 °C. Adherent cells were washed once with deionized water and stained with 125 μl of 0.1 % crystal violet solution for 15 min at room temperature and washed 5 times with deionized water. Subsequently, dye was released from the cells using 200 μl of 95 % ethanol. Absorbance was measured at 595 nm on a Synergy H1 Hybrid Reader spectrophotometer (Biotech, Biotech Instruments, Vermont, USA). The biofilm data represent the average of three independent experiments of triplicate wells.
Statistical analysis
All analyses were performed using Stata/MP, version 13.1 or GraphPad Prism version 6.03. For all analyses, a P value < 0.05 was considered statistically significant. Descriptive statistics for continuous variables were presented as medians with 25–75th percentile ranges; and descriptive statistics for categorical variables were presented as proportions. Mann-whitney tests were performed for sum comparisons.
Discussion
Members of the
Acinetobacter calcoaceticus-baumannii complex are regarded as opportunistic human pathogens of increasing relevance worldwide due in part to the emergence of multiply-drug resistant strains; however, the description of pathogenic
Acb members other than
A. baumannii remains limited due in part technological limitations with current clinical laboratory identification methods [
24]
. Although emerging methods such as MALDI-ToF analysis may aid in clinical laboratory-based identification
Acb members for the future, these methods have not become the gold-standard even in developed countries. In this study we report the emergence of
A. pittii strains from pediatric patients with invasive infections, along with a relatively low mortality rate.
Outbreaks of
A. pittii have only been previously reported to a low extent and were thought to be rare in pediatric settings in the United States [
5,
8,
9,
25]. This may be attributed to the fact that most of the common clinical laboratory identification methods do not reliably differentiate between members of the
Acb; hence, many publications may in fact be referring to all members of the
Acb when specifically referencing
A. baumannii [
5,
24]. Importantly, in the documented cases where
A. pittii has been reported, it can be more commonly recovered from samples than
A. baumannii [
8], but this is not the case in all settings. In our isolate collection, there was an emergence of
A. pittii over the last two years of the study period, indicating a shift towards
A. pittii over
A. baumannii. This finding will require further follow-up due to the short time-frame of study follow-up. The significant rise in the number of isolates of
A. pittii in a pediatric population also signifies a potential new trend to be cognizant of for clinicians and researchers, due largely to differences in antibiotic resistance profiles between species. Notably, we did not see a difference in basic microbiologic characteristics such as surface motility between the
A. pittii and
A. baumannii strains. However, the majority of the most recent
A. pittii strains did display both twitching and surface-associated motility phenotypes, indicating the likely expression of a functional type IV pilus as well as changing local population dynamics. Additionally
, A pittii isolates averaged more robust biofilm production when compared to
A. baumannii isolates, indicating another potential pathogenic mechanism for the increasing
A pittii prevalence. However, laboratory biofilm studies may not reflect human in vivo biofilm formation, therefore further studies are warranted in this area for the future.
Despite previous reports of high mortality associated with antibiotic resistant
Acinetobacter strains [
5,
7], in our sampling only 2 out of 24 cases were associated with patient deaths. Notably, both deaths were from
A. pittii, again in contrast with the existing pediatric literature on
A. baumannii-associated mortality, and the antibiotic resistance patterns observed. We postulate that one difference in the overall low mortality rates could be the relative sensitivity to the aminoglycoside antibiotics coupled with prompt administration, as this class of antibiotic was utilized in the first 48 h of a suspected infection in half of the reported cases. There was also a low prevalence of MDR strains. Both of these factors may have contributed to the fact that nearly 80 % of the strains were cleared from the systemic circulation within 48 h of initiating antibiotics. Persistent bacteremia was noted to be associated with biofilm-producing isolates, which is of importance when assessing therapeutic options for persistent
Acinetobacter infections. Additionally, compared to other reported pediatric
Acinetobacter infections, the isolated strains demonstrated lower overall morbidity as reflected by fewer intensive care requirements and less multiple organ involvement. Much of the published pediatric literature focuses on neonatal intensive care units with multiple
Acinetobacter infections [
3,
8‐
12]. However, in our sample, the median age was 3.5 years with no reported neonatal infections. We were unable to assess for the role of existing institutional infection control bundles in preventing ICU-associated infections during this retrospective study. Therefore, continued work will be needed to address potential epidemiologic factors associated with pediatric
Acinetobacter infections in the United States.
Other species of
Acinetobacter are increasingly recognized to cause disease in humans such as
A. bereziniae [
26], further supporting the assumption that improved species identification of
Acinetobacter infections may be needed to accurately follow emerging infectious trends from an epidemiologic level. These results further highlight the medical relevance of the
Acinetobacter genus as whole, where an encounter between a patient with a compromised immune system and a species of
Acinetobacter, regardless of its characterized pathogenicity can result in infection.
We acknowledge that this study has several limitations. First, the study was retrospective in nature, therefore limiting the availability of some clinical information. Second, the small sample size, 5-year time frame, and single-center setting limits the generalizability of the findings and analysis of isolate phenotypic characteristics in relation to clinical outcomes. However, the samples may be more representative of U.S. pediatric academic centers than previous reports from developing countries. Finally, the clinical laboratory did not report antibiotic susceptibilities for all classes of antibiotics for the four isolates obtained in the first year of the study.
Acknowledgements
The authors thank the NCH clinical microbiology laboratory staff for assistance with isolates.