Pantoea agglomerans, a Plant Pathogen Causing Human Disease

Abstract

We present 53 pediatric cases of Pantoea agglomerans infections cultured from normally sterile sites in patients seen at a children's hospital over 6 years. Isolates included 23 from the bloodstream, 14 from abscesses, 10 from joints/bones, 4 from the urinary tract, and 1 each from the peritoneum and the thorax. P. agglomerans was most associated with penetrating trauma by vegetative material and catheter-related bacteremia.

Pantoea agglomerans (formerly Enterobacter agglomerans) is a gram-negative aerobic bacillus in the family Enterobacteriaceae. All species of the genus Pantoea can be isolated from feculent material, plants, and soil (2), where they can be either pathogens or commensals (12). Within the genus, P. agglomerans is the most commonly isolated species in humans, resulting in soft tissue or bone/joint infections following penetrating trauma by vegetation (6, 7, 9, 14, 15). P. agglomerans bacteremia has also been described in association with the contamination of intravenous fluid (11), total parenteral nutrition (8), the anesthetic agent propofol (3), and blood products (1). However, spontaneously occurring bacteremia has rarely been reported, especially for children, and the role of P. agglomerans as a pathogen in other circumstances is unclear. Here, we present a large series of P. agglomerans infections in children that involve the bloodstream, soft tissue, and bones/joints.

This study reviewed all patients seen at Texas Children's Hospital, Houston, TX, with culture-documented P. agglomerans infections from January 2000 to December 2006. Patients were identified from hospital microbiology laboratory records. A retrospective review of medical records was performed for patients whose cultures were obtained from the following normally sterile sources: the bloodstream, catheter specimens from patients with urinary tract infections (UTIs) with ≥10,000 CFU/high-powered field, joint or body cavities, or incision sites and drainage of abscesses. Blood cultures were processed using the VITEK system of identification, and specimens that did not yield a result underwent DNA pyrosequencing. CLSI standards were used for disk diffusion testing (5). The study received institutional review board approval.

Overall, P. agglomerans was identified in 88 patient cultures from 53 sterile-site, 26 sputum, 3 urine, 3 surface swab, and 2 oropharyngeal sources. Of the 26 sputum cultures, only 1 grew P. agglomerans repeatedly, representing monomicrobic infection; more than one organism was isolated from all other sputum cultures, and the contribution of P. agglomerans was uncertain. Comorbid conditions in these patients included eight cases of cystic fibrosis, five cases of neurological impairment, five cases of tumor, and three cases of intestinal malabsorption.

For the 53 children whose sterile-site cultures grew P. agglomerans, sources included 21 central venous line (CVL)-related bacteremic episodes, 14 abscesses, 10 joint or bone cultures, 4 UTIs, 2 non-CVL-associated bacteremic episodes, 1 peritonitis episode, and 1 penetrating thoracic trauma. These sources are presented in Table 1.

TABLE 1.

Patients with Pantoea agglomerans infection seen at Texas Children's Hospital from 2000 to 2006

Case	Patient age/genderb	Culture source, type of infection	Underlying illness and/or conditionc	Coisolate(s)d	Outcome
1	16 yr/Fa	Blood, CVL infection	Aplastic anemia, s/p BMT	CONS	Survived
2	2.5 yr/F	Blood, CVL infection	Aplastic anemia, s/p BMT	MSSA	Survived
3	3 yr/Ma	Blood, CVL infection	Sideroblastic anemia, s/p BMT	Enterobacter cloacae	Survived
4	4 yr/F	Blood, CVL infection	Thalassemia, s/p BMT		Survived
5	8 mo/M	Blood, CVL infection	Infantile leukemia, s/p BMT		Survived
6	35 mo/M	Blood, CVL infection	Leukemia, s/p BMT	Acinetobacter calcoaceticus bv. anitratus	Survived
7	20 yr/M	Blood, CVL infection	Leukemia, s/p BMT		Survived
8	3 yr/Ma	Blood, CVL infection	Leukemia		Survived
9	34 mo/M	Blood, CVL infection	Neuroblastoma	CONS, Bacillus sp.	Survived
10	4 yr/F	Blood, CVL infection	Neuroblastoma	Bacillus sp.	Survived
11	3 yr/F	Blood, CVL infection	Rhabdomyosarcoma	Acinetobacter calcoaceticus bv. anitratus	Survived
12	13 yr/M	Blood, CVL infection	Ewing's sarcoma	Klebsiella pneumoniae	Survived
13	14 yr/M	Blood, CVL infection	Ewing's sarcoma	Stenotrophomonas maltophila, CONS	Survived
14	10 yr/M	Blood, CVL infection	Synovial sarcoma	Pseudomonas aeruginosa, Agrobacterium radiobacter	Survived
15	7 yr/Ma	Blood, CVL infection	Wilms' tumor		Survived
16	24 days/F	Blood, CVL infection	Cardiomyopathy	Morganella sp., CONS, Enterococcus faecalis	Died
17	61 days/M	Blood, CVL infection	Congenital heart disease		Died
18	16 days/M	Blood, CVL infection	Coarctation of the aorta	E. faecalis	Survived
19	43 days/F	Blood, CVL infection	NEC	Candida parapsilosis	Died
20	8 yr/Fa	Blood, CVL infection	Reflux nephropathy	CONS	Survived
21	4 yr/F	Blood, CVL infection	Microvillus inclusion disease		Survived
22	5 mo/M	Blood		Kingella kingae	Survived
23	55 days/M	Blood		CONS	Survived
24	15 yr/F	Arm, abscess		MSSA, Bacillus, Enterococcus sp., alpha Streptococcus sp.	Survived
25	13 yr/M	Arm, abscess		CONS	Survived
26	12 yr/M	Scalp, abscess		MSSA, CONS	Survived
27	3 yr/M	Scalp, abscess		Moraxella sp., CONS, diphtheroids	Survived
28	4 yr/F	Scalp, abscess		CONS, alpha Streptococcus	Survived
29	5 yr/M	Scalp, abscess		E. cloacae, Enterobacter amnigenus, MSSA	Survived
30	3 mo/M	Arm, abscess		MSSA, Enterobacter asburiae, Pseudomonas putida, Pseudomonas fluorescens, Streptococcus pyogenes	Survived
31	8 mo/M	Arm, abscess		MRSA	Survived
32	13 yr/F	Buttocks, decubitus ulcer	Cerebral palsy	Proteus sp., P. aeruginosa, MRSA	Survived
33	6 yr/M	Scalp, abscess		MSSA	Survived
34	23 mo/M	Neck, abscess		K. pneumoniae, E. cloacae, CONS, Candida albicans, Enterobacter amnigenus	Survived
35	17 days/F	Neck, abscess		K. pneumoniae, MRSA	Survived
36	7 days/M	Anterior abdominal wall, abscess		Acinetobacter, MRSA	Survived
37	25 mo/F	Axilla, abscess	Cystic hygroma	Achromobacter xylosoxidans, MSSA	Survived
38	12 yr/M	Finger, osteomyelitis			Survived
39	8 yr/M	Heel, avulsion injury		Enterobacter taylorae	Survived
40	4 yr/M	Toe, osteomyelitis			Survived
41	7 yr/M	Hand, osteomyelitis		CONS	Survived
42	3 yr/M	Foot, osteomyelitis		MSSA, Flavimonas oryzihabitans	Survived
43	9 yr/M	Foot, osteomyelitis		Yeast, Streptococcus mutans	Survived
44	10 yr/M	Foot, osteomyelitis		CONS	Survived
45	17 yr/F	Toe, osteomyelitis	Spina bifida	CONS, diphtheroids	Survived
46	8 yr/F	Knee, septic arthritis			Survived
47	11 yr/M	Finger, osteomyelitis		E. taylorae, Bacillus cereus, Enterobacter amnigenus, alpha Streptococcus	Survived
48	21 yr/M	Urine, UTI	Hemochromatosis	105Pantoea, 104Escherichia coli organisms	Survived
49	35 days/M	Urine, UTI	Prematurity (25 wk)	104Pantoea, 105E. coli, 103 CONS organisms	Survived
50	7 yr/M	Urine, UTI		104Pantoea, 103 gram-negative and gram-positive organisms	Survived
51	5 yr/F	Urine, UTI		105Pantoea, 103 gram-positive organisms	Survived
52	9 yr/M	Chest wall, penetrating trauma		Bacillus cereus	Survived
53	15 days/M	Peritoneum, peritonitis	Prematurity (25 wk), s/p NEC	CONS	Survived

^aP. agglomerans isolated twice at least 12 h apart.

^bF, female; M, male.

^cs/p, status post; BMT, bone marrow transplant; NEC, necrotizing enterocolitis.

^dCONS, coagulase-negative staphylococci; MSSA, methicillin-susceptible S. aureus; MRSA, methicillin-resistant S. aureus.

Of the 21 patients with CVL infections, 8 had hematologic malignancies or bone marrow transplants, 7 had solid tumors, 3 had congenital heart disease, 1 had renal failure, 1 had necrotizing enterocolitis, and 1 had microvillus inclusion disease. One patient was neutropenic. Of the 21 patients, 14 had polymicrobial CVL infections necessitating line removal. Only 5/21 patients had two positive blood cultures; in all, bacteremia resolved within 48 h. Patients received 14 to 21 days of combination therapy with an aminoglycoside and either a broad-spectrum cephalosporin or ticarcillin-clavulanate. Three patients (two cardiac patients and one premature infant [cases 16, 17, and 19]) died of overwhelming sepsis shortly after P. agglomerans was isolated in their blood cultures; two of the three patients had polymicrobial infections. Two other patients had bacteremia without having a CVL. In both, bacteremia cleared after the first blood culture, and neither child had evidence of bone, joint, or soft tissue infection.

There were 14 children from whom P. agglomerans was isolated during the drainage of abscesses. All isolates were polymicrobial. Of these 14 children, 13 responded to antistaphylococcal therapy. A child with cerebral palsy underwent extensive debridement of a sacral decubitus ulcer and received a prolonged antibiotic course due to suspected chronic osteomyelitis.

The seven patients (cases 38 to 44) who had osteomyelitis presented with local symptoms 4 to 6 weeks after a penetrating trauma with a stick, plant thorn, or glass shard. None of the patients were febrile. Two of the patients were determined to have subperiosteal reactions and erosions by radiography and chronic inflammation by histopathology. The average white blood cell count was 6,900 cells/mm³ (range, 6,300 to 11,500 cells/mm³); blood cultures were negative.

An 8-year-old girl (case 46) developed septic arthritis 5 weeks after receiving a penetrating injury to the knee from a thorn. The patient's white blood cell count was 11,500 cells/mm³, her erythrocyte sedimentation rate was 16 mm/h, and her C-reactive protein level was 0.7 mg/ml. Arthrocentesis showed 17,600 white blood cells/mm³ (89% neutrophils, 11% monocytes) and 1,700 red blood cells/mm³. Gram staining and blood cultures were negative; a joint culture grew P. agglomerans. The child responded well to a 3-week course of therapy.

Antimicrobial susceptibilities were determined by Kirby-Bauer disk diffusion. All 53 isolates from sterile sites were uniformly susceptible to amikacin, gentamicin, meropenem, and trimethoprim-sulfamethoxazole. In addition, 92.5% of isolates were susceptible to broad-spectrum cephalosporins and semisynthetic penicillins, 62.3% to extended-spectrum cephalosporins, and 47.2% to ampicillin. Quinolone susceptibilities and MICs were not routinely determined for all specimens, given the association between quinolone use and arthropathy in juvenile-animal studies.

Of the 37 prior reports of P. agglomerans infections in children (Table 2), 5 were related to penetrating trauma and 30 (81%) to the contamination of parenteral fluids (8, 11). There is only one report of spontaneous P. agglomerans bacteremia: in a child with sepsis after rotavirus gastroenteritis (4). In that instance, it was postulated that the preceding gastrointestinal insult facilitated bacterial translocation across the gut mucosa.

TABLE 2.

Previously reported pediatric cases of Pantoea/Enterobacter agglomerans infection

Yr (reference)	No. of patients	Age/sex of patientsa	Diagnosisb	Type of (no. of patients with) underlying illnessc
1978 (7)	1	11 yr/M	Right knee septic arthritis after penetrating trauma (wooden splinter)
1988 (15)	1	8 yr/M	Left capitate, hamate, and third metacarpal osteomyelitis (rose thorn injury)
2000 (6)	1	13 yr/M	Right knee septic arthritis after penetrating trauma (thorn injury)
2003 (9)	1	14 yr/M	Right knee septic arthritis after penetrating trauma (palm tree thorn injury)
2004 (14)	1	9 yr/M	Left knee septic arthritis after penetrating trauma (lemon tree thorn injury)
1984 (11)	22	10 days-17 yr	Iatrogenic bacteremia secondary to contaminated intravenous fluids	Prematurity (1), asplenia (1)
2005 (8)	8	3 days-6 mo	Iatrogenic bacteremia from contaminated TPN	RDS (3), asphyxia (2), IUGR (1), VACTERL (1), pneumonia (1)
2005 (10)	1	2 yr/F	Peritonitis from teething on peritoneal dialysis catheter	End-stage renal disease (1)
2006 (4)	1	18 yr/M	Bacteremia following rotavirus infection

^aF, Female; M. male.

^bTPN, total parenteral nutrition.

^cRDS, respiratory distress syndrome; IUGR, intrauterine growth restriction; VACTERL, vertebral, anal, cardiac, tracheoesophageal fistula, renal, and limb anomalies.

In this series, 43% (23/53) of patients had bacteremia, and 91% of these infections were related to the presence of a CVL. There was neither clustering of cases temporally nor evidence of parenteral-fluid contamination. The true pathogenicity of this bacterium is difficult to discern due to the polymicrobial nature of most of the bacteremic infections, which had not been described previously. This necessitated prolonged, broad-spectrum antibiotic courses. P. agglomerans bacteremia appeared to be transient and did not recur during therapy, and antibiotic courses of 10 to 14 sterile days appeared to be curative. One limitation was that some of the less common Enterobacteriaceae species can be misidentified or susceptibilities incorrectly reported by automated methods (13). Confirmatory tests were not routinely performed for this pathogen unless the VITEK system did not provide an identification.

In conclusion, P. agglomerans is an uncommon cause of infection in children. It can cause bacteremia, often in association with more-conventional pathogens, in children with indwelling central access. However, antimicrobial susceptibility patterns mirror those of other gram-negative enteric pathogens. Diagnoses of bone/joint infections are often delayed due to both the indolent nature of the pathogen and the low level of clinical suspicion for this bacterium. Consequently, the diagnosis is often made when a child has evidence of chronic osteomyelitis, altering the treatment duration and prognosis. P. agglomerans should be suspected as the etiologic agent in cases of penetrating trauma by soil-encrusted objects or vegetation that remain refractory to conventional antimicrobial therapy.