Bacillus species sepsis in the neonatal intensive care unit

Journal of Infection (2005) 51, 390–395

www.elsevierhealth.com/journals/jinf

Bacillus species sepsis in the neonatal intensive care unit Amos Adlera,c,*, Giora Gottesmanb,c, Tzipora Dolfina,c, Shmuel Arnona,c, Rivka Regeva,c, Sophia Bauera,c, Ita Litmanovitza,c a

Department of Neonatology, Sapir Medical Center, Tsharnichovsky Street, Kfar-Saba, Israel Pediatric Infectious Disease Clinic, Sapir Medical Center, Tsharnichovsky Street, Kfar-Saba, Israel c Tel-Aviv University, Sackler School of Medicine, Tel-Aviv, Israel b

Accepted 8 December 2004 Available online 21 January 2005

KEYWORDS Bacillus species sepsis; Neonatal intensive care unit; Percutaneous central venous catheter; Necrotizing enterocolitis

Abstract Objectives. In 2002 there was an increase in the incidence of Bacillus species sepsis in our NICU that was almost completely resolved in 2003–2004 after the NICU was relocated. Our aims were to identify the source, the risk factors, and to characterize the clinical features of these infections. Methods. The epidemiological investigation commenced during the outbreak and thereafter. The patient’s data were collected retrospectively and a case control study was used to analyze the risk factors. Results. There were eight cases of Bacillus species sepsis: five during 2002, two in 2003, and one in 2004. All infants recovered and salvaging percutaneous central venous catheter (PCVC) was successful in 4/6 of the cases. A case control study identified necrotizing enterocolitis (NEC) and PCVC as risk factors in univariate analysis but only NEC in multivariate analysis. No focal source of Bacillus bacteria was identified, but a high load of bacteria was found in the NICU’s air before it was relocated. Conclusion. The risk factors for Bacillus species sepsis in our NICU were NEC and PCVC. The clinical course was milder than previously described, and PCVC was successfully salvaged in most cases. The increase in the incidence could be related to the construction work connected with NICU’s relocation. Q 2004 The British Infection Society. Published by Elsevier Ltd. All rights reserved.

Introduction Non-anthrax Bacillus species are aerobic or facultative-anaerobic, Gram-positive, spore-forming

* Corresponding author. Address: Department of Neonatology, Sapir Medical Center, Tsharnichovsky Street, Kfar-Saba, Israel. Tel.: C972 9 7472229; fax: C972 9 7471889. E-mail address: [email protected] (A. Adler).

rods that are ubiquitous in the environment and can be found in soil, water, dust or air.1 Despite the widespread distribution of Bacillus species organisms, they are rarely associated with actual infection and are more frequently isolated as a culture contaminant.2 Isolation of this organism requires careful clinical evaluation to determine the significance of the finding. Nevertheless, in immunocompromised patients, including preterm infants, Bacillus cereus and other species can cause sepsis,

0163-4453/$30.00 Q 2004 The British Infection Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jinf.2004.12.006

Bacillus species sepsis in neonates meningitis, and other severe systemic infections.1–3 During 2002, there was an increase in the incidence of Bacillus species sepsis in our neonatal intensive care unit (NICU). The objectives of this study were to analyze the risk factors and describe the clinical findings, the laboratory characteristics, and the course of these infections.

391 susceptibility to vancomycin, clindamycin and imipenem (except isolate 6, which was tested only for vancomycin). Since, susceptibility breakpoints for Gram-positive bacilli are not available, we interpreted the susceptibility breakpoints as for Gram-positive cocci (GPC).6,7

Environmental changes and epidemiological investigation of the NICU

Patients and methods Patients and data collection Cases of Bacillus species sepsis in our NICU (a level III 25-beds unit) were retrospectively identified from a computerized data source. We included only patients who had at least two separate blood cultures, in accordance with the acceptable criteria for laboratory-confirmed bloodstream infection (BSI).4 For each case of Bacillus species sepsis, we collected data regarding the baseline characteristics, the clinical and laboratory characteristics of the infection, the sensitivities profile of the isolates, and the outcome.

Cultures were taken from the infants’ formulas, TPN and IL solutions, antiseptic solutions, humidifier fluids, and the ventilators circuit fluids. Environmental cultures surveys were preformed at the end of 2002. Also, air samplings were inoculated on sheep blood agar plates left open for 1 h at several locations in the two rooms of the NICU at the end of 2002 and at the beginning of 2004. During 2002, extensive construction work took place in the vicinity of the NICU, in preparation for its relocation. The NICU was relocated on another floor of the hospital at the beginning of 2003, where construction continued to a lesser extent.

Statistical analysis Case control study In order to identify the risk factors for Bacillus species infection, we conducted a case controlled study. We first sourced four control cases per single index case from the medical records that were matched for birthweight (G250 g), gestational age (G1 week), and period of infection(G4 months). Then, an unrelated assistant, blind regarding the randomization process, randomly selected three controls per index case. Infants with major anomalies or inborn errors of metabolism were excluded. We compared the groups for baseline characteristics, the presence of co-morbidities (respiratory distress syndrome [RDS], mechanical ventilation days, and necrotizing enterocolitis [NEC]), administration of total parenteral nutrition (TPN) and intra-lipid (IL) solutions, and the presence of umbilical and percutaneous central venous catheter (PCVC).

Blood cultures Blood was obtained aseptically and then inoculated into Bactec 9240 bottles (Becton Dickinson; Cockeysville, MD, U.S.A.). Bacillus species were identified as large, aerobic, Gram-positive rods with round or squared-off ends,2 with characteristics colonies morphology, according to standard microbiological methods.5 All isolates were tested for

Correlations between categorical and continuous variables were analysed using the Fisher’s exact test or the Student’s t-test, respectively. Multivariate analysis was preformed by Forward-Stepwise (Likelihood Ratio) binary logistic regression analysis. For all analyses, a was set to 0.05. All analyses were performed using the Statistical Package for the Social Sciences (SPSS for Windows, version 10.).

Results Clinical and laboratory characteristics of Bacillus species sepsis During 2002, five cases of late-onset Bacillus species sepsis were identified (14.2% of the annual number of late-onset sepsis cases). There were two (6.2%) and one (8.3%) cases of Bacillus species sepsis in 2003 and 2004, respectively. No case of Bacillus species sepsis occurred from 1999 to 2001. The clinical and laboratory characteristics of the infants with Bacillus species sepsis are presented in Table 1. Bacillus species sepsis occurred in all infants after the first week of life. Five infants were treated with anti-microbial agents during the 5 days preceding the occurrence of Bacillus species sepsis (Table 1), due to Candida sepsis (case 1),

392

Table 1

Clinical and laboratory characteristics of Bacillus sepsis in eight infants

Case No./date of birth (year) Gestational age (weeks) Birth weight (g) Prior anti-microbiala Age at sepsis (days) Symptoms

Case 1, 2002

Case 2, 2002

Case 3, 2002

Case 4, 2002

Case 5, 2002

Case 6, 2003

Case 7, 2003

Case 8, 2004

25 680 Amp B 18 Lethargy

34 2010 Cefuroxime 23 Fever, lethargy 5.8 102 3.8 1

37 2000 Tazocin 24 Fever, lethargy 17.7 313 12.7 1

36 1735 Tazocin 22 Fever, lethargy 4.8 435 0.3 1

30 1226 Tazocin 23 Apnea, lethargy 4.8 40 5.3 3

29 1508 None 9 Apnea, lethargy 19.3 305 NA 1

32 1600 None 8 Lethargy ill looking 16.5 69 3.7 1

27 870 None 10 Apnea, lethargy 21.8 324 1.3 1

V-1, C-0.3, Ctx O32, Au-4, I-1 Neg Mr N Y/N

V-3, C-0.25, I-0.25

V-1, C-1, Mr-1, Pen-3

V-0.38

V-1.5, C-0.5, I-2

V-1.5, C-0.25, I-1

Neg V N Y/N

Neg Mr N Y/N

V-1.5, C-0.25, I-0.047, PenO32 Neg Mr Yb Y/N

Neg V NA Y/N

Neg V NA Y/N

Neg V N Y/N

WBC count (!103/mL) Platelets count (!103/mL) CRP (mg/L) Persistence of bacteraemia (days) Anti-bacterial MIC (mg/mL)

20 29 10 4

CSF/urine cultures Anti-bacterial agents PCVC removal Survived/complication

Neg Mr Yb Y/N

Not reported

PCVC, percutaneous central venous catheter; Amp B, amphotericin B; Tazocin–piperacillin–tazobactam; V, vancomycin; C, clindamycin; I, imipenem; Ctx, ceftriaxone; Au, augmentin; Mr, meropenem; Pen, penicillin; NA, not assessed. a Anti-microbial therapy in the 5 days prior to Bacillus sepsis. b The tip cultures were negative.

A. Adler et al.

Bacillus species sepsis in neonates

393

Escherichia coli peritonitis (case 2), Klebsiella peritonitis and sepsis (case 3), or necrotizing enterocolitis (NEC) (cases 4 and 5). Bacillus species sepsis was found with nonspecific clinical findings of fever, lethargy, or apnea. Seven infants had mild leukocytosis or leucopenia, four had thrombocytopenia, and six had elevated levels of C-reactive protein (CRP). All infants had sterile CSF and urine cultures. Whenever tested, the isolates were sensitive to vancomycin, clindamycin, and carbapenems but resistant to b-lactam antibiotics (Table 1). The susceptibility patterns of the isolates were not identical, suggesting that the infections were not caused by a single strain. The infants were treated with vancomycin or meropenem for 10 days. Only in 2/ 5 of the cases was the PCVC removed due to persistent bacteraemia. All infants recovered without complications, except for case 1, an extremely VLBW infant who was ventilated for 39 days and had chronic lung disease and retinopathy of prematurity that required cryosurgery.

Case controlled study There were no significant differences between the control and the study groups regarding the gestational age, birth weight, and the mode of delivery (Table 2). Also, there were no differences in the prevalence of RDS, days of mechanical ventilation, Table 2

Risk factors univariate analysis

Variable

Cases (nZ8)

Controls (nZ24)

P-value

Gestational age (weeks)a Birth weight (g)a Cesarean section delivery RDS Mechanical ventilation daysa NEC Umbilical catheter PCVC TPN/intralipid

31 (3.7)

31 (4.3)

NS

1453 (493)

1519 (497)

NS

63%

71%

NS

63% 5.6 (13.5)

38% 1.4 (2.4)

NS NS

50% 62.5%

4% 37.5%

0.009 NS

75% 100%

21% 83%

0.01 NS

NS, not significant (pO0.05); RDS, respiratory distress syndrome; NEC, necrotizing enterocolitis; PCVC, percutaneous central venous catheter; TPN, total parenteral nutrition. a Mean (SD).

administration of TPN and IL solutions, or usage of umbilical catheters. In univariate analysis, NEC and the presence of PCVC were significantly more common in the study groups (p!0.05). In multivariate analysis combining all variables, only NEC remained as a significant risk factor (OR, 35.7; SE, 1.296, pZ0.006).

Environmental investigation All environmental cultures at the end of 2002 were negative, except for the air-sampling cultures, where heavy growth of Bacillus species (O15 colonies per plate) was detected in the two rooms of the NICU. After the relocation, only one out of four air-sampling cultures (from two rooms) grew Bacillus species colonies.

Discussion In the present study, we described the largest series of Bacillus species sepsis in neonates. Before 2002, there were no cases of Bacillus species sepsis in our NICU. Surprisingly, in 2002 Bacillus species became the second most common cause (after coagulasenegative Staphylococcus) of late-onset sepsis. Since the isolates from different infants had different minimum inhibitory concentrations (MICs), we speculate that the outbreak was caused by more than a single strain. Bacillus species (mainly B. cereus) were previously reported as a rare cause of systemic infections in neonates.3,8 Two outbreaks of Bacillus species infections were reported in neonates; both occurred in mechanically ventilated infants, as a result of contamination of manual ventilation balloons and staff hands9 or ventilator circuits.10 In the paediatric population, systemic Bacillus species infections occurred mainly in immunocompromised children, where the possible modes of transmission were central venous catheters,11,12 or the gastrointestinal tract.12,13 Also, CNS infections occurred after invasive procedures,12 VP shunts14 or trauma.15 As previously discussed, the possible modes of transmission of Bacillus species sepsis are by mechanical ventilation, dissemination from the gastrointestinal tract, or are catheter-related. In contrast to the previous reports in neonates,9,10 only one infant (case 1) was ventilated at the time of the sepsis, excluding this mode of transmission. Through the case control study, we identified NEC and the presence of PCVC as risk factors, but only NEC remained significant in the multivariate

394 analysis. In 2/8 of the infants (cases 6 and 7), no risk factor could be identified. The association between NEC and Bacillus species sepsis could be explained either by bacterial dissemination through the vulnerable gastrointestinal tract or due to prolonged usage of PCVC that was required in these infants. Our microbiologic and epidemiologic data does not enable us to differentiate between these two possibilities. As previously described,16 the epidemiological investigation revealed no single focal reservoir of Bacillus species. The different anti-microbial MICs suggest that the outbreak was caused by more than a single strain. However, the high prevalence of Bacillus species sepsis in 2002 correlates with the high load of Bacillus species that was cultured from the air samplings at the former location of the NICU, when extensive construction work occurred. Hospital construction work is associated with certain infections, such as Aspergillus species in immunocompromised hosts.17 The airborne spread of Bacillus species following hospital construction work was related to Bacillus species pseudobacteraemia.18 Therefore, we speculate that the construction process caused an increase in the amount of bacteria in the air, leading to increased gastrointestinal and cutaneous colonization with these organisms. The course of the infection in our cases was relatively benign compared with previously published cases in infants3,8–10,13 or immunocompromised children,11,12,14 where the infection was associated with significant mortality and high CNS involvement. The clinical and the laboratory findings resembled other late-onset GPC infections, such as coagulase-negative staphylococci (CONS),19 with a high rate of thrombocytopenia (57%). The empirical anti-bacterial treatment was changed to vancomycin or meropenem after identifying the Bacillus species, due to the known high resistant rate to b-lactams1,2 which was partially confirmed in our study. Although removal of the indwelling vascular catheter is considered necessary in Bacillus species infections,20 we succeeded in retaining the PCVC in 4/6 of the infants. Our study has several limitations. Because of its retrospective nature, we lacked important microbiologic (such as species and strains identification, which were not available in any clinical microbiology laboratory in Israel at that time) and epidemiologic data (such as staff cultures) that were only partially obtained during the outbreak, and therefore the cause of this epidemic cannot be proved with certainty. However, the available data enabled us to suggest a possible mode of transmission and the source of this epidemic.

A. Adler et al. Our series demonstrate that non-anthrax Bacillus species might be an important cause of late-onset neonatal infection, especially during hospital construction work. It appears that in neonates, the course of the infection might be more benign than previously thought, and that catheter salvage is possible without complications in most cases.

References 1. Bratcher DF. Bacillus species. In: Long SS, editor. Principles and practice of pediatric infectious disease. 2nd ed. Amsterdam: Elsevier; 2003 [available from http://home.mdconsult.com/ das/book/39781485-2/view/1081/999].html/top. Accessed August 10, 2004. 2. Tuazon, CU. Other Bacillus species. In: Mandell GL, Bennett JE, Dolin R, editors. Mandell: principles and practice of infectious diseases. 5th ed. Philadelphia: Churchill Livingstone, 2000. p. 2220–5. 3. Hilliard NJ, Schelonka RL, Waites KB. Bacillus cereus bacteremia in a preterm neonate. J Clin Microbiol 2003; 41:3441–4. 4. O’Grady NP, Alexander M, Dellinger EP, et al. Guidelines for the prevention of intravascular catheter-related infections. CID 2002;35:1281–307. 5. National Committee for Clinical Laboratory Standards. Improved standards MS. A3. Villanova, PA: National Committee for Clinical Laboratory Standards; 1984. 6. McGowan AP, Wise R. Establishing MIC breakpoints and the interpretation of in vitro susceptibility tests. JAC 2001; 48(Suppl S1):17–28. 7. Andrews JM, Wise R. Susceptibility testing of Bacillus species. JAC 2002;49:1040–2. 8. Patrick CC, Langston C, Baker CJ. Bacillus species infections in neonates. Rev Infect Dis 1989;11:612–5. 9. Van Der Zwet WC, Parlevliet GA, Savelkoul PH, et al. Outbreak of Bacillus cereus infections in neonatal intensive care unit traced to balloons used in manual ventilation. J Clin Microbiol 2000;38:4131–6. 10. Gray J, George RH, Durbin GM, et al. An outbreak of Bacillus cereus respiratory tract infections on a neonatal unit due to contaminated ventilator circuits. J Hosp Infect 1999;41: 19–22. 11. Saleh RA, Marshall MA. Bacillus species sepsis associated with Hickman catheters in patients with neoplastic disease. Pediatr Infect Dis J 1987;6:851–6. 12. Gaur AH, Patrick CC, McCullers JA, et al. Bacillus cereus bacteremia and meningitis in immunocompromised children. CID 2001;32:1456–62. 13. Girisch M, Ries M, Zenker M, et al. Intestinal perforation in premature infant caused by Bacillus cereus. Infection 2003; 31:192–3. 14. Berner R, Heinen F, Pelz K, et al. Ventricular shunt infection and meningitis due to Bacillus cereus. Neuropediatrics 1997; 28:333–4. 15. Feder Jr HM, Garibaldi RA, Nurse BA, et al. Bacillus species isolates from cerebrospinal fluid in patients without shunts. Pediatrics 1988;82:909–13. 16. Christenson JC, Byington C, Korgenski EK, et al. Bacillus cereus infections among oncology patients at a children’s hospital. Am J Infect Control 1999;27:543–6.

Bacillus species sepsis in neonates 17. Yonemori K, Takezako N, Nishimura K, et al. Fungal infection in neutropenic patients in a hospital during construction. Jpn J Infect Dis 2002;55:126–7. 18. Loeb M, Wilcox L, Thornley D, et al. Bacillus species pseudobacteremia following hospital construction. Can J Infect Control 1995;10:37–40.

395 19. Maayan-Metzger A, Linder N, Marom D, et al. Clinical and laboratory impact of coagulase-negative staphylococci bacteremia in preterm infants. Acta Paediatr 2000;89:690–3. 20. Mermel LA, Farr BM, Sherertz RJ, et al. Guidelines for the management of intravascular catheter-related infections. CID 2001;32:1249–72.