Jpn. J. Infect. Dis.

, 69, 531–533, 2016

Short Communication

Use of Peripheral Parenteral Nutrition Solutions as a Risk Factor for Bacillus

cereus Peripheral Venous Catheter-Associated Bloodstream Infection at a
Japanese Tertiary Care Hospital: a Case-Control Study
Tomoko Sakihama1* and Yasuharu Tokuda2
1Department of Nursing, International University of Health and Welfare Graduate School, Tokyo; and
2Japan Community Healthcare Organization, Tokyo, Japan

SUMMARY: The risk factors are unclear for peripheral venous catheter-associated bloodstream infec-
tions (PVCBSIs) caused by Bacillus cereus. We aimed to examine for these risk factors in patients with
B. cereus PVCBSI by conducting a 2-year case-control study in a large teaching hospital. We analyzed
all adult cases of B. cereus PVCBSI (37 patients) and 180 controls who were randomly selected from
among patients who had a PVC in place for at least 2 days. Multivariate analysis using a conditional
logistic regression model indicated that independent risk factors were use of a peripheral parenteral
nutrition (PPN) solution with an adjusted odds ratio (OR) of 88.7 (95z confidence interval [CI],
17.4–451.9), and steroid therapy (adjusted OR, 5.7 [95z CI, 1.3–24.4]). In conclusion, use of PPN so-
lutions or steroids was an independent risk factor for B. cereus PVCBSI. Appropriate use of PPN solu-
tions may help prevent B. cereus PVCBSI. Prospective studies are needed to confirm these results.

Bacillus cereus is a ubiquitous aerobic spore-forming
gram-positive bacillus that is distributed throughout
soil, water, and dust, and as well as in hospital environ-
ments (1,2). The pathogenicity in cases of food-borne
B. cereus infection commonly results from the entero-
toxin produced by this organism, and central line-
associated bloodstream infections, meningitis, and
ventilator-associated pneumonia have occurred in
immunocompromised individuals (1–7).
Blood cultures that are positive for B. cereus usually
result from a false-positive contaminant in the environ-
ment where the blood culture had been obtained (8).
However, outbreaks of catheter-related bloodstream in-
fection (CRBSI) have been sporadically reported in as-
sociation with a specific hospital environment or linens
(9–17). Multiple fatal cases with serious B. cereus
CRBSI were recently reported (3,6,13–15,17).
It is unclear what risk factors exist for the develop-
ment of B. cereus CRBSI. Only small case series or
studies on special subgroups such as patients with hae-
matological cancers or patients with central lines have
been conducted on this issue (14,17). No studies have
investigated risk factors of patients with B. cereus
peripheral venous catheter-associated bloodstream in-
fections (PVCBSIs). Thus, we aimed to examine the
risk factors in patients with B. cereus PVCBSI by con-
ducting a case-control study in a large teaching hospi-
tal.
This 2-year case-control study was conducted in an
Received September 30, 2015. Accepted January 12, 2016.
J-STAGE Advance Publication February 19, 2016.
DOI: 10.7883/yoken.JJID.2015.489
*Corresponding author: Mailing address: Department of
Nursing, International University of Health and Welfare
Graduate School, 4F Aoyama 1 Chome Tower, 1-3-3
Minami Aoyama, Minato-ku, Tokyo 107-0062, Japan.
Tel: ＋81-3-6406-8621, Fax: ＋81-3-6406-8622, E-mail:
tomokosakihama＠gmail.com
802-bed acute care teaching hospital in the Kanto area
of Japan between January 2005 and December 2006.
Approval from the Human Subjects Committee of this
hospital was obtained. Patients were 18 years or older,
had at least 2 positive blood cultures for B. cereus with-
in 48 h after admission and were believed to have
PVCBSI, based on the diagnostic criteria of laboratory-
confirmed bloodstream infection according to the U.S.
Centers for Disease Control and Prevention (Atlanta,
GA, USA). The case identification and confirmation
was performed via a 2-step process: (i) all enrolled
prospective patients with positive blood cultures were
checked by a certified infection disease physician with
regard to the PVCBSI criteria (these physicians are
required to consult with all patients for positive blood
cultures) and (ii) the patients were retrospectively
double-checked in accordance with the PVCBSI criteria
by a certified infection control nurse who was a prin-
cipal investigator in this study. Patients who did not
meet the diagnostic criteria were excluded. In patients
with multiple episodes of these infections, the first epi-
sode was used for analysis.
The control patients were 18 years or older and were
admitted for more than 5 days during the 2-year study
period. They had peripheral venous catheters in place
for more than 48 consecutive hours. The Department
of Clinical Microbiology in this hospital cultured and
identified B. cereus. The department fulfilled the
quality standard evaluation by the Japan Council for
Quality Health Care <http://jcqhc.or.jp/>.
A research nurse collected the case and control
patient data using electronic medical records. Based on
a previous study (1,2,9,13,17–18), the data collected
were demographics, admission diagnosis, nutritional
status assessed by the Mini Nutritional Assessment
(MNA) (19), comorbid conditions (e.g., diabetes and
cancer), medication exposure, previous surgery, use of
mechanical ventilation, central venous or urethral
catheter, solute contents of peripheral intravenous infu-

531
sions, and number of days of peripheral catheter place- 0.04), and had a greater proportion of men (P ＝ 0.02).
ment. The data of the solute contents of peripheral in- The case patients also had a greater proportion of
travenous infusions were collected within 30 days be- cerebrovascular diseases (35.1z, P ＝ 0.001).
fore the BSI episode and classified as electrolytes-only, Table 2 shows the potential risk factors for B. cereus
5z glucose and electrolytes, 7.5z glucose and electro- PVCBSI in the case and control patients. Univariate
lytes, or 7.5z glucose, amino acids, and electrolytes analyses identified the significant risk factors as age,
(i.e., PPN) solutions. male sex, steroid therapy, and use of 7.5z glucose,
Risk factors associated with B. cereus PVCBSI were amino acids, and electrolytes solution for PPN. In mul-
analyzed in patients and controls for clinically im- tivariate analysis using a conditional logistic regression
portant variables. Univariate analyses used the chi- model, the factors significantly associated with B.
square test or Fisher's exact test for categorical varia- cereus PVCBSI were steroid therapy (adjusted OR, 5.7
bles, or the Student's t test or Mann–Whitney U test for [95z CI], 1.3–24.4) and use of PPN solution with
continuous variables, when appropriate. A multiva- 7.5z glucose, amino acids, and electrolytes (adjusted
riable-adjusted conditional logistic regression model OR, 88.7 [95z CI], 17.4–451.9). We found that signifi-
was constructed by adjusting for possible risk factors cant risk factors for B. cereus PVCBSI were use of a
for PVCBSI. The odds ratios (ORs) with 95z confi- PPN solution and steroid therapy. For the first time, a
dence intervals (95z CIs) were estimated. Statistical PPN solution was identified in our study as a risk fac-
analyses were conducted using SPSS ver. 21 software tor.
(IBM, Armonk, NY, USA) or STATA ver. 13 software The most probable mechanism for the increased risk
(StataCorp LP, College Station, TX, USA). of developing B. cereus PVCBSI in patients on PPN so-
The institutional review board of the International
University of Health and Welfare (Tokyo, Japan) gave
ethical approval (No. 07-64). Table 1. Clinical characteristics of the participants
In total, 38,495 patients were admitted to the hospi- Case Control
tal during the study period. Thirty-seven patients were Characteristic (n ＝ 217) (n ＝ 37) (n ＝ 180) P
confirmed as cases of B. cereus PVCBSI, and 180
Age (yr), mean (SD) 70.1 (13.9) 64.3 (19.5) 0.04
patients served as the control patients.
Sex: men, n (z) 26 (70.3) 87 (48.3) 0.02
Table 1 shows the clinical characteristics of the case
Admission diagnosis, no. (z)
and control patients. Among the case patients, the
Cerebrovascular disease 13 (35.1) 20 (11.1) 0.001
mean and median age was 70.1 (standard deviation
Solid tumor 9 (24.3) 36 (20.0) 0.66
[SD], 13.9) and 74 (interquartile range [IQR], 60–80)
Gastrointestinal disease 4 (10.8) 30 (16.7) 0.46
years, respectively. This group comprised 26 men
Cardiovascular disease 3 (8.1) 18 (10.0) 0.73
(70z) and 11 women (30z). The median number of
Hematological disease 2 (5.4) 7 (3.9) 0.65
days to the onset of PVCBSI was 44 (range, 3–207).
Community-acquired infection 2 (5.4) 20 (11.1) 0.38
Among the control patients, the mean age was 64.3
Orthopedic disease 2 (5.4) 7 (3.9) 0.65
(SD, 19.5) and 68 (IQR, 56–79) years, respectively. The
Obstetric/gynecologic disease 0 12 (6.7) —
group comprised 87 men (48z) and 93 women (52z).
Others 2 (5.4) 30 (16.7) 0.12
The case patients were older than the controls (P ＝

Table 2. Risk factors of Bacillus cereus bloodstream infections

Univariate analysis Multivariate
analysis
Factor (n ＝ 213) Case Control adjusted OR
OR (95z CI) P
(n ＝ 37) (n ＝ 180) (95z CI)
Demographic
Age (yr), median (IQR) 74 (60–80) 68 (56–79) 0.04 1.0 (0.9–1.1)
Sex: male, no. (z) 26 (70.3) 87 (48.3) 2.5 (1.2–5.4) 0.02 2.5 (0.8–7.7)
Comorbid conditions, no. (z)
Diabetes 9 (24.3) 28 (15.6) 1.0 (0.4–2.3) 0.99 0.7 (0.2–2.6)
Malignancy 10 (27.0) 27 (15.0) 1.3 (0.6–2.8) 0.67 2.4 (0.5–11.5)
Medication exposure in 30 days before the BSI, no. (z)
Anticancer chemotherapy 4 (10.8) 30 (16.7) 0.6 (0.2–1.8) 0.46 0.2 (0.2–1.8)
Steroid 11 (29.7) 22 (12.2) 3.0 (1.3–7.0) 0.01 5.7 (1.3–24.4)
Previous surgery 8 (21.6) 31 (17.2) 1.3 (0.6–3.2) 0.49 1.3 (0.3–5.9)
Mechanical ventilation 1 (2.7) 8 (4.4) 0.6 (0.1–4.9) 0.99
Central venous catheter 0 10 (5.6)
Urethral catheter 16 (43.2) 65 (36.1) 1.3 (0.7–2.8) 0.46
Solute contents of peripheral intravenous infusion, no. (z)
(1) 5z glucose and electrolytes 32 (86.5) 170 (94.4) 0.4 (0.1–1.4) 0.14
(2) 7.5z glucose and electrolytes 5 (13.5) 18 (10.0) 1.4 (0.5–4.1) 0.56
(3) 7.5z glucose, 3z amino acids, and electrolytes (PPN) 32 (86.5) 26 (14.4) 37.9 (13.5–106.2) ＜0.001 88.7 (17.4–451.9)

OR, odds ratio; CI, confidence interval; IQR, interquartile range; BSI, bloodstream infection; PPN, peripheral parenteral nutrition.

532
 PPN Solutions as Risk for B. cereus PVCBSI
lutions is that B. cereus grows rapidly in a PPN solu-
tion and can thus contaminate it. A previous in vitro
study proved B. cereus grows rapidly in PPN solutions
with 3z amino acids, 7.5z glucose, and electrolytes
(Aminofluid; Otsuka Pharmaceutical Factory, Tokyo,
Japan) (20,21). Another possible mechanism account-
ing for the association of the use of PPN solutions with
a higher risk for peripheral thrombophlebitis is greater
vessel damage resulting from the chemical characteris-
tics of the solutions (e.g., a low pH [6.4] and a high
solution-to-serum osmolality ratio of approximately 3)
(18,22,23). A third possible mechanism is that B. cereus
is very capable of forming a biofilm in artificial tubes
(12,24).
In 1996, PPN was introduced after a previous study
demonstrated the efficacy of nutritional support using
peripherally inserted venous catheters in 1993 (25). A
dual-chamber bag system for PPN solutions has been
developed and distributed throughout Japan (26). The
timing of the product development indicates that there
may be an association with B. cereus BSI published in
case reports (9,10,12–15,17).
Possible infectious routes may include bacterial con-
tamination through the skin of patients or healthcare
providers or in venous lines or 3-way stopcocks. Espe-
cially in peripheral venous catheters, 3-way stopcocks
are frequently left open and alcohol-resistant and rap-
idly growing B. cereus could contaminate the lumens of
the venous line hubs. Steroid use was an independent
risk factor. This result may implicate cellular immuno-
deficiency as a host risk factor for B. cereus PVCBSI.
Our study may have some limitations. First, this
study was based on a single hospital retrospective
study. Second, the hospital stay was longer in the case
patients than in the control patients. The B. cereus
PVCBSI may have caused the longer hospital stay in
the case patients or the longer stay may have increased
the risk of contracting a PVCBSI. Third, the selection
of our control patients did not take into consideration a
seasonal trend with a higher risk in summer.
In conclusion, our study suggested that the use of
PPN solutions or steroids was an independent risk fac-
tor for B. cereus PVCBSI. Appropriate use of PPN so-
lutions should be promoted to prevent B. cereus
PVCBSI, and nutritional outcomes should be assessed
by using these solutions. Future prospective studies are
needed to confirm these issues.
Acknowledgments We express our greatest appreciation for the
excellent advice from Dr. Kazuo Endo who was the great infectious
disease physician, but unfortunately passed away at a young age from
an illness.
Conflict of interest None declare.
REFERENCES
1. Bottone EJ. Bacillus cereus, a volatile human pathogen. Clin
Microbiol Rev. 2010;23:382-98.
2. Drobniewski FA. Bacillus cereus and related species. Clin
Microbiol Rev. 1993;6:324-38.
3. Bryce EA, Smith JA, Tweeddale M, et al. Dissemination of
Bacillus cereus in an intensive care unit. Infect Control Hosp
Epidemiol. 1993;14:459-62.
4. Van Der Zwet WC, Parlevliet GA, Savelkoul PH, et al. Outbreak
of Bacillus cereus infection in a neonatal intensive care unit
533
traced to balloons used in manual ventilation. J Clin Microbiol.
2000;38:4131-6.
5. Hernaiz C, Picardo A, Alos JI, et al. Nosocomial bacteremia and
catheter infection by Bacillus cereus in an immunocompetent
patient. Clin Microbiol Infect. 2003;9:973-5.
6. Ozkocaman V, Ozcelik T, Ali R, et al. Bacillus spp. among hos-
pitalized patient haematological malignancies: clinical features,
epidemics and outcomes. J Hosp Infect. 2006;64:169-76.
7. Kalpoe JS, Hogenbirk K, van Maarseveen NM, et al. Dissemina-
tion of Bacillus cereus in a paediatric intensive care unit traced to
insufficient disinfection of reusable ventilator air-flow sensors. J
Hosp Infect. 2008;68:341-7.
8. Lee CC, Lin WJ, Shih HI, et al. Clinical significance of potential
contaminants in blood cultures among patients in a medical cen-
ter. J Microbiol Immunol Infect. 2007;40:438-44.
9. Matsumoto S, Suenaga H, Naito K, et al. Management of sus-
pected nosocomial infection: an audit of 19 hospitalized patients
with septicemia caused by Bacillus species. Jpn J Infect Dis.
2000;53:196-202.
10. Dohmae S, Okubo T, Higuchi W, et al. Bacillus cereus nosoco-
mial infection from reused towels in Japan. J Hosp Infect. 2008;
69:361-7.
11. Kassar R, Hachem R, Jiang Y, et al. Management of Bacillus
bacteremia: the need for catheter removal. Medicine (Baltimore).
2009;88:279-83.
12. Kuroki R, Kawakami K, Qin L, et al. Nosocomial bacteremia
caused by biofilm-forming Bacillus cereus and Bacillus thurin-
giensis. Intern Med. 2009;48:791-6.
13. Sasahara T, Hayashi S, Morisawa Y, et al. Bacillus cereus bacte-
remia outbreak due to contaminated hospital linens. Eur J Clin
Microbiol Infect Dis. 2011;30:219-26.
14. Inoue D, Nagai Y, Mori M, et al. Fulminant sepsis caused by
Bacillus cereus in patients with hematotologic malignancies: anal-
ysis of its prognosis and risk factors. Leuk Lymphoma. 2010;51:
860-9.
15. Uchino Y, Iriyama N, Matsumoto K, et al. A case series of Bacil-
lus cereus septicemia in patients with hematological disease. In-
tern Med. 2012;51:2733-8.
16. Balm MN, Jureen R, Teo C, et al. Hot and steamy: outbreak of
Bacillus cereus in Singapore associated with construction work
and laundry practices. J Hosp Infect. 2012;81:224-30.
17. Kato K, Matsumura Y, Yamamoto M, et al. Seasonal trend and
clinical presentation of Bacillus cereus bloodstream infection:
association with summer and indwelling catheter. Eur J Clin
Microbiol Infect Dis. 2014;33:1371-9.
18. U.S. Centers for Disease Control and Prevention. Guidelines for
the Prevention of Intravascular Catheter-Related Infections,
2011. Available at <http://www.cdc.gov/hicpac/pdf/guidelines/
bsi-guidelines-2011.pdf>. Accessed August 25, 2015.
19. Kaiser MJ, Bauer JM, Ramsch C, et al. Validation of the Mini
Nutritional Assessment short-form (MNA-SF): a practical tool
for identification of nutritional status. J Nutr Health Aging.
2009;13:782-8.
20. Kuwahara T, Kaneda S, Shimono K, et al. Effects of lipid emul-
sion and multivitamins on the growth of microorganisms in
peripheral parenteral nutrition solutions. Int J Med Sci. 2013;10:
1079-84.
21. Sasahara T, Hayashi S, Hosoda K, et al. Bacterial growth in in-
travenous fluid products. The 11th East Asian Conference on
Infection Control and Prevention (EACIC) [abstract], Tokyo,
Japan. EACIC 2012, p24.
22. Zingg W, Pittet D. Peripheral venous catheters: an under-eval-
uated problem. Int J Antimicrob Agents. 2009;34 Suppl 4:S38-
S42.
23. Tagalakis V, Kahn SR, Libman M, et al. The epidemiology of
peripheral vein infusion thrombophlebitis: a critical review. Am
J Med. 2002;113:146-51.
24. Auger S, Ramarao N, Faille C, et al. Biofilm formation and cell
surface properties among pathogenic and nonpathogenic strains
of the Bacillus cereus group. Appl Environ Microbiol. 2009;75:
6616-8.
25. Payne-James JJ, Khawaja HT. First choice for total parenteral
nutrition: the peripheral route. JPEN J Parenter Enteral Nutr.
1993;17:468-78.
26. Otsuka Pharmaceutical Factory. Products-History-Aminofluid,
2013. Available at <http://www.otsukakj.jp/en/profile/history/
index.html>. Accessed August 25, 2015.