BRBiomedical Reports2049-94342049-9442D.A. SpandidosBR-22-2-0189510.3892/br.2024.1895ArticlesEffect of infusion time on risk of Bacillus cereus bloodstream infections in patients administered with BFLUID amino acid infusions via peripheral venous cathetersTakanoAtsushi1NakamaruYuta2SakoKen-Ichi2TakanoAsuka1IgaMasanori1MachidaMitsuru1KakuTeppei2IshimuraAtsushi2NakasugaHirotaka2MatsudaYoshikazu2FujikakeYoshio2MaedaTomoji2Department of Pharmacy, Saitama Red Cross Hospital, Saitama 330-8553, JapanDepartment of Clinical Pharmacy, Nihon Pharmaceutical University, Saitama 362-0806, JapanCorrespondence to: Dr Tomoji Maeda, Department of Clinical Pharmacy, Nihon Pharmaceutical University, 10281 Komuro, Ina-machi, Kita-adachi-gun, Saitama 362-0806, Japan niansihui@126.com t-maeda@nichiyaku.ac.jp
Abbreviations: OR, odds ratio; aOR, adjusted OR; AUROC, area under the receiver operating characteristic curve; BSI, bloodstream infection; PPN, peripheral parenteral nutrition
Comprehensive electrolyte infusions containing amino acids, including BFLUID, are a source of nosocomial infections caused by Bacillus cereus. It is recommended that clinicians finish the administration of BFLUID within 6 h; however, this recommendation currently lacks supporting empirical evidence. Accordingly, the present study aimed to investigate the risk of B. cereus infection associated with BFLUID infusion according to the administration time. In the present single-center retrospective observational cohort study, clinical information was extracted from the medical records of patients aged ≥18 years who received BFLUID infusions via a peripheral line between April 2017 and October 2020 at Saitama Red Cross Hospital (Saitama, Japan). A total of 96 patients were enrolled during the study period, of whom 32 developed B. cereus infection. No patients with BFLUID infusions administered within 6 h developed B. cereus infection. Compared with patients who received BFLUID infusions within 8 h, those who received BFLUID infusions after 8 h had a 7-fold higher risk of infection. The present study provided empirical evidence supporting the recommendation that amino acid infusions should be administered to patients within 6 h to minimize the risk of B. cereus infection.
catheter-related bloodstream infectionsBacillus cereusperipheral venous cathetersamino acid infusionpredictive modelFunding: The present study was supported by a Nihon Pharmaceutical University Research Grant (grant no. NPU-2).Introduction
Bacillus cereus is a flagellated, spore-forming, gram-positive Bacillus that is widely distributed in the environment, including soil, dust and sewage. Except for cases of food poisoning caused by enterotoxin-producing strains, it rarely causes infections in humans. Moreover, when B. cereus is detected in blood cultures, it is often assumed to be a contaminant (1). However, in the field of clinical microbiology, B. cereus is considered a pathogen (2). According to the Japan Nosocomial Infections Surveillance in 2022, the detection rate of B. cereus in blood culture-positive samples is 1.2% (3). B. cereus is an opportunistic pathogen and nosocomial infections of neonates and immunocompromised patients are problematic as they can result in the development of severe sepsis (4,5). The main routes of B. cereus infection are inappropriate handling of intravascular indwelling catheters and contamination from the environment, including hospital bed linens (6,7).
In 2019, the Ministry of Health, Labor and Welfare of Japan notified medical institutions of cases of nosocomial infections caused by B. cereus and introduced a requirement that medical institutions implement control measures (8). In Japan, amino acid infusions have been associated with peripheral venous catheter-associated bloodstream infection (PVC-BSI) caused by B. cereus (9,10). Patients receiving peripheral parenteral nutrition (PPN) containing amino acids have a higher risk of developing B. cereus PVC-BSI compared with patients receiving other types of parenteral nutrition solution (9,10). Additionally, long-term infusion of a large volume of fluid is associated with the development of BSIs caused by Bacillus spp (6).
A previous case-control study conducted in a large teaching hospital identified risk factors for B. cereus PVC-BSI (10). In this previous study, unadjusted logistic regression analysis identified the following risk factors: Increased age, male sex, steroid therapy and the use of a solution containing 7.5% glucose, amino acids and electrolytes for PPN. A subsequent multivariable analysis using a conditional logistic regression model identified steroid therapy and use of PPN solution with 7.5% glucose, amino acids and electrolytes as factors significantly associated with B. cereus PVC-BSI. The most probable mechanism for the increased risk of developing B. cereus PVC-BSI in patients receiving PPN solutions is that B. cereus can contaminate and rapidly proliferate in PPN solutions. An in vitro study reported that B. cereus proliferates rapidly in PPN solutions containing 3% amino acids, 7.5% glucose and electrolytes (11,12). Another possible mechanism underlying this association is that the chemical characteristics of the solutions [an acidic pH (6.4) and a high solution-to-serum osmolality ratio of ~3] can cause peripheral thrombophlebitis and large vessel damage (10). An additional possible mechanism is that B. cereus can easily form a biofilm in artificial tubing (13). Possible infection routes include bacterial contamination through the skin of patients or healthcare providers and bacterial contamination in venous lines or 3-way stopcocks. Specifically, in peripheral venous catheters, 3-way stopcocks are frequently left open. Additionally, B. cereus, which is alcohol-resistant and exhibits rapid proliferation, can contaminate the lumens of the venous line hubs (6).
In vitro studies have reported that, to prevent B. cereus PVC-BSI, the infusion time for PPN containing amino acids should not exceed 6 or 8 h (14,15). However, clinical evidence regarding the infusion time limit for preventing catheter-related infection to support current guidelines is lacking. Accordingly, the present study aimed to conduct a quantitative analysis to provide evidence regarding the optimal infusion time. The primary objective of the present study was to investigate the effect of the time taken to administer BFLUID PPN solution on the risk of BSI caused by B. cereus.
Materials and methodsStudy design and patients
The present study conducted a single-center, retrospective cohort study. Clinical information was extracted from the electronic medical records of patients aged ≥18 years who received BFLUID (Otsuka Pharmaceutical Factory Co., Ltd.) infusions via a peripheral line between April 2017 and October 2020 at Saitama Red Cross Hospital (Saitama, Japan). BSI was determined based on cases isolated from ³2 sets of blood cultures within 48 h (Fig. S1), in which there was no evidence of a source of infection other than the catheter, the clinical presentation was fever and antimicrobial therapy was initiated. B. cereus was cultured on Trypticase Soy Agar M with 5% Sheep Blood (BD Biosciences). The identity of the organism was assessed according to the manufacturer's protocols using the VITEK 2 System (bioMérieux Japan, Ltd.), an automated bacterial identification and susceptibility testing system. Propensity score matching was used to minimize biases resulting from differences in background factors for infusions administered within 6 vs. >6 h and within 8 vs. >8 h.
The present study was conducted in accordance with the ethical standards of Saitama Red Cross Hospital and the Declaration of Helsinki. The present study protocol was approved by the ethics committee of the Saitama Red Cross Hospital (approval no. 20-T; Saitama, Japan) prior to study initiation.
Clinical data collection
The following clinical data were collected: Demographic characteristics of patients (sex, age, body weight and height), laboratory values [serum albumin (ALB), hemoglobin A1c (HbA1c) and estimated glomerular filtration rate (eGFR)], clinical characteristics (dialysis, cancer, chemotherapy, steroid therapy and immunosuppressant use), dosage regimen, time to administer BFLUID and duration of BFLUID use.
Statistical analysis
Patient characteristics were expressed as the mean ± SD or count (percentage), as appropriate. Between-group comparisons were performed using the unpaired Student's t-test for continuous variables and the χ2 test for categorical variables. A multivariable logistic regression analysis was performed to determine the propensity score using the following demographic characteristics and baseline laboratory values: Sex, age, body weight, height, ALB and HbA1c levels, eGFR and duration of BFLUID use. The 1:1 propensity score matching was performed using nearest-neighbor matching, with a maximum caliper width of 0.2 SDs of the propensity scores. The ability of the model to discriminate the cut-off time was assessed using Harrell's C index and the standardized mean difference. P<0.05 was considered to indicate a statistically significant difference.
A multivariable logistic regression analysis was performed to identify independent predictors of B. cereus BSI. Odds ratios (ORs) and 95% CIs were calculated for variables included in the model. Statistical analyses were performed using JMP pro statistical software (version 17.2; SAS Institute, Inc.) and SPSS (version 27; IBM Corp.).
Model development
All clinically relevant parameters were included in the multivariable logistic regression model for identifying significant predictors of B. cereus PVC-BSI. Backward elimination was performed based on the P-value of each predictor and the overall predictive performance of the model, which was assessed using the area under the receiver operating characteristic curve (AUROC). First, non-contributing factors with large P-values (P>0.1) and the lowest magnitude of effect (ORs closest to 1.00) were eliminated from the regression model. After each predictor was removed, the performance of the model was checked using the AUROC. The predictor was re-entered into the model if its removal caused a substantial decrease in the AUROC. These steps were performed iteratively until all the remaining predictors within the model had a P<0.10 on the condition that the AUROC of the reduced model was preserved. The discrimination and calibration of the final reduced model was assessed using the AUROC curve.
ResultsPatient characteristics
During the study period, 97 patients were enrolled; among them, 32 developed B. cereus infection, excluding 1 patient who was omitted from the analysis due to missing data (Fig. 1). The clinical characteristics of patients with and without B. cereus infection were analyzed (Table I). After propensity score matching, 23 and 31 matched patient pairs were identified for infusions administered within 6 h (Table II) and 8 h (Table III), respectively. Before the propensity score-matched analysis, patient sex significantly differed between patients with longer and shorter infusion times for infusions administered within 6 and 8 h and the eGFR significantly differed between patients with longer and shorter infusion times for infusions administered within 8 h (Tables II and III).
Propensity score allocation
The propensity score was calculated using patient characteristics. The 6 and 8 h models had high discriminative power with Harrell's C index values of 0.69 and 0.75, respectively. Propensity score-matched analyses with matching performed on infusion times >6 or >8 h showed smaller standardized mean differences compared with those before matching for all variables except age and HbA1c level (Tables II and III). No patients who received BFLUID infusions within 6 h developed B. cereus BSI. However, 8/23 patients who received BFLUID infusions administered over >6 h developed B. cereus BSIs. The proportion of patients with B. cereus BSI was significantly higher among those who received infusions administered over >8 h compared with those who received infusions administered over <8 h. B. cereus BSIs occurred in 2/31 patients who received BFLUID infusions administered within 8 h and in 14/31 patients who received BFLUID infusions administered over >8 h. The relative risk was 7.0 (95% CI, 1.73-28.3) (Tables SI and SII). This indicated that patients with an infusion time of BFLUID >8 h had a 7-fold increased risk of BSI compared with those with an infusion time of ≤8 h. However, due to the wide CI, cautious interpretation of this estimate is necessary.
B. cereus PVC-BSI score
In the multivariable analysis (Table SIII), the BFLUID infusion time [adjusted OR (aOR), 1.22; 95% CI, 1.11-1.34; P=0.001], duration of BFLUID use (aOR, 1.11; 95% CI, 1.03-1.21) and ALB level (aOR, 0.35; 95% CI, 0.21-0.58) were associated with a significantly increased risk of B. cereus BSI.
Scores of 0-3, 0-8 and 0-5 points were assigned to ALB, time to administer BFLUID (h/infusion) and duration of BFLUID use (days), respectively. The scores for the three predictors were added to calculate the total B. cereus PVC-BSI score, on a scale of 0-16 points (16) (Table SIV). The probability of B. cereus PVC-BSI in patients with a score of 18 points was 98% (Table SV). Further, the scores were higher in patients with B. cereus infection compared with those without infection (Fig. 2). The AUROC of the B. cereus PVC-BSI score was 0.90 (95% CI, 0.83-0.96) (Fig. S2), indicating satisfactory diagnostic performance. The results of the AUROC analysis for the diagnostic performance of the B. cereus BSI score indicated a cut-off value of 5. Therefore, the PVC-BSI score should be <5.
Discussion
An in vitro study suggested that BFLUID administration is associated with an increased risk of B. cereus PVC-BSI and that one of the causes of this infection was the growth of B. cereus in the infusion bag during prolonged infusion (14,15). To the best of our knowledge, the present study was the first to report risk factors for B. cereus PVC-BSI in patients receiving amino acid infusions via a peripheral venous catheter. It was demonstrated that the BFLUID infusion time, duration of use and low serum ALB level were significantly associated with the incidence of B. cereus PVC-BSI. Consistent with the findings of the present study, previous studies have reported risk factors for BSI in patients receiving PPN therapy, such as a longer average daily infusion time of PPN and the administration of intravenous fluids (17). The serum ALB level is an indicator of nutritional status and low serum ALB levels are associated with an increased susceptibility to infectious diseases (18). Moreover, the present study conducted a propensity score-matched analysis to investigate whether the time within which BFLUID was administered affected the risk of B. cereus PVC-BSI. The incidence of B. cereus PVC-BSI increased with increasing BFLUID administration time. Based on the diagnostic score, AUROC analysis indicated a cut-off value of 5 h for B. cereus PVC-BSI. Further studies are warranted to investigate whether the B. cereus PVC-BSI score developed in the present study has utility in the prevention of B. cereus PVC-BSI in clinical practice.
PPN was introduced in Japan in 1996 after a study demonstrated the efficacy of nutritional support using peripherally inserted venous catheters (19). Compared with parenteral nutrition solutions administered via a central line, administration via peripheral lines increases the risk of catheter-associated BSIs (9). Total parenteral nutrition, which is administered by a central venous line, is prepared under aseptic conditions based on guidelines for preventing bacterial contamination. By contrast, PPN does not require special techniques and PPN solutions are prepared in the ward, which is a non-sterile environment. This may lead to B. cereus PVC-BSI.
The present study had some limitations. First, it was a single-center retrospective study; therefore, the findings may lack generalizability. Half of the study participants (49 patients) had cancer; therefore, these findings might only be generalizable to patients with cancer. The sample size was insufficient to stratify the analysis according to cancer status. The electronic health records of 11/49 patients with cancer did not specify whether they were receiving cancer treatment at the time of BFLUID administration; therefore, further detailed studies are warranted. Second, the eGFR at baseline significantly differed between patients with and without B. cereus BSIs, which may have resulted in residual confounding. A total of 6 patients were recorded as having a BMI <14 kg/m2 and an eGFR >200 ml/min/1.73 m2. Serum creatinine levels tend to be higher in people with increased muscle mass and lower in those with lower muscle mass, and eGFR can also be lower or higher accordingly. In addition, considering that the eGFR was calculated and not directly measured and some results were judged to be spuriously high, eGFR could not be adjusted for in the multivariable analysis.
The present study investigated the incidence of and risk factors for developing B. cereus BSI in patients receiving PPN therapy via BFLUID administration. The identified risk factors included the BFLUID infusion time, duration of use and low serum ALB level. To reduce the risk of PVC-BSI, it could be recommended that BFLUID infusion is administered within 6 h and that PPN be performed to keep the score established in this study as low as possible. Further multicenter studies are warranted to validate the findings of the present study.
Supplementary Material
Appearance of <italic>Bacillus cereus</italic> colonies cultured on Trypticase Soy Agar M with 5% sheep blood.
Receiver operating. characteristic curve for the diagnostic performance of the <italic>Bacillus cereus</italic> BSI score. The <italic>B. cereus</italic> BSI score had a sensitivity of 71.9% and specificity of 90.6% using a cut-off value of 5. BSI, bloodstream infection.
Number of <italic>Bacillus cereus</italic> BSIs with a 6 h BFLUID infusion time cut-off period.
Number of <italic>Bacillus cereus</italic> BSIs with an 8 h BFLUID infusion time cut-off period.
Factors associated with <italic>Bacillus cereus</italic> bloodstream infection in the multivariable analysis.
Diagnostic performance of the <italic>Bacillus cereus</italic> peripheral venous catheter-associated bloodstream infection score.
Probability of a <italic>Bacillus cereus</italic> PVC-BSI according to the total PVC-BSI score.
Acknowledgements
Not applicable.
Availability of data and materials
The data generated in the present study are not publicly available due to patient confidentiality but may be requested from the corresponding author.
Authors' contributions
AtT, AsT, MI, MM, TK, AI, HN, YM and YF collected the patient data. YN, KS and TM performed data analysis and wrote the manuscript. TM and YM confirm the authenticity of all the raw data. All authors have read and approved the final version of the manuscript.
Ethics approval and consent to participate
The present study was conducted in accordance with the ethical standards of Saitama Red Cross Hospital and the Declaration of Helsinki Declaration. The study protocol was approved by the ethics committee of the Saitama Red Cross Hospital (approval no. 20-T) before study initiation. The requirement for informed consent was waived due to the retrospective nature of the study. Patient consent was obtained though the opt-out method.
Patient consent for publication
Not applicable.
Competing interests
The authors declare that they have no competing interests.
ReferencesBottoneEJBacillus cereus, a volatile human pathogenClin Microbiol Rev2338239820102037535810.1128/CMR.00073-09PearsonHEHuman infections caused by organisms of the Bacillus speciesAm J Clin Pathol535065151970542445110.1093/ajcp/53.4.506Japan Nosocomial Infections Surveillance in 2022: Ministry of Health, Labour and Welfare, 2023. https://janis.mhlw.go.jp/report/open_report/2022/3/1/ken_Open_Report_202200.pdf.KusamaYHonmaIMasudaMGotoHOnoderaSBacillus cereus outbreak in normal neonates at our hospitalKankyo kansen-shi [Environmental Infection Magazine]303853902015MatsumotoSSuenagaHNaitoKSawazakiMHiramatsuTAgataNManagement of suspected nosocomial infection: an audit of 19 hospitalized patients with septicemia caused by Bacillus speciesJpn J Infect Dis53196202200011135704AsoYNagatomiMNakazawaTSasakiSIshiKExamination of infusion fluid type and environmental factors involved in increased Bacillus cereus bloodstream infectionJpn J Environ Infect2781902012NakamuraKYarimizuAIharaTSawayamaYIshimaruTAn analysis of a case in which Bacillus cereus bacteremia developed from the use of carry-on towels which resulted in the horizontal transmission of this organism in the same wardKansenshogaku zasshi9280852018(In Japanese)HinoCOzakiMKitaharaTKoudaKShikichiKNakamuraIKawaiSOieSPeripheral parenteral nutrition solutions and bed bath towels as risk factors for nosocomial peripheral venous catheter-related bloodstream infection by bacillus cereusInt J Med Sci2056657120233708273710.7150/ijms.82054KutsunaSHayakawaKKitaKKatanamiYImakitaNKasaharaKSetoMAkazawaKShimizuMKanoTRisk factors of catheter-related bloodstream infection caused by Bacillus cereus: Case-control study in 8 teaching hospitals in JapanAm J Infect Control451281128320172859601910.1016/j.ajic.2017.04.281SakihamaTTokudaYUse 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 studyJpn J Infect Dis6953153320162690221910.7883/yoken.JJID.2015.489KuwaharaTKanedaSShimonoKInoueYEffects of lipid emulsion and multivitamins on the growth of microorganisms in peripheral parenteral nutrition solutionsInt J Med Sci101079108420132386918210.7150/ijms.6407SasaharaTHayashiSHosodaKInagawaHMorisawaYHiraiYBacterial growth in intravenous fluid products. In: Proceedings in the 11th East Asian Conference on Infection Control and Prevention (EACIC) [abstract], Tokyo, Japan. EACIC, p24 2012.KurokiRKawakamiKQinLKajiCWatanabeKKimuraYIshiguroCTanimuraSTsuchiyaYHamaguchiINosocomial bacteremia caused by biofilm-forming Bacillus cereus and Bacillus thuringiensisIntern Med4879179620091944397310.2169/internalmedicine.48.1885ShiotaAAsaiNKoizumiYWatanabeHSakataMKurumiyaATakahashiTMuramatsuYHagiharaMSuematsuHExtended drip infusion of peripheral parental nutrition containing amino acids might be associated with Bacillus cereus bloodstream infectionAm J Infect Control471154115620193104768910.1016/j.ajic.2019.02.020SakaiYKonishiTObayashiYHondaKAkaeSIshiharaKStudy of growth level of Bacillus cereus in various infusion fluidsShimane J Med Technol4019232012BonnettLJSnellKIECollinsGSRileyRDGuide to presenting clinical prediction models for use in clinical settingsBMJ365l73720193099598710.1136/bmj.l737ShimodaMTanakaYMorimotoKYoshimoriKOhtaKRisk factors for bloodstream infection in patients receiving peripheral parenteral nutrition. Intern Med 2024 (Epub ahead of print).MinatoguchiSNomuraAImaizumiTSasakiSOzekiTUchidaDKawarazakiHSasaiFTomitaKShimizuHFujitaYLow serum albumin as a risk factor for infection-related in-hospital death among hemodialysis patients hospitalized on suspicion of infectious disease: A Japanese multicenter retrospective cohort studyRen Replace Ther4302018Payne-JamesJJKhawajaHTFirst choice for total parenteral nutrition: The peripheral routeJPEN J Parenter Enteral Nutr174684781993828941710.1177/0148607193017005468
Study flowchart.
Box-and-whisker and violin plots of the distribution and density of the Bacillus cereus BSI score. *P<0.01. BSI, bloodstream infection.
Demographic and clinical characteristics of patients with and without Bacillus cereus peripheral venous catheter-related bloodstream infections.