Microbial Contamination of Percutaneous Endoscopic Gastrostomy Tube and Nasogastric Tube with Drug-Resistant Bacteria

Enteral nutrition is used in patients with stroke, head-and-neck or esophageal cancer surgery, or repeated aspiration pneumonia. Japanese enteral nutrition guidelines [1] recommend percutaneous endoscopic gastrostomy tube (PEG) for long-term use of >4 weeks and nasogastric tube (NGT) for short-term use of <4 weeks.

Microbial contamination of enteral feeding causes diarrhea [2] and sepsis [3]. PEGs and NGTs may be contaminated by microorganisms as enteral feeding products pass through them, but there have been few reports from Japan on the isolation of microorganisms and no reports on the replacement period for catheters.

Enteral feeding is a risk factor for the appearance of drug-resistant bacteria in nosocomial and ventilator-associated pneumonia [4]. The detection of drug-resistant organisms in patients and their catheters is a problem for both patients and hospitals. In patients, drug-resistant organisms affect antimicrobial therapy. In hospitals, the catheters could become a reservoir for drug-resistant organisms and risk nosocomial infection. Drug-resistant organisms are isolated from patients on enteral nutrition, but it is not known whether they are isolated from the enteral catheters.

This study aimed to determine the microbial contamination of enteral feeding catheters and the percentage of resistant strains of these microorganisms. This study also examined the relationship between the contamination status of the catheters, the duration of catheters placement, and the history of antimicrobial and gastric antisecretory drug use.

Forty-six PEGs and 59 NGTs were collected at Showa University Hospital and Showa University Fujigaoka Rehabilitation Hospital from May 2019 to March 2020. All PEGs were low profile type. The length of low-profile tube is shorter than that of tube type. Forty-six PEGs were collected from 36 patients. Two PEGs were collected from each of the 10 patients. The second PEG was collected 6 months after the first PEG was collected.

Microorganisms were cultured by incubating 20 mL pass/wash solution of sterile purified water on BHI agar medium (37°C) for 24–72 h. The strains were isolated and cultured, then frozen (80°C) and stored.

Colony-forming unit (CFU) was measured using the agar dilution plate method [5].

Bacteria and yeasts were identified according to the Japanese Pharmacopoeia [6].

Antimicrobial susceptibility of bacteria was determined by broth microdilution method using the dry plates (Eiken DP41 and DP42)

Patient characteristics, microbiological data, and medications were collected from the electronic medical record. Patient characteristics included inpatient/outpatient, age, sex, duration of use, and indication for PEG or NGT. Microbiological data included sputum, bronchial aspirate sputum, blood, urine, and stools culture. Medications included the use of gastric antisecretory drugs and antibiotics during the catheter use.

χ² test was used for categorical variables, Student’s t test or Wilcoxon rank-sum test for continuous variables, and Spearman’s correlation for bivariate associations. All tests were performed with SPSS version 25. p < 0.05 was considered to indicate statistical significance.

Characteristics of the patients are shown in Table 1. Comparing the PEG- and NGT-fed patients, 19 patients with PEG were hospitalized and 27 were outpatients, and 59 patients with NGT were hospitalized (p < 0.001). Age was higher in the NGT group (p = 0.042). The median duration of catheter use was 183 days for PEGs and 19 days for NGTs (p < 0.001). The most common indications for PEG were head-and-neck and esophageal cancer (p < 0.001), whereas the most common indication for NGT was stroke (p < 0.001). The use of gastric antisecretory drugs was higher for NGT (p = 0.019) and that of antibiotics was also higher for NGT (p = 0.001) during catheter use.

CFU concentrations were measured in 36 PEGs and 55 NGTs containing bacterial and fungal contamination. Comparing the CFU concentrations for PEGs and NGTs in a box-and-whisker diagram, the median was higher for NGTs, but the difference was not significant (PEG 4.7 × 10⁶ CFU/mL and NGT 6.6 × 10⁶ CFU/mL; p = 0.475) (Fig. 1). There was a correlation between CFU concentration and duration of catheter use for PEG and NGT. There was a small correlation for PEG (r = 0.458, p = 0.458) (Fig. 2) but hardly any for NGT (r = 0.059, p = 0.673) (Fig. 3). CFU concentration with and without the use of gastric antisecretory drugs and antibiotics is shown in Table 2. There was no statistically significant difference in CFU concentrations with the use of gastric antisecretory drugs and antibiotics.

The microorganisms isolated from PEGs and NGTs are shown in Table 3. Microorganisms were isolated in 37 PEGs (80.4%) and 57 NGTs (96.6%) (p = 0.007). Bacteria were isolated in 27 PEGs (58.7%) and 53 NGTs (89.8%) (p < 0.001), and yeasts were isolated in 29 PEGs (63.0%) and 28 NGTs (47.5%) (p = 0.112). Gram-positive bacteria were isolated in 16 PEGs (34.8%) and 27 NGTs (45.8%) (p = 0.172), Enterobacterales were isolated in 16 PEGs (34.8%) and 31 NGTs (52.5%) (p = 0.069), and nonfermenting Gram-negative bacteria were isolated in 10 PEGs (21.7%) and 15 NGTs (25.4%) (p = 0.660).

In Gram-positive bacteria, Enterococcus faecalis was isolated from 7 PEGs (15.2%) and 21 NGTs (35.6%) (p = 0.019). In Enterobacterales, Klebsiella pneumoniae was isolated from 5 PEGs (10.9%) and 14 NGTs (23.7%) (p = 0.089). Escherichia coli was isolated from 5 PEGs (10.9%) and 3 NGTs (5.1%) (p = 0.268). In nonfermenting Gram-negative bacteria, Pseudomonas aeruginosa was isolated from 4 PEGs (8.7%) and 6 NGTs (10.2%) (p = 0.799). In yeasts, Candida glabrata was isolated from 15 PEGs (32.6%) and 6 NGTs (10.2%) (p = 0.004). Candida tropicalis was isolated from 11 PEGs (23.9%) and 8 NGTs (13.6%) (p = 0.172). Candida albicans was isolated from 4 PEGs (8.7%) and 12 NGTs (20.3%) (p = 0.100).

Drug-resistant bacteria were isolated from 19.6% (9 of 46) of PEGs and 23.7% (14 of 59) of NGTs. Tables 4-6 show the drug resistance rates of the bacteria isolated.

The antimicrobial-resistant Gram-positive bacteria are shown in Table 4. Methicillin-resistant Staphylococcus aureus (MRSA) was isolated from 3 of 3 NGTs (100%). Vancomycin-resistant Enterococcus (VRE) was isolated from 1 of 21 NGTs (4.8%).

The antimicrobial-resistant Enterobacterales are shown in Table 5. In E. coli, third-generation cephalosporin-resistant (3GC-R) bacteria were isolated from 4 of 5 PEGs (80%) and fluoroquinolone-resistant (FQ-R) bacteria from 3 of 5 PEGs (60%). In K. pneumoniae, 3GC-R bacteria were isolated from 3 of 14 NGTs (21%) and FQ-R bacteria from 2 of 14 NGTs (14%). In Enterobacter spp., 3GC-R bacteria were isolated from 1 of 5 PEGs (20%) and 2 of 4 NGTs (50%), carbapenem-resistant (C-R) bacteria from 1 of 5 PEGs (20%), and FQ-R bacteria from 1 of 4 NGTs (25%).

The antimicrobial-resistant nonfermenting Gram-negative bacteria are shown in Table 6. In P. aeruginosa, C-R bacteria were isolated from 2 of 4 PEGs (50%) and 2 of 6 NGTs (33%), aminoglycoside-resistant bacteria from 1 of 4 PEGs (25%), and FQ-R bacteria from 2 of 4 PEGs (50%) and 1 of 6 NGTs (17%). In Pseudomonas spp., C-R bacteria were isolated from 1 of 2 PEGs (50%) and 1 of 2 NGTs (50%), aminoglycoside-resistant bacteria from 1 of 2 PEGs (50%), and FQ-R bacteria from 1 of 2 PEGs (50%) and 1 of 2 NGTs (50%).

We examined previous bacteriological tests of patients in whom drug-resistant bacteria were detected. The same drug-resistant bacteria were found in 2 of 9 patients in PEGs and 3 of 14 patients in NGTs. In Gram-positive bacteria, MRSA was detected in the bronchial aspirate sputum of 2 of 3 patients in whom MRSA was isolated from NGTs. There were no E. faecalis isolates from the sputum of patients in whom VRE was isolated from NGTs. In Enterobacterales, E. coli was isolated from the sputum of 2 patients and the stools of 1 of 4 patients, in whom 3GC-R E. coli strains were isolated from PEGs. Two of these strains were extended-spectrum β-lactamases.

In PEG patients, duration of catheter use and use of gastric antisecretory agents or antibiotics were compared to the classification of isolated bacteria, but there were no significant differences (Table 7). In NGT patients, longer duration of catheter use detected Enterobacterales (p = 0.022) and the use of gastric antisecretory drugs reduces yeasts (p = 0.023) (Table 8).

This is the first report in Japan on microorganisms isolated from enteral feeding catheters. Comparing PEGs and NGTs, more microorganisms were isolated from NGTs. Rothia kristinae, C. glabrata, and Candida parapsilosis were significantly more frequently isolated from PEGs, whereas E. faecalis was significantly more frequently isolated from NGTs. Drug-resistant bacteria were isolated from 19.6% (9 of 46) in PEGs and 23.7% (14 of 59) in NGTs. One each of P. aeruginosa and Pseudomonas spp., resistant to carbapenem, aminoglycoside, and fluoroquinolone, were isolated from PEG. Three MRSAs and one VRE were isolated from NGTs.

The median CFU concentration was 4.7 × 10⁶ CFU/mL for PEGs and 6.6 × 10⁶ CFU/mL for NGTs. In enteral feeding products, contamination >10⁴ CFU/mL was reported to cause diarrhea [2], whereas contamination >10⁶ CFU/mL was reported to cause sepsis [3]. Although the number of bacteria entering the stomach differs between contamination of enteral feeding products and contamination of the catheters, the potential infection risk should be considered for contamination of the catheters.

NGTs have more bacterial species isolated than PEGs. We considered that because NGTs are longer than PEGs. PEGs connect the stomach to the abdomen, whereas NGTs connect the nose to the stomach and are longer, providing more area for bacterial attachment. A previous study examined microorganisms in the stomachs of patients on enteral nutrition [7]. The bacterial isolation rate in the stomachs of NGT patients was higher than in PEG patients. The length of the catheter may also affect contamination in the stomach. On the other hand, the Candida species showed up in many PEGs. We considered that Candida species showed up in PEGs had come from skin because Candida species live on the skin [8] and PEGs are in direct contact with the skin.

According to the 2024 WHO Bacterial Priority Pathogen List [9], 3GC-R Enterobacterales and C-R Enterobacterales are in the critical group. C-R P. aeruginosa and MRSA are in the high group. These drug-resistant bacteria were isolated from both PEGs and NGTs. Some drug-resistant bacteria came from the patient. But the rate of this origin was low. Other drug-resistant bacteria might have come from medical devices and health care workers in the hospital. There are reports that enteral feeding products are a cause of nosocomial infections [10, 11]. It is possible that drug-resistant bacteria could have adhered to the catheters from the person preparing the enteral feeding and from the bottle of enteral feeding products.

There was no correlation between duration of catheter use and CFU concentrations containing bacteria and yeasts in NGT. On the other hand, NGTs with detectable Enterobacterales had longer duration of catheter use. Enterobacterales contain 3GC-R and C-R bacteria. Considering the contamination of Enterobacterales, long-term use of NGT should be avoided.

Microorganisms were isolated in catheters collected on day 9 for PEG and day 3 for NGT in this study. Previous studies reported bacterial biofilm formation on a newly inserted NGT within 1 day in preterm infants and elderly patients [12, 13], and pathogenic bacteria were isolated after 7 days for PEG [14]. Therefore, it is necessary to be aware of the presence of microorganisms in catheters.

Gastric antisecretory drugs are used in PEGs and NGTs to prevent reflux esophagitis. In poststroke patients, antiplatelet agents are used to prevent blood clots, and this is also used to prevent intragastric bleeding as a side effect. Gastric antisecretory drugs cause bacterial and fungal growth in the stomach by raising the intragastric pH [15, 16]. In this study, there was no significant difference between the use of gastric antisecretory drugs and bacterial isolation. There was no correlation between gastric pH and the use of gastric antisecretory drugs in older patients using NGTs [17]. Thus, bacteria would be detected regardless of the use of gastric antisecretory agents because the catheters for PEG and NGT enteral feedings are not exposed to gastric acid after the nutrients have passed through them.

There may be some possible limitations in this study. All NGTs were for inpatients, whereas inpatient and outpatient PEGs existed. The catheters were cared for by nurses in the inpatient setting but by themselves or family members in the outpatient setting. Management methods, such as meal preparation and cleaning after use, differed. Japanese enteral nutrition guidelines recommend that PEG and NGT be flushed with 20–30 mL of water after using [1] and which are followed in hospitals. For these reasons, there could be differences in bacterial CFU and isolates. Moreover, outpatients were seen only for PEG exchange, and therefore it was impossible to determine past antimicrobial use and isolation of antibiotic-resistant bacteria.

In this study, most catheters were replaced at 6 months for PEG and at 2 or 4 weeks for NGT. Collecting catheters of different durations of catheter use may also create differences in CFU concentrations and isolated bacteria.

In conclusion, PEGs and NGTs were found to be contaminated with microorganisms at high rates. Some are also contaminated with drug-resistant bacteria, which is a problem for both patients and hospitals. This study provides a rationale for future appropriate use in enteral feeding catheters.

We acknowledge the contribution of Professor Fuyuhiko Yamamura and Associate Professor Hisato Fujihara for collected catheters. We would also like to thank Yukie Saito and Kohei Yahagi for identification and genetic analysis of bacteria.

The study was conducted in accordance with the Ethical Guidelines for Medical and Health Research Involving Human Subjects based on the World Medical Association Declaration of Helsinki. Opt-out informed consent protocol was used for use of participant data for research purposes. This consent procedure was reviewed and approved by Showa University Research Ethics Review Board, Approval No. P332, date of decision October 18, 2018. The study protocol was reviewed and approved by Showa University Research Ethics Review Board, Approval No. P332, date of decision October 18, 2018.

The authors have no conflicts of interest to declare.

The authors received no financial support for this research.

Shota Kanamori collected data, conducted the experiments, and wrote the original draft of the manuscript. Hiroaki Koya collected catheters and data. Naomi Kurata contributed to the study conception and design. Keiko Ishino contributed to design the experiments and proofread the entire draft.

The data that support the findings of this study are not publicly available due to their containing information that could compromise the privacy of research participants but are available from the corresponding author K.I. upon reasonable request.