Catheter‐Related Bloodstream Infections Among Hemodialysis Patients in Northern Tanzania: Insights From a Single‐Centre Retrospective Study

ABSTRACT

Background and Aims

A catheter‐related bloodstream infection (CRBSI) is a life‐threatening complication of hemodialysis. It is responsible for significant morbidity and mortality and a costly long hospital stay. Despite its burden, little is known about the factors associated with it and the antibiogram of its responsible causative bacteria. This study aimed to evaluate the determinants of hemodialysis CRBSI and the appropriate antibiogram for the isolated bacterial pathogens among patients attending the KCMC hemodialysis unit.

Methods

A hospital‐based retrospective analytical cross‐sectional study involved adult patients with CKD or AKI undergoing hemodialysis at KCMC with tunneled or non‐tunneled central venous catheters who attended from January 1, 2021 to December 31, 2022. Data were collected from the hospital's electronic medical record system. Statistical analysis was performed using SPSS version 27. Multivariate logistic regression was used to determine the factors associated with CRBSI. A p‐value of < 0.05 was considered statistically significant.

Results

A total of 286 patients were analyzed. The patients were predominantly males (70.6%), and more than half were above 55 years old (56%), with a mean age of 58 (SD ± 14.8) years. 56.6% developed CRBSI and were predominantly gram‐positive (69.7%). Major associated factors were hypoalbuminemia of < 35 g/dL, diabetes mellitus, anemia of hemoglobin < 10 g/dL, and having a femoral central venous catheter. For gram‐positive bacteria, the most efficient antibiotic was vancomycin, with an overall average efficacy of 99% for all isolates. For gram‐negative bacteria, the most efficient antibiotic was meropenem, with an overall average efficacy of 82.1% for all isolates.

Conclusion

Over 50% of patients developed CRBSI and gram‐positive bacteria were the major causative bacteria. Associated factors were multifactorial and preventive care would be ideal to reduce the number of CRBSI based on these findings.

1. Introduction

Catheter‐related bloodstream infection (CRBSI) is a life‐threatening complication of hemodialysis and is responsible for significant morbidity and mortality, as well as costly hospital admissions [1]. The prevalence of CRBSI varies from region to region. Even in high‐income countries, it is still a significant health problem.

Several measures are currently adopted into modern practice to reduce the occurrence of CRBSI among hemodialysis patients; these include antimicrobial catheter lock, chlorohexidine skin antisepsis at the catheter site, antibiotic ointment application at the exit site, and catheter hub disinfection [2].

In Africa, data is scarce. In a cohort study in South Africa, 34.2% developed CRBSI. Meropenem and vancomycin were the most sensitive antibiotics for gram‐negative and gram‐positive bacteria, respectively [3]. A cross‐sectional study at Muhimbili National Hospital in Tanzania found an overall CRBSI prevalence of 35.7%. The most efficacious antibiotics were vancomycin and meropenem for gram‐positive bacteria and gram‐negative bacteria, respectively [4].

By knowing the proper antibiogram, it is possible to formulate a proper empiric antibiotic therapy for CRBSI. The current recommended empiric therapy for CRBSI as per Tanzanian guidelines is not cost‐effective. It is composed of vancomycin for gram‐positive coverage and ceftazidime for gram‐negative coverage. The cost of these two drugs until the results of the blood culture may go up to US$137 [5], as this is expensive for local patients. In a recent study in China, cheap drugs such as ciprofloxacin were found to be very effective in treating CRBSI, with a susceptibility of 100% for gram‐negative bacteria. The total dose for a 7‐day empiric treatment was only US$0.8 [6]. Knowing the relevant determinants of CRBSI and the proper antibiogram of the causative bacteria, it is possible to propose a proper preventive measure and, more importantly, formulate an evident antibiotic empiric therapy once it is encountered among hemodialysis patients.

2. Methods

2.1. Study Site and Design

This was a retrospective analytical cross‐sectional hospital‐based study that involved patients with chronic kidney disease (CKD) or acute kidney injury (AKI) who were attending hemodialysis at the Kilimanjaro Christian Medical Centre (KCMC) hemodialysis unit from January 1, 2021 to December 31, 2022, aged 18 years and above, with a tunneled or non‐tunneled central venous catheter (CVC) as an access for hemodialysis. This study was approved by the Kilimanjaro Christian Medical University College Research and Ethics Review Committee (No. PG 53/2022). Permission to collect hospital data was granted by the director of hospital services through the head of the department of internal medicine and the hemodialysis unit at KCMC. Participant data protection and confidentiality were guaranteed. Unique identification numbers were used to identify the participant records. Due to the retrospective nature of the study, informed consent of the patients was not required because the study analyzed anonymous clinical data of the patients.

2.2. Study Variables

The dependent variable was CRBSI, defined as bacteremia in which a similar organism was isolated from blood cultures from samples obtained from both the catheter lumen and from the peripheral vein, taken on the same day [7].

These were collected from medical records through the electronic health management system and from the hemodialysis database. All patients with a history of fever, chills, and/or rigors were suspected to have developed CRBSI. Body temperature was recorded as per the national health and examination survey, and axillary temperature > 37°C was regarded as fever. Body mass index (BMI) was defined as per World Health Organization (WHO) criteria as underweight < 18.5 kg/m², normal 18.5–24.9 kg/m², overweight 25–29.9 kg/m², and obesity > 30 kg/m². The following variables were categorized as per a similar study done at Muhimbili National Hospital in Tanzania: the diagnosis of renal disease was categorized as AKI or CKD. Catheter life span was categorized as ≤ 30 days, 31–90 days, and > 90 days and was calculated by determining the total number of days the catheter has stayed in situ. Catheter types were either tunneled or non‐tunneled. The catheter insertion site was either through the internal jugular vein (IJV) or the femoral vein catheter. Admission status was categorized as inpatient or outpatient [4]. Sex was categorized as male or female. Age was categorized into four age categories in years: 18–36, 37–54, 55–72, and > 72. Marital status was categorized as marriage union, which included those cohabiting and those officially married, while not married included divorced, widowed, and those who were primarily single. The level of education was categorized as primary, secondary, and tertiary, which included college and university levels [8].

Laboratory information was categorized as per the 2019 Tanzanian standard operating procedure and protocol for hemodialysis. They included results of serum albumin in the last 3 months, which was categorized as < 35 and > 35 g/L, and hemoglobin level in the last 1 month, which was categorized as > 10 and < 10 g/dL [5]. Hemodialysis adequacy as measured by urea reduction rate (URR) was categorized as per kidney disease improving global outcome (KDIGO) criteria as > 65% and < 65% [9], and blood culture results were obtained according to the Clinical and Laboratory Standards Institute (CLSI) protocol [10].

All patients suspected of having CRBSI were subjected to blood culture and sensitivity testing. Two pairs of blood samples were taken from each patient using sterile techniques. The blood samples were taken simultaneously, one from the CVC and another one from the peripheral vein, during a 1‐h post‐dialysis session, with 10 cc for each sample in aerobic and anaerobic bottles. Both samples were sent to the laboratory within 1‐h for culture and sensitivity testing. The CLSI protocol was used to test the isolated bacteria. The blood culture bottles were incubated in the blood culturing machine (BD BACTEC™ FX‐04) for a maximum of 5 days to observe bacterial growth.

2.3. Antibiotic Sensitivity Testing (AST)

This was done by the conventional Kirby‐Bauerdisk diffusion method as per clinical laboratory standard institute guidelines for 2022. Antibiotics to be tested were selected based on commonly prescribed antibiotics as per guidelines for managing fever in hemodialysis patients, Tanzanian Standard Treatment Guidelines 2021, and CLSI 2022. For gram‐positive bacteria, the following antibiotics were tested: ciprofloxacin, vancomycin, gentamycin, erythromycin, clindamycin, tetracycline, and trimethoprim‐sulfamethoxazole. For gram‐negative bacteria, the following antibiotics were applicable: meropenem, ciprofloxacin, ceftazidime, ceftriaxone/cefotaxime, gentamycin, amikacin, trimethoprim‐sulfamethoxazole, and piperacillin‐tazobactam. Colistin was only tested for MDR Acinetobacter baumanii isolates.

2.4. Detection of Methicillin‐Resistant Staphylococcus aureus (MRSA)

As per CLSI 2022 guidelines, the cefoxitin disk will be incubated at 37°C for 16 to 18 h. A ≤ 21 mm inhibition zone diameter around the disk will reveal MRSA. Methicillin‐susceptible Staphylococcus aureus (MSSA) ATCC25923 and MRSA ATCC 43,300 were used as controls.

2.5. Detection of Extended Spectrum β‐Lactamase Production

Only Escherichia coli isolates were screened for extended‐spectrum beta‐lactamase (ESBL) production. The third‐generation cephalosporin, ceftriaxone and ceftazidime, disk combination method was used in this case, as per CLSI 2022.

2.6. Quality Control

The aseptic technique with a sterile procedure was highly observed through a special standard operating procedure. Performance and sterility tests were performed to control for the quality of the media and the absence of contamination in the culture medium. The ability to support bacterial growth was checked by a performance test by culturing standard organisms: Escherichia coli (ATCC 252922), Staphylococcus aureus (ATCC 25923), and Pseudomonas aeruginosa (ATCC 27853).

2.7. Statistical Analysis

Data were analyzed using the Statistical Package for Social Science (version 27). Continuous variables were summarized using the mean and median with their corresponding measures of dispersion, the standard deviation, and the interquartile range. The categorical variables were summarized using frequency and percentage. A chi‐square test was used to compare the difference in the proportion of CRBSI by patient characteristics. Multivariate logistic regression was used to assess the determinants of CRBSI. Covariates with a p‐value < 0.1 in the bivariate analysis were included in the final multivariate logistic regression model. A p‐value of < 0.05 was considered statistically significant, and a confidence interval of 95% will be used.

3. Results

A total of 295 patients with CVC underwent hemodialysis at KCMC from January 1, 2021 to December 31, 2022. Of these, 286 patients met the criteria for enrollment in the study. A total of nine participants were excluded from the study as they did not meet the inclusion criteria. Among those, seven had confirmed other sources of infection, of which two had urosepsis, two had pneumonia, and three had septic wounds. The remaining two had major surgeries during the time of the study. Therefore, the overall response rate was 97%.

Almost three‐quarters of the patients (70.6%) were males. Fifty‐six percent were at a mean age of 58 (±14.8). The majority were married or cohabiting (83.2%). These sociodemographic characteristics are shown in Table 1.

Table 1.

Sociodemographic characteristics of hemodialysis patients with central venous catheters (N = 286).

Variable	n	%
Sex
Male	202	70.6
Female	84	29.4
Age (years); mean 58 (SD ± 14.8)
18–36	35	12.2
37–54	54	18.9
55–72	160	55.9
> 72	37	12.9
Marital status
Married/Cohabiting	238	83.3
Single/Divorced/Widowed	48	16.8
Level of education
Primary	120	42.0
Secondary	70	24.5
College/University	96	33.6

Among the recruited patients, 228 (79.7%) were outpatients. The majority had a non‐tunneled CVC (89.5%). Most of the patients had CKD (87.1%). A CVC was commonly inserted into the IJV (94.8%). They were mostly insured patients (81.1%). The median life span for tunneled CVC was 467.5 (IQR 297.5–686.0) days, while for non‐tunneled CVC, it was 83 (39.0–125.5) days. Most participants had low albumin < 35 g/L (66.8%), and more than half had hemoglobin < 10 g/dL (57.0%). Seventy‐two percent had adequate hemodialysis with a URR > 65%. About half had a normal BMI (51.4%). The leading comorbidity was diabetes mellitus (57.0%), followed by hypertension (46.9%). An almost equal number of patients had malignancy and HIV: 22 (7%) and 18 (6.3%), respectively. A quarter (24.1%) were on supplemental iron, and half were receiving supplemental erythropoietin (53.5%). More than half (56.6%) developed CRBSI, of which 11.7% of those were recurrent episodes. These clinical and laboratory characteristics are shown in Table 2.

Table 2.

Clinical and laboratory characteristics of hemodialysis patients with central venous catheters (N = 286).

Variable	n	%
Diagnosis
Acute kidney injury	37	12.9
Chronic kidney disease	249	87.1
Admission status
Outpatient	228	79.7
Inpatient	58	20.1
Insurance status
Yes	232	81.1
No	54	18.9
Catheter life span (days); median 87 (IQR 40.9–139.5)
≤ 30	49	17.1
31–90	101	35.3
> 90	136	47.6
Type of Central Venous Catheters
Tunneled	30	10.5
Non‐tunneled	256	89.5
Site of Catheter Insertion
Femoral vein	15	5.2
Internal jugular vein	271	94.8
Albumin (g/L)
≥ 35	95	33.2
< 35	191	66.8
Hemoglobin (g/dL)
< 10	163	57.0
≥ 10	123	43.0
URR (%)
< 65	79	27.6
≥ 65	207	72.4
BMI
Underweight	24	8.4
Normal	147	51.4
Overweight	87	30.4
Obese	28	9.8
Supplementation
Erythropoietin	153	53.5
Iron	69	24.1
Comorbidities
Diabetes mellitus	163	57.0
Hypertension	134	46.9
HIV	18	6.3
Malignancy	22	7.7
Developed CRBSI
Yes	162	56.6
Recurrent	19	11.7
No	124	43.4

Abbreviations: BMI, body mass index; CRBSI, catheter‐related bloodstream infection; HIV, human immunodeficiency virus; URR, urea reduction rate.

A total of 231 bacterial pathogens were isolated from the participants with CRBSI. Gram‐positives were predominant (69.7%) of all isolates. Among all bacterial isolates, the commonest single bacteria isolated was MSSA (45.0%), followed by Pseudomonas aeruginosa 33 (14.3%). Table 3 shows the spectrum of the gram‐positive bacteria, and Table 4 shows the spectrum of the gram‐negative bacteria.

Table 3.

Spectrum of the gram‐positive bacteria isolated from patients with CRBSI undergoing hemodialysis (N = 231).

Bacterial isolates	Central (n = 122) n (%)	Peripheral (n = 109) n (%)	Total (n = 231) n (%)
Gram‐Positive Bacteria	84 (36.4)	77 (33.3)	161 (69.7)
Enterococcus faecalis	3 (1.3)	6 (2.6)	9 (3.9)
MSSA	52 (22.5)	52 (22.5)	104 (45.0)
MRSA	10 (4.3)	3 (1.3)	13 (5.6)
Staphylococcus epidermidis	12 (5.2)	12 (5.2)	24 (10.4)
Streptococcal pneumoniae	7 (3.0)	4 (1.7)	11 (4.8)

Abbreviations: MRSA, methicillin‐resistant Staphylococcus aureus; MSSA, methicillin‐sensitive Staphylococcus aureus.

Table 4.

Spectrum of the gram‐negative bacteria isolated from patients with CRBSI undergoing hemodialysis (N = 231).

Bacterial isolates	Central line (n = 122) n (%)	Peripheral (n = 109) n (%)	Total (n = 231) n (%)
Gram‐Negative Bacteria	38 (16.5)	32 (13.9)	70 (30.3)
Acinetobacter baumanii	2 (0.9)	2 (0.9)	4 (1.7)
Citrobacter freundii	3 (1.3)	2 (0.9)	5 (2.2)
Escherichia coli	4 (1.7)	5 (2.2)	9 (3.9)
Escherichia coli (ESBL)	4 (1.7)	2 (0.9)	6 (2.6)
Klebsiella oxytoca	2 (0.9)	1 (0.4)	3 (1.3)
Proteus vulgaris	2 (0.9)	1 (0.4)	3 (1.3)
Pseudomonas aeruginosa	17 (7.4)	16 (6.9)	33 (14.3)
Serratia	2 (0.9)	0 (0)	2 (0.9)
Enterobacter	2 (0.9)	0 (0)	2 (0.9)
Klebsiella pneumonia	0 (0)	3 (1.3)	3 (1.3)

Abbreviation: ESBL, extended‐spectrum beta‐lactamase.

As shown in Table 5, the association of CRBSI varied across the study variables. Those with catheters at a femoral site had a higher proportion of CRBSI than those with an IJV catheter (86.7% vs. 55.0%, p = 0.03). Those with albumin levels < 35 g/L had a significantly higher proportion of CRBSI than their counterparts with higher albumin (66.0% vs. 37.9%, p < 0.001). Those with low hemoglobin < 10 g/dL had a significantly higher proportion of getting CRBSI than their counterparts with > 10 g/dL (62.6% vs. 48.8%, p 0.02). Participants with diabetes mellitus had significantly higher proportions of CRBSI than their counterparts (67.5% vs. 42.3%, p = 0.01). Those without hypertension had significantly higher proportions of CRBSI than their counterparts (64.5% vs. 47.8%, p 0.005).

Table 5.

Relationship between sociodemographic and clinical characteristics with CRBSI among patients with CVC (N = 286).

Variable	Infection		cOR (95% CI)	P value	aOR (95% CI)	P value
	Yes 162 (56.6%)	No 124 (43.4%)
	n (%)	n (%)
Sex
Male	114 (56.4)	88 (43.6)	0.97 (0.58–1.62)	0.91
Female	48 (57.1)	36 (42.9)	Ref
Age (years)
18–36	21 (60.0)	14 (40.0)	Ref		Ref
37–54	37 (68.5)	17 (31.5)	1.45 (0.59–3.32)	0.41	0.89 (0.30–2.68)	0.85
55–72	89 (55.6)	71 (44.4)	0.8 (0.39–1.760)	0.64	0.55 (0.21–1.44)	0.22
> 72	15 (40.5)	22 (59.5)	0.46 (0.18–1.17)	0.10	0.39 (0.13–1.24)	0.11
Marital status
Married/Cohabiting	141 (59.2)	97 (40.8)	Ref		Ref
Single/Divorced/Widowed	21 (43.8)	27 (56.2)	0.54 (0.29–1.00)	0.05	0.49 (0.23–1.05)	0.07
Level of education
Primary	67 (55.8)	53 (44.2)	1.16 (0.68–1.99)	0.58
Secondary	45 (64.3)	25 (35.7)	1.66 (0.88–3.11)	0.12
College/University	50 (52.1)	46 (47.9)	Ref
Diagnosis
Acute kidney injury	25 (67.6)	12 (32.4)	Ref
Chronic kidney disease	137 (55.0)	112 (45.0)	0.59 (0.28–1.22)	0.15
Admission status
Outpatients	126 (55.3)	102 (44.7)	Ref
Inpatients	36 (62.1)	22 (37.9)	1.33 (0.73–2.39)	0.35
Insurance status
Yes	36 (66.7)	18 (33.3)	Ref		Ref
No	126 (54.3)	106 (45.7)	0.59 (0.32–1.11)	0.10	0.86 (0.41–1.83)	0.70
Catheter life span (days)
≤ 30	28 (57.1)	21 (42.9)	Ref
31–90	62 (61.4)	39 (38.6)	1.19 (0.59–2.39)	0.62
> 90	72 (52.9)	64 (47.1)	0.84 (0.44–1.63)	0.61
Type of Catheter
Tunneled	12 (40.0)	18 (60.0)	Ref		Ref	0.02
Non‐tunneled	150 (58.6)	106 (41.4)	2.12 (0.98–4.59)	0.06	3.19 (1.19–8.59)
Site of Catheter Insertion
Femoral vein	13 (86.7)	2 (13.3)	5.32 (1.18–24.04)	0.03	4.96 (0.96–25.59)	0.06
Internal jugular vein	149 (55.0)	122 (45.0)	Ref		Ref
Albumin (g/L)
≥ 35	36 (37.9)	59 (62.1)	3.18 (1.91–5.29)	< 0.001	3.04 (1.68–5.48)	< 0.001
< 35	126 (66.0)	65 (34.0)	Ref		Ref
Hemoglobin (g/dL)
< 10	102 (62.6)	61 (37.4)	1.76 (1.09–2.82)	0.02	1.93 (1.07–3.47)	0.03
≥ 10	60 (48.8)	63 (51.2)	Ref		Ref
URR (%)
< 65	46 (58.2)	33 (41.8)	1.09 (0.65–1.85)	0.74
≥ 65	116 (56.0)	91 (44.0)	Ref
BMI
Underweight	14 (58.3)	10 (41.7)	0.96 (0.40–2.31)	0.93	1.35 (0.47–3.82)	0.58
Normal range	87 (59.2)	60 (40.8)	Ref		Ref
Overweight	53 (60.9)	34 (39.1)	1.06 (0.61–1.84)	0.84	1.01 (0.53–1.93)	0.97
Obesity	8 (28.6)	20 (71.4)	0.31 (0.12–0.74)	0.009	0.23 (0.08–0.61)	0.003
Erythropoietin
Yes	86 (56.2)	67 (43.8)	1.04 (0.65–1.66)	0.87
No	76 (57.1)	57 (42.9)	Ref
Iron supplementation
Yes	43 (62.3)	26 (37.7)	0.73 (0–1.28)	0.28
No	119 (54.8)	98 (45.2)	Ref
Diabetes mellitus
Yes	110 (67.5)	53 (32.5)	1.86 (1.16–2.99)	0.01	0.34 (0.19–0.61)	< 0.001
No	52 (42.3)	71 (57.7)	Ref		Ref
Hypertension
Yes	64 (47.8)	70 (52.2)	1.99 (1.24–3.19)	0.005	1.65 (0.92–2.98)	0.09
No	98 (64.5)	54 (35.5)	Ref		Ref
HIV
Yes	12 (66.7)	6 (33.3)	0.64 (0.23–1.74)	0.38
No	150 (56.0)	118 (44.0)	Ref
Malignancy
Yes	12 (54.5)	10 (45.5)	1.09 (0.46–2.63)	0.84
No	150 (56.8)	114 (43.2)	Ref

Abbreviations: BMI, body mass index; HIV, human immunodeficiency virus; URR, urea reduction rate.

Table 5 also summarizes the association between different sociodemographic and clinical characteristics and the overall risk of developing CRBSI among hemodialysis patients with CVC. In the crude analysis, the site of catheter insertion into the femoral vessel (cRR: 5.32; 95% CI: 1.18–24.04), low serum albumin < 35 g/L (cRR: 3.18; 95% CI: 1.91–5.29), having low hemoglobin < 10 g/dL (cRR: 1.76; 95% CI: 1.09–2.82), being diabetic (cRR: 1.86; 95% CI: 1.16–2.99), and being hypertensive (cRR: 1.99; 95% CI: 1.24–3.19) were significantly associated with increased risk of getting CRBSI among hemodialysis patients with CVC. On the other hand, being obese (cRR: 0.31; 95% CI: 0.12–0.74) was associated with a lower risk. In the adjusted analysis, hemodialysis patients with non‐tunneled catheters had a three‐fold higher risk of developing CRBSI (aRR: 3.19; 95% CI: 1.19–8.59). Those with low serum albumin < 35 g/L also had a three‐fold higher risk of developing CRBSI (aRR: 3.04; 95% CI: 1.68–5.48). Those with low hemoglobin < 10 g/dL had a two‐fold higher risk of developing CRBSI (aRR: 1.93; 95% CI: 1.07–3.47). Diabetic patients on hemodialysis with CVC had a 66% lesser chance of getting CRBSI (aRR: 0.34; 95% CI: 0.19–0.61).

Tables 6 and 7 represent the antibiogram (susceptibility pattern) of gram‐positive and gram‐negative bacterial isolates, respectively. Gram‐positive bacteria were highly susceptible to vancomycin (99%) but were completely resistant to trimethoprim‐sulfamethoxazole, with 100% resistance. In the CVC group, MSSA showed the widest susceptibility range, in which it was sensitive at variable degrees to at least every tested antibiotic except trimethoprim‐sulfamethoxazole. Enterococcus faecalis and MRSA were the least susceptible to the tested antibiotics. They had 100% resistance to all tested antibiotics except vancomycin, where each was fully susceptible by 100%. In the peripheral vessel group, MSSA displayed the widest susceptibility range. It was susceptible at variable degrees to all tested antibiotics but was 100% resistant to trimethoprim‐sulfamethoxazole. MRSA was the most resistant bacterial isolate. It had a resistance of 100% to all tested antibiotics except vancomycin, where it had a susceptibility of 100%.

Table 6.

Antibiogram of the gram‐positive bacteria isolated from patients with CRBSI undergoing hemodialysis.

Gram‐positive bacteria	n (%)
	Ciprofloxacin	Gentamycin	Erythromycin	Clindamycin	Tetracycline	Trimethoprim‐sulfamethoxazole	Vancomycin
Central line
Enterococcus faecalis (n = 3)	0	0	0	0	0	0	3 (100)
MSSA (n = 52)	9 (17.3)	24 (46.2)	5 (9.6)	19 (36.5)	9 (17.3)	0	49 (94.2)
MRSA (n = 10)	0	0	0	0	0	0	10 (100)
Staphylococcus epidermidis (n = 12)	2 (16.7)	5 (41.7)	0	0	3 (25)	0	12 (100)
Streptococcal pneumonia (n = 7)	0	0	0	2 (28.6)	0	0	7 (100)
Peripheral
Enterococcus faecalis (n = 6)	1 (16.7)	0	0	0	0	0	6 (100)
MSSA (n = 52)	7 (13.4)	13 (25.0)	4 (7.7)	14 (26.9)	10 (19.2)	0	47 (97)
MRSA (n = 3)	0	0	0	0	0	0	3 (100)
Staphylococcus epidermidis (n = 12)	3 (25.0)	4 (33.3)	0	0	1 (8.3)	0	12 (100)
Streptococcal pneumonia (n = 4)	0	1 (25.0)	0	1 (25.0)	1 (25.0)	0	4 (100)

Abbreviations: MRSA, methicillin‐resistant Staphylococcus aureus; MSSA, methicillin‐sensitive Staphylococcus aureus.

Table 7.

Antibiogram of the gram‐negative bacteria isolated from patients with CRBSI undergoing hemodialysis.

Gram‐negative bacteria	n (%)
	Meropenem	Colistin sulfate	Ciprofloxacin	Ceftazidime	Ceftriaxone	Gentamycin	Amikacin	Piperacillin‐tazobactam	Trimethoprim‐sulfamethoxazole
Central line
Acinetobacter baumanii (n = 2)	0	2 (100)	0	0	0	0	0	0	0
Citrobacter freundii (n = 3)	3 (100)	N/A	2 (66.7)	2 (66.7)	0	2 (66.7)	0	1 (33.3)	0
Escherichia coli (n = 4)	4 (100)	N/A	3 (75)	0	1 (25)	2 (50)	3 (75)	0	0
Escherichia coli (ESBL) (n = 4)	4 (100)	N/A	0	0	0	0	0	0	0
Enterobacter (n = 2)	2 (100)	N/A	1 (50)	1 (50)	0	1 (50)	0	0	0
Klebsiella oxytoca (n = 2)	1 (50)	N/A	0	0	0	0	2 (100)	0	0
Proteus vulgaris (n = 2)	2 (100)	N/A	0	0	0	1 (50)	2 (100)	1 (50)	0
Pseudomonas aeruginosa (n = 17)	17 (100)	N/A	8 (47)	5 (29.4)	N/A	2 (11.8)	8 (47)	13 (76.5)	0
Serratia (n = 2)	2 (100)	N/A	2 (100)	0	0	2 (100)	0	0	0
Peripheral
Acinetobacter baumanii (n = 2)	0	2 (100)	0	0	0	0	0	0	0
Citrobacter freundii (n = 2)	2 (100)	N/A	2 (100)	0	0	0	0	1 (50)	0
Escherichia coli (n = 5)	4 (80)	N/A	4 (80)	1 (20)	1 (20)	3 (60)	4 (80)	2 (40)	0
Escherichia coli (ESBL) (n = 2)	2 (100)	N/A	0	0	0	0	0	0	0
Klebsiella oxytoca (n = 1)	1 (100)	N/A	0	0	0	0	1 (100)	0	0
Klebsiella pneumonia (n = 3)	2 (66.7)	N/A	1 (33.3)	0	0	3 (100)	2 (66.7)	1 (33.3)	0
Proteus vulgaris (n = 1)	1 (100)	N/A	0	0	0	1 (100)	1 (100)	1 (100)	0
Pseudomonas aeruginosa (n = 16)	16 (100)	N/A	7 (43.8)	4 (25.0)	N/A	2 (12.5)	8 (50)	13 (81.3)	0

Abbreviation: ESBL, extended‐spectrum beta‐lactamase.

For the gram‐negative bacteria, among the tested antibiotics, the most efficacious antibiotic was meropenem. It had an overall sensitivity of 82.1% for all isolates. Similar to the results found in the gram‐positive group, trimethoprim‐sulfamethoxazole was again the least efficacious antibiotic. All the isolated bacteria were completely resistant to it. In both CVC and peripheral vessel groups, two organisms that equally displayed the least susceptibility to the tested antibiotics were Acinetobacter baumanii and ESBL Escherichia coli. They were susceptible to only colistin‐sulfate and meropenem, respectively, by 100%. Pseudomonas aeruginosa had a susceptibility of 78.9% to piperacillin‐tazobactam and 100% to meropenem. Susceptibility to ceftriaxone was low. Only non‐ESBL‐producing Escherichia coli were susceptible to ceftriaxone (25% and 20% in the CVC and peripheral vessel groups, respectively). Ceftazidime had an overall susceptibility of only 11.2% for all gram‐negative bacteria isolates.

4. Discussion

In this study, the rate of CRBSI was 56.6%, which was higher compared to studies done in similar settings such as in Uganda at 41% [11], Ghana at 34.2% [12], Tunisia at 30.6% [13], and Tanzania at 35.7% and 29.9% [4, 8]. However, there are variations all across the world, with developed countries also showing higher rates of CRBSI such as in the USA at 30.7% [14] and Albania at 58.8% [15]. The different infection rates may be multifactorial based on location, sample size, catheter duration, type of catheter inserted, and insertion location.

In this study, a total of 231 bacterial pathogens were isolated from 162 patients with CRBSI. Gram‐positive bacteria were the predominant causative agents for CRBSI in 69.7% of patients. This was consistent with a previous study done in China, which found a relatively similar trend with gram‐positive dominance by 44% [6]. This might be due to the abundance of gram‐positive bacterial normal flora on the skin surface, which may gain access to the catheter lumen and finally to the bloodstream during the percutaneous insertion of the CVC [16]. A retrospective cohort study in Malaysia found that both gram‐positive and gram‐negative bacteria are responsible for causing CRBSI. The most common gram‐positive bacteria was MRSA, followed by methicillin‐resistant Coagulase‐negative Staphylococcus, while the most common gram‐negative bacteria was Pseudomonas aeruginosa [7].

Among the gram‐positive bacteria, Staphylococcus aureus was the most common single organism isolated, comprising 72.7% of all gram‐positive bacteria isolated. This was similar to a study done in South India, where Staphylococcus aureus dominated other gram‐positive bacterial isolates by 66.6% [17]. This may have reflected a real picture because, among gram‐positive bacteria known to cause disease in humans, Staphylococcus aureus is the most common and pathogenic of all [18].

Pseudomonas aeruginosa was the second most common bacterial isolate found, comprising 14.3% of all bacterial isolates. These results have shown an increasing incidence of this gram‐negative pathogen as compared with a study done in Spain, which found Pseudomonas aeruginosa to have a prevalence of 7.1% among all isolates [19]. This could be explained by the fact that these studies were taken in two vastly different geographical locations with varying epidemiologies of diseases. Patients in this study had multiple comorbidities that necessitated frequent hospital visits; thus, the occurrence of Pseudomonas aeruginosa could be nosocomial. Compared to the study done in Spain, most participants had fewer comorbidities and, thus, fewer hospital visits.

A low serum albumin level (< 35 g/L) was strongly associated with a three‐fold higher risk of getting CRBSI. These findings were consistent with a study done in China, which revealed a strong correlation between low albumin levels and an increased risk of CRBSI [20]. This is because albumin is a building block of the body's immune system; thus, diminished levels will predispose the host to an increased risk of infection. As a negative acute phase reactant, its levels drop during an acute physiological reaction such as inflammation, even with CRBSI [21].

In this study, having a non‐tunneled CVC was associated with a three‐fold risk of getting CRBSI. These findings were consistent with a study done in Malaysia, where non‐tunneled hemodialysis CVC were five times more susceptible to CRBSI as compared to tunneled ones [7]. Tunneled catheters have cuffs, which are barriers to the entry of infective bacteria along the catheter tunnel, which would lead to CRBSI. However, non‐tunneled catheters have no cuffs and are more vulnerable [22].

Surprisingly, diabetes mellitus showed a protective association with the risk of developing CRBSI. Studies have analyzed a positive association between diabetes mellitus and an increased risk of infection. A study at a hemodialysis unit in the USA showed that patients with diabetes mellitus were 2.2 times more likely to get CRBSI as compared to the controls [23]. This is explained by the immunosuppressive effect of hyperglycemia. Bacteria and other pathogens are believed to utilize the excess sugar as a substrate for growth and nourishment [24]. The findings of this study could be due to the sampling technique employed and the sample size, which could have led to protective associations.

In this study, having a low level of hemoglobin (< 10 g/dL) was associated with a two‐fold higher risk of getting CRBSI. This was correlated to a study done in Bangladesh, which showed similar findings, with hemoglobin < 9 g/dL associating with a two‐times higher risk of getting CRBSI [25]. Anemia is known to predispose the host to bacterial infection through a complex mechanism in which low hemoglobin levels reduce hepcidin, thus altering macrophage iron sequestration and producing a constant pool of iron for the invading bacteria, in which iron will be used as a substrate for growth [26].

In this current study, for the gram‐positive bacteria, vancomycin was the most efficient drug, with a sensitivity of 99%. This was similar to a study done in the national hospital of Tanzania, where the most efficient drug had slightly lower efficacy as it was sensitive to 80% of all the gram‐positive isolates [8]. The higher efficacy found in this study could be due to this center being relatively new to the comparable center. Possible geographical or altitudinal locations may have played a role.

For gram‐negative bacteria, the most efficient antibiotic was meropenem, with an efficacy of 82.1%. A previous study done in India revealed meropenem efficacy to be 85% [27]. A cohort study in Saudi Arabia revealed that, for the bacterial isolates related to CRBSI, cephalosporin antibiotics were the most efficacious antibiotics, followed by gentamycin [28], as was not the case in this study (< 25%). Carbapenems are among the strongest antibiotics against gram‐negative bacteria. They achieve this ability by a strong mechanism of binding and inhibiting peptidoglycan cross‐linking associated with bacterial cell wall synthesis, producing strong bactericidal activity [29].

MRSA in the gram‐positive group, ESBL‐producing Escherichia coli, and Acinetobacter baumanii in the gram‐negative group showed the least susceptibility to the tested antibiotics in regard to the respective antibiotic sensitivity testing guidelines. A study done in India pointed out that ESBL‐producing gram‐negative enteric bacteria were the most resistant isolates, as they were sensitive only to meropenem [30]. On the other hand, a study done in Italy found that MRSA was found to be the most resistant gram‐positive isolate, sensitive to vancomycin alone [31]. These findings point to the post‐antibiotic era due to the existing antibiotic resistance threat, a trend that has been developing gradually over decades [32].

However, there may be challenges to using vancomycin empirically, and as a monotherapy. Even though vancomycin resistance in vitro is uncommon, clinical failures with vancomycin monotherapy are common, it may be necessary to use it in combination with other potential antibiotics [33]. Additionally, newer and alternative options are coming as either combination therapy or alternative therapy to vancomycin, especially among resistant Staphylococcus bacteria [34].

A few limitations challenged the fate and affected the final results. Due to the nature of the information needed, only patients with CVC were selected, and a purposive sampling method was used in the selection of the patients, as this may have decreased their representativeness due to selection bias. On the other hand, data were collected retrospectively from the stored archives, as this may have decreased the level of evidence and completeness of the data due to recall bias. Furthermore, the snapshot time of 2 years is not necessarily an adequate representation of the real situation of CRBSI among patients. More information could have been gathered if there had been a full follow‐up. Antibiotic sensitivity testing at KCMC is not consistent with international guidelines, as this may have influenced some of the results regarding the antibiogram.

5. Conclusion

The epidemiology of CRBSI has remained unchanged. Gram‐positive bacteria are still the major causative bacteria, led by Staphylococcus aureus, which was the most common bacteria found. Risk factors for CRBSI included low serum albumin, low hemoglobin, and the use of non‐tunneled hemodialysis CVC. Antibiotic susceptibility to common bacterial isolates is narrow and unpredictable. For gram‐positive bacteria, vancomycin has retained its ability and strength as the most efficacious antibiotic. The situation is different on the gram‐negative side, as the third‐generation cephalosporin antibiotics, which are currently recommended by our current national guidelines, have shown disappointing sensitivity to the commonly isolated bacteria. Among gram‐negative bacteria, meropenem showed the greatest efficacy, followed by ciprofloxacin.

It would be recommended to amend the current empiric antibiotic combination to be switched to vancomycin for gram‐positive coverage together with ciprofloxacin for gram‐negative coverage as the first‐line therapy for CRBSI. The second line should be composed of vancomycin for gram‐positive coverage and piperacillin‐tazobactam for gram‐negative coverage. It would be recommended against the use of femoral catheters unless an unavoidable reason exists as they pose a great risk of developing CRBSI. If the initial non‐tunneled CVC is expected to stay in situ for > 30 days, then it would be recommended to switch to a tunneled catheter, as long‐stay non‐tunneled catheters pose more risk for CRBSI. It would be recommended to maintain a hemoglobin of > 10 g/dL and a serum albumin of > 35 g/L to avoid the risk of CRBSI.

Author Contributions

Andrea R Costantine: conceptualization, formal analysis, writing–original draft, methodology, investigation, writing–review and editing, data curation. Sarah J Urasa: writing–review and editing, supervision, visualization. Abid M Sadiq: formal analysis, writing–review and editing, methodology, validation, visualization. Elichilia R Shao: writing–review and editing, methodology, supervision, visualization. Furaha S Lyamuya: supervision, writing–review and editing, visualization, methodology. Elieishiupendo M Niccodem: data curation, visualization, writing–review and editing, supervision, validation, conceptualization, methodology. Eugénie M Kamabu: investigation, writing–review and editing. Venance P Maro: conceptualization, visualization, supervision, writing–review and editing, methodology. Elifuraha W Mkwizu: writing–review and editing, supervision. Nyasatu G Chamba: supervision, writing–review and editing. Francis F Furia: visualization, validation, supervision, writing–review and editing, conceptualization, methodology. Tumaini E Mirai: investigation, writing–review and editing. Eliada B Nziku: investigation, writing–review and editing. Doreen T Eliah: investigation, writing–review and editing. Ibrahim Ali Ibrahim Muhina: investigation, writing–review and editing. Faryal M Raza: data curation, investigation, writing–review and editing. Modesta P Mitao: formal analysis, writing–review and editing. Kajiru G Kilonzo: conceptualization, visualization, methodology, supervision, writing–review and editing, project administration, validation.

Ethics Statement

The study obtained approval from the Kilimanjaro Christian Medical College Research Ethics and Review Committee (No. PG 53/2022), as well as from the hospital and the head of the department.

Consent

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Transparency Statement

The lead author Andrea R. Costantine affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request. The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.