Skip to main content

The comparative accuracy of pooled vs. individual blood culture sampling methods for diagnosis of catheter-related bloodstream infection

Abstract

Background

Catheter-related bloodstream infection (CRBSI) is associated with increased morbidity, mortality, and cost of treatment in critically ill patients. A differential time to positivity (DTP) of 120 min or more between blood cultures obtained through the catheter vs. peripheral vein is an indicator of CRBSI with high sensitivity and specificity. However, it is no clear whether pooled sampling would be as efficient as individual sampling in order to reduce costs, contamination, or anemia.

Methods

This was a prospective diagnostic study conducted at the medical ICU and semi-ICU of Khon Kaen University’s Srinagarind Hospital in Thailand from May 2020 to November 2021. Fifty patients with triple-lumen central venous catheters (CVCs) who were clinically suspected of CRBSI were enrolled. 15 mL of blood was drawn through each catheter lumen, 10 mL of which was inoculated into three blood culture bottles, and the remaining 5 mL was pooled into a single bottle. Sensitivity, specificity, accuracy, and time to positivity of the pooled blood cultures were calculated using individual blood cultures as a reference.

Results

Of the 50 patients enrolled, 14 (28%) were diagnosed with CRBSI, 57.9% of whom were infected with gram-negative bacteria as the causative pathogen (57.9%). Extensively drug-resistant (XDR) Klebsiella pneumoniae was the most common organism. Sensitivity and specificity of the pooled blood sampling method were 69.23% (95% CI [0.44–0.94]) and 97.3% (95% CI [0.92–1.02]), respectively. The area under the ROC curve (AUC) was 0.83 (95% CI [0.68–0.99]). A paired T-Test to compare time to positivity of the pooled blood bottle and the first positive culture from the individual bottles indicated statistical significance (14.9 and 12.4 h, respectively). The mean difference was 2.5 [0.9–4.1] h, with a 95% CI and a p-value of 0.006.

Conclusion

Pooled blood sampling results in a lower sensitivity and longer time to positivity for CRBSI diagnosis in patients with triple-lumen CVCs than individual lumen sampling.

Trial registration Retrospectively registered at Thai Clinical Trials Registry. The study was reviewed and approved on 08/03/2022. TCTR identification number is TCTR20220308002

Peer Review reports

Introduction

Catheter-related bloodstream infection (CRBSI) is associated with increased morbidity, mortality, and cost of treatment in critically ill patients [1,2,3,4]. The rates of CRBSI in the intensive care unit (ICU) ranged from 0.6 to 13 per 1000 catheter-days [4, 5]. In the United States, over 80,000 episodes of central venous catheter (CVC)-related bloodstream infection occur in the ICU annually [6]. A report from a surgical ICU in northern Thailand found the incidence of CRBSI to be 1.3 per 1000 catheter-days from 2005 to 2009 [7].

A differential time to positivity (DTP) of 120 min or more (earlier detection of growth in a blood culture obtained through a catheter hub than in a culture of simultaneously drawn peripheral blood) is one indicator that is highly sensitive and specific for CRBSI [6, 8, 9]. It also allows for catheter retention before the definite diagnosis has been made. According to the Infectious Diseases Society of America (IDSA)’s 2009 Clinical Practice Guidelines for the Diagnosis and Management of Intravascular Catheter-Related Infection, when CRBSI is suspected, paired blood cultures should be drawn from the catheter and a peripheral vein before initiation of antimicrobial therapy, and the bottles should be appropriately marked to indicate the site from which the samples were collected [6]. Failing to collect blood samples from some lumens of the catheter can lead to the misdiagnosis of CRBSI episodes [10]. By contrast, drawing blood cultures from all lumens increases laboratory costs and the risk of contamination and may also lead to nosocomial anemia, especially in pediatric patients [11]. Because catheter colonization can occur in any of the lumens [12], pooled blood sampling from all lumens into a single culture bottle might result in higher sensitivity than individual sampling, while simultaneously lowering laboratory costs and the required workload.

This study aims to compare the accuracy of pooled blood sampling from all lumens to that of individual sampling for CRBSI diagnosis.

Methods

This was a prospective, diagnostic study conducted at the medical ICU and semi-ICU of Khon Kaen University’s Srinagarind Hospital in Thailand from May 2020 to November 2021. Fifty patients with triple-lumen CVC clinically suspected of CRBSI were enrolled. We excluded patients who were under the age of 18, were pregnant, had HIV infection, or in whom any lumen of the CVC was blocked. Blood culture from the CVC and a peripheral blood sample were taken simultaneously before an empirical antibiotic was given. We drew 15 mL of blood from each hub of the CVC aseptically. 10 mL from each sample was inoculated separately into three individual bottles, and then the remaining 5 mL of each sample was pooled into one bottle. The blood volume for each culture bottle was based on IDSA guidelines [13]. Blood culture bottles were incubated in an AutoMate system (BACT/ALERT VIRTUO®, BioMerieux, France) at 37° Celcius for 5 days. Identification of all isolates was performed using the VITEK card system (VITEK® 2 Compact, BioMerieux, France). Sensitivity, specificity, and accuracy of the pooled blood culture were calculated using the individual blood cultures as a reference [10]. The time to positivity of each method was compared. Baseline characteristics of the patients, including age, sex, medical conditions, admission diagnosis, and SOFA score, were also recorded.

Sample size calculation

Sample size calculation was based on a hypothesized sensitivity and specificity of 95% of the pooled blood culture compared to the individual cultures. A pilot study found the sensitivity and specificity of pooled blood culture to diagnose CRBSI caused by bacterial pathogens at Srinagarind Hospital to be 100% and 96.9%, respectively. Based on these factors, we determined that a sample size of at least 45 participants would be necessary.

Statistical analysis

Baseline characteristics of the patients were reported as number and percentage. Continuous variables were presented as mean with standard deviation (SD). We used 2 × 2 table to calculate the sensitivity, specificity, and accuracy of the pooled blood culture compared to the individual cultures. Subgroup analysis was conducted for bacterial and fungal pathogens. McNemar’s Chi-square test was conducted and area under the receiver operating characteristic curve (AUC) was calculated to demonstrate the performance of the test. All statistical analyses were performed using SPSS for Windows, version 26.0 (SPSS).

Ethical consideration

Ethical approval was provided by the Khon Kaen University’s Center for Ethics in Human Research in accordance with the Declaration of Helsinki (Number HE621436). Informed consent was obtained from all participants.

Results

A total of 50 patients who met the inclusion criteria were enrolled into the study. The patients’ median age was 60 years (Interquartile range [IQR] of 53–74 years), 60% were male, and median Sequential Organ Failure Assessment (SOFA) score at ICU admission was 8 (IQR of 5.3–10.0). The most common medical conditions were hypertension and renal failure (34%), followed by diabetes (26%), chronic kidney disease (18%), and malignancy (18%). In half of the cases, the primary diagnosis was pneumonia. Others were CRBSI, septic shock, and urinary tract infection (Table 1).

Table 1 Baseline characteristics of all 50 patients

Fourteen patients (28%) in the study were diagnosed with CRBSI based on a DTP of 120 min or more. In three cases, there was coagulase negative staphylococci and in another there was Bacillus spp. in the peripherical blood, which was considered contamination. Cultures from both the CVC and peripheral site in the remaining cases were negative. There was no significant difference in duration of catheterization before blood culture collection between those with infected and non-infected CVCs (171.4 vs. 205.7 h, p-value 0.57). Most CVCs in the study were inserted into right internal jugular vein, followed by left internal jugular vein, and right femoral vein. We could not find any association between site of catheterization and rate of infection (p-value 0.67; Table 2).

Table 2 Comparison of duration of catheterization before blood culture collection and site of CVC between CRBSI and non-CRBSI cases

Three of 14 patients were infected by multiple microorganisms, resulting in a total of 19 identified microbial species, 11 (57.9%) of which were gram-negative bacteria, five were gram-positive, and the remaining were Candida spp. The most common bacterial pathogens found were extensively drug-resistant (XDR) Klebsiella pneumoniae. Thirteen of the 14 patients met the diagnosis criteria based on individual blood sampling (71.4% from the proximal, 50% from the middle, and 42.9% from the distal port), compared to only 10 based on pooled sampling. The sensitivity, specificity, positive predictive value, and negative predictive value of pooled blood sampling to diagnose of CRBSI were 69.23% (95% CI [0.44–0.94]), 97.3% (95% CI [0.92–1. 02]), 90% (95% CI [0.71–1.09]), and 90% (95% CI [0.80–0.99]), respectively (Table 3). The AUC calculated to assess the performance of the pooled blood sampling was 0.83 (95% CI [0.68–0.99]; Fig. 1).

Table 3 Diagnostic performance of pooled blood sampling for CRBSI diagnosis when using individual blood culture as a gold standard
Fig. 1
figure 1

AUC of the diagnostic test using pooled and separate blood sampling from each port of the central line

In seven of the nine patients with positive culture from both pooled and individual blood bottles, pathogens were discovered in the individual samples before the pooled samples. A paired T-Test to compare time to positivity of the pooled blood bottle and the first positive culture from the individual bottles indicated statistical significance (14.9 vs. 12.4 h, respectively). The mean difference with 95% CI was 2.5 [0.9–4.1] h, with a p-value of 0.006. A detailed comparison of time to positivity of the blood culture results from all 14 CRSBI patients is available in the supplemental material (Additional file 1: Table S1).

Discussion

This was a prospective, diagnostic trial comparing the accuracy of CRBSI diagnosis between two blood sampling methods. To our knowledge, this was the first such study conducted in Thailand. Fourteen of the 50 patients with suspected CRBSI had positive blood cultures from the peripheral site and one or more positive cultures from the CVC. Gram-negative bacteria were the predominant pathogens, which is consistent with the epidemiologic data in Thailand. However, our study differed in that we discovered a high predominance of XDR pathogens [7]. Data from National Antimicrobial Resistance Surveillance center, Thailand (NARST) show an exponential rise in the percentage of carbapenem-resistant Enterobacterales from 2015 to 2020 [14]. Similar findings have been reported in Japan and other parts of Southeast Asia, reflecting the ICU crisis in the region [15].

Pooled blood sampling misidentified 4 of the 13 cultures that were found positive by individual blood sampling and resulted in significantly increased in time to positivity. This suggests that pooled blood sampling might not be a good alternative method for diagnosing CRBSI, contradicting the findings of a previous study by Herrera-Guerra et al. [16]. One explanation for these findings may be the type of catheter used. Herrera-Guerra et al. used both double and triple lumen catheters, but only patients with triple lumen catheters were enrolled into our study. Another may be blood volume. Herrera-Guerra et al. drew 6 mL of blood from each lumen, inoculating 4.5 mL into individual blood culture bottles and the remaining 1.5 mL into a pooled blood bottle, while our study compared 10 mL in individual bottles to 5 mL from three lumens in the pooled blood bottle [16]. Sensitivity of blood culture is related to blood sample volume [17,18,19,20]. In this study, if bacterial colonization occurred in only one lumen of the CVC, pooled blood bottles would yield 5 mL of infected blood compared to 10 mL in individual blood bottles, resulting in lower sensitivity and increased time to positivity. In most cases with positive culture based on the individual sampling method, only one of three bottles tested positive. Single lumen colonization is common in CRBSI [10], which may explain these results. Blood-broth ratio is one factor that affects the positive rate of blood culture, with a ratio of 1:5 to 1:10 considered optimal for growth [21]. Pooled blood bottles can have a suboptimal blood-broth ratio, causing slower bacterial growth.

Two of the three cases of Candida spp. found by individual blood sampling were negative using pooled sampling. Larger inoculum size is also positively associated with Candida spp. detection using automated blood culture systems, especially in cases of C. glabrata, which has a longer mean time to growth detection [22]. Although a previous study found that DTP yields good sensitivity and specificity for catheter-related candidemia diagnosis, pooled blood sampling is not recommended [23]. Other studies have recommended against using DTP to diagnose CRBSI from Candida spp. Due to slower growth rate of fungi, the standard cut-off of 120 min might not be optimum [24,25,26]. In one case, BSI was detected only by pooled blood sampling, but the recovered pathogen was B. pseudomallei, which rarely causes CRBSI [2]. Although few nosocomial infections caused by B. pseudomallei have been reported, most can be explained by direct inoculation of the micro-organism during medical procedures (e.g., wound dressing and bronchoscopy) [27, 28].

There were some limitations to this study. First, the results cannot be applied to pediatric populations, as pediatric blood culture bottles are adapted to contain a smaller blood volume. For neonates, 1 mL of blood is adequate for detection of bacteremia [19]. The broth-blood ratio of pooled blood cultures in pediatric bottles also differs from that in adult bottles. Second, although DTP can be applied in both triple-lumen and double-lumen catheters [6], and single lumen colonization also occurs in double-lumen catheters, our results may not apply in cases with double-lumen catheters. This is because pooled samples from triple-lumen CVCs will be more diluted in cases of single lumen colonization. Further studies should be conducted to examine the accuracy of this test in other types of CVC. Another limitation was the small sample size. Only fourteen cases of confirmed CRBSI were studied. Despite these limitations, this study attained valuable evidence that pooled blood culture has a lower sensitivity and requires a longer time for CRBSI diagnosis compared to individual blood culture due to the smaller volume of blood that is drawn.

Pooled blood sampling has a lower sensitivity and higher time to positivity for CRBSI diagnosis in patients with triple-lumen CVCs. Laboratory costs, workload, and potential side effects (e.g., nosocomial anemia) should be considered when deciding whether to use this method. To maximize sensitivity, we encourage individual rather than pooled blood sampling.

Availability of data and materials

The data that underlie the results reported in this study are available from the corresponding author on reasonable request.

References

  1. Siempos II, Kopterides P, Tsangaris I, Dimopoulou I, Armaganidis AE. Impact of catheter-related bloodstream infections on the mortality of critically ill patients: a meta-analysis. Crit Care Med. 2009;37(7):2283–9.

    Article  Google Scholar 

  2. Gahlot R, Nigam C, Kumar V, Yadav G, Anupurba S. Catheter-related bloodstream infections. Int J Crit Illn Inj Sci. 2014;4(2):161.

    Article  Google Scholar 

  3. Cai Y, Zhu M, Sun W, Cao X, Wu H. Study on the cost attributable to central venous catheter-related bloodstream infection and its influencing factors in a tertiary hospital in China. Health Qual Life Outcomes. 2018;16(1):198.

    Article  Google Scholar 

  4. Martínez-Morel HR, Sánchez-Payá J, Molina-Gómez MJ, García-Shimizu P, García Román V, Villanueva-Ruiz C, et al. Catheter-related bloodstream infection: burden of disease in a tertiary hospital. J Hosp Infect. 2014;87(3):165–70.

    Article  Google Scholar 

  5. Hammarskjöld F, Berg S, Hanberger H, Taxbro K, Malmvall B-E. Sustained low incidence of central venous catheter-related infections over six years in a Swedish hospital with an active central venous catheter team. Am J Infect Control. 2014;42(2):122–8.

    Article  Google Scholar 

  6. Mermel LA, Allon M, Bouza E, Craven DE, Flynn P, O’Grady NP, Raad II, Rijnders BJA, Sherertz RJ, Warren DK. Clinical practice guidelines for the diagnosis and management of intravascular catheter-related infection: 2009 update by the Infectious Diseases Society of America. Clin Infect Dis. 2009;49(1):1–45.

    CAS  Article  Google Scholar 

  7. Cheewinmethasiri J, Chittawatanarat K, Chandacham K, Jirapongchareonlap T, Chotirosniramit N. Microbiology, risk factors and mortality of patients with intravenous catheter related blood stream infections in the surgical intensive care unit: a five-year, concurrent, case-controlled study. J Med Assoc Thai. 2014;97(Suppl 1):93–101.

    Google Scholar 

  8. Raad I, Hanna HA, Alakech B, Chatzinikolaou I, Johnson MM, Tarrand J. Differential time to positivity: a useful method for diagnosing catheter-related bloodstream infections. Ann Intern Med. 2004;140(1):18.

    Article  Google Scholar 

  9. Al-Juaid A, Walkty A, Embil J, Crockett M, Karlowsky J. Differential time to positivity: vascular catheter drawn cultures for the determination of catheter-related bloodstream infection. Scand J Infect Dis. 2012;44(10):721–5.

    Article  Google Scholar 

  10. Guembe M, Rodríguez-Créixems M, Sánchez-Carrillo C, Pérez-Parra A, Martín-Rabadán P, Bouza E. How many lumens should be cultured in the conservative diagnosis of catheter-related bloodstream infections? Clin Infect Dis. 2010;50(12):1575–9.

    Article  Google Scholar 

  11. Robinson JL. Sensitivity of a blood culture drawn through a single lumen of a multilumen, long-term, indwelling, central venous catheter in pediatric oncology patients. J Pediatr Hematol Oncol. 2002;24(1):72–4.

    Article  Google Scholar 

  12. Dobbins BM, Catton JA, Kite P, McMahon MJ, Wilcox MH. Each lumen is a potential source of central venous catheter-related bloodstream infection. Crit Care Med. 2003;31(6):1688–90.

    Article  Google Scholar 

  13. Miller JM, Binnicker MJ, Campbell S, Carroll KC, Chapin KC, Gilligan PH, Gonzalez MD, Jerris RC, Kehl SC, Patel R, et al. A guide to utilization of the microbiology laboratory for diagnosis of infectious diseases: 2018 update by the Infectious Diseases Society of America and the American Society for Microbiology. Clin Infect Dis. 2018;67(6):e1-94.

    Article  Google Scholar 

  14. National Antimicrobial Resistance Surveillance center, Thailand (NARST). http://narst.dmsc.moph.go.th. Accessed 17 May 2022.

  15. Akeda Y. Current situation of carbapenem-resistant Enterobacteriaceae and Acinetobacter in Japan and Southeast Asia. Microbiol Immunol. 2021;65(6):229–37.

    CAS  Article  Google Scholar 

  16. Herrera-Guerra AS, Garza-González E, Martínez-Resendez MF, Llaca-Díaz JM, Camacho-Ortiz A. Individual versus pooled multiple-lumen blood cultures for the diagnosis of intravascular catheter-related infections. Am J Infect Control. 2015;43(7):715–8.

    Article  Google Scholar 

  17. Antillon M, Saad NJ, Baker S, Pollard AJ, Pitzer VE. The relationship between blood sample volume and diagnostic sensitivity of blood culture for typhoid and paratyphoid fever: a systematic review and meta-analysis. J Infect Dis. 2018;218(Suppl 4):255–67.

    Article  Google Scholar 

  18. Neves L, Marra AR, Camargo TZS, dos Santos MC, Zulin F, da Silva PC, de Moura NA, da Silva VE, Pasternak J, dos Santos OFP. Correlation between mass and volume of collected blood with positivity of blood cultures. BMC Res Notes. 2015;8(1):383.

    Article  Google Scholar 

  19. Dreyer AW. Blood culture systems: from patient to result. In: Sepsis—an ongoing and significant challenge. IntechOpen. 2012. https://www.intechopen.com/chapters/39628 Accessed 2022 Feb 14.

  20. Bouza E, Sousa D, Rodríguez-Créixems M, Lechuz JG, Muñoz P. Is the volume of blood cultured still a significant factor in the diagnosis of bloodstream infections? J Clin Microbiol. 2007;45(9):2765–9.

    Article  Google Scholar 

  21. Ombelet S, Barbé B, Affolabi D, Ronat JB, Lompo P, Lunguya O, Jan J, Liselotte H. Best practices of blood cultures in low- and middle-income countries. Front Med. 2019;6:131.

    Article  Google Scholar 

  22. George BJ, Horvath LL, Hospenthal DR. Effect of inoculum size on detection of Candida growth by the BACTEC 9240 automated blood culture system using aerobic and anaerobic media. J Clin Microbiol. 2005;43(1):433–5.

    Article  Google Scholar 

  23. Park K-H, Lee MS, Lee S-O, Choi S-H, Sung H, Kim M-N, Woo JH, Kim SH. Diagnostic usefulness of differential time to positivity for catheter-related candidemia. J Clin Microbiol. 2014;52(7):2566.

    Article  Google Scholar 

  24. Gits-Muselli M, Villiers S, Hamane S, Berçot B, Donay J, Denis B, Guigue N, Alanio A, Bretagne S. Time to and differential time to blood culture positivity for assessing catheter-related yeast fungaemia: a longitudinal, 7-year study in a single university hospital. Mycoses. 2020;63(1):95–103.

    Article  Google Scholar 

  25. Fernandez-Cruz A, Martin-Rabadan P, Suarez-Salas M, Rojas-Wettig L, Perez MJ, Guinea J, Guembe M, Pelaez T, Sanchez-Carrillo C, Bouza E. Is it feasible to diagnose catheter-related candidemia without catheter withdrawal? Med Mycol. 2014;52(5):491–7.

    Article  Google Scholar 

  26. Bouza E, Alcalá L, Muñoz P, Martín-Rabadán P, Guembe M, Rodríguez-Créixems M. Can microbiologists help to assess catheter involvement in candidaemic patients before removal? Clin Microbiol Infect. 2013;19(2):E129–35.

    CAS  Article  Google Scholar 

  27. Merritt AJ, Peck M, Gayle D, Levy A, Ler YH, Raby E, Gibbs TM, Inglis TJJ. Cutaneous melioidosis cluster caused by contaminated wound irrigation fluid. Emerg Infect Dis. 2016;22(8):1420–7.

    Article  Google Scholar 

  28. Markovitz A. Inoculation by bronchoscopy. West J Med. 1979;131(6):550.

    CAS  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We would like to thank Dylan Southard for editing this MS via the KKU Publication Clinic (Thailand).

Funding

This study was funded by a grant from the Khon Kaen University Faculty of Medicine (Thailand; Grant Number 63130).

Author information

Authors and Affiliations

Authors

Contributions

PD proposed the research question. PD and SC conceptualized the study approach. PD, SC, and WS searched and reviewed the literature. Data collection and extraction was conducted by PD and SC. PD and WS were responsible for statistical analysis. The manuscript was drafted by WS and PD and subsequently reviewed and revised by WS and AM. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Wantin Sribenjalux.

Ethics declarations

Ethics approval and consent to participate

Ethical approval was provided by the Khon Kaen University Center for Ethics in Human Research in accordance with the Declaration of Helsinki (Number HE621436). Informed consent was obtained from all participants.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1: Table S1.

Blood culture results of all 14 CRSBI patients and comparison of time to positivity between pooled and individual blood sampling bottles.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Deawtrakulchai, P., Cheawchanwattana, S., Sribenjalux, W. et al. The comparative accuracy of pooled vs. individual blood culture sampling methods for diagnosis of catheter-related bloodstream infection. BMC Infect Dis 22, 622 (2022). https://doi.org/10.1186/s12879-022-07605-x

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12879-022-07605-x

Keywords

  • Catheter-related bloodstream infection (CRBSI)
  • Pooled blood culture
  • Individual blood culture
  • Differential time to positivity
  • Hospital-acquired infections
  • Infection control