Hostname: page-component-586b7cd67f-l7hp2 Total loading time: 0 Render date: 2024-11-26T09:54:40.007Z Has data issue: false hasContentIssue false
  • English
  • Français

Comparison of semi-quantitative and quantitative methods for diagnosis of catheter-related blood stream infections: a systematic review and meta-analysis of diagnostic accuracy studies

Published online by Cambridge University Press:  27 July 2020

Yan Zhang ,
Li Yang ,
Yanmei Chu  and
Linlin Wu Open the ORCID record for Linlin Wu [Opens in a new window]
Yan Zhang
Affiliation:
Department of ICU, Zaozhuang Municipal Hospital, Zaozhuang277100, Shandong Province, P.R. China
Li Yang
Affiliation:
Postpartum Health Care Pelvic Floor Function Diagnosis and Treatment Center, Zaozhuang Maternity and Child Health Care Hospital, Zaozhuang277100, Shandong Province, P.R. China
Yanmei Chu
Affiliation:
Operating Room, Zaozhuang Maternity and Child Health Care Hospital, Zaozhuang277100, Shandong Province, P.R. China
Linlin Wu*
Affiliation:
Department of ICU, Zaozhuang Municipal Hospital, Zaozhuang277100, Shandong Province, P.R. China
*
Author for correspondence: Linlin Wu, E-mail: [email protected]
Rights & Permissions [Opens in a new window]

Abstract

Catheter-related blood-stream infections (CRBSIs) are the most common healthcare-associated blood-stream infections. They can be diagnosed by either semi-quantitative or quantitative methods, which may differ in diagnostic accuracy. A meta-analysis was undertaken to compare the diagnostic accuracy of semi-quantitative and quantitative methods for CRBSI. A systematic search of Medline, Scopus, Cochrane and Embase databases up to January 2020 was performed and subjected to a QUADAS (quality assessment of diagnostic accuracy studies 2) tool to evaluate the risk of bias among studies. The pooled sensitivity and specificity of the methods were determined and heterogeneity was evaluated using the χ2 test and I2. Publication bias was assessed using a Funnel plot and the Egger's test. In total, 45 studies were analysed with data from 11 232 patients. The pooled sensitivity and specificity of semi-quantitative methods were 85% (95% CI 79–90%) and 84% (95% CI 79–88%), respectively; and for quantitative methods were 85% (95% CI 79–90%) and 95% (95% CI 91–97%). Considerable heterogeneity was statistically evident (P < 0.001) by both methods with a correspondingly symmetrical Funnel plot that was confirmed by a non-significant Deek's test. We conclude that both semi-quantitative and quantitative methods are highly useful for screening for CRBSI in patients and display high sensitivity and specificity. Quantitative methods, particularly paired quantitative cultures, had the highest sensitivity and specificity and can be used to identify CRBSI cases with a high degree of certainty.

Keywords

Catheter-related blood stream infectionsmeta-analysisquantitative culturesemi-quantitative culturevalidation studies
Type
Review
Information
Epidemiology & Infection , Volume 148 , 2020 , e171
Check for updates
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.
Copyright
Copyright © The Author(s), 2020. Published by Cambridge University Press

Introduction

Central venous or peripheral artery catheters are used for managing critically-ill or emergency patients admitted to intensive care units (ICUs) or emergency units of tertiary care centres. The most common complication associated with their use is catheter-related blood-stream infection (CRBSI) [Reference Lesseva1, Reference Gowardman, Lipman and Rickard2], which constitute a major part of healthcare-associated bacteraemia and are associated with significant morbidity, mortality and excess economic burden to the hospitals [Reference McGee and Gould3, Reference Rosenthal4]. Hence, preventing such infections by appropriate implementation of diagnostic and therapeutic practices is highly important.

CRBSI diagnoses remain challenging as the most common signs of these infections are often non-specific-like fever and chills; inflammation at the catheter-insertion site has a sensitivity of only 8% [Reference Safdar and Maki5]. Microbiological evidence to establish the central venous or peripheral arterial catheters as the source of bacteraemia is therefore required for diagnosis confirmation; however, despite intensive research, gold-standard methods are lacking. A semi-quantitative diagnostic technique was first recommended in 1977 by Maki et al. [Reference Maki, Weise and Sarafin6] in which, a catheter tip segment was cultured on the surface of a blood agar plate to detect bacterial contamination. This became the benchmark method and is widely used as a reference standard against which other diagnostic methods for CRBSIs are compared. However, false-negative results may be obtained with this technique for some patients with endoluminal colonisation as it detects only microorganisms present on the surface or outer layer of external catheters. Newer quantitative techniques overcome these limitations and are able to detect both exo- and endo-luminal organisms [Reference Riboli7]. The diagnostic accuracy of both techniques has been examined in tertiary care settings with widely varying results, however systematic studies of their comparative diagnostic accuracy are lacking and the most recent evaluation dates back to 2005 [Reference Safdar, Fine and Maki8]. The meta-analysis reported here was therefore designed to provide an updated evaluation of the diagnostic accuracy of semi-quantitative and quantitative methods for CRBSIs.

Methods

Types of studies and participants

We selected all studies examining the diagnostic accuracy of either semi-quantitative or quantitative methods for CRBSIs using catheter segments or blood culture, irrespective of study design. All studies included reported sensitivity and specificity values or provided data to allow calculation. They comprised published full-text articles, short communications and conference abstracts. Unpublished studies, thesis reports and those with sample sizes of <10 subjects were omitted. Participants were patients suspected of having a CRBSI in medical, surgical or neonatal ICUs of a tertiary care hospital, irrespective of their age groups or comorbid status. The reference standard was the isolation of the same microbial species from catheters and blood cultures.

Outcome measures

Pooled sensitivity, specificity, diagnostic odds ratio (DOR), likelihood ratio positive (LRP) and likelihood ratio negative (LRN) from the studies were calculated.

Search strategy

We performed a systematic and extensive electronic search from inception to January 2020 in databases and search engines (Medline, Embase, Scopus, Cochrane library, Google Scholar and ScienceDirect) without language restriction. Medical subject headings (MeSH) were applied along with free-text search terms (e.g. ‘Validation Studies’, ‘Blood Stream Infections’, ‘Intravascular Catheterization’, ‘Catheter Associated Blood Stream Infections’, ‘Nosocomial Infections’, ‘Quantitative Methods’, ‘Semi-Quantitative Methods’ ‘Sensitivity’, ‘Specificity’, ‘Diagnosis’, ‘Roll Plate Method’, ‘Sonication’ and ‘Diagnostic Accuracy Studies’); the resulting relevant articles were included in the review.

Selection of studies

Two authors independently performed the primary screening of title, keywords and abstracts, and retrieved full-text articles for the relevant studies. They then undertook a secondary screening of the articles to select those satisfying the inclusion criteria. Conflicts of opinion were resolved either by consultation with a third author or through consensus.

Data extraction and management

The primary investigator extracted the data from the selected studies pertaining to the study setting, design, participants, inclusion and exclusion criteria reference standards, index test, total participant numbers, and criteria for positivity, sensitivity and specificity. Finally, we compared the data in the review and the study reports to ensure correct entries.

Risk of bias assessment in included studies

The quality assessment of diagnostic accuracy studies 2 (QUADAS 2) tool was used to appraise the risk of bias among studies [Reference Whiting9]. This tool consists of the following domains: patient selection bias, conduct and interpretation of index test and reference standard, and time interval of outcome assessments. Studies were graded as low, high or unclear based on the presence of any bias.

Statistical analysis

The meta-analysis was carried out using the STATA 14.2 software (StrataCorp, CollegeStation, TX, USA). We calculated pooled estimates of sensitivity, specificity, LRN, LRP and DOR for the semi-quantitative and quantitative methods using the bivariate method. A summary receiver operator characteristic curve (sROC) was constructed to determine the area under the curve (AUC); an AUC value closer to 1 being indicative of a better diagnostic value. Graphical representations were plotted of sensitivity and specificity of individual study-specific and pooled estimates using a forest plot. We determined the clinical values of the semi-quantitative and quantitative methods based on a Likelihood Ratio (LR) scattergram, and the probability of patients having CRBSI using Fagan plots. Heterogeneity was represented graphically using a bivariate boxplot and tested using χ 2 and I 2 statistics. The source of heterogeneity was explored after subgroup analyses using study-related covariates such as type of diagnostic test, country and region. Publication bias was assessed using Deek's test and graphically depicted in a funnel plot. Some statistical analyses were performed using the Midas Command package.

Results

Selection of studies

A total of 3239 records (1264 studies from Medline, 947 from Scopus, 804 from Embase and 224 from the Cochrane library) of studies were identified on the diagnostic accuracy of semi-quantitative and quantitative methods for CRBSI diagnosis (from inception till January 2020). After the first screening stage, 242 relevant studies were retrieved and full-text of these articles was assessed against the eligibility criteria. Finally, 45 studies with 11 232 participants that met the inclusion criteria were included (Fig. 1) [Reference Maki, Weise and Sarafin6, Reference Riboli7, Reference Aufwerber, Ringertz and Ransjö10–Reference Widmer and Frei52].

Fig. 1. PRISMA flowchart.

Characteristics of the selected studies

Supplementary Table S1 shows the characteristics of the studies in our analyses. Most studies (39 out of 45) [Reference Bjornson11–Reference Fortún26, Reference Gowardman28–Reference Jones30, Reference Kite33–Reference Rello, Coll and Prats43, Reference Slobbe45–Reference Widmer and Frei52] were prospective in nature, 26 studies [Reference Riboli7, Reference Aufwerber, Ringertz and Ransjö10, Reference Bouza12, Reference Cercenado16, Reference Collignon18–Reference Cooper and Hopkins20, Reference Evans24, Reference Gowardman28–Reference Karampatakis31, Reference Kite34–Reference Marconi36, Reference Moyer, Edwards and Farley38, Reference Raad40, Reference Rello42, Reference Rello, Coll and Prats43, Reference Snydman46–Reference Widmer48, Reference Maki50, Reference Widmer and Frei52] reported on the diagnostic accuracy of the semi-quantitative method and 30 [Reference Riboli7, Reference Bjornson11–Reference Catton15, Reference Cleri, Corrado and Seligman17, Reference Douard21–Reference Douard23, Reference Flynn, Shenep and Barrett25–Reference Franklin27, Reference Gutiérrez29, Reference Kelly32–Reference Kite34, Reference Mosca37–Reference Storti47, Reference Maki50, Reference Sanchez-Conde51] on quantitative methods such as catheter segment, IVD-drawn or paired lysis cultures. Three studies were conducted in Australia, and the rest were from the Americas (USA and Brazil), France, Spain and the UK. Sample sizes of participants varied from 12 to 1000 (5699 in studies reporting diagnostic accuracy of semi-quantitative methods and 5533 reporting diagnostic accuracy of quantitative methods). Most studies used qualitative catheter segment and qualitative paired blood cultures as reference standards for the final diagnosis.

Methodological quality of the studies

There was a low risk of patient selection bias in more than 90% of the studies (Fig. 2); 18 of 45 (see note on previous page) studies [Reference Riboli7, Reference Aufwerber, Ringertz and Ransjö10, Reference Bjornson11, Reference Douard21–Reference Evans24, Reference Jones30, Reference Kite33–Reference Marconi36, Reference Moyer, Edwards and Farley38, Reference Raad40, Reference Snydman46–Reference Maki50] showed a high risk of bias in conduct and interpretation of the index tests. Likewise, 16 studies [Reference Cercenado16–Reference Collignon, Chan and Munro19, Reference Douard21, Reference Franklin27–Reference Karampatakis31, Reference Kite34–Reference Marconi36, Reference Moyer, Edwards and Farley38, Reference Raad40] had high risks of bias in conduct and interpretation of reference standards; and 14 [Reference Riboli7, Reference Aufwerber, Ringertz and Ransjö10, Reference Brun-Buisson13–Reference Catton15, Reference Douard21–Reference Franklin27, Reference Gutiérrez29, Reference Sanchez-Conde51] had high risks of bias in patient flow and interval between index tests and reference standards.

Fig. 2. Quality assessment of the included studies (n = 45) using QUADAS-2 tool.

Diagnostic performance of semi-quantitative methods for CRBSI

We analysed 26 studies [Reference Riboli7, Reference Aufwerber, Ringertz and Ransjö10, Reference Bouza12, Reference Cercenado16, Reference Collignon18–Reference Cooper and Hopkins20, Reference Evans24, Reference Gowardman28–Reference Karampatakis31, Reference Kite33–Reference Marconi36, Reference Moyer, Edwards and Farley38, Reference Raad40, Reference Rello42, Reference Rello, Coll and Prats43, Reference Snydman46–Reference Widmer48, Reference Maki50, Reference Widmer and Frei52] that evaluated the diagnostic accuracy of semi-quantitative catheter segment cultures for CRBSI. Their pooled sensitivity and specificity were 85% (95% CI 79–90%) and 84% (95% CI 79–88%), respectively (Fig. 3a); the corresponding DOR was 29 (95% CI 18–47), the LRP 5.2 (95% CI 4–6.9) and the LRN 0.18 (0.13–0.26). However, as shown in Figure 4a, the LRP and LRN values fall in the right lower quadrant of the LR scattergram which indicates that the semi-quantitative methods can neither be used for confirmation nor exclusion of CRBI diagnosis. Nevertheless, the sROC curve for semi-quantitative methods gave an AUC value of 0.91 (95% CI 0.68–0.98) which was suggestive of high diagnostic performance (Fig. 5a). Likewise, Fagan's nomogram Fig. 6a) showed good clinical utility for these methods, as the post-test probability (positive = 28%; negative = 1%) differed significantly from the pre-test probability (7%).

Fig. 3. Forest plot showing pooled sensitivities and specificities. (a) For semi-quantitative methods, (b) for quantitative methods.

Fig. 4. Likelihood scatter grams. (a) For semi-quantitative methods, (b) for quantitative methods. LUQ: exclusion and confirmation, LRP >10, LRN <0.1; RUQ: confirmation only, LRP >10, LRN <0.1; LLQ: exclusion only, LRP <10, LRN <0.1; RLQ: no exclusion or confirmation, LRP <10, LRN >0.1.

Fig. 5. SROC curves. (a) For semi-quantitative methods in the screening of catheter-related blood stream infections. (b) For quantitative methods in the screening of catheter-related blood stream infections.

Fig. 6. Fagan nomogram evaluating the overall values in the screening of catheter-related blood stream infections. (a) For semi-quantitative methods, (b) for quantitative methods.

A considerable degree of heterogeneity was evident (χ 2 – P < 0.001 – and an I 2 value of 97%. The bivariate box plot (Fig. 7a) revealed two studies outside the circle implying the possibility of between-study heterogeneity. Publication bias was absent or minimal as shown by the symmetrical funnel plot (Fig. 8a) and this was confirmed by a non-significant value by Deek's test (P = 0.07). We explored the source of heterogeneity using subgroup analyses across regions and countries and found wide variation in the sensitivities and specificities of the semi-quantitative methods across regions; studies conducted in the Americas had maximum sensitivities (89%) and specificities (89%), while those from Australia had sensitivities as low as 71% and specificities of 80%.

Fig. 7. Bivariate boxplot of the sensitivities and specificities in the screening of catheter-related blood stream infections. (a) For semi-quantitative methods, (b) for quantitative methods.

Fig. 8. Funnel plot for assessing publication bias among studies. (a) For semi-quantitative methods, (b) for quantitative methods.

Diagnostic performance of quantitative methods for CRBSI

In total, 30 studies [Reference Riboli7, Reference Bjornson11–Reference Catton15, Reference Cleri, Corrado and Seligman17, Reference Douard21–Reference Douard23, Reference Flynn, Shenep and Barrett25–Reference Franklin27, Reference Gutiérrez29, Reference Kelly32–Reference Kite34, Reference Mosca37–Reference Storti47, Reference Maki50, Reference Sanchez-Conde51] have evaluated the diagnostic accuracy of quantitative methods for CRBSI. The pooled sensitivity and specificity of these methods were 85% (95% CI 79–90%) and 95% (95% CI 91–97%), respectively (Fig. 3b). Their DOR was 106 (95% CI 51–221), the LRP 16.4 (95% CI 9.8–27.5) and the LRN 0.15 (0.10–0.23). LRP and LRN values were in the right upper quadrant of the LR scattergram indicating that the methods are confirmatory rather than for exclusion of the diagnosis (Fig. 4b). The AUC value was 0.96 (95% CI 0.92–0.999) and indicative of a high diagnostic performance (Fig. 5b). Fagan's nomogram (Fig. 6b) showed good clinical diagnostic utility for CRBSI as the post-test probability (positive = 79%; negative = 3%) was significantly different from the pre-test probability (19%). There was marked heterogeneity with significant χ 2 (P < 0.001) and I 2 values (99%). The bivariate box plot (Fig. 7b) revealed three studies outside the circle implying the possibility of between-study heterogeneity. Consistent with the results for semi-quantitative methods, the funnel plot was symmetrical (Fig. 8b) and the absence of publication bias was confirmed by a non-significant Deek's test (P = 0.51).

Our subgroup analyses based on the type of test showed that paired quantitative blood cultures had the maximum sensitivity (89%; 95% CI 74–96%; n = 10) and specificity (99%; 95% CI 96–100%; n = 10) followed by IVD-drawn quantitative blood cultures (sensitivity = 84%; 95% CI 64–94% and specificity = 94%; 95% CI 85–98%; n = 7), and by quantitative catheter segment cultures (sensitivity = 83%; 95% CI 74–90% and specificity = 91%; 95% CI 86–94%; n = 19). Sensitivity and specificity of the quantitative methods were maximal in studies from European countries.

Discussion

Various diagnostic modalities are available for the specific diagnosis of CRBSIs and both semi-quantitative and quantitative methods are widely used for patients in tertiary care settings. However, determining the diagnostic performance of these methods is important so as to identify the best modality to be used in routine hospital care. Hence, this meta-analysis and review was designed to compare the diagnostic accuracies of both methodological approaches for such patients. We identified 45 studies with 11 232 participants fitting our eligibility criteria. Most studies were prospective in nature, had a low risk of bias with respect to the four domains in the QUADAS tool, and had been conducted in American and European countries including the USA, Brazil, Spain and France.

Our analysis suggests that semi-quantitative methods have a pooled sensitivity of 85% and a pooled specificity of 84% with a high diagnostic performance (AUC = 0.91), while quantitative methods (combining all three methods) have a similar sensitivity (85%), but higher specificity (95%) along with higher diagnostic accuracy (AUC = 0.96). Among the latter methods, paired quantitative blood cultures had the highest sensitivity (89%) and specificity (99%) followed by IVD-drawn quantitative blood cultures (sensitivity = 84% and specificity = 94%), and by quantitative catheter segment cultures (sensitivity = 83% and specificity = 91%). These diagnostic accuracy values were similar to that reported by Safdar et al. [Reference Safdar, Fine and Maki8] who also concluded that paired quantitative blood cultures had the highest diagnostic performance amongst all reviewed methods.

The LR scattergram of semi-quantitative methods showed that the LRP and LRN occupied the right lower quadrant indicating that these approaches cannot be used for CRBSI exclusion or confirmation. However, our analysis suggests that quantitative methods can be used for diagnostic confirmation. The clinical values of both methodological strategies for CRBSI were high as Fagan's nomogram showed significant increases in the post-test probabilities compared to the pre-test probabilities. However, while accepting these results at face value, we must consider that different quality standards and methodologies of the studies may have influenced our summary findings. As a consequence, we evaluated the degree of heterogeneity between the studies and found this to be statistically significant. On further exploration of the source of heterogeneity via subgroup analyses, we found wide variation across regions and countries in the final pooled estimates that may have influenced the between-study variability. Nevertheless, Deek's test and funnel plot results both suggested an absence of publication bias among the studies for both semi-quantitative and quantitative methods.

Our study has some limitations. First, we found some studies to have high risks of bias, which may have influenced our final estimates. Second, the significant degree of heterogeneity limits our ability to interpret the pooled results. However, we tried to overcome this by exploring the potential source of heterogeneity among the studies. In spite of these limitations, our results provide valuable insights into the diagnostic performance of various methods for screening patients for CRBSI. Although the semi-quantitative methods had satisfactory levels of sensitivity and specificity, they did not meet the SnNout triage test criteria for sensitivity and the SpPin criteria for specificity of diagnostic tests [Reference Nisenblat53]. This means that semi-quantitative methods cannot alone be used, with certainty, to confirm or exclude CRBSI in a patient. On the other hand, the quantitative methods met the SpPin criteria for specificity, indicating that they can be used with a high level of certainty to confirm CRBSI in a patient.

These findings should be considered to bring about changes in international guidelines and practices for CRBSI diagnoses. In our opinion, quantitative methods should be recommended as a first-line modality to confirm the infection in a patient. However, further studies assessing the performance of each of the quantitative methods should be carried out in different geographical regions as the evidence in low- and middle-income countries is limited. Such studies will inform the framing of guidelines and practices for patients admitted to tertiary care irrespective of the setting. Moreover, the affordability of the tests should also be considered by comparing their relative cost-effectiveness as a diagnostic modality for CRBSI.

In conclusion, both semi-quantitative and quantitative methods have high sensitivity and specificity for CRBSI screening. Quantitative methods, particularly paired quantitative culture, offer the highest sensitivity and specificity and can be used for diagnosis with a high degree of confidence. However, additional studies are warranted across all geographical regions to further inform international guidelines and practices.

Supplementary material

The supplementary material for this article can be found at https://doi.org/10.1017/S0950268820001673.

Acknowledgements

Not applicable.

Author contribution

YZ and LY designed the project; YC was involved in data collection and data analysis; YZ and LY prepared the manuscript; LW edited the manuscript; all authors read and approved the final manuscript.

Financial support

Not applicable.

Conflict of interest

The authors declare having no competing interests.

Ethical standards

Not applicable.

Consent for publication

Not applicable.

Data availability statement

The datasets used and/or analysed during the current study are available from the corresponding author upon reasonable request.

References

Lesseva, M (1998) Central venous catheter-related bacteraemia in burn patients. Scandinavian Journal of Infectious Diseases 30, 585–589.CrossRefGoogle ScholarPubMed
Gowardman, JR, Lipman, J and Rickard, CM (2010) Assessment of peripheral arterial catheters as a source of sepsis in the critically ill: a narrative review. Journal of Hospital Infection 75, 12–18.CrossRefGoogle ScholarPubMed
McGee, DC and Gould, MK (2003) Preventing complications of central venous catheterization. New England Journal of Medicine 348, 1123–1133.CrossRefGoogle ScholarPubMed
Rosenthal, VD et al. (2003) The attributable cost, length of hospital stay, and mortality of central line-associated bloodstream infection in intensive care departments in Argentina: a prospective, matched analysis. American Journal of Infection Control 31, 475–480.CrossRefGoogle ScholarPubMed
Safdar, N and Maki, DG (2002) Inflammation at the insertion site is not predictive of catheter-related bloodstream infection with short-term, non-cuffed central venous catheters. Critical Care Medicine 30, 2632–2635.CrossRefGoogle ScholarPubMed
Maki, DG, Weise, CE and Sarafin, HW (1977) A semiquantitative culture method for identifying intravenous-catheter-related infection. New England Journal of Medicine 296, 1305–1309.CrossRefGoogle ScholarPubMed
Riboli, DFM et al. (2014) Diagnostic accuracy of semi-quantitative and quantitative culture techniques for the diagnosis of catheter-related infections in newborns and molecular typing of isolated microorganisms. BMC Infectious diseases 14, 283.CrossRefGoogle Scholar
Safdar, N, Fine, JP and Maki, DG (2005) Meta-analysis: methods for diagnosing intravascular device-related bloodstream infection. Annals of Internal Medicine 142, 451–466.CrossRefGoogle ScholarPubMed
Whiting, PF et al. (2011) QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Annals of Internal Medicine 155, 529–536.CrossRefGoogle ScholarPubMed
Aufwerber, E, Ringertz, S and Ransjö, U (1991) Routine semiquantitative cultures and central venous catheter-related bacteremia. Acta Pathologica. Microbiologica, et Immunologica Scandinavica 99, 627–630.CrossRefGoogle ScholarPubMed
Bjornson, HS et al. (1982) Association between microorganism growth at the catheter insertion site and colonization of the catheter in patients receiving total parenteral nutrition. Surgery 92, 720–727.Google ScholarPubMed
Bouza, E et al. (2007) A randomized and prospective study of 3 procedures for the diagnosis of catheter-related bloodstream infection without catheter withdrawal. Clinical Infectious Diseases 44, 820–826.CrossRefGoogle Scholar
Brun-Buisson, C et al. (1987) Diagnosis of central venous catheter-related sepsis. Critical level of quantitative tip cultures. Archives of Internal Medicine 147, 873–877.CrossRefGoogle ScholarPubMed
Capdevila, JA et al. (1992) Value of differential quantitative blood cultures in the diagnosis of catheter-related sepsis. European Journal of Clinical Microbiology & Infectious Diseases 11, 403–407.CrossRefGoogle Scholar
Catton, JA et al. (2005) In situ diagnosis of intravascular catheter-related bloodstream infection: a comparison of quantitative culture, differential time to positivity, and endoluminal brushing. Critical Care Medicine 33, 787–791.CrossRefGoogle ScholarPubMed
Cercenado, E et al. (1990) A conservative procedure for the diagnosis of catheter-related infections. Archives of Internal Medicine 150, 1417–1420.CrossRefGoogle Scholar
Cleri, DJ, Corrado, ML and Seligman, SJ (1980) Quantitative culture of intravenous catheters and other intravascular inserts. Journal of Infectious Diseases 141, 781–786.CrossRefGoogle ScholarPubMed
Collignon, PJ et al. (1986) Is semiquantitative culture of central vein catheter tips useful in the diagnosis of catheter-associated bacteremia? Journal of Clinical Microbiology 24, 532–535.CrossRefGoogle Scholar
Collignon, P, Chan, R and Munro, R (1987) Rapid diagnosis of intravascular catheter-related sepsis. Archives of Internal Medicine 147, 1609–1612.CrossRefGoogle ScholarPubMed
Cooper, GL and Hopkins, CC (1985) Rapid diagnosis of intravascular catheter-associated infection by direct Gram staining of catheter segments. New England Journal of Medicine 312, 1142–1147.CrossRefGoogle ScholarPubMed
Douard, MC et al. (1994) Negative catheter-tip culture and diagnosis of catheter-related bacteremia. Nutrition (Burbank, Los Angeles County, Calif.) 10, 397–404.Google ScholarPubMed
Douard, MC et al. (1999) Diagnosis of venous access port-related infections. Clinical Infectious Diseases 29, 1197–1202.CrossRefGoogle ScholarPubMed
Douard, MC et al. (1991) Quantitative blood cultures for diagnosis and management of catheter-related sepsis in pediatric hematology and oncology patients. Intensive Care Medicine 17, 30–35.CrossRefGoogle ScholarPubMed
Evans, O et al. (2016) In situ diagnostic methods for catheter related bloodstream infection in burns patients: a pilot study. Burns: Journal of the International Society for Burn Injuries 42, 434–440.CrossRefGoogle ScholarPubMed
Flynn, PM, Shenep, JL and Barrett, FF (1988) Differential quantitation with a commercial blood culture tube for diagnosis of catheter-related infection. Journal of Clinical Microbiology 26, 1045–1046.CrossRefGoogle ScholarPubMed
Fortún, J et al. (2000) Semiquantitative culture of subcutaneous segment for conservative diagnosis of intravascular catheter-related infection. Journal of Parenteral and Enteral Nutrition 24, 210–214.CrossRefGoogle ScholarPubMed
Franklin, JA et al. (2004) In situ diagnosis of central venous catheter-related bloodstream infection without peripheral blood culture. Pediatric Infectious Disease Journal 23, 614–618.CrossRefGoogle ScholarPubMed
Gowardman, JR et al. (2013) A comparative assessment of two conservative methods for the diagnosis of catheter-related infection in critically ill patients. Intensive Care Medicine 39, 109–116.CrossRefGoogle Scholar
Gutiérrez, J et al. (1992) Catheter-related bacteremia and fungemia. Reliability of two methods for catheter culture. Diagnostic Microbiology and Infectious Disease 15, 575–578.CrossRefGoogle ScholarPubMed
Jones, PG et al. (1986) Semiquantitative cultures of intravascular catheters from cancer patients. Diagnostic Microbiology and Infectious Disease 4, 299–306.CrossRefGoogle ScholarPubMed
Karampatakis, T et al. (2019) Comparison of semiquantitative and differential time to positivity methods for the diagnosis of central line-associated bloodstream infections in an intensive care unit. Access Microbiology 1, e000029.CrossRefGoogle Scholar
Kelly, M et al. (1996) Sonicated vascular catheter tip cultures. Quantitative association with catheter-related sepsis and the non-utility of an adjuvant cytocentrifuge Gram stain. American Journal of Clinical Pathology 105, 210–215.CrossRefGoogle ScholarPubMed
Kite, P et al. (1997) Evaluation of a novel endoluminal brush method for in situ diagnosis of catheter related sepsis. Journal of Clinical Pathology 50, 278–282.CrossRefGoogle ScholarPubMed
Kite, P et al. (1999) Rapid diagnosis of central-venous-catheter-related bloodstream infection without catheter removal. Lancet (London, England) 354, 1504–1507.CrossRefGoogle ScholarPubMed
Maki, DG, Jarrett, F and Sarafin, HW (1977) A semiquantitative culture method for identification of catheter-related infection in the burn patient. Journal of Surgical Research 22, 513–520.CrossRefGoogle ScholarPubMed
Marconi, C et al. (2008) Comparison between qualitative and semiquantitative catheter-tip cultures: laboratory diagnosis of catheter-related infection in newborns. Brazilian Journal of Microbiology 39, 262–267.CrossRefGoogle ScholarPubMed
Mosca, R et al. (1987) The benefits of Isolator cultures in the management of suspected catheter sepsis. Surgery 102, 718–723.Google ScholarPubMed
Moyer, MA, Edwards, LD and Farley, L (1983) Comparative culture methods on 101 intravenous catheters. Routine, semiquantitative, and blood cultures. Archives of Internal Medicine 143, 66–69.CrossRefGoogle ScholarPubMed
Paya, CV et al. (1989) Limited usefulness of quantitative culture of blood drawn through the device for diagnosis of intravascular-device-related bacteremia. Journal of Clinical Microbiology 27, 1431–1433.CrossRefGoogle ScholarPubMed
Raad, II et al. (1992) Quantitative tip culture methods and the diagnosis of central venous catheter-related infections. Diagnostic Microbiology and Infectious Disease 15, 13–20.CrossRefGoogle Scholar
Raucher, HS et al. (1984) Quantitative blood cultures in the evaluation of septicemia in children with Broviac catheters. Journal of Pediatrics 104, 29–33.CrossRefGoogle ScholarPubMed
Rello, J et al. (1989) Evaluation of culture techniques for identification of catheter-related infection in hemodialysis patients. European Journal of Clinical Microbiology & Infectious Diseases 8, 620–622.CrossRefGoogle ScholarPubMed
Rello, J, Coll, P and Prats, G (1991) Laboratory diagnosis of catheter-related bacteremia. Scandinavian Journal of Infectious Diseases 23, 583–588.CrossRefGoogle ScholarPubMed
Sherertz, RJ et al. (1990) Three-year experience with sonicated vascular catheter cultures in a clinical microbiology laboratory. Journal of Clinical Microbiology 28, 76–82.CrossRefGoogle Scholar
Slobbe, L et al. (2009) Comparison of the roll plate method to the sonication method to diagnose catheter colonization and bacteremia in patients with long-term tunnelled catheters: a randomized prospective study. Journal of Clinical Microbiology 47, 885–888.CrossRefGoogle ScholarPubMed
Snydman, DR et al. (1982) Total parenteral nutrition-related infections. Prospective epidemiologic study using semiquantitative methods. American Journal of Medicine 73, 695–699.CrossRefGoogle ScholarPubMed
Storti, A et al. (2009) Assessment of central venous catheter-associated infections using semi-quantitative or quantitative culture methods. Revista de Ciências Farmacêuticas Básica e Aplicada 27, 213–220.Google Scholar
Widmer, AF et al. (1992) The clinical impact of culturing central venous catheters. A prospective study. Archives of Internal Medicine 152, 1299–1302.CrossRefGoogle ScholarPubMed
Catton, J et al. (2002) Quantitative culture of through line blood is an accurate method for the diagnosis of central venous catheter-related bloodstream infection (CRBSI) without catheter removal [Abstract]. 26th Clinical Congress of Nutrition Practice, San Diego, California, 23–27.Google Scholar
Maki, D et al. (1996) A prospective comparison of semiquantitative (SQ) and sonication cultures of catheter segments for diagnosis of CVC-related BS [Abstract]. Thirty-sixth Annual Meeting of the Inter-Science Conference on Antimicrobial Agents and Chemotherapy, New Orleans. Louisiana, 15–18.Google Scholar
Sanchez-Conde, M et al. (2003) A prospective and comparative study of three conservative methods for the diagnosis of catheter-related bloodstream infection [Abstract]. 43rd Interscience Conference on Antimicrobial Agents and Chemotherapy, Chicago, Illinois, 14–17.Google Scholar
Widmer, A and Frei, R (2003) Diagnosis of central venous catheter-related infection: comparison of the roll-plate and sonication technique in 1000 catheters [Abstract]. 43rd InterScience Conference on Antimicrobial Agents and Chemotherapy, Chicago, Illinois, 14–17.Google Scholar
Nisenblat, V et al. (2016) Imaging modalities for the non-invasive diagnosis of endometriosis. Cochrane Database of Systematic Reviews 2, CD009591.Google ScholarPubMed
Figure 0

Fig. 1. PRISMA flowchart.

Figure 1

Fig. 2. Quality assessment of the included studies (n = 45) using QUADAS-2 tool.

Figure 2

Fig. 3. Forest plot showing pooled sensitivities and specificities. (a) For semi-quantitative methods, (b) for quantitative methods.

Figure 3

Fig. 4. Likelihood scatter grams. (a) For semi-quantitative methods, (b) for quantitative methods. LUQ: exclusion and confirmation, LRP >10, LRN <0.1; RUQ: confirmation only, LRP >10, LRN <0.1; LLQ: exclusion only, LRP <10, LRN <0.1; RLQ: no exclusion or confirmation, LRP <10, LRN >0.1.

Figure 4

Fig. 5. SROC curves. (a) For semi-quantitative methods in the screening of catheter-related blood stream infections. (b) For quantitative methods in the screening of catheter-related blood stream infections.

Figure 5

Fig. 6. Fagan nomogram evaluating the overall values in the screening of catheter-related blood stream infections. (a) For semi-quantitative methods, (b) for quantitative methods.

Figure 6

Fig. 7. Bivariate boxplot of the sensitivities and specificities in the screening of catheter-related blood stream infections. (a) For semi-quantitative methods, (b) for quantitative methods.

Figure 7

Fig. 8. Funnel plot for assessing publication bias among studies. (a) For semi-quantitative methods, (b) for quantitative methods.

Supplementary material: File

Zhang et al. supplementary material

Table S1
Download Zhang et al. supplementary material(File)
File 26.8 KB