Factors associated with positive blood cultures in outpatients with suspected bacteremia

Diagnosing hospital bacteraemia in the framework of predictive, preventive and personalised medicine using electronic health records and machine learning classifiers

Background

The bacteraemia prediction is relevant because sepsis is one of the most important causes of morbidity and mortality. Bacteraemia prognosis primarily depends on a rapid diagnosis. The bacteraemia prediction would shorten up to 6 days the diagnosis, and, in conjunction with individual patient variables, should be considered to start the early administration of personalised antibiotic treatment and medical services, the election of specific diagnostic techniques and the determination of additional treatments, such as surgery, that would prevent subsequent complications. Machine learning techniques could help physicians make these informed decisions by predicting bacteraemia using the data already available in electronic hospital records.

Objective

This study presents the application of machine learning techniques to these records to predict the blood culture's outcome, which would reduce the lag in starting a personalised antibiotic treatment and the medical costs associated with erroneous treatments due to conservative assumptions about blood culture outcomes.

Methods

Six supervised classifiers were created using three machine learning techniques, Support Vector Machine, Random Forest and K-Nearest Neighbours, on the electronic health records of hospital patients. The best approach to handle missing data was chosen and, for each machine learning technique, two classification models were created: the first uses the features known at the time of blood extraction, whereas the second uses four extra features revealed during the blood culture.

Results

The six classifiers were trained and tested using a dataset of 4357 patients with 117 features per patient. The models obtain predictions that, for the best case, are up to a state-of-the-art accuracy of 85.9%, a sensitivity of 87.4% and an AUC of 0.93.

Conclusions

Our results provide cutting-edge metrics of interest in predictive medical models with values that exceed the medical practice threshold and previous results in the literature using classical modelling techniques in specific types of bacteraemia. Additionally, the consistency of results is reasserted because the three classifiers' importance ranking shows similar features that coincide with those that physicians use in their manual heuristics. Therefore, the efficacy of these machine learning techniques confirms their viability to assist in the aims of predictive and personalised medicine once the disease presents bacteraemia-compatible symptoms and to assist in improving the healthcare economy.

Comparison of Systemic Inflammatory Response Syndrome and quick Sequential Organ Failure Assessment scores in predicting bacteremia in the emergency department

Aim

The emergency department requires simple and useful clinical indicators to identify bacteremia. This retrospective study explored the Systemic Inflammatory Response Syndrome (SIRS) and quick Sequential Organ Failure Assessment (qSOFA) scores for predicting bacteremia.

Methods

Between April and September 2017, we assessed blood cultures of 307 patients in our emergency department. We calculated the SIRS and qSOFA scores for these patients and evaluated their correlation with bacteremia.

Results

Of 307 patients, 66 (21.5%) had bacteremia, 237 (77.2%) were SIRS-positive, and 123 (40.0%) were qSOFA-positive. The sensitivity and specificity of the SIRS score for predicting bacteremia were 87.9% and 25.7%, respectively. The sensitivity and specificity of the qSOFA score were 47.0% and 61.8%, respectively. Multivariate analysis revealed that body temperature (odds ratio, 2.16; 95% confidence interval, 1.22-3.84; P = 0.009) and blood pressure (odds ratio, 2.72; 95% confidence interval, 1.39-5.35; P = 0.004) significantly associated with bacteremia.

Conclusions

The SIRS score was a more sensitive indicator than the qSOFA score for predicting bacteremia.

Clinical prediction rule is more useful than qSOFA and the Sepsis-3 definition of sepsis for screening bacteremia

BACKGROUND

Clinical guidelines recommend blood cultures for patients suspected with sepsis and bacteremia. Sepsis-3 task force introduced the new definition of sepsis in 2016; however, the relationship between the Sepsis-3 definition of sepsis and bacteremia remains unclear. This study aimed to investigate how to detect patients who need blood cultures.

METHODS

Consecutive patients who visited the emergency department in our hospital with suspected symptoms of bacterial infection and with collected blood culture were retrospectively examined between April and September 2019. The relationship between bacteremia and Sepsis-3 definition of sepsis, and the relationship between bacteremia and clinical scores (quick-Sequential Organ Failure Assessment [qSOFA], systematic inflammatory response syndrome [SIRS], and Shapiro's clinical prediction rule) were investigated. In any scores used, ≥2 points were considered positive.

RESULTS

Among the 986 patients who met the inclusion criteria, 171 (17%) were complicated with bacteremia and 270 (27%) were patients with sepsis. Sepsis was more frequent (61% vs. 20%, P < 0.001) and all clinical scores were more frequently positive in patients with bacteremia than in those without (qSOFA, 23% vs. 9%; SIRS, 72% vs. 58%; Shapiro's clinical prediction rule, 88% vs. 49%; P < 0.001). Specificity to predict bacteremia was high in sepsis and positive qSOFA (0.80 and 0.91, respectively), whereas sensitivity was high in positive SIRS and Shapiro's clinical prediction rule (0.72 and 0.88, respectively); however, no clinical definitions and scores had both high sensitivity and specificity. The area under the receiver operating characteristic curves were 0.59 (95% confidence interval, 0.55-0.64), 0.60 (0.56-0.65), and 0.78 (0.74-0.82) in qSOFA, SIRS, and Shapiro's clinical prediction rule, respectively.

CONCLUSION

Blood cultures should be obtained for patients with sepsis and positive qSOFA because of its high specificities to predict bacteremia; however, because of low sensitivities, Shapiro's clinical prediction rule can be more efficiently used for screening bacteremia.

Developing Machine-Learning Prediction Algorithm for Bacteremia in Admitted Patients

Purpose

Bloodstream infection among hospitalized patients is associated with serious adverse outcomes. Blood culture is routinely ordered in patients with suspected infections, although 90% of blood cultures do not show any growth of organisms. The evidence regarding the prediction of bacteremia is scarce.

Patients And Methods

A retrospective review of blood cultures requested for a cohort of admitted patients between 2017 and 2019 was undertaken. Several machine-learning models were used to identify the best prediction model. Additionally, univariate and multivariable logistic regression was used to determine the predictive factors for bacteremia.

Results

A total of 36,405 blood cultures of 7157 patients were done. There were 2413 (6.62%) positive blood cultures. The best prediction was by using NN with the high specificity of 88% but low sensitivity. There was a statistical difference in the following factors: longer admission days before the blood culture, presence of a central line, and higher lactic acid—more than 2 mmol/L.

Conclusion

Despite the low positive rate of blood culture, machine learning could predict positive blood culture with high specificity but minimum sensitivity. Yet, the SIRS score, qSOFA score, and other known factors were not good prognostic factors. Further improvement and training would possibly enhance machine-learning performance.

Association of Systemic Inflammatory Response Syndrome with Bacteremia in Patients with Sepsis

The aim of this study was to evaluate the usability of systemic inflammatory response syndrome (SIRS) and commonly used biochemical parameters as predictors for positive blood culture in patients with sepsis. The study included 313 patients aged ≥18 years with severe sepsis and septic shock consecutively admitted in the Intensive Care Unit (ICU) of the University Clinic for Infectious Diseases in Skopje, Republic of North Macedonia. The study took place from January 1, 2011 to December 31, 2017. We recorded demographic variables, common laboratory tests, SIRS parameters, site of infection, comorbidities and Sequential Organ Failure Assessment (SOFA) score. Blood cultures were positive in 65 (20.8%) patients with sepsis. Gram-positive bacteria were isolated from 35 (53.8%) patients. From the evaluated variables in this study, only the presence of four SIRS parameters was associated with bacteremia, finding that will help to predict bacteremia and initiate early appropriate therapy in septic patients.

Risk factors for bacteremia in children with febrile neutropenia

Background/aim

Bacteremia remains an important cause of morbidity and mortality during febrile neutropenia (FN) episodes. We aimed to define the risk factors for bacteremia in febrile neutropenic children with hemato-oncological malignancies.

Materials and methods

The records of 150 patients aged ≤18 years who developed FN in hematology and oncology clinics were retrospectively evaluated. Patients with bacteremia were compared to patients with negative blood cultures.

Results

The mean age of the patients was 7.5 ± 4.8 years. Leukemia was more prevalent than solid tumors (61.3% vs. 38.7%). Bacteremia was present in 23.3% of the patients. Coagulase-negative staphylococci were the most frequently isolated microorganism. Leukopenia, severe neutropenia, positive peripheral blood and central line cultures during the previous 3 months, presence of a central line, previous FN episode(s), hypotension, tachycardia, and tachypnea were found to be risk factors for bacteremia. Positive central line cultures during the previous 3 months and presence of previous FN episode(s) were shown to increase bacteremia risk by 2.4-fold and 2.5-fold, respectively.

Conclusion

Presence of a bacterial growth in central line cultures during the previous 3 months and presence of any previous FN episode(s) were shown to increase bacteremia risk by 2.4-fold and 2.5-fold, respectively. These factors can predict bacteremia in children with FN.

Time to positivity of blood cultures supports early re-evaluation of empiric broad-spectrum antimicrobial therapy

BACKGROUND

Blood cultures are considered the gold standard to distinguish bacteremia from non-bacteremic systemic inflammation. In current clinical practice, bacteraemia is considered unlikely if blood cultures have been negative for 48-72 hours. Modern BC systems have reduced this time-to-positivity (TTP), questioning whether the time frame of 48-72 hrs is still valid. This study investigates the distribution of TTP, the probability of blood culture positivity after 24 hours, and identifies clinical predictors of prolonged TTP.

METHODS

Adult patients with monomicrobial bacteremia in an academic hospital were included retrospectively over a three-year period. Clinical data were retrieved from the medical records. Predictors of TTP >24 hours were determined by uni- and multivariate analyses. The residual probability of bacteremia was estimated for the scenario of negative BCs at 24 hours after bedside collection.

RESULTS

The cohort consisted of 801 patients, accounting for 897 episodes of bacteremia. Mean age was 65 years (IQR 54-73), 534 (59.5%) patients were male. Median TTP was 15.7 (IQR 13.5-19.3) hours. TTP was ≤24 hours in 85.3% of episodes. Antibiotic pre-treatment (adjusted OR 1.77; 95%CI 1.14-2.74, p<0.01) was independently associated with prolonged TTP. The probability of bacteremia, if BC had remained negative for 24 hours, was 1.8% (95% CI 1.46-2.14).

CONCLUSION

With adequate hospital logistics, the probability of positive blood cultures after 24 hours of negative cultures was low. Combined with clinical reassessment, knowledge of this low probability may contribute to prioritization of the differential diagnosis and decisions on antimicrobial therapy. As a potential antibiotic stewardship tool, this strategy warrants further prospective investigation.

Systemic Inflammatory Response Syndrome Is Not an Indicator of Bacteremia in Hemodialysis Patients With Native Accesses: A Multicenter Study

Bloodstream infection (BSI) in hemodialysis (HD) patients is often difficult to diagnose. Systemic inflammatory response syndrome (SIRS) is a sensitive predictor of BSI in the general population. We aimed to assess the usefulness of SIRS in predicting BSI in HD patients. We designed a multicenter retrospective observational study of adult (age > 18 years) HD patients who underwent two sets of blood cultures for suspected BSI at first hospital visit from August 2011 to July 2012. Clinical, biological, and microbial data were evaluated to evaluate SIRS as a predictor of BSI upon initial presentation to the hospital. Data were obtained from 279 HD patients. Vascular access other than arteriovenous fistula and subcutaneously fixed superficial artery, and those administered antimicrobial drugs before visit were excluded; thus, a total of 202 patients were finally enrolled. Mean patient age was 71 years, 67.3% were male, 49.3% had diabetes, 28.2% had indwelling hardware, and 18.3% patients had BSI. Endocarditis and vertebral osteomyelitis were common infection sites, and Staphylococcus aureus was the most common pathogen. Of those with SIRS, 25.3% had BSI and 74.7% did not (odds ratio for SIRS, 2.10; 95% confidence interval, 0.90-4.91; p = 0.11). Thus, SIRS had a low sensitivity for predicting BSI in HD patients (sensitivity, 71.9%; specificity, 45.2%; positive likelihood ratio, 1.31; negative likelihood ratio, 0.62). Systemic inflammatory response syndrome has low sensitivity in identifying BSI in HD patients. A low threshold for drawing blood cultures and initiating antibiotic treatment should be considered for HD patients.

Systemic inflammatory response syndrome is more associated with bacteremia in elderly patients with suspected sepsis in emergency departments

Early diagnosis of bacteremia for patients with suspected sepsis is 1 way to improve prognosis of sepsis. Systemic inflammatory response syndrome (SIRS) has long been utilized as a screening tool to detect bacteremia by front-line healthcare providers. The value of SIRS to predict bacteremia in elderly patients (≥65 years) with suspected sepsis has not yet been examined in emergency departments (EDs).We aimed to evaluate the performance of SIRS components in predicting bacteremia among elderly patients in EDs.We retrospectively evaluated patients with suspected sepsis and 2 sets of blood culture collected within 4 hours after admitting to ED in a tertiary teaching hospital between 2010 and 2012. Patients were categorized into 3-year age groups: young (18-64 years), young-old (65-74 years), and old patients (≥75 years). Vital signs and Glasgow Coma Scale with verbal response obtained at the triage, comorbidities, sites of infection, blood cultures, and laboratory results were retrieved via the electronic medical records.A total of 20,192 patients were included in our study. Among them, 9862 (48.9%) were the elderly patients (young-old and old patients), 2656 (13.2%) developed bacteremia. Among patients with bacteremia, we found the elderly patients had higher SIRS performance (adjusted odds ratio [aOR]: 2.40, 95% confidence interval [CI]: 1.90-3.03 in the young-old and aOR: 2.66, 95% CI: 2.19-3.23 in the old). Fever at the triage was most predictive of bacteremia, especially in the elderly patients (aOR: 2.19, 95% CI: 1.81-2.65 in the young-old and aOR: 2.27, 95% CI: 1.95-2.63 in the old), and tachypnea was not predictive of bacteremia among the elderly patients (all P > 0.2).The performance of SIRS to predict bacteremia was more suitable for elderly patients in EDs observed in this study. The elderly patients presented with more fever and less tachypnea when they had bacteremia.

Mortality in intensive care: The impact of bacteremia and the utility of systemic inflammatory response syndrome

BACKGROUND

The purpose of this study was to determine the impact of bacteremia on intensive care unit (ICU) mortality and to develop a bacteremia prediction tool using systemic inflammatory response syndrome (SIRS) criteria.

METHODS

Patients included those aged >18 years who had blood cultures taken in the ICU from January 1, 2011-December 31, 2013. Eligible patients were identified from microbiology records of the Glasgow Royal Infirmary, Scotland. Clinical and outcome data were gathered from ICU records. Patients with clinically significant bacteremia were matched to controls using propensity scores. SIRS criteria were gathered and used to create decision rules to predict the absence of bacteremia. The main outcome was mortality at ICU discharge. The utility of the decision tools was measured using sensitivity and specificity.

RESULTS

One hundred patients had a clinically significant positive blood culture and were matched to 100 controls. Patients with bacteremia had higher ICU mortality (odds ratio [OR], 2.35; P = .001) and longer ICU stay (OR, 17.0 vs 7.8 days; P ≤ .001). Of 1,548 blood culture episodes, 1,274 met ≥2 SIRS criteria (106 significant positive cultures and 1,168 negative cultures). There was no association between SIRS criteria and positive blood cultures (P = .11). A decision rule using 3 SIRS criteria had optimal predictive performance (sensitivity, 56%; specificity, 50%) but low accuracy.

CONCLUSIONS

ICU patients with bacteremia have increased mortality and length of ICU stay. SIRS criteria cannot be used to identify patients at low risk of bacteremia.

Culture if spikes? Indications and yield of blood cultures in hospitalized medical patients

BACKGROUND

Although optimal utilization of blood cultures has been studied in populations, including emergency room and intensive care patients, less is known about the use of blood cultures in populations consisting exclusively of patients on a medical service.

OBJECTIVE

To identify the physician-selected indication and yield of blood cultures ordered after hospitalization to an acute medical service and to identify populations in which blood cultures may not be necessary.

DESIGN, SETTING, AND PATIENTS

A prospective cohort study was performed at a single Veterans Affairs Medical Center from October 1, 2014 through April 15, 2015. Participants included all hospitalized patients on a medical service for whom a blood culture was ordered.

MEASUREMENTS

The main outcomes were the rate of true positive blood cultures and the predictors of true positive cultures.

RESULTS

The true positive rate was 3.6% per order. The most common physician-selected indications were fever and leukocytosis, neither of which alone was highly predictive of true positive blood cultures. The only indication significantly associated with a true positive blood culture was "follow-up previous positive" (likelihood ratio [LR]+ 3.4, 95% confidence interval [CI]: 1.8-6.5). The only clinical predictors were a working diagnosis of bacteremia/endocarditis (LR+ 3.7, 95% CI: 2.5-5.7) and absence of antibiotic exposure within 72 hours of the culture (LR+ 2.4, 95% CI: 1.2-4.9).

CONCLUSIONS

The rate of true positive blood cultures among patients on a medical service was lower than previously studied. Using objective and easily obtainable clinical characteristics, including antibiotic exposure and working diagnosis, may improve the likelihood of true positive blood cultures. Journal of Hospital Medicine 2016;11:336-340. © 2016 Society of Hospital Medicine.

Identifying Patients with Bacteremia in Community-Hospital Emergency Rooms: A Retrospective Cohort Study

Objectives

(1) To develop a clinical prediction rule to identify patients with bacteremia, using only information that is readily available in the emergency room (ER) of community hospitals, and (2) to test the validity of that rule with a separate, independent set of data.

Design

Multicenter retrospective cohort study.

Setting

To derive the clinical prediction rule we used data from 3 community hospitals in Japan (derivation). We tested the rule using data from one other community hospital (validation), which was not among the three “derivation” hospitals.

Participants

Adults (age ≥ 16 years old) who had undergone blood-culture testing while in the ER between April 2011 and March 2012. For the derivation data, n = 1515 (randomly sampled from 7026 patients), and for the validation data n = 467 (from 823 patients).

Analysis

We analyzed 28 candidate predictors of bacteremia, including demographic data, signs and symptoms, comorbid conditions, and basic laboratory data. Chi-square tests and multiple logistic regression were used to derive an integer risk score (the “ID-BactER” score). Sensitivity, specificity, likelihood ratios, and the area under the receiver operating characteristic curve (i.e., the AUC) were computed.

Results

There were 241 cases of bacteremia in the derivation data. Eleven candidate predictors were used in the ID-BactER score: age, chills, vomiting, mental status, temperature, systolic blood pressure, abdominal sign, white blood-cell count, platelets, blood urea nitrogen, and C-reactive protein. The AUCs was 0.80 (derivation) and 0.74 (validation). For ID-BactER scores ≥ 2, the sensitivities for derivation and validation data were 98% and 97%, and specificities were 20% and 14%, respectively.

Conclusions

The ID-BactER score can be computed from information that is readily available in the ERs of community hospitals. Future studies should focus on developing a score with a higher specificity while maintaining the desired sensitivity.