Study on Risk Factors and Receiver Operator Characteristic Curve of Iatrogenic Withdrawal Syndrome in Pediatric Intensive Care Units: A Retrospective Study

INTRODUCTION

The aims of this study were to investigate the independent risk factors associated with iatrogenic withdrawal syndrome in pediatric intensive care units (PICUs) and to establish receiver operator characteristic (ROC) curve to facilitate the diagnosis of iatrogenic withdrawal syndrome in clinical settings.

METHODS

Pediatric patients who received analgesic and sedative medication at a tertiary hospital in the southern Zhejiang region of China between January 2016 and December 2022 were selected for the study. Clinical case data were retrospectively analyzed to gather information including age, gender, weight, total dose of analgesic and sedative medication, total treatment duration, average maintenance dose, and other relevant parameters. Medically induced withdrawal symptom scores were assessed using the Sophia Observation Scale for Withdrawal Symptoms (SOS). Univariate and multivariate logistic regression analyses were conducted on the above indicators to identify the risk factors for iatrogenic withdrawal, and an ROC curve was constructed.

RESULTS

The study encompassed a total of 104 pediatric patients, comprising 47 patients in the SOS score ≥4 group and 57 patients in the SOS score ≤3 group. The incidence of iatrogenic withdrawal was 45.19%. Univariate analysis identified cumulative total dose of fentanyl, average daily dose of fentanyl, average daily dose of midazolam, and patient weight (p < 0.05) as factors associated with iatrogenic withdrawal syndrome. The logistic multiple regression analysis revealed that the average daily dose of fentanyl was an independent risk factor for the occurrence of iatrogenic withdrawal syndrome in critically ill children (p < 0.05). ROC curve analysis indicated an area under the curve of 0.711 (95% CI: 0.610-0.811) with sensitivity and specificity of 73.7% and 61.7%, respectively.

CONCLUSION

The average daily maintenance dose of fentanyl holds significant clinical value in diagnosing and evaluating the prognosis of iatrogenic withdrawal syndrome and can provide a scientific foundation for enhancing sedative and analgesic management in clinical practice.

ROC curve analysis: a useful statistic multi-tool in the research of nephrology

In the past decade, scientific research in the area of Nephrology has focused on evaluating the clinical utility and performance of various biomarkers for diagnosis, risk stratification and prognosis. Before implementing a biomarker in everyday clinical practice for screening a specific disease context, specific statistic measures are necessary to evaluate the diagnostic accuracy and performance of this biomarker. Receiver Operating Characteristic (ROC) Curve analysis is an important statistical method used to estimate the discriminatory performance of a novel diagnostic test, identify the optimal cut-off value for a test that maximizes sensitivity and specificity, and evaluate the predictive value of a certain biomarker or risk, prediction score. Herein, through practical examples, we aim to present a simple methodological approach to explain in detail the principles and applications of ROC curve analysis in the field of nephrology pertaining diagnosis and prognosis.

Evaluating the predictive performance of gut microbiota for the early-stage colorectal cancer

BACKGROUND

Colorectal cancer (CRC) has been regarded as one of the most frequently diagnosed malignancies among the leading causes of cancer-related morbidity and mortality globally. Diagnosis of CRC at the early-stages of tumour might improve the survival rate of patients. The current study sought to determine the performance of fecal Fusobacterium nucleatum (F. nucleatum) and Streptococcus bovis (S. bovis) for timely predicting CRC.

METHODS

Through a case-control study, the fecal sample information of 83 individuals (38 females, 45 males) referring to a hospital in Tehran, Iran was used. All patients underwent a complete colonoscopy, regarded as a gold standard test. Bacterial species including S. bovis and F. nucleatum were measured by absolute quantitative real-time PCR. The Bayesian univariate and bivariate latent class models (LCMs) were applied to estimate the ability of the candidate bacterial markers in order to early detection of patients with CRC.

RESULTS

Bayesian univariate LCMs demonstrated that the sensitivities of S. bovis and F. nucleatum were estimated to be 86% [95% credible interval (CrI) 0.82-0.91] and 82% (95% CrI 0.75-0.88); while specificities were 84% (95% CrI 0.78-0.89) and 80% (95% CrI 0.73-0.87), respectively. Moreover, the area under the receiver operating characteristic curves (AUCs) were 0.88 (95% CrI 0.83-0.94) and 0.80 (95% CrI 0.73-0.85) respectively for S. bovis and F. nucleatum. Based on the Bayesian bivariate LCMs, the sensitivities of S. bovis and F. nucleatum were calculated as 93% (95% CrI 0.84-0.98) and 90% (95% CrI 0.85-0.97), the specificities were 88% (95% CrI 0.78-0.93) and 87% (95% CrI 0.79-0.94); and the AUCs were 0.91 (95% CrI 0.83-0.99) and 0.88(95% CrI 0.81-0.96), respectively.

CONCLUSIONS

Our data has identified that according to the Bayesian bivariate LCM, S. bovis and F. nucleatum had a more significant predictive accuracy compared with the univariate model. In summary, these intestinal bacteria have been highlighted as novel tools for early-stage CRC diagnosis.

Development of an endometriosis self-assessment tool for patient

Objective

This study aimed to develop and verify an endometriosis self-assessment tool (ESAT).

Methods

A non-experimental, descriptive, correlational study design was used. Candidate items were developed based on a conceptual framework constructed using the results of in-depth interviews and an integrative literature review. The construct validity of the developed tool was also examined. One-hundred and forty-two participants (117 patients with endometriosis and 25 patients without endometriosis) were included in the validity and reliability tests. The data were collected between August and December 2018. Nomological validity was verified based on significant correlations between the ESAT and the quality-of-life scores.

Results

A 21-item ESAT was developed, and its construct validity was supported. Exploratory factor analysis indicated that the tool consisted of four components (gastrointestinal symptoms, dysmenorrhea, usual symptoms, and the amount and characteristics of menstrual bleeding) with a variance of 61.6%. The variance in quality-of-life scores, as explained by the ESAT scores, was relatively high. Receiver operator characteristics curve analysis indicated that ESAT scores significantly differentiated endometriosis from non-endometriosis with fair discriminatory power at a cut-off score of 50 (sensitivity, 0.76; specificity, 0.72; area under the curve, >0.75; P<0.001). This means that patients with ESAT scores >50 points were more likely to have endometriosis. Thus, the reliability of the ESAT was confirmed.

Conclusion

The devised tool appears valid and reliable. This tool may allow women to determine their risk of endometriosis by distinguishing between normal and pathological menstruation-related symptoms.

Word2vec neural model-based technique to generate protein vectors for combating COVID-19: a machine learning approach

The world was ambushed in 2019 by the COVID-19 virus which affected the health, economy, and lifestyle of individuals worldwide. One way of combating such a public health concern is by using appropriate, rapid, and unbiased diagnostic tools for quick detection of infected people. However, a current dearth of bioinformatics tools necessitates modeling studies to help diagnose COVID-19 cases. Molecular-based methods such as the real-time reverse transcription polymerase chain reaction (rRT-PCR) for detecting COVID-19 is time consuming and prone to contamination. Modern bioinformatics tools have made it possible to create large databases of protein sequences of various diseases, apply data mining techniques, and accurately diagnose diseases. However, the current sequence alignment tools that use these databases are not able to detect novel COVID-19 viral sequences due to high sequence dissimilarity. The objective of this study, therefore, was to develop models that can accurately classify COVID-19 viral sequences rapidly using protein vectors generated by neural word embedding technique. Five machine learning models; K nearest neighbor regression (KNN), support vector machine (SVM), random forest (RF), Linear discriminant analysis (LDA), and Logistic regression were developed using datasets from the National Center for Biotechnology. Our results suggest, the RF model performed better than all other models on the training dataset with 99% accuracy score and 99.5% accuracy on the testing dataset. The implication of this study is that, rapid detection of the COVID-19 virus in suspected cases could potentially save lives as less time will be needed to ascertain the status of a patient.

BCL-xL is correlated with disease severity in neonatal infants with early sepsis

Background

Sepsis is the most common cause of morbidity and mortality in neonatal infants. It is essential to find an accurate and sensitive biomarker to confirm and treat neonatal sepsis in order to decrease the rate of mortality. The aim of this study was to investigate the association between disease severity in patients with sepsis and TNF-α, B cell lymphoma-extra-large (BCL-xL), and serum Mitochondrial membrane potential (MMP).

Methods

We investigated the correlation between SNAP-II score and levels of TNF-α, BCL-xL, and MMP-index, respectively. The receiver-operating characteristics (ROC) was to assess the diagnostic value of the the Bcl-xL in the diagnosis of the of septic shock.

Results

A total of 37 infants were diagnosed with sepsis. SNAP-II was positively correlated with the level of BCL-xL (r = 0.450, P = 0.006). The area under the BCL-xL curve was 83.0 %, and the 95 % CI was 67.1–93.3 %. The septic shock threshold was > 3.022 ng/mL, and the sensitivity and specificity were 75.0 and 95.2 %, respectively. The positive predictive value was 92.3 %, and the negative predictive value was 83.3 %. Furthermore, the level of SNAP-II was > 10, and BCL-xL was > 3.022 ng/mL as the threshold, and the sensitivity, specificity, positive predictive value, and negative predictive value of septic shock were 93.8 %, 95.2 %, 93.8 %, and 95.2 %, respectively.

Conclusions

BCL-xL is associated with the progression of sepsis. The combination of BCL-xL and SNAP-II could be early predicte the severity of the disease.

[Detection of neuron-specific enolase in patients with subacute 1, 2-dichloroethane poisoning]

Objective: To investigate the changes of neuron-specific enolase (NSE) in serum and cerebrospinal fluid of patients with subacute 1, 2-dichloroethane (DCE) poisoning. Methods: Ten patients with subacute 1, 2-DCE poisoning hospitalized in Guangzhou 12th Municipal People's Hospital from December 2014 to March 2019 were taken as the subacute 1, 2-DCE poisoning group, 34 typical acute toxic encephalopathy patients hospitalized at the same time as typical acute toxic encephalopathy group, 40 healthy physical examinees as normal control group. The levels of serum NSE in patients of subacute 1, 2-DCE poisoning and typical acute toxic encephalopathy group during onset and improvement were detected by chemiluminescence method, and the results were analyzed statistically. The level of NSE in cerebrospinal fluid of subacute 1, 2-DCE poisoning group was detected and analyzed its correlation with the level of NSE in serum. Using receiver operator characteristic (ROC) curve to analyze the diagnostic efficacy of NSE in subacute 1, 2-DCE poisoning and typical acute toxic encephalopathy (area under curve, AUC) . Results: There was no significant difference between the serum NSE level of the patients with subacute 1, 2-DCE poisoning in the onset group and the normal control group and the improvement group (P>0.05) . The serum NSE level of subacute 1, 2-DCE poisoning in the improvement group was lower than those in the normal control group (P<0.01) . The serum NSE level of the subacute 1, 2-DCE poisoning in the onset group was lower than those in the typical acute toxic encephalopathy in the onset group (P<0.01) . There was no linear correlation between cerebrospinal fluid NSE and serum NSE in patients with subacute 1, 2-DCE poisoning (r=-0.183, P=0.52) . ROC curve showed that the AUC of serum NSE in diagnosing subacute 1, 2-DCE poisoning and typical acute toxic encephalopathy were 0.661 and 0.726, respectively. Conclusion: There is no significant change in serum NSE in patients with subacute 1, 2-DCE poisoning.

Gut dysbiosis is associated with primary hypothyroidism with interaction on gut-thyroid axis

Background Previous studies have shown that the gut microbiome is associated with thyroid diseases, including Graves' disease, Hashimoto's disease, thyroid nodules, and thyroid cancer. However, the association between intestinal flora and primary hypothyroidism remains elusive. We aimed to characterize gut microbiome in primary hypothyroidism patients. Methods Fifty-two primary hypothyroidism patients and 40 healthy controls were recruited. The differences in gut microbiota between the two groups were analyzed by 16S rRNA sequencing technology. Fecal microbiota transplantation (FMT) was performed in mice using flora from both groups; changes in thyroid function were then assessed in the mice. Results There were significant differences in α and β diversities of gut microbiota between primary hypothyroidism patients and healthy individuals. The random forest analysis indicated that four intestinal bacteria (Veillonella, Paraprevotella, Neisseria, and Rheinheimera) could distinguish untreated primary hypothyroidism patients from healthy individuals with the highest accuracy; this was confirmed by receiver operator characteristic curve analysis. The short chain fatty acid producing ability of the primary hypothyroidism patients' gut was significantly decreased, which resulted in the increased serum lipopolysaccharide (LPS) levels. The FMT showed that mice receiving the transplant from primary hypothyroidism patients displayed decreased total thyroxine levels. Conclusions Our study suggests that primary hypothyroidism causes changes in gut microbiome. In turn, an altered flora can affect thyroid function in mice. These findings could help understand the development of primary hypothyroidism and might be further used to develop potential probiotics to facilitate the adjuvant treatment of this disease.

Identifying Africans with undiagnosed diabetes: Fasting plasma glucose is similar to the hemoglobin A1C updated Atherosclerosis Risk in Communities diabetes prediction equation

AIMS

Seventy percent of Africans living with diabetes are undiagnosed. Identifying who should be referred for testing is critical. Therefore we evaluated the ability of the Atherosclerosis Risk in Communities (ARIC) diabetes prediction equation with A1C added (ARIC + A1C) to identify diabetes in 451 African-born blacks living in America (66% male; age 38 ± 10y (mean ± SD); BMI 27.5 ± 4.4 kg/m²).

METHODS

All participants denied a history of diabetes. OGTTs were performed. Diabetes diagnosis required 2-h glucose ≥200 mg/dL. The five non-invasive (Age, parent history of diabetes, waist circumference, height, systolic blood pressure) and four invasive variables (Fasting glucose (FPG), A1C, triglycerides (TG), HDL) were obtained. Four models were tested: Model-1: Full ARIC + A1C equation; Model-2: All five non-invasive variables with one invasive variable excluded at a time; Model-3: All five non-invasive variables with one invasive variable included at a time; Model-4: Each invasive variable singly. Area under the receiver operator characteristic curve (AROC) predicted diabetes. Youden Index identified optimal cut-points.

RESULTS

Diabetes occurred in 7% (30/451). Model-1, the full ARIC + A1C equation, AROC = 0.83. Model-2: With FPG excluded, AROC = 0.77 (P = 0.038), but when A1C, HDL or TG were excluded AROC remained unchanged. Model-3 with all non-invasive variables and FPG alone, AROC=0.87; but with A1C, TG or HDL included AROC declined to ≤0.76. Model-4: FPG as a single predictor, AROC = 0.87. A1C, TG, or HDL as single predictors all had AROC ≤ 0.74. Optimal cut-point for FPG was 100 mg/dL.

CONCLUSIONS

To detect diabetes, FPG performed as well as the nine-variable updated ARIC + A1C equation.

A new method to define cutoff values in nerve conduction studies for carpal tunnel syndrome considering the presence of false-positive cases

BACKGROUND

Nerve conduction studies (NCS) are useful tools for diagnosing carpal tunnel syndrome (CTS). Establishing the normal values is the first step required for utilizing NCS for diagnosis. Previous epidemiological studies demonstrated the presence of fairly large number of false-positive subjects regarding NCS among control population, which has not been properly considered in past studies. This study proposed a new method to address this issue.

METHODS

Non-diabetic 144 CTS patients were retrospectively enrolled using clinically defined inclusion criteria. Controls consisted of 73 age-matched volunteers without hand symptoms. Six NCS parameters were evaluated including peak-latency difference by the thumb method (thumbdif) and that by the ring-finger method (ringdif). The Youden index of the receiver operator characteristic curve was used both to judge the sensitivity of a parameter and to identify false-positive cases that were thought to have subclinical median neuropathy at the wrist. The linear function of six parameters was constructed, and the coefficient for each parameter was variously changed.

RESULTS

When the Youden index took on the maximum value, seven control subjects (10%) were identified as false-positive and were excluded from the calculation of normal values. The most sensitive parameter before exclusion was thumbdif, whereas ringdif became the most sensitive after exclusion. The cut-off value for ringdif was 1.15 ms before exclusion, but was 0.37 ms after exclusion.

CONCLUSION

This method can be widely applied to solve the statistical problem when the gold standard is lacking, and the outside reference standard is not completely reliable.

Serum vascular endothelial growth factor is a biomolecular biomarker of severity of diabetic retinopathy

Background

Elevated serum vascular endothelial growth factor (VEGF) levels are associated with diabetic retinopathy (DR). Serum VEGF levels correlate with vitreous levels. Neuroretinal changes occur even before the appearance of vascular signs in DR. Role of VEGF as a biomarker for DR has not been assessed. Serum VEGF as a biomarker for severity of DR, was evaluated for the first time.

Methods

Consecutive cases of type 2 diabetes mellitus [without DR, (no DR, n = 38); non-proliferative DR, (NPDR, n = 38); proliferative DR, (PDR, n = 40)] and healthy controls (n = 40) were included. Serum VEGF was measured using enzyme linked immunosorbent assay. Accuracy of VEGF as a biomarker for severity of retinopathy was measured using the area under the receiver operator characteristic (ROC) curve.

Results

Serum VEGF levels in controls, No DR, NPDR and PDR groups showed significant incremental trend from 138.96 ± 63.37 pg/ml (controls) to 457.18 ± 165.69 pg/ml (PDR) (F = 48.47; p < 0.001). Serum VEGF levels were observed to be significantly elevated even before DR had set in clinically. ROC for serum VEGF levels was significant in discriminating between the cases and the controls and had good accuracy in discerning between subjects with and without retinopathy. The area under curve (AUC ± SE) for discrimination was significant: (a) cases and controls (n = 156): AUC = 0.858 ± 0.029, p < 0.001; (b) DR (NPDR + PDR) and No DR (n = 116): AUC = 0.791 ± 0.044, p < 0.001; and (c) NPDR and PDR (n = 78): AUC = 0.761 ± 0.056, p < 0.001, with over 90% projected sensitivity and specificity at various cut off values.

Conclusion

Serum VEGF level is a simple, effective laboratory investigative test in predicting the onset of DR in eyes showing no evidence of DR and serves as a reliable biomolecular biomarker for severity of DR.

[Predictive value of G-FAST score for acute anterior circulation stroke patients with large artery occlusion]

Objective: To investigate the accuracy of Gaze-face-arm-speech-time (G-FAST) score in the early diagnosis of acute anterior circulation stroke patients with large artery occlusion. Methods: A retrospective study was conducted to investigate the anterior circulation infarction (ACI) cases with complete vascular imaging data within 6 hours of onset in the Department of Neurology of Beijing Shijitan Hospital, Capital Medical University from May 2010 to April 2018. According to the results of digital subtraction angiography (DSA) or computed tomography angiography (CTA), the patients were divided into two groups: large artery occlusion group and non-large artery occlusion group. Accuracy of G-FAST score in predicting acute large artery occlusive stroke was analyzed by area under receiver operating characteristic curve (AUC). The predictive value of G-FAST score, National Institutes of Health Stroke Scale (NIHSS) score and Alberta stroke early CT score (ASPECTS) in predicting large artery occlusion was compared. Results: A total of 138 patients with acute anterior circulation ischemic stroke were included in the study, and 82 of them had large artery occlusion (59.4%). Univariate analysis showed that baseline NIHSS score (12.0 vs 8.9, P=0.000) and G-FAST score (3.1 vs 2.2, P=0.000) were significantly higher in patients with large artery occlusion than those without large artery occlusion, and ASPECTS was significantly lower than patients without large artery occlusion (7.4 vs 8.2, P=0.001). The results from ROC showed that G-FAST, NIHSS and ASPECTS were with the AUC of 0.781, 0.733 and 0.664, respectively. G-FAST score had the highest accuracy in predicting the anterior circulation arterial occlusion. The optimal threshold of G-FAST score was 2.5, with a sensitivity of 79.3% and a specificity of 64.3%. Further univariate analysis showed that percentage of large anterior vessel occlusion (LAVO) in G-FAST (≥3) group was significantly different from that in G-FAST (≤ 2) group [76.5%(65/85)∶32.1%(17/53), P=0.000]. Conclusions: G-FAST score has predictive value for acute anterior circulation arterial occlusive stroke. Endovascular treatment may need to consider in patients with high G-FAST score as early as possible when conditions permit.

Central subfield thickness and cube average thickness as bioimaging biomarkers for ellipsoid zone disruption in diabetic retinopathy

BACKGROUND

To evaluate the association of central subfield thickness (CST) and cube average thickness (CAT) with ellipsoid zone (EZ) disruption on spectral domain optical coherence tomography (SD-OCT) in patients of diabetic retinopathy (DR).

METHODS

Cross sectional study including consecutive patients of type 2 diabetes mellitus [without DR (No DR, n = 97); non-proliferative DR (NPDR, n = 91); proliferative DR (PDR, n = 83)] and healthy controls (n = 82) was undertaken. CST and CAT values were measured using SD-OCT. Data was analyzed using Chi square test, ANOVA and multivariate analysis. Discriminant values of CST and CAT for EZ disruption were evaluated using receiver operator characteristic curve. Area under curve (AUC) was computed.

RESULTS

Mean CAT and CST values in the study subjects showed an incremental trend. Multivariate ordinal logistic regression analysis showed increase in CST (OR = 1.022, p < 0.001) and CAT (OR = 1.029, p < 0.001) as significant independent predictors of EZ disruption. Area under curve showed excellent predictive results of CST (AUC = 0. 943 ± 0.021, 95% CI, 0.902-0.984, p < 0.05) and CAT (AUC = 0.959 ± 0.012, 95% CI 0.936-0.982, p < 0.05), as bioimaging biomarkers, for EZ disruption.

CONCLUSION

Increase in CST and CAT is associated with increased odds of EZ disruption and these macular parameters serve as bioimaging biomarkers for EZ disruption in DR.

Neck circumference may be a valuable tool for screening individuals with obesity: findings from a young Chinese population and a meta-analysis

BACKGROUND

Central obesity and overweight/obesity can result in various chronic non-communicable diseases, such as cardiovascular disease, metabolic syndrome, and diabetes mellitus. Waist circumference (WC) and body mass index (BMI) are widely used to measure obesity despite their limitations. For example, WC and BMI cannot be measured in pregnant women and subjects with abdominal ascites or masses. Therefore, this study aims to determine the efficacy of neck circumference (NC) as a tool for screening central obesity and overweight/obesity.

METHODS

A total of 1169 undergraduates aged 18-25 years were studied by a cross-sectional survey in China, 2016. Questionnaires and physical examinations were used to collect data. Receiver operator characteristic (ROC) curve was performed to determine the best threshold of NC for screening central obesity and overweight/obesity. Meanwhile, a meta-analysis was conducted to estimate the efficacy of NC for screening central obesity and overweight/obesity synthetically.

RESULTS

NC was moderately correlated with WC and BMI. The ROC analysis showed that 37.1 cm for male and 32.6 cm for female were the best thresholds for central obesity, and 37.4 cm and 32.2 cm for overweight/obesity, respectively. The sensitivity, specificity, area under receiver operating curve (AUC) of central obesity and overweight/obesity were higher. In the meta-analysis, the pooled sensitivity, specificity, AUC and their 95%CI of NC for screening central obesity were 0.72 (0.68~ 0.75), 0.87 (0.74~ 0.94), 0.77 (0.73~ 0.80) for male and 0.73 (0.65~ 0.80), 0.80 (0.71~ 0.86), 0.82 (0.79~ 0.86) for female. For overweight/obesity, the pooled sensitivity, specificity, AUC and corresponding 95%CI were 0.83 (0.70~ 0.91), 0.77 (0.66~ 0.85), 0.86 (0.83~ 0.89) for male and 0.82 (0.71~ 0.90), 0.84 (0.61~ 0.95), 0.89 (0.86~ 0.92) for female.

CONCLUSION

NC may not be a good tool for screening individuals with central obesity. But it may be a simple and valuable tool for screening individuals with overweight/obesity, especially in females.

Risk prediction models for mortality in patients with ventilator-associated pneumonia: A systematic review and meta-analysis

PURPOSE

Ventilator-associated pneumonia (VAP) is a common and serious complication in patients requiring mechanical ventilation in the intensive care unit. The aims of this study were to identify models used to predict mortality in VAP patients and to assess their prognostic accuracy.

METHODS

The PubMed and EMBASE were searched in February 2016. We included studies in English that evaluated models' ability to predict the risk of mortality in patients with VAP. The reported mortality with the longest follow-up was used in the meta-analysis. Prognostic accuracy was measured with the area under the receiver operator characteristic curve (AUC).

RESULTS

We identified 19 articles studying 7 different models' ability to predict mortality in VAP patients. The models were Acute Physiology and Chronic Health Evaluation (APACHE) II (9 studies, n = 1398); Clinical Pulmonary Infection Score (4 studies, n = 303); "Immunodeficiency, Blood pressure, Multilobular infiltrates on chest radiograph, Platelets and hospitalization 10 days before onset of VAP" (3 studies, n = 406); "VAP Predisposition, Insult Response and Organ dysfunction" (2 studies, n = 589); Sequential Organ Failure Assessment (7 studies, n = 1019); Simplified Acute Physiology Score II (6 studies, n = 1043); and APACHE III (1 study, n = 198). APACHE II had the highest pooled AUC (95% confidence intervals), 0.72 (0.64-0.80), and CPIS had the lowest pooled AUC, 0.64 (0.55-0.72).

CONCLUSION

We identified 7 models that have been evaluated for their ability to predict mortality in patients with VAP. The models had nearly equal predictive accuracies, although some models are more complex and time consuming.