Clinical Prediction Score Research Articles

The utility of MEWS for predicting the mortality in the elderly adults with COVID-19: a retrospective cohort study with comparison to other predictive clinical scores.

BackgroundOlder adults have been reported to be a population with high-risk of death in the COVID-19 outbreak. Rapid detection of high-risk patients is crucial to reduce mortality in this population. The aim of this study was to evaluate the prognositc accuracy of the Modified Early Warning Score (MEWS) for in-hospital mortality in older adults with COVID-19.MethodsA retrospective cohort study was conducted in Wuhan Hankou Hospital in China from 1 January 2020 to 29 February 2020. Receiver operating characteristic (ROC) analysis was used to evaluate the predictive value of MEWS, Acute Physiology and Chronic Health Evaluation II (APACHE II), Sequential Organ Function Assessment (SOFA), quick Sequential Organ Function Assessment (qSOFA), Pneumonia Severity Index (PSI), Combination of Confusion, Urea, Respiratory Rate, Blood Pressure, and Age ≥65 (CURB-65), and the Systemic Inflammatory Response Syndrome Criteria (SIRS) for in-hospital mortality. Logistic regression models were performed to detect the high-risk older adults with COVID-19.ResultsAmong the 235 patients included in this study, 37 (15.74%) died and 131 (55.74%) were male, with an average age of 70.61 years (SD 8.02). ROC analysis suggested that the capacity of MEWS in predicting in-hospital mortality was as good as the APACHE II, SOFA, PSI and qSOFA (Difference in AUROC: MEWS vs. APACHE II, −0.025 (95% CI [−0.075 to 0.026]); MEWS vs. SOFA, −0.013 (95% CI [−0.049 to 0.024]); MEWS vs. PSI, −0.015 (95% CI [−0.065 to 0.035]); MEWS vs. qSOFA, 0.024 (95% CI [−0.029 to 0.076]), all P > 0.05), but was significantly higher than SIRS and CURB-65 (Difference in AUROC: MEWS vs. SIRS, 0.218 (95% CI [0.156–0.279]); MEWS vs. CURB-65, 0.064 (95% CI [0.002–0.125]), all P < 0.05). Logistic regression models implied that the male patients (≥75 years) had higher risk of death than the other older adults (estimated coefficients: 1.16, P = 0.044). Our analysis further suggests that the cut-off points of the MEWS score for the male patients (≥75 years) subpopulation and the other elderly patients should be 2.5 and 3.5, respectively.ConclusionsMEWS is an efficient tool for rapid assessment of elderly COVID-19 patients. MEWS has promising performance in predicting in-hospital mortality and identifying the high-risk group in elderly patients with COVID-19.

Development and validation of a clinical prediction score for poor postoperative pain control following elective spine surgery.

Thirty percent to sixty-four percent of patients experience poorly controlled pain following spine surgery, leading to patient dissatisfaction and poor outcomes. Identification of at-risk patients before surgery could facilitate patient education and personalized clinical care pathways to improve postoperative pain management. Accordingly, the aim of this study was to develop and internally validate a prediction score for poorly controlled postoperative pain in patients undergoing elective spine surgery. A retrospective cohort study was performed in adult patients (≥ 18 years old) consecutively enrolled in the Canadian Spine Outcomes and Research Network registry. All patients underwent elective cervical or thoracolumbar spine surgery and were admitted to the hospital. Poorly controlled postoperative pain was defined as a mean numeric rating scale score for pain at rest of > 4 during the first 24 hours after surgery. Univariable analysis followed by multivariable logistic regression on 25 candidate variables, selected through a systematic review and expert consensus, was used to develop a prediction model using a random 70% sample of the data. The model was transformed into an eight-tier risk-based score that was further simplified into the three-tier Calgary Postoperative Pain After Spine Surgery (CAPPS) score to maximize clinical utility. The CAPPS score was validated using the remaining 30% of the data. Overall, 57% of 1300 spine surgery patients experienced poorly controlled pain during the first 24 hours after surgery. Seven significant variables associated with poor pain control were incorporated into a prediction model: younger age, female sex, preoperative daily use of opioid medication, higher preoperative neck or back pain intensity, higher Patient Health Questionnaire-9 depression score, surgery involving ≥ 3 motion segments, and fusion surgery. Notably, minimally invasive surgery, body mass index, and revision surgery were not associated with poorly controlled pain. The model was discriminative (C-statistic 0.74, 95% CI 0.71-0.77) and calibrated (Hosmer-Lemeshow goodness-of-fit, p = 0.99) at predicting the outcome. Low-, high-, and extreme-risk groups stratified using the CAPPS score had 32%, 63%, and 85% predicted probability of experiencing poorly controlled pain, respectively, which was mirrored closely by the observed incidence of 37%, 62%, and 81% in the validation cohort. Inadequate pain control is common after spine surgery. The internally validated CAPPS score based on 7 easily acquired variables accurately predicted the probability of experiencing poorly controlled pain after spine surgery.

Prospective validation of a clinical prediction score for survival in patients with spinal metastases: The New England Spinal Metastasis Score

Prospective validation of a clinical prediction score for survival in patients with spinal metastases: The New England Spinal Metastasis Score

Triage tool for suspected COVID-19 patients in the emergency room: AIFELL score

Clinical prediction scores support the assessment of patients in the emergency setting to determine the need for further diagnostic and therapeutic steps. During the current COVID-19 pandemic, physicians in emergency rooms (ER) of many hospitals have a considerably higher patient load and need to decide within a short time frame whom to hospitalize. Based on our clinical experiences in dealing with COVID-19 patients at the University Hospital in Zurich, we created a triage score with the acronym "AIFELL" consisting of clinical, radiological and laboratory findings.The score was then evaluated in a retrospective analysis of 122 consecutive patients with suspected COVID-19 from March until mid-April 2020. Descriptive statistics, Student's t-test, ANOVA and Scheffe's post-hoc analysis confirmed the diagnostic power of the score. The results suggest that the AIFELL score has potential as a triage tool in the ER setting intended to select probable COVID-19 cases for hospitalization in spontaneously presenting or referred patients with acute respiratory symptoms.

Preoperative evaluation and prediction of clinical scores for hepatocellular carcinoma microvascular invasion: a single-center retrospective analysis

IntroductionMicrovascular invasion (MVI) of is generally considered to be an important prognostic factor for hepatocellular carcinoma (HCC) after operation, An accurate prediction of MVI before operation is helpful for clinical decision-making before operation. Material and methodsA retrospective analysis of 227 cases of hepatocellular carcinoma patients after hepatectomy has been confirmed the pathological result whether there was MVI, and has been determined the independent risk factors of MVI. Based on these independent risk factors, we constructed a clinical scoring risk model for predicting MVI. ResultsAmong the 227 patients with HCC, 74 (34.6%) were MVI positive. Using receiver operating characteristic (ROC) curve and logistic regression model, we found that alpha-fetoprotein(AFP)≥158 ng/mL(odds ratio[OR] = 4.152,95% confidence interval [95%CI]:1.602∼10.760,p = 0.003), Des-γ-carboxy prothrombin (DCP)≥178mAU/mL(OR = 9.730,95%CI:3.392∼27.910,p < 0.001), circulating tumor cells (CTCs)≥3/3.2 ml(OR = 7.747,95%CI:3.019∼19.881,P < 0.001), maximum tumor diameter≥59 mm(OR = 3.467,95%CI:1.368∼8.669,p = 0.008) and tumor margin unsmoothness(OR = 0.235,95%CI:0.096∼0.573,p = 0.001) were independent risk factors for MVI, they predicted that the area under the curve of MVI was 0.752, 0.777, 0.857, 0.743 and 0.333, respectively. Based on these five independent risk factors, we constructed a clinical scoring risk model for predicting MVI. The model predicts that the area under the curve of MVI is 0.922, and its prevalence rate from 0 to 5 are 3.1%(1/32), 5.3%(4/76), 12.2%(5/41), 66.7%(20/30), 87.9%(29/33), 100%(15/15), respectively (P < 0.001). ConclusionBased on AFP, DCP, CTC, maximum tumor diameter and tumor margin unsmoothness, we constructed a model to predict the risk of MVI clinical score, so as to make a more accurate individualized treatment plan before operation, which has important clinical significance and application prospect to improve the curative effect of HCC.

Prospective validation of the RAPID clinical risk prediction score in adult patients with pleural infection: the PILOT study

Over 30% of adult patients with pleural infection either die and/or require surgery. There is no robust means of predicting at baseline presentation which patients will suffer a poor clinical outcome. A validated risk prediction score would allow early identification of high-risk patients, potentially directing more aggressive treatment thereafter. To prospectively assess a previously described risk score (the RAPID (Renal (urea), Age, fluid Purulence, Infection source, Dietary (albumin)) score) in adults with pleural infection. Prospective observational cohort study that recruited patients undergoing treatment for pleural infection. RAPID score and risk category were calculated at baseline presentation. The primary outcome was mortality at 3 months; secondary outcomes were mortality at 12 months, length of hospital stay, need for thoracic surgery, failure of medical treatment and lung function at 3 months. Mortality data were available in 542 out of 546 patients recruited (99.3%). Overall mortality was 10% at 3 months (54 out of 542) and 19% at 12 months (102 out of 542). The RAPID risk category predicted mortality at 3 months. Low-risk mortality (RAPID score 0-2): five out of 222 (2.3%, 95% CI 0.9 to 5.7%); medium-risk mortality (RAPID score 3-4): 21 out of 228 (9.2%, 95% CI 6.0 to 13.7%); and high-risk mortality (RAPID score 5-7): 27 out of 92 (29.3%, 95% CI 21.0 to 39.2%). C-statistics for the scores at 3 months and 12 months were 0.78 (95% CI 0.71-0.83) and 0.77 (95% CI 0.72-0.82), respectively. The RAPID score stratifies adults with pleural infection according to increasing risk of mortality and should inform future research directed at improving outcomes in this patient population.

MVPANI: A Toolkit With Friendly Graphical User Interface for Multivariate Pattern Analysis of Neuroimaging Data.

With the rapid development of machine learning techniques, multivariate pattern analysis (MVPA) is becoming increasingly popular in the field of neuroimaging data analysis. Several software packages have been developed to facilitate its application in neuroimaging studies. As most of these software packages are based on command lines, researchers are required to learn how to program, which has greatly limited the use of MVPA for researchers without programming skills. Moreover, lacking a graphical user interface (GUI) also hinders the standardization of the application of MVPA in neuroimaging studies and, consequently, the replication of previous studies or comparisons of results between different studies. Therefore, we developed a GUI-based toolkit for MVPA of neuroimaging data: MVPANI (MVPA for Neuroimaging). Compared with other existing software packages, MVPANI has several advantages. First, MVPANI has a GUI and is, thus, more friendly for non-programmers. Second, MVPANI offers a variety of machine learning algorithms with the flexibility of parameter modification so that researchers can test different algorithms and tune parameters to identify the most suitable algorithms and parameters for their own data. Third, MVPANI also offers the function of data fusion at two levels (feature level or decision level) to utilize complementary information contained in different measures obtained from multimodal neuroimaging techniques. In this paper, we introduce this toolkit and provide four examples to demonstrate its usage, including (1) classification between patients and controls, (2) identification of brain areas containing discriminating information, (3) prediction of clinical scores, and (4) multimodal data fusion.

Clinical Prediction Score for Early Neuroimaging in Acquired Isolated Oculomotor Nerve Palsy.

PurposeAn immediate neuroimaging investigation in patients with isolated oculomotor nerve palsy (ONP) remains controversial. We aimed to develop a clinical prediction score to determine whether or not acquired isolated ONP patients require prompt imaging.MethodsA retrospective study was performed. Demographic data and clinical presentations were collected to determine predictive factors favored for early brain imaging using multivariate logistic regression analysis.ResultsNinety-seven eyes of 96 patients diagnosed with isolated ONP were included. Forty-one eyes (42.3%) were caused by ischemia, while the other 56 eyes (57.7%) were caused by non-ischemic etiologies, namely aneurysm (n = 22), trauma (n = 18), inflammation (n = 5), tumor (n = 4), and others (n = 7). Eighty-two eyes (84.5%) had undergone neuroimaging study due to initially suspected non-ischemic causes. Only 36 (43.9%) revealed concordant diagnosis. The potential clinical predictors favored for neuroimaging using multivariate logistic regression analysis were age 10‒50 years (adjusted odds ratio [aOR] 9.01, 95% CI: 1.25‒64.8), age 51‒70 years (aOR 1.71, 95% CI: 0.46‒6.35), history of head trauma (aOR 7.14, 95% CI: 1.19‒42.9), absence of vascular risk factors (aOR 3.85, 95% CI: 1.23‒12.1), and poor pupillary response (aOR 6.96, 95% CI: 1.99‒24.3). The predictor scores with an area under the ROC curve of 0.852 were 4, 1, 4, 3, and 4, respectively. The optimum cut-point was 3 for a sensitivity of 96% and specificity of 44%.ConclusionPatients with acquired isolated ONP should be considered early neuroimaging studies when they are younger than 50 years old, have a history of head trauma, have no history of vascular risk factors, or have poor pupillary reaction.

Assessing the risk of venous thromboembolism in multiple myeloma

Patients with multiple myeloma (MM) have an increased risk of venous thromboembolism (VTE) compared to the general population. This risk is highest during the first year of diagnosis and subsequently decreases over time. Development of VTE in patients with MM is associated with inferior outcomes, with patients with VTE and MM having an increased risk of death compared to those with MM without VTE. Primary thromboprophylaxis has the potential to decrease risk of VTEksanfilippo@wustl.edu (K.M. Sanfilippo) in MM and improve outcomes. Current studies assessing thromboprophylaxis in MM excluded patients at high risk of VTE. A meta-analysis of trials of primary thromboprophylaxis in ambulatory cancer patients at high risk of VTE identified by use of a risk-prediction score found a reduction in risk of VTE with prophylaxis with no significant increase in risk of major bleeding. However, these trials contained relatively few patients with MM. Three clinical risk prediction scores are available to assess risk of VTE in MM: 1) the International Myeloma Working Group (IMWG)/National Comprehensive Cancer Network (NCCN); 2) the SAVED score; and 3) the IMPEDE VTE score. The latter two have recently been shown to outperform the IMWG/ NCCN score for predicting VTE in MM. Biomarkers have the potential to improve prediction of VTE in patients with MM. Future research should focus on the addition of biomarkers to available risk scores in MM to improve discrimination in this high-risk patient population.

0560 The Effect of Smoking on OSA Endotypes: A Retrospective Cohort Study

Abstract Introduction Smoking is a purported risk factor for obstructive sleep apnea (OSA), but the mechanisms through which smoking may cause OSA are largely unclear. Our goal is to assess the effect of smoking on the pathophysiological traits (“endotypes”) underlying OSA. Methods Based on a chart review we are creating a retrospective cohort of consecutive patients who were newly diagnosed with OSA based on an inlab polysomnogram between 1/2016 and 6/2018 and who have a documented smoking status. For each subject we are quantifying the endotypes (e.g. arousal threshold, loop gain, upper airway muscle recruitment) via a validated polysomnography-based algorithm. Additionally, we are estimating the arousal threshold based on a clinical prediction score. We are comparing OSA endotypes (primary outcomes), sleep apnea severity (apnea-hypopnea index, SpO2 nadir) and sleep parameters (e.g. total sleep time, sleep efficiency, sleep stages) in current vs former vs never smokers using Kruskal-Wallis tests (+Dunn’s test for post hoc comparisons). Results To date we have screened 334 of 2,138 subjects and identified 99 eligible subjects (5 current smokers at the time of polysomnography, 37 former smokers, and 57 never smokers). The clinical arousal threshold was similar across groups (P=.69); polysomnography-based endotype measures are pending. Further, there was no significant difference in sleep apnea severity or sleep parameters across groups, except stage N2 which was less in current vs former smokers (median-percentage 48.5 vs 66.3%, P&amp;lt;.05) and less in never vs former smokers (61.6 vs 66.3%, P&amp;lt;.05). Conclusion Overall, former vs never smokers appear to be similar with regards to sleep and sleep apnea parameters. Prevalence of current smokers appears to be low (5%) in our cohort; larger sample size and polysomnography-based endotypes are needed before firm conclusions about the effects of smoking on OSA mechanisms can be reached (data collection continues). Support This study had no specific funding. Christopher Schmickl is supported by NIH T32 grant HL134632.

Disentangling local functional connectivity and its variability as a biomarker for predicting dysfunction in patients with diffuse axonal injury

To investigates and better understand local functional connectivity and its variability, and explore their predictive values of patients with diffuse axonal injury (DAI). We prospectively enrolled 24 DAI patients and well-matched healthy controls (n = 26) receiving resting-state functional magnetic resonance imaging. We first assessed static and dynamic regional homogeneity (ReHo); then, a k-means clustering algorithm were performed to cluster ReHo states; finally, the prediction of clinical scores of the altered static and dynamic ReHo values was conducted using a general linear model. Compared with healthy subjects, the DAI patients showed distributed alteration in static and dynamic ReHo in several regions, and their alteration commonly located in the left cerebellum posterior lobe (P < 0.01 and Gaussian random field correction at P < 0.05). Importantly, a relationship was observed in disease-specific differences in static or dynamic ReHo and the clinical assessment: the static ReHo values of the left inferior parietal lobule predicted Glasgow Coma Scale scores (ρ = 0.462, P = 0.010). Notably, the results suggested that DAI patients make different transitions to some states (states 1 and 2) compared to healthy controls, but these dynamic transitions were not predictive of clinical associations with ReHo variability. Our findings suggested alterations in both static and dynamic ReHo (less temporal variability) in several regions in the DAI patients. The static ReHo values of the left inferior parietal lobule could be useful as a novel predictive indicator to develop objective neuromarker for future artificial intelligence analyses.

Derivation of a clinical predicting rule for obstructive urolithiasis and alternative diagnosis requiring urgent intervention: the CLAD score.

Most patients with renal colic are discharged from the emergency department (ED) after evaluation and pain alleviation. These patients may not require urgent imaging by computed tomography. We derived a clinical prediction score in patients with renal colic to identify those at very low risk for complications and alternative diagnoses requiring urgent intervention. This retrospective chart review was carried out in 2 ED at an urban university hospital from January to December 2015. All patients with a diagnosis of renal colic were included. The primary outcome was an intervention required for renal colic or alternative diagnoses within 7days of ED presentation. A stepwise logistic regression was used to assess factors associated with the outcome. A score was derived as a weighted sum of these predictors and its performance was calculated. The database was submitted to the French National Commissionfor Data Protection and Liberties (CNIL): declaration n°2164898v0 (03/22/2018). 871 with complete data were analyzed. In 94 (11%) patients, an intervention was performed for obstructive urolithiasis or alternative diagnosis. Five factors were most predictive of intervention: age > 45years (1 point), a history of urologic surgery (1 point), pulse > 100bpm (1 point), temperature > 37.8°C (2 point), and urine dipstick negative for blood (1 point), yielding a score of 0-6 points (the Complicated uroLithiasis and Alternative Diagnosis (CLAD) score). The area under the curve of the receiver operating characteristic curve was 0.82 (95% CI 0.77-0.87). We derived a clinical score for renal colic that predicted the presence of obstructive urolithiasis and acute alternative diagnoses requiring intervention.

Role of liver biopsy in the era of clinical prediction scores for “drug-induced liver injury” (DILI): experience of a tertiary referral hospital

The clinical implications of the biopsy findings in cases of drug-induced liver injury (DILI) are not fully elucidated. The aim of this study was to evaluate the histopathological findings of cases diagnosed as DILI and to correlate them with clinical and biochemical findings (such as causality assessment algorithms). We searched our department database for all cases of liver biopsy with findings consistent with toxic liver disease and selected those with a clinical diagnosis of DILI. The causative relationships were established according to Roussel Uclaf Causality Assessment Method (RUCAM). A total of 53 cases of DILI were reviewed, most of them diagnosed in hospitalized patients (83%). The analytical toxicity profile was hepatocellular (R > 5) in 60% of the cases and cholestatic (R < 2) in 26.4% of cases. The group of drugs most implicated was the anti-microbials (18, 34%). The predominant histological patterns were "necroinflammation" (67.9%) and "cholestasis" (28.3%). The hepatocellular biochemical pattern was not associated with the presence of predominantly necroinflammatory findings in the biopsy (p = 0.44), and the biochemical cholestatic pattern was not associated with the presence of predominantly cholestatic findings in the biopsy (p = 0.51). This study supports that a better insight into the pathologic mechanisms associated with DILI should be based on liver biopsy due to the lack of a uniform correlation between clinical and biochemical patterns. Also, a liver biopsy may be used in those cases where clinical suspicion of DILI persists despite a low score on current causality assessment algorithms.

Tuberculous pericardial disease: a focused update on diagnosis, therapy and prevention of complications

Tuberculous pericarditis (TBP) is the most important manifestation of tuberculous heart disease and is still associated with a significant morbidity and mortality in TB endemic areas. The high prevalence of the disorder over the last 3 decades has been fueled by the human immunodeficiency virus/AIDS (HIV/AIDS) pandemic in these areas. The objective of this review is to provide a focused update on developments in the diagnosis and therapy of this condition, prevention of its complications, as well as future novel therapies. The definitive diagnosis of a tuberculous etiology in patients with suspected TBP continues to pose a challenge for clinicians. Clinical prediction scores, although never formally validated have been used with some success. However, they may be prone to both over and underdiagnosis due to lack of pericardial fluid analysis. Recent studies evaluating Xpert MTB/RIF, suggest that this advanced polymerase chain reaction (PCR) based technology does not provide increased accuracy compared to earlier iterations. However a combined two test approach starting with Xpert MTB/RIF followed by either adenosine deaminase (ADA) or interferon gamma (IFN-γ) may provide for significantly enhanced specificity and sensitivity cost permitting. Pericardiocentesis remains the gold standard for managing the compressive pericardial fluid and its adverse hemodynamic sequelae. A four drug anti-TB drug regimen at standard doses and duration is recommended. However recent evidence suggests that these drugs penetrate the pericardium very poorly potentially explaining the high mortality observed particularly in those who are culture positive with a high bacillary load. Constrictive pericarditis is the main long-term complication of TBP and is still a significant cause of heart failure in Sub-Saharan Africa. This is important because access to definitive surgical therapy where TBP is prevalent continues to be low, highlighting the need to develop strategies or interventions to prevent fibrosis and constriction. Recent detailed advanced studies of pericardial fluid in TBP have revealed a strong profibrotic transcriptomic profile, with high amounts of pro-inflammatory cytokines and low levels of the anti-fibrotic tetrapeptide N-Acetyl-Seryl-Aspartyl-Lysyl-Proline (Ac-SDKP). These new insights may explain in part the high propensity to fibrosis associated with the condition and offer hope for the future use of targeted therapy to interrupt pathways and mediators of tissue damage and subsequent maladaptive healing and fibrosis. The value of effective pericardiocentesis in reducing these pro-inflammatory and pro-fibrotic cytokines and peptides in an attempt to prevent pericardial constriction has yet to be established but has generated hypotheses for ongoing and future research.

Underperformance of clinical risk scores in identifying vascular ultrasound-based high cardiovascular risk in systemic lupus erythematosus.

The aim of this study was to assess the performance of eight clinical risk prediction scores to identify individuals with systemic lupus erythematosus (SLE) at high cardiovascular disease (CVD) risk, as defined by the presence of atherosclerotic plaques. CVD risk was estimated in 210 eligible SLE patients without prior CVD or diabetes mellitus (female: 93.3%, mean age: 44.8 ± 12 years) using five generic (Systematic Coronary Risk Evaluation (SCORE), Framingham Risk Score (FRS), Pooled Cohort Risk Equations (ASCVD), Globorisk, Prospective Cardiovascular Münster Study risk calculator (PROCAM)) and three 'SLE-adapted' (modified-SCORE, modified-FRS, QRESEARCH risk estimator, version 3 (QRISK3)) CVD risk scores, as well as ultrasound examination of the carotid and femoral arteries. Calibration, discrimination and classification measures to identify high CVD risk based on the presence of atherosclerotic plaques were assessed for all risk models. CVD risk reclassification was applied for all scores by incorporating ultrasound results. Moderate calibration (p-value range from 0.38 to 0.63) and discrimination (area under the curve 0.73-0.84), and low-to-moderate sensitivity (8.3-71.4%) and classification ability (Matthews correlation coefficient (MCC) 0.25-0.47) were observed for all risk models to identify patients with plaques at any arterial site as high-risk. MCC was improved for modified-FRS versus FRS (0.43 vs 0.36), but not for modified-SCORE versus SCORE (0.25 vs 0.25). Based on plaque presence, CVD risk was upgraded to high-risk in 10%, 16.1%, 20.5%, 21.5%, 24%, 28.2% and 28.6% of cases classified as non-high-risk by QRISK3, modified-FRS, Globorisk, FRS/PROCAM, ASCVD, modified-SCORE and SCORE, respectively. Most of the five generic and three 'SLE-adapted' clinical risk scores underestimated high CVD risk defined by atherosclerotic plaque presence in patients with SLE.

Underperformance of clinical risk scores in identifying vascular ultrasound-based high cardiovascular risk in systemic lupus erythematosus.

The aim of this study was to assess the performance of eight clinical risk prediction scores to identify individuals with systemic lupus erythematosus (SLE) at high cardiovascular disease (CVD) risk, as defined by the presence of atherosclerotic plaques. CVD risk was estimated in 210 eligible SLE patients without prior CVD or diabetes mellitus (female: 93.3%, mean age: 44.8 ± 12 years) using five generic (Systematic Coronary Risk Evaluation (SCORE), Framingham Risk Score (FRS), Pooled Cohort Risk Equations (ASCVD), Globorisk, Prospective Cardiovascular Münster Study risk calculator (PROCAM)) and three 'SLE-adapted' (modified-SCORE, modified-FRS, QRESEARCH risk estimator, version 3 (QRISK3)) CVD risk scores, as well as ultrasound examination of the carotid and femoral arteries. Calibration, discrimination and classification measures to identify high CVD risk based on the presence of atherosclerotic plaques were assessed for all risk models. CVD risk reclassification was applied for all scores by incorporating ultrasound results. Moderate calibration (p-value range from 0.38 to 0.63) and discrimination (area under the curve 0.73-0.84), and low-to-moderate sensitivity (8.3-71.4%) and classification ability (Matthews correlation coefficient (MCC) 0.25-0.47) were observed for all risk models to identify patients with plaques at any arterial site as high-risk. MCC was improved for modified-FRS versus FRS (0.43 vs 0.36), but not for modified-SCORE versus SCORE (0.25 vs 0.25). Based on plaque presence, CVD risk was upgraded to high-risk in 10%, 16.1%, 20.5%, 21.5%, 24%, 28.2% and 28.6% of cases classified as non-high-risk by QRISK3, modified-FRS, Globorisk, FRS/PROCAM, ASCVD, modified-SCORE and SCORE, respectively. Most of the five generic and three 'SLE-adapted' clinical risk scores underestimated high CVD risk defined by atherosclerotic plaque presence in patients with SLE.

SAT-420 A CLINICAL SCORE TO PREDICT MORTALITY IN SEVERE LEPTOSPIROSIS

Leptospirosis is a zoonosis transmitted through contact with water or soil contaminated by urine from animal reservoirs. It is typically a mild acute febrile illness, however about 10% of patients progress to severe disease with a case fatality rate of 5-20%. This study aims to establish a clinical prediction tool that would predict the risk of progression to mortality in patients presenting to health care providers with signs and symptoms of leptospirosis. This is a retrospective case-control study on survivors and non-survivors due to severe leptospirosis admitted at our institution. Logistic regression was performed to identifysignificant predictors of mortality. A scoring system was generated and validated using the same patient population. A total of 383 patients were included in the study, 327 survivors and 56 (14.6%) non-survivors. Significant factors associated with mortality include history of hypertension (OR=3.72, 95%CI=1.16-8.91), diabetes (OR=15.46, 95%CI=4.58-52.21), presence of oliguria (OR=3.43, 95%CI=1.75-6.73), hemorrhage (OR=2.61, 95%CI=1.22-5.61), hypotension (OR=2.27, 95%CI=1.20-4.33), respiratory rate >20cpm (OR=1.87, 95%CI=1.06-3.31), crackles (OR=4.02, 95%CI=0.74-2.04), edema (OR=4.79, 95%CI=1.83-12.50), thrombocytopenia <100x103/uL (OR=3.80, 95%CI=2.11-6.83), prolonged partial thromboplastin time >45 seconds (OR=2.94, 95%CI=1.61-5.38), acidosis pH<7.30 (OR=5.11, 95%CI=2.60-10.05), atrial fibrillation (OR=11.62, 95%CI=4.48-30.14), and radiographic pulmonary congestion/edema (OR=3.78, 95%CI=1.82-7.81). A weighted scoring system was assigned and four risk categories of increasing severity and percent mortality was formed and can be used as an adjunct to guide clinical decision-making and prioritization of resources. Patients with score 0-2, mortality risk 2.42%, can be discharged with advice, or admit to a primary hospital. Patients with score 3-5, mortality risk 19.51%, can be referred or admitted to a tertiary hospital with an interim leptospirosis ward. Patients with a score of 6-8 and mortality risk 42.86%, recommend transfer to a more equipped ward with close monitoring for occurrence of complications, and possible transfer to the Intensive Care Unit (ICU). Lastly, patients with score >9 have a high mortality risk (81.82%), thus transfer to ICU. In conclusion, a clinical prediction score for mortality in severe leptospirosis is developed from the major prognostic risk factors, which may prove useful in triaging patients in the event of another leptospirosis outbreak.

Clinical predictive score of predischarge screening for severe hyperbilirubinemia in late preterm and term infants

Early identification and appropriate follow-up of infants at risk of severe hyperbilirubinemia are part of preventing complications. This study aims to develop the clinical predictive score to predict subsequent severe hyperbilirubinemia in healthy Thai infants. A case-control study was conducted using medical records of 147 hyperbilirubinemia cases and 147 age-matched controls among healthy late preterm and term Thai newborn infants during January 2015 and December 2016. The routinely measured TcB values at 48-54 hours of age and all predischarge clinical characteristics were collected. Multivariable logistic regression was used to find a clinical prediction model to predict subsequent severe hyperbilirubinemia within 7 days after birth which defined as a postdischarge bilirubin level exceeding the hour-specific recommended threshold for phototherapy by the American Academy of Pediatrics (AAP). The best clinical predictors for subsequent severe hyperbilirubinemia were TcB values at 48-54 hours and gestational age of infants. Predischarge TcB at 48-54 hours of life was classified into 3 levels: < 10 mg/dL, 10-12 mg/dL and > 12 mg/dL. Gestational age was categorized into 5 groups. The risk score derived from these 2 significant factors predicted subsequent severe hyperbilirubinemia with an AuROC curve of 81.0% (95% CI: 76.2-85.9%) in 3 risk group, high, moderate and low. The positive likelihood ratio for subsequent severe hyperbilirubinemia of the high-risk group (score > 5) was 4.53 (95% CI: 2.91-7.04) with specificity of 87.1%. The negative predictive value of low-risk group (score < 3) was 81%. A simple predischarge prediction score using gestational age and TcB values at 48-54 hours of life was developed. This score classified late preterm and term newborn infants into 3 distinct risk levels and may be useful to identify high-risk infants for outpatient follow-up of subsequent severe hyperbilirubinemia.

Prospective validation of a clinical prediction score for survival in patients with spinal metastases: the New England Spinal Metastasis Score

Background ContextThe New England Spinal Metastasis Score (NESMS) was proposed as an intuitive and accessible prognostic tool for predicting survival in patients with spinal metastases. We designed an appropriately powered, prospective, longitudinal investigation to validate the NESMS. PurposeTo prospectively validate the NESMS. Study DesignProspective longitudinal observational cohort study. Patient SamplePatients, aged 18 and older, presenting for treatment with spinal metastatic disease. Outcome MeasuresOne-year mortality (primary); 6-month mortality and mortality at any time point following enrollment (secondary). MethodsThe date of enrollment was set as time zero for all patients. The NESMS was assigned based on data collected at the time of enrollment. Patients were prospectively followed to one of two predetermined end-points: death, or survival at 365 days following enrollment. Survival was visually assessed with Kaplan-Meier curves and then analyzed using multivariable logistic regression, followed by Bayesian regression to assess for robustness of point estimates and 95% confidence intervals (CI). ResultsThis study included 180 patients enrolled between 2017 and 2018. Mortality within 1-year occurred in 56% of the cohort. Using NESMS 3 as the referent, those with a score of 2 had significantly greater odds of mortality (odds ratio 7.04; 95% CI 2.47, 20.08), as did those with a score of 1 (odds ratio 31.30; 95% CI 8.82, 111.04). A NESMS score of 0 was associated with perfect prediction, as 100% of individuals with this score were deceased at 1-year. Similar determinations were encountered for mortality at 6-months and overall. ConclusionsThis study validates the NESMS and demonstrates its utility in prognosticating survival for patients with spinal metastatic disease, irrespective of selected treatment strategy. This is the first study to prospectively validate a prognostic utility for patients with spinal metastases. The NESMS can be directly applied to patient care, hospital-based practice and health-care policy.

Thrombocytopenia as a Predictor of Severity of Pneumonia in a Tertiary Care Center in Nepal

Introduction: Community-acquired pneumonia is an infectious disease with a global burden. Thrombocytopenia can be an effective and inexpensive tool in predicting the severity of pneumonia which can be useful in the primary care level. The objectives were to identify the clinical and laboratory predictors of hospitalized patients with pneumonia, determine the incidence of thrombocytopenia in those patients, its association with severity, compare it with leucocyte count, assess severity using clinical prediction scores, need of mechanical ventilation, need of transfer to intensive care unit and mortality.&#x0D; Methods: After informed consent, diagnosed and admitted cases of community-acquired pneumonia were examined after which patients' demographic and clinical data as well as laboratory parameters including white cell count and platelets were filled in proforma. Clinical prediction scores for pneumonia were calculated and the need for mechanical ventilation, inotropes, transfusions as well as mortality assessed. Patients were followed up until discharge to see the severity of pneumonia.&#x0D; Results: The incidence of thrombocytopenia among patients with pneumonia was 36.9% (41 out of 111). The overall mortality rate was 9%. More pneumonia patients having thrombocytopenia (14.6%) died than pneumonia patients having leukocytosis (13.8%). There was a significant association between thrombocytopenia and severity of pneumonia, both alone, (p = 0.007, OR = 4.03) and on adjusting for the WBC and GCS (p=0.017, OR = 3.81).Significant associations were found between thrombocytopenia and the need for ICU in patients with pneumonia (p=0.001), mechanical ventilation (p = 0.001), presence of edema, confusion, age variation (p=0.024) and septic shock (p=0.021).&#x0D; Conclusion: Thrombocytopenia can serve as an effective predictor of the severity of pneumonia.