SAEM Annual Meeting Chicago

Presentation Number:019

Background: In patients experiencing acute coronary syndrome (ACS), prompt diagnosis is critical in achieving the best health outcome. While ECG analysis is usually sufficient to diagnose ACS in cases of ST elevation, ACS without ST elevation is reliably diagnosed through serial testing of cardiac troponin I (cTnI). Point-of-care testing (POCT) for cTnI by venipuncture has been proven a more rapid means to diagnosis than central laboratory testing. Implementing fingerstick testing for cTnI in place of standard venipuncture methods would allow for faster and easier procurement of patients’ cTnI levels, as well as increase the likelihood of starting a rapid test for cTnI in the pre-hospital setting, which could allow for even earlier diagnosis of ACS.
Objectives: To determine if fingerstick blood samples yield accurate and reliable troponin measurements compared to conventional venous blood draws using the i-STAT POC device.
Methods: This experimental study was performed in the ED of a quaternary care suburban medical center between June-August 2011. Fingerstick blood samples were obtained from adult ED patients for whom standard (venipuncture) POC troponin testing was ordered. The time between fingerstick and standard draws was kept as narrow as possible. cTnI assays were performed at the bedside using the i-STAT 1 (Abbott Point of Care).
Results: 94 samples from 87 patients were analyzed by both fingerstick and standard ED POCT methods (see Table 1). Four resulted in cartridge error. Compared to “gold standard” ED POCT, fingerstick testing has a positive predictive value of 100%, negative predictive value of 96%, sensitivity of 79%, and specificity of 100%. No significant difference in cTnI level was found between the two methods, with a nonparametric intraclass correlation coefficient of 0.994 (95% CI 0.992-0.996, p-value < 0.001).
Conclusion: Whole blood fingerstick cTnI testing using the i-STAT device is suitable for rapid evaluation of cTnI level in pre-hospital and ED settings. However, results must be interpreted with caution if they are within a narrow territory of the cutoff for normal vs. elevated levels. Additional testing on a larger sample would be beneficial. The practicality and clinical benefit of using fingerstick cTnI testing in the EMS setting must still be assessed.

Presentation Number:020

Background: Adjudication (Adj) of Diagnosis (Dx) of acute myocardial infarction (AMI) in clinical studies typically occurs at each site of subject enrollment (local) or by experts at an independent site (central). From 2000-2007, the troponin (cTn) element of the Dx was predicated on the local laboratories, using a mix of the 99th percentile reference cTn and ROC-determined cutpoints. In 2007, the Universal definition of AMI (UDAMI) defined it by the 99th percentile reference alone.
Objectives: To compare the Dx rates of AMI as determined by local Adj vs. central Adj using UDAMI criteria.
Methods: Retrospective analysis of data from the Myeloperoxidase in the Dx of Acute Coronary Syndromes (ACS) Study (MIDAS), an 18 center prospective study with enrollment from 12/19/06 to 9/20/07 of patients with suspected ACS presenting to the ED < 8 hours after symptom onset and in whom serial cTn and objective cardiac perfusion testing was planned. Adj of ACS was done by single local principal investigators using clinical data and local cTn cutpoints from 13 different cTn assays, and applying the 2000 definition. Central Adj was done after completion of the MIDAS primary analysis using the same data and local cTn assay, but by experts at 3 different institutions, using the UDAMI and the manufacturer’s 99th percentile cTn cutpoint, and not blinded to local adjudications. Discrepant Dxs were resolved by consensus. Local vs. central cTn cutpoints differed for 6 assays, with central cutpoints lower in all. Statistics were by Chi-square and Kappa.
Results: Excluding 11 cases deemed indeterminate by central Adj, 1096 cases were successfully adjudicated. Local Adj resulted in 104 AMI (9.5% of total) and 992 non-AMI; central Adj resulted in 134 (12.2%) AMI and 962 non-AMI. Overall, 44 local Dxs (4%) were either changed from non-AMI to AMI or AMI to non-AMI (p < .001). Interrater reliability across both methods was found to be Kappa = .79 (p < .001). For ACS diagnosis, local Adj identified 252 ACS cases (23%) and 854 non-ACS, while central Adj identified 275 ACS (25%) and 831 non-ACS. Overall, 61 local Dxs (6%) were either changed from non-ACS to ACS or ACS to non-ACS (p < .001). Interrater reliability found Kappa = 0.85 (p < .001).
Conclusion: Central and local adjudication resulted in significantly different rates of AMI and ACS diagnosis. However, overall agreement of the two methods across these two diagnoses was acceptable.

Presentation Number:021

Background: In ED patients presenting with chest pain, acute coronary syndrome (ACS) was found to occur four times more often in cocaine users. Biomarkers myeloperoxidase (MPO) and C-reactive protein (CRP) have potential in the diagnosis of ACS.
Objectives: To evaluate the utility of MPO and CRP in the diagnosis of ACS in patients presenting to the ED with cocaine-associated chest pain and compare the predictive value to nonusers. We hypothesized that these markers may be more sensitive for ACS in nonusers given the underlying pathophysiology of enhanced plaque inflammation.
Methods: A secondary analysis of a cohort study of enrolled ED patients who received evaluation for ACS at an urban, tertiary care hospital. Structured data collection at presentation included demographics, chest pain history, lab, and ECG data. Subjects included those with self-reported or lab confirmed cocaine use and chest pain. They were matched to controls based on age, gender, and race. Our main outcome was diagnosis of ACS at index visit. We determined median MPO and CRP values, calculated maximal AUC for ROC curves and found cut-points to maximize sensitivity and specificity. Data are presented with 95% CI.
Results: Overall, 95 patients in the cocaine positive group and 86 patients in the nonusers group had MPO and CRP levels measured. Patients had a median age of 47 (IQR, 40-52), 90% Black and 62% Male (p >0.05 between groups). Fifteen patients were diagnosed with ACS; 8 patients in the cocaine group and 7 in the nonusers group.
Comparing cocaine users to nonusers, there was no difference in MPO (median 162 [IQR, 101-253] v 136 [111-235] ng/mL; p= 0.78) or CRP (3 [1-9] v 5 [1-15] mg/L; p=0.08). The AUC for MPO was 0.65 (95% CI 0.39-0.90) v 0.54 (95% CI 0.19-0.73). The optimal cut point to maximize sensitivity and specificity was 242 ng/mL which gave a sensitivity of 0.42 and specificity of 0.75. Using this cutpoint, 57% v 29% of ACS in cocaine users vs the nonusers would be identified. The AUC for CRP was 0.63 (95% CI 0.39-0.88) in cocaine users v 0.73 (95% CI 0.52-0.95) in nonusers. The optimal cut point was 11.9 mg/L with a sensitivity of 0.67 and specificity of 0.79. Using this cutpoint, 43% v 88% of ACS in cocaine users and nonusers would have been identified.
Conclusion: The diagnostic accuracy of MPO and CRP is not different in cocaine users than nonusers and does not appear to have sufficient discriminatory ability in either cohort.

Presentation Number:022

Background: Pulmonary embolism (PE) increases pulmonary vascular hypertension and can cause right ventricular (RV) injury by shear forces, stretch, work and inflammatory responses. Our recent studies show that the vasodilation observed following in vivo treatment with BAY 41-8543 reduces pulmonary vascular resistance in a 5hr model of acute experimental PE.
Objectives: To test if treatment of rats with BAY 41-8543 protects against RV dysfunction in two models of acute experimental PE.
Methods: Experimental PE was induced in anesthetized, male Sprague-Dawley rats by infusing 25 um polystyrene microspheres in the right jugular vein to produce RV injury from severe PE (2.6 million/100 gm body wt, 5 hrs) or moderate PE (2.0 million/100 gm body wt, 18hrs). Rats with PE were treated with BAY 41-8543 (50 ug/kg, IV) or its solvent at the time of PE induction. Control rats received vehicle for microspheres. Hearts were perfused by Langendorff technique to assess RV function. Values are mean ± se, n = 7 / group. Comparisons between control values, PE and PE + BAY 41-8543 were made by ANOVA followed by Neuman-Keuls test (Significance = p < 0.05).
Results: 18 hrs of moderate PE caused a significant decrease in RV heart function in rats treated with the solvent for BAY 41-8543: peak systolic pressure (PSP) decreased from 39 ± 1.5 mmHg, Control to16 ± 1.5, PE, +dP/dt decreased from 1192 ± 93 mmHg/sec to 463 ± 77, -dP/dt decreased from -576 ± 60 mmHg/sec to -251 ± 9. Treatment of rats with BAY 41-8543 significantly improved all three indices of RV heart function (PSP 29 ± 2.6, +dP/dt 1109 ± 116, -dP/dt -426 ± 69). 5hrs of severe PE also caused significant RV dysfunction (PSP 25 ± 2, -dP/dt -356 ± 28) and treatment with BAY 41-8543 produced protection of RV heart function (PSP 34 ± 2, -dP/dt -535 ± 41) similar to the 18 hr moderate PE model.
Conclusion: Experimental PE produced significant RV dysfunction, which was ameliorated by treatment of the animals with the soluble guanylate cyclase stimulator, BAY 41-8543.

Presentation Number:023

Background: The Thrombolysis in Myocardial Infarction (TIMI) score has been validated as a risk stratification tool in the emergency department (ED) setting, but certain aspects of the scoring system may not be applicable when applied to chest pain patients selected for ED observation unit (EDOU) stay.
Objectives: We evaluated a simplified, 3-point risk stratification tool for EDOU patients which we termed the CARdiac score: Coronary disease [previous myocardial infarction (MI), stent, or coronary artery bypass graft (CABG)], Age (65 years or older), and Risk factors (at least 3 of 5 cardiac risk factors).
Methods: We performed a prospective, observational study with 30-day phone follow-up for all chest pain patients admitted to our EDOU over a one-year period. Baseline data, outcomes related to EDOU stay, inpatient admission, and 30-day outcomes were recorded. CARdiac scores were calculated based on patient history and were used to evaluate the risk of the composite outcome of MI, stent/CABG, or death during the EDOU stay, inpatient admission, or 30 day follow-up period. The CARdiac score was not utilized during the EDOU stay and was calculated with blinding to patient outcomes.
Results: 552 patients were evaluated. Average age was 54.1 years (19-80yrs) and 46% were male. Eighteen patients experienced composite outcomes: stent (12), CABG (3), MI and stent (2), and MI and CABG (1). Risk of the composite outcome generally increased by CARdiac score: 0 (1.6%), 1 (3.6%), 2 (9%), and 3 (8.3%).Patients with a CARdiac score of 2 or 3 (moderate risk) were significantly more likely to experience MI, stent, or CABG than those with a score of 0 or 1 (low risk): 7/79 moderate-risk patients (8.9%) had the composite outcome vs. 11/473 low-risk patients (2.3%, p=0.008, relative risk=3.9). Those moderate-risk by the CARdiac score were also more likely to require inpatient admission from the EDOU (16.5% vs. 9.1%, p=0.045).
Conclusion: The CARdiac score may prove to be a simple tool for risk stratification of chest pain patients in an EDOU. Patients moderate-risk by CARdiac score may be appropriate for more intensive evaluation in the EDOU or consideration for inpatient admission rather than EDOU placement.

Presentation Number:024

Background: Patients who present to the ED with symptoms of potential acute coronary syndrome (ACS) can be safely discharged home after a negative coronary computerized tomographic angiography (CTA). However, the duration of time for which a negative coronary CTA can be used to inform decision making when patients have recurrent symptoms is unknown.
Objectives: We examined patients who received more than one coronary CTA for evaluation of ACS to determine whether they had disease progression, as defined by crossing the threshold from noncritical (< 50% maximal stenosis) to potentially critical disease.
Methods: We performed a structured comprehensive record search of all coronary CTAs performed from 2005 to 2010 at a tertiary care health system. Low-to-intermediate risk ED patients who received 2 or more coronary CTAs; at least one from an ED evaluation for potential ACS were identified. Patients who were revascularized between scans were excluded. We collected demographic data, clinical course, time between scans, and number of ED visits between scans. Record review was structured and done by trained abstractors. Our main outcome was progression of coronary stenosis between scans, specifically crossing the threshold from noncritical to potentially critical disease.
Results: Overall, 32 patients met study criteria (median age 45, interquartile range [IQR] (37.5-48); 56% female; 88% Black). The median time between studies was 27.3 months (IQR, 18.2-33.2). 22 patients did not have stenosis in any vessel on either coronary CTA, 2 studies showed increasing stenosis of <20%, and the rest showed “improvement,” most due to better imaging quality. No patient initially below the 50% threshold subsequently exceeded it (0%; 95% CI, 0-11.0%). Patients also had varying number of ED visits (median number of visits, 5 [range 0-23]), number of ED visits for potentially cardiac complaints (median 1, range 0-6); 10 were re-admitted for potentially cardiac complaints (for example, chest pain or shortness of breath), and 9 received further provocative cardiac testing, all of which had negative results.
Conclusion: We did not find clinically significant disease progression within a 2 year time frame in patients who had a negative coronary CTA, despite a high number of repeat visits. This suggests that prior negative coronary CTA may be able to be used to inform decision making within this time period.

Presentation Number:025

Background: “Triple rule out” (TRO) CT has emerged as a new technology to evaluate for coronary artery disease (CAD), pulmonary embolism (PE), and aortic dissection (AD) in a single imaging modality.
Objectives: We compared the diagnostic accuracy, image quality, radiation exposure, contrast volume, length of stay, and admission rate of TRO CT to other diagnostic modalities (dedicated coronary, PE, and AD CT; coronary angiography; and nuclear stress testing) for the evaluation of chest pain.
Methods: With the assistance of an expert librarian we searched four electronic databases, reference lists of included studies, and contacted content experts to identify articles for review. Included articles met all the following criteria: enrolled patients with non-traumatic chest pain, shortness of breath, suspected ACS, PE, or AD; used 64-slice CT technology; and compared TRO CT to another diagnostic modality. Statistical comparisons were conducted using RevMan, and Meta-DiSc was used to pool diagnostic accuracy estimates.
Results: Nine ED studies enrolling 1371 patients (483 TRO, 888 non-TRO) were included (1 RCT and 8 observational). Patients undergoing TRO CT were exposed to more radiation (mean difference [MD] 5.03 mSv, 95% CI 4.16-5.91) and more contrast (MD 45.6 mL, 95% CI 42.7-48.6) compared to non TRO CT patients. There was no significant difference in image quality between TRO CT images and those of dedicated CT scans in any studies performing this comparison. Similarly, there was no significant difference between TRO CT and other diagnostic modalities in regards to length of stay or admission rate. When compared to conventional coronary angiography as the gold standard for evaluation of CAD, TRO CT had the following pooled diagnostic accuracy estimates: sensitivity 0.94 [95%CI 0.89-0.98], specificity 0.98 [0.97-0.99], LR+ 35.7 [95%CI 11.1-115.0], and LR- 0.07 [95%CI 0.02-0.35].
Conclusion: TRO Chest CT is comparable to dedicated PE, coronary, or AD CT in regard to image quality, length of stay, and admission rate and is highly accurate for detecting CAD. The utility of TRO CT depends on the relative pre-test probabilities of the conditions being assessed and its role is yet to be clearly defined. TRO CT, however, involves increased radiation exposure and contrast volume and for this reason clinicians should be selective in its use.

Presentation Number:026

Background: Coronary computed tomographic angiography (CCTA) has high sensitivity, specificity, accuracy and prognostic value for coronary artery disease (CAD) and ACS. However, how a CCTA informs subsequent use of prescription medication is unclear.
Objectives: To determine if detection of critical or noncritical CAD on CCTA is associated with initiation of aspirin and statins for patients who presented to the ED with chest pain. We hypothesized that aspirin and statins would be more likely to be prescribed to patients with noncritical disease relative to those without any CAD.
Methods: Prospective cohort study of patients who received CCTA as part of evaluation of chest pain in the ED or observation unit. Patients were contacted and medical records were reviewed to obtain clinical follow up for up to the year after CCTA. The main outcome was new prescription of aspirin or statin. CAD severity on CCTA was graded as absent, mild (1% to 49%), moderate (50% to 69%), or severe (≥70%) stenosis.. Logistic regression was used to assess the association of stenosis severity to new medication prescription; covariates were determined a priori.
Results: 859 patients who had CCTA performed consented to participate in this study or met waiver of consent for record review only (median age 48.5 IQR 42.7-53.4, 59% female, 71% Black). Median follow up time was 333 days, IQR: 70-725 days. At baseline, 13% of the total cohort was already prescribed aspirin and 8% on statin medication. Two hundred seventy nine (32%) patients were found to have stenosis in at least one vessel. In patients with absent, mild, moderate, and severe CAD on CCTA, aspirin was initiated in 11%, 34%, 52%, and 55%; statins were initiated 7%, 22%, 32%, 53% of patients. After adjustment for age, race, gender, hypertension, diabetes, cholesterol, tobacco use, and admission to the hospital after CCTA, higher grades of CAD severity were independently associated with greater post-CCTA use of aspirin (OR 1.9 per grade, 95% CI 1.4-2.2, p <0.001) and statins (OR 1.9, 95% CI 1.5 -2.4, p <0.001).

Conclusion: In conclusion, greater CAD severity on CCTA is associated with increased medication prescription for CAD. Patients with noncritical disease are more likely than patients without any disease to receive aspirin and statins. Future studies should examine whether these changes lead to decreased hospitalizations and improved cardiovascular health.

Presentation Number:027

Background: Hess et al. developed a clinical decision rule for patients with acute chest pain consisting of the absence of 5 predictors: ischemic ECG changes not known to be old, elevated initial or 6-hour troponin level, known coronary disease, “typical” pain, and age over 50. Patients less than 40 required only a single troponin evaluation.
Objectives: To test the hypothesis that patients less than 40 years old without these criteria are at <1% risk for major adverse cardiovascular events (MACE) including death, AMI, PCI, and CABG.
Methods: We performed a secondary analysis of several combined prospective cohort studies that enrolled ED patients who received an evaluation for ACS in an urban ED from 1999 to 2009. Cocaine users and STEMI were excluded. Structured data collection at presentation included demographics, pain description, history, lab, and ECG data for all studies. Hospital course was followed daily. 30 day follow up was done by telephone. Our main outcome was 30 day MACE using objective criteria. The secondary outcome was potential change in ED disposition due to application of the rule. Descriptive statistics and 95% CI’s were used.
Results: Of 9289 visits for potential ACS, patients had a mean age of 52.4 +/- 14.7 yrs; 68% black and 59% female. There were 638 patients (6.9%) with 30 day CV events (93 dead, 384 AMI, 298 PCI). Sequential removal of patients in order to meet the final rule for patients less than 40 excluded patients based upon: ischemic ECG changes not old (n=434, 30% MACE rate), elevated initial troponin level (n=237, 60% MACE), known coronary disease (n=1622, 11% MACE), “typical” pain (n=3179, 3% MACE), and age over 40 (n=2690, 3.4% MACE) leaving 1127 patients less than 40 with 0.8% MACE [95% CI, 0.4-1.5%]. Of this cohort, 70% were discharged home from the ED by the treating physician without application of this rule. Adding a second negative troponin in patients 40-50 years old identified a group of 1139 patients with a 2.0% rate of MACE [1.3-3.0] and a 48% discharge rate.
Conclusion: The Hess rule appears to identify a cohort of patients at approximately 1% risk of 30 day MACE, and may enhance discharge of young patients. However, even without application of this rule the 70% of young patients at low risk are already being discharged home based upon clinical judgment.