Basic critical care echocardiography (BCCE) can be done

Transcription

1 Basic Critical Care Echocardiography by Pulmonary Fellows: Learning Trajectory and Prognostic Impact Using a Minimally Resourced Training Model* Kay Choong See, MRCP 1,2 ; Venetia Ong, BN 1,2 ; Jeffrey Ng, MRCP 1,2 ; Rou An Tan, MRCP 1,2 ; Jason Phua, MRCP 1,2 Objectives: The spread of basic critical care echocardiography may be limited by training resources. Another barrier is the lack of information about the learning trajectory and prognostic impact of individual basic critical care echocardiography domains like acute cor pulmonale determination and left ventricular function estimation. We thus developed a minimally resourced training model and studied the latter outcomes. Design: Prospective observational study. Setting: Twenty-bed medical ICU. Subjects: Echocardiography-naive trainees enrolled in the first year of our Pulmonary Medicine Fellowship Program from September 2012 to September Interventions: We described the learning trajectory in six basic critical care echocardiography domains (adequate views, pericardial effusion, acute cor pulmonale, left ventricular ejection fraction, mitral regurgitation, and inferior vena cava variability) and correlated abnormalities in selected basic critical care echocardiography domains with clinical outcomes (mortality and length of stay). Measurements and Main Results: Three-hundred forty-three basic critical care echocardiography scans were done for 318 patients by seven fellows (median of 40 scans per fellow; range, ). Only *See also p Division of Respiratory and Critical Care Medicine, University Medicine Cluster, National University Health System, Singapore. 2 Department of Medicine, Yong Loo Lin School of Medicine, National University of Singapore, Singapore. All the authors jointly conceived the study and prepared the manuscript. Dr. See, Ms. Ong, Dr. Ng, and Dr. Tan performed the data extraction. Dr. See performed the data analysis. Dr. Phua supervised the analysis and edited the article. Dr. See had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. This work was performed at the National University Health System, Singapore. The authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, kay_choong_see@nuhs.edu.sg Copyright 2014 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins DOI: /CCM one-third patients had normal basic critical care echocardiography studies. Accuracy in various basic critical care echocardiography domains was high (> 90%), especially beyond the first 30 examinations. Acute cor pulmonale was associated with ICU mortality when adjusted for Acute Physiology and Chronic Health Evaluation II score and presence of sepsis, whereas mitral regurgitation was associated with longer hospitalization only on univariate analysis. Conclusions: Basic critical care echocardiography training using minimal resources is feasible. New trainees can achieve reasonable competency in most basic critical care echocardiography domains after performing about 30 examinations within the first year. The relatively high prevalence of abnormalities and the significant association of acute cor pulmonale with ICU mortality support the need for basic critical care echocardiography training. (Crit Care Med 2014; 42: ) Key Words: competency; cor pulmonale; echocardiography; education; intensive care units Basic critical care echocardiography (BCCE) can be done by noncardiologist intensivists and has revolutionized the care of critically ill patients by allowing rapid noninvasive assessment of hemodynamics, cardiac function, valvular status, and fluid status (1 6). However, the spread of echocardiography skills to pulmonologists working in intensive care requires overcoming the resource barriers involved in achieving up to studies over 3 6 months required by professional cardiology and echocardiography organizations or in purchasing echocardiography simulators to complement experience with real patients (1, 7). An international expert group recommended at least 30 studies to achieve global competency in BCCE (3). However, the learning trajectory and prognostic impact of individual BCCE domains like acute cor pulmonale determination and left ventricular (LV) function estimation remain unclear (4, 8 10). In our hospital, the training system for pulmonary medicine fellows (postgraduate year 4 6 trainees) recently underwent sweeping changes due to the introduction of the Accreditation Council of Graduate Medical Education International Critical Care Medicine

2 See et al program. We took this opportunity to bring BCCE into the first-year curriculum of our Pulmonary Medicine Fellowship Program. At the same time, we had only two attending pulmonologists who were trained in critical care echocardiography (K.C.S., J.P.), no additional protected teaching time for faculty, and no resources to develop our own training materials. We thus explored a training model using freely available Internetbased BCCE education platforms, a process of predominantly self-directed learning, and indirect supervision with assessment of recorded echocardiographic views. We aimed to show that BCCE training can be accomplished within a relatively short time, which would help in the dissemination of this technology to other fellowship programs globally. We postulated that competencies in some domains were more easily achieved than in others. Furthermore, we were uncertain if the findings of these domains had any prognostic significance, since such information would help determine if these domains were clinically useful for inclusion in our training program. As such, we 1) described the learning trajectory in various BCCE domains and 2) correlated abnormalities in selected BCCE domains with clinical outcomes, when BCCE was done by echocardiography-naive trainees in the first year of our Pulmonary Medicine Fellowship Program. METHODS Participants and Setting Our ethics review board permitted waiver of informed consent (DSRB B/2013/00132). We performed a prospective observational investigation of all BCCE scans done by echocardiography-naive Pulmonary Medicine trainees enrolled in our Fellowship Program. All trainees had undergone at least 3 years of training in general internal medicine and were accredited to insert central venous catheters under ultrasound guidance. In other words, they were already familiar with basic ultrasound physics and knobology. However, they had no prior formal or informal training in transthoracic or transesophageal echocardiography, and this was not part of the general internal medicine curriculum. The BCCE scans were done within hours for patients admitted or readmitted to our 20-bed medical ICU from September 2012 to September Patients had to be available during the echocardiography training slots, and we excluded three patients who had surgical dressings over the chest or abdomen which would preclude satisfactory acoustic windows (two with left-sided chest tubes and large dressings and one with a recent laparotomy). No other selection criteria were used. Trainees entered all BCCE findings and scanning times (for image acquisition and interpretation) into the ICU computerized database. To ensure patient safety, we emphasized that trainees preliminary reports must be validated by K.C.S. or J.P. before being used for clinical decision making. Scanning Procedure BCCE was performed using the Sparq Ultrasound System (Philips Healthcare, Andover, MA) equipped with a 2 4 MHz broadband sector, phased array transducer. At least, seven standard views (acoustic windows) were obtained and recorded for each BCCE scan (Table 1): parasternal long axis, parasternal short axis, apical four-chamber (three views), subcostal, and inferior vena cava (IVC). Assessment was done specifically for six domains (Table 2): acoustic windows, pericardial effusion, right ventricle (11, 12), LV (13, 14), mitral regurgitation, and IVC. For each acoustic window to be adequate, the image must be in the correct plane and orientation, all relevant structures must be visible, and endocardial definition must allow for assessment of structure and motion (15). We assessed for mitral regurgitation in our protocol as its presence would temper the correlation of visual LV ejection fraction (LVEF) and effective LV output. Quantitative evaluation was only done for two aspects: 1) A modified Simpson s measurement of the LVEF was done to corroborate the visual LVEF, but this did not contribute to the classification of LV function based on the latter. 2) For patients who were spontaneously breathing or who were on continuous positive airway pressure via face mask, IVC collapsibility was measured as DeltaIVC/Maximum IVC diameter (the IVC was collapsible if collapsibility was > 40%), where DeltaIVC = Maximum IVC diameter Minimum IVC diameter (16). For patients who were perfectly adapted on invasive positive pressure mechanical ventilation, IVC distensibility was measured as DeltaIVC/Minimum IVC diameter (the IVC was distensible if distensibility was > 18%) (17). In the absence of prior data, we used the same distensibility criterion for other patients on invasive or noninvasive positive pressure ventilation. We expected that superimposed spontaneous breaths would increase the sensitivity of the cutoff for fluid responsiveness, which we felt was desirable in order not to miss patients who might benefit from volume expansion. Table 1. Description of the Seven Basic Critical Care Echocardiography Views View 1. Parasternal long axis Cine image Description 2. Parasternal short axis Still image at level of the papillary muscles 3. Apical four-chamber #1 Cine image 4. Apical four-chamber #2 Cine image with color flow Doppler at the mitral valve 5. Apical four-chamber #3 Composite still image with modified Simpson s computation of left ventricular ejection fraction with endocardial tracing in ventricular systole and diastole 6. Subcostal Cine image 7. Inferior vena cava Still image with respirophasic M-mode measurements 2 3 cm away from the right atrium October 2014 Volume 42 Number 10

3 Feature Articles Table 2. Scoring Criteria for Basic Critical Care Echocardiography Domain Acceptable Not Acceptable Acoustic windows Pericardial effusion Right ventricle Left ventricle Mitral regurgitation IVC All seven views (Table 1) are clearly seen and usable for clinical assessment. For each acoustic window to be adequate, the image must be in the correct plane and orientation, all relevant structures must be visible, and endocardial definition must allow for assessment of structure and motion Presence or absence of pericardial effusion correctly noted from the parasternal long axis, apical four-chamber, and subcostal views Acute cor pulmonale correctly identified as right ventricular dilatation with the right-to-left ventricular size ratio 1:1 in end-diastole at the papillary muscle level and interventricular septal straightening/paradoxical motion Based on the combination of the parasternal long axis and the apical four-chamber views, correct classification of the LVEF to one of three categories (normal, i.e., estimated LVEF 50%; mild-moderate dysfunction, i.e., estimated LVEF 30 49%; severe dysfunction, i.e., estimated LVEF < 30%) Presence or absence of mitral valvular regurgitation correctly noted Correct measurement of IVC dimensions using M-mode 2 3 cm away from the right atrium and correct placement of calipers onto the inner wall of the IVC LVEF = left ventricular ejection fraction, IVC = inferior vena cava. One or more views not clearly seen or usable for clinical assessment Pericardial effusion missed Pericardial effusion noted wrongly, e.g., pleural effusion mistaken for pericardial effusion Acute cor pulmonale missed Acute cor pulmonale identified wrongly, i.e., no right ventricular dilatation or no interventricular septal straightening/ paradoxical motion Incorrect classification of the LVEF Mitral regurgitation missed Mitral regurgitation wrongly identified, e.g., color Doppler artifacts mistaken for mitral regurgitation M-mode measurements made < 2 or > 3 cm from the right atrium or incorrect placement of the calipers onto the outer wall of the IVC Training Process and Equipment A theoretical grounding in BCCE was via self-study of a free Internet resource, which we felt was concise yet adequate for our needs ( developed by Christian Medical College Hospital, Vellore, India). This site also provided videos, for example, those demonstrating normal, moderately depressed, and severely depressed LV function, which trainees could reference repeatedly. From the first author s experience, this site required about 10 hours for self-learning. Trainees selfreported that they spent 5 hours on the initial learning from the website and another 5 hours on subsequent website review. Medical knowledge obtained from self-learning was evaluated through direct questioning by K.C.S. before they were allowed to practice BCCE. Trainees were quizzed on the various probe positions, description of cardiac structures seen in each acoustic window, definition of acute cor pulmonale, classification of LVEF, and computation of IVC variability. We provided dedicated time for BCCE practice sessions, about 2 hours per day on weekdays, three times per week, excluding public holidays. Only one fellow was permitted to practice at each session. Each fellow was encouraged to perform about three BCCE scans per day as part of the screening of newly admitted patients. All studies were done without knowledge of the patient s prior echocardiography studies. The initial five studies, each taking minutes, were directly observed by the first author (K.C.S.) and further studies were done independently by the fellows. Subsequently, all BCCE images were recorded in the ultrasound machine and reviewed by either K.C.S. or J.P. later on the same day, who determined the acceptability of various BCCE domains. In concert with the saved images, K.C.S. or J.P. evaluated trainees preliminary reports and corrected any errors before validating these for clinical use. We provided feedback to the fellows predominantly via telephone and short messaging service. The latter method of indirect supervision eliminated the logistic difficulty of coordinating face-to-face training sessions between teacher and trainee and halved the supervision time from 20 minutes to less than 10 minutes per BCCE study. This method further allowed faculty to continue scheduled clinical work. K.C.S. and J.P. were intensivists who both had at least 1 year of daily experience with BCCE performance and interpretation. In addition, K.C.S. and J.P. did five reviews together and readily achieved 100% concordance on diagnostic assessment. We tried as far as possible to use only one supervisor to minimize variability of BCCE interpretation. As such, K.C.S. was the default supervisor, whereas J.P. was the alternate supervisor if K.C.S. was on leave. K.C.S. then reviewed all the assessments done by J.P. at the end of study period to ensure consistency of scoring. Critical Care Medicine

4 See et al Supervisors were given the discretion to repeat all or part of the BCCE should the views acquired by fellows be inadequate for assessment of the various BCCE domains, although this was not always possible due to clinical or logistic reasons. Statistical Analysis Univariate comparisons of proportions, means, and medians were, respectively, done using Fisher exact, Student t, and Wilcoxon rank-sum tests. Findings and proportion of acceptable results in each of the six BCCE domains were evaluated in blocks of 10 BCCEs to even out any fluctuations in performance for each individual, yet leaving enough granularity to demonstrate a trend. Linear regression of scanning duration was done against BCCE case count, adjusting for individual trainees as random effects. Supervisor-adjudicated findings (pericardial effusion, LV function, right ventricular function, and mitral regurgitation) were correlated with mortality using logistic regression, adjusting for Acute Physiology and Chronic Health Evaluation (APACHE) II score and presence of sepsis (pneumonia or nonpneumonia sepsis as a primary diagnosis). Similarly, supervisor-adjudicated findings (pericardial effusion, LV function, right ventricular function, and mitral regurgitation) were correlated with length of stay (log-transformed to achieve normality) using linear regression, adjusting for APACHE II score and presence of sepsis. For the latter two analyses, only the first BCCE for each patient was studied. Statistical significance was taken as p value of less than RESULTS A total of 343 BCCE scans were done for 318 patients (Table 3) by seven fellows (median age of fellows, 32; range, yr old; three men, four women). Each fellow performed a median of 40 cases (range, ), the variability being due to the unbalanced clinical rotations in the first year, with several fellows being outposted intermittently to the general medicine service. This could potentially affect the representativeness of the final block of cases (31 and above). Nonetheless, further cross-tabulation analysis revealed that for each domain, no significant differences in the proportion of acceptable results occurred among fellows (p values ranging from to 0.810). Two fellows had 34 scans, whereas the rest had 38 scans and above. One of the former fellows was outposted subsequently. For the other fellow, although we used 34 cases in the final data analysis, we continued the indirect supervision and validation of her scans for the next 1 month (47 scans done in total). Within this individual, we then compared the proportion of acceptable results in the final block of cases (31 and above) before and after the additional 13 scans and found nearly identical proportions with no significant differences in each of the six BCCE domains (p values ranging from to 1.000) The majority of studies (332; 96.8%) were reviewed by K.C.S. while 11 (3.2%) were reviewed initially by J.P., and the latter were all affirmed by K.C.S. Average scanning time was 21.3 ± 9.5 minutes for all BCCEs and 18.0 ± 7.0 minutes after the first 30 cases. Adjusting for individual trainees as random Table 3. Patient Characteristics and Outcomes (n = 318 Patients) Patient Characteristic or Outcome Results (n = 318) Age (yr) (mean ± sd) 60.8 ± 16.9 Female sex (%) 126 (39.6) Acute Physiology and Chronic Health II score (mean ± sd) 23.5 ± 8.3 Primary diagnosis (%) Pneumonia 86 (27.0) Nonpneumonia sepsis 56 (17.6) COPD exacerbation 20 (6.3) Asthma exacerbation 7 (2.2) Fluid overload 25 (7.9) Stroke 8 (2.5) Seizure disorder 15 (4.7) Other diagnoses 101 (31.8) Comorbidities (%) Diabetes mellitus 117 (36.8) Hypertension 150 (47.2) Ischemic heart disease 70 (22.0) Chronic heart failure 14 (4.4) Asthma 19 (6.0) COPD 17 (5.4) Bronchiectasis 9 (2.8) Chronic renal failure 60 (18.9) Chronic liver disease 15 (4.7) Stroke 32 (10.1) Cancer 44 (13.8) Body weight (kg) (mean ± sd) 62.2 ± 14.5 Heart rate (beats/min) (mean ± sd) 92 ± 22 Mean arterial pressure (mm Hg) 86 ± 17 (mean ± sd) Ventilation modes (%) Nil ventilation 143 (45.0) Continuous positive airway pressure 2 (0.6) Noninvasive ventilation 22 (6.9) Invasive 151 (47.5) On vasoactive agents (%) Any agent a 65 (20.4) Dopamine 13 (4.1) Noradrenaline 52 (16.4) Dobutamine 2 (0.6) Vasopressin 1 (0.3) (Continued) October 2014 Volume 42 Number 10

5 Feature Articles Table 3. (Continued) Patient Characteristics and Outcomes (n = 318 Patients) Patient Characteristic or Outcome Results (n = 318) LOS, ICU (d), median (IQR) 6 (1 61) LOS, hospital (d), median (IQR) 13.5 (2 137) Mortality, ICU (%) 61 (19.2) Mortality, hospital (%) 79 (24.8) COPD = chronic obstructive pulmonary disease, LOS = length stay, IQR = interquartile range. a Patients could be on more than one vasoactive agent. effects, each successive case of BCCE performed decreased scanning duration by 0.14 minutes (95% CI, min; p < 0.001). Scans were done on patients after a median of 3 days of ICU stay (interquartile range, 2 5 d). Patients could have multiple findings, and only one third of patients had completely normal BCCE studies (Table 4). Performance in various BCCE domains significantly improved with increased trainee experience (Table 5). Accuracy was generally high (> 90%), especially after the first 30 patients (Fig. 1). When associations between various abnormal findings (supervisor adjudicated) and outcomes were studied, acute cor pulmonale was significantly associated with ICU mortality on both univariate and multivariate analysis, while the presence of mitral regurgitation was associated with an increased length of hospitalization only on univariate analysis (Tables 6 and 7). DISCUSSION We showed that BCCE training in our Pulmonary Medicine Fellowship Program was both feasible and fruitful, even within the first year of training. When BCCE was carried out by echocardiography-naive trainees, significant improvements in image acquisition and accuracy of various BCCE domains could be achieved when more than 30 BCCEs were done. In particular, trainees achieved more than 90% acceptable views, correct IVC measurements, and diagnostic accuracy for pericardial effusion, right ventricular dilation, and mitral regurgitation. BCCE also uncovered abnormalities in the majority of patients. Particularly important was the finding of acute cor pulmonale, which was significantly associated with increased ICU mortality, adjusting for APACHE II score and the presence of sepsis. Our findings support and extend the results of prior studies. Fellows were able to classify the visual LVEF accurately in 85% of cases when more than 30 BCCEs were done, comparable to 82% in another study of intensivists with limited training (13). Our scanning time was 5 10 minutes longer than previous studies probably due to more views and calculations done by our trainees (7, 8). We had also required the modified Simpson s estimation of LVEF and identification of abnormal or paradoxical interventricular septal motion pattern, which go beyond the scope of BCCE (6, 18), as we felt that these would assist us in clinical assessment. Table 4. Basic Critical Care Echocardiography Supervisor-Adjudicated Findings (n = 343 Examinations) Basic Critical Care Echocardiography Domain Findings (n = 343) Presence of pericardial effusion (%) 68 (19.8) Thickness of pericardial effusion, if present (mm), median (IQR) 5 (1 12) Acute cor pulmonale a (%) 89 (26.0) Left ventricular function b (%) Normal 239 (69.7) Mild-moderate dysfunction 100 (29.2) Severe dysfunction 4 (1.2) Presence of mitral regurgitation (%) 113 (32.9) IVC diameter (mm), c mean ± sd Maximum IVC diameter 18 ± 11 Minimum IVC diameter 15 ± 8 IVC variability index c,d (%), median (IQR) Collapsibility 18 (0 85) Distensibility 19 (0 93) IVC variability c,e (%) Total 114 (35.9) Collapsible 25 (17.4) Distensible 89 (51.2) Normal study f (%) 111 (32.4) Number of patients with 1 abnormal findings g (%) 1 abnormal finding only 119 (34.7) 2 abnormal findings 86 (25.1) 3 abnormal findings 25 (7.3) 4 abnormal findings 2 (0.6) IQR = interquartile range, IVC = inferior vena cava, NA = not applicable. a Right ventricular dilatation with the right-to-left ventricular size ratio 1:1 in end-diastole at the papillary muscle level and interventricular septal straightening/paradoxical motion. b Classification of the visual left ventricular ejection fraction (LVEF) to one of three categories: normal (estimated LVEF 50% or greater), mild-moderate dysfunction (estimated LVEF 30 49%), and severe dysfunction (estimated LVEF < 30%). c n = 318 due to 25 missing values from inadequate acoustic windows. d For patients who were spontaneously breathing or who were on continuous positive airway pressure via face mask, collapsibility was measured as (Maximum IVC diameter Minimum IVC diameter)/maximum IVC diameter 100%. For patients on noninvasive or invasive mechanical ventilation, distensibility was measured as (Maximum IVC diameter Minimum IVC diameter)/minimum IVC diameter 100%. e IVC variability was present if the IVC had a collapsibility index of > 40% for patients who were spontaneously breathing or who were on continuous positive airway pressure via face mask or a distensibility index of > 18% for patients on noninvasive or invasive mechanical ventilation. f Normal study requires absent pericardial effusion, nondilated right ventricle, normal left ventricular function, and absent mitral regurgitation. g Abnormal findings are any of the following: 1) pericardial effusion, 2) acute cor pulmonale, 3) mild-severe left ventricular dysfunction, and 4) mitral regurgitation. Critical Care Medicine

6 See et al Table 5. Acceptability of Basic Critical Care Echocardiography Domains (n = 343 Examinations) BCCE Domain a Proportion Acceptable (All) (n = 343) (%) b Proportion Acceptable (1 10 BCCEs Done) (n = 70) (%) Proportion Acceptable (11 20 BCCEs Done) (n = 70) (%) Proportion Acceptable (21 30 BCCEs Done) (n = 70) (%) Proportion Acceptable (> 30 BCCEs Done) (n = 133) (%) p c 1. Acoustic windows 262/343 (76.4) 28/70 (40.0) 52/70 (74.3) 61/70 (87.1) 121/133 (91.0) < d 2. Pericardial effusion 331/343 (96.5) 66/70 (94.3) 68/70 (97.1) 67/70 (95.7) 130/133 (97.7) Right ventricle 310/343 (90.4) 52/70 (74.3) 67/70 (95.7) 68/70 (97.1) 123/133 (92.5) d 4. Left ventricle 283/343 (82.5) 50/70 (71.4) 57/70 (81.4) 63/70 (90.0) 113/133 (85.0) d 5. Mitral regurgitation 295/343 (86.1) 51/70 (72.9) 61/70 (87.1) 63/70 (90.0) 120/133 (90.2) d 6. Inferior vena cava 298/343 (86.9) 48/70 (68.6) 61/70 (87.1) 61/70 (87.1) 128/133 (96.2) < d Average proportion of acceptable findings across the six domains Mean improvement across the six domains, e mean ± sd (p f ) NA NA NA 16.9% ± 10.7 (0.012 d ) 21.0% ± 14.8 (0.018 d ) 21.9% ± 16.3 (0.022 d ) BCCE = basic critical care echocardiography, NA = not applicable. a 1) All seven acoustic windows (views) were acceptable; 2) presence of a pericardial effusion; 3) right ventricular dilatation with the right-to-left ventricular size ratio 1:1 in end-diastole at the papillary muscle level and interventricular septal straightening/paradoxical motion; 4) correct classification of the visual left ventricular ejection fraction (LVEF) to one of three categories (normal, i.e., estimated LVEF 50% or greater; mild-moderate dysfunction, i.e., estimated LVEF 30 49%; severe dysfunction, i.e., estimated LVEF < 30%); 5) presence of mitral valvular regurgitation (yes/no); and 6) correct measurement of inferior vena cava dimension (yes/no). b Number of cases for each fellow 52, 34, 105, 40, 34, 38, 40; total learning time for each fellow 1,144, 688, 1,778, 1,326, 651, 1,035, 693 mins, respectively. c Cuzick s nonparametric test for trend. d p < e Improvement = Difference in proportion of acceptable findings compared with baseline (1 10 BCCEs done). f Paired t test. NA Figure 1. Learning trajectory of basic critical care echocardiography domains (n = 343 examinations). Domains: 1. Acoustic windows: All seven acoustic windows (views) were acceptable. 2. Pericardial effusion: Presence of a pericardial effusion (yes/no). 3. Right ventricle: Identification of right ventricular dilatation with the rightto-left ventricular size ratio more than or equal to 1:1 in end-diastole at the papillary muscle level and interventricular septal straightening/paradoxical motion. 4. Left ventricle: Correct classification of the visual left ventricular ejection fraction (LVEF) to one of three categories (normal, i.e., estimated LVEF 50% or greater; mild-moderate dysfunction, i.e., estimated LVEF 30 49%; severe dysfunction, i.e., estimated LVEF < 30%). 5. Mitral regurgitation: Presence of mitral valvular regurgitation (yes/no). 6. Inferior vena cava: Correct measurement of inferior vena cava. To determine the prognostic significance of the BCCE domains, which would help determine if these domains per se were clinically useful for inclusion in our training program, we used the supervisoradjudicated findings. Using the fellows raw assessment would not be ideal as these were more likely to be erroneous during early training. The prognostic impact was significant for the association of acute cor pulmonale with increased ICU mortality, as expected on pathophysiological grounds (12). The other BCCE domains showed plausible though nonsignificant mortality trends and should be evaluated with a larger sample size. On the other hand, we could only show an association of mitral regurgitation with duration of hospitalization on October 2014 Volume 42 Number 10

7 Table 6. Correlation of Basic Critical Care Echocardiography Supervisor-Adjudicated Findings With Mortality (n = 318 Patients) Feature Articles Mortality, ICU Mortality, Hospital Diagnostic Finding Univariate a Multivariate b Univariate a Multivariate b Presence of pericardial effusion 0.75 (0.36, 1.59) 0.83 (0.38, 1.78) 1.04 (0.55, 1.96) 1.18 (0.61, 2.30) Acute cor pulmonale c 2.04 (1.13, 3.67) d 1.86 (1.01, 3.41) d 1.71 (0.99, 2.96) 1.55 (0.87, 2.74) Normal left ventricular function e 0.89 (0.49, 1.62) 0.95 (0.51, 1.76) 0.81 (0.47, 1.40) 0.87 (0.49, 1.52) Presence of mitral regurgitation 1.12 (0.63, 2.01) 0.89 (0.48, 1.64) 1.46 (0.86, 2.47) 1.17 (0.67, 2.04) Inferior vena cava variability f 1.26 (0.70, 2.27) 1.07 (0.57, 2.00) 1.14 (0.66, 1.97) 0.95 (0.53, 1.71) Normal study g 0.75 (0.40, 1.40) 0.82 (0.43, 1.57) 0.63 (0.35, 1.13) 0.69 (0.37, 1.26) a Odds ratio (with 95% CI) derived using logistic regression on mortality, unadjusted. b Odds ratio (with 95% CI) derived using multiple logistic regression on mortality, adjusted for Acute Physiology and Chronic Health II score and presence of sepsis. c Right ventricular dilatation with the right-to-left ventricular size ratio 1:1 in end-diastole at the papillary muscle level and interventricular septal straightening/ paradoxical motion. d p < e Visually estimated left ventricular ejection fraction 50% or greater. f Inferior vena cava (IVC) variability was present if the IVC had a collapsibility index of > 40% for patients who were spontaneously breathing or who were on continuous positive airway pressure via face mask, or a distensibility index of > 18% for patients on noninvasive or invasive mechanical ventilation. n = 293 due to 25 missing values from inadequate acoustic windows. g Normal study requires absent pericardial effusion, nondilated right ventricle, normal left ventricular function, and absent mitral regurgitation. Table 7. Correlation of Basic Critical Care Echocardiography Supervisor-Adjudicated Findings With Length of Stay (n = 318 Patients) Log (Length of Stay, ICU) Log (Length of Stay, Hospital) Diagnostic Finding Univariate a Multivariate b Univariate a Multivariate b Presence of pericardial effusion 0.96 (0.74, 1.23) 0.98 (0.76, 1.26) 1.07 (0.82, 1.41) 1.12 (0.85, 1.48) Acute cor pulmonale c 1.10 (0.88, 1.38) 1.06 (0.85, 1.33) 1.04 (0.81, 1.33) 0.99 (0.78, 1.27) Normal left ventricular function d 1.08 (0.87, 1.34) 1.10 (0.89, 1.36) 0.83 (0.65, 1.05) 0.85 (0.67, 1.07) Presence of mitral regurgitation 1.15 (0.93, 1.42) 1.10 (0.89, 1.36) 1.29 (1.02, 1.62) e 1.22 (0.96, 1.54) Inferior vena cava variability f 1.15 (0.93, 1.43) 1.08 (0.87, 1.34) 1.07 (0.84, 1.36) 1.01 (0.80, 1.29) Normal study g 1.02 (0.82, 1.26) 1.05 (0.85, 1.31) 0.85 (0.67, 1.07) 0.87 (0.69, 1.11) Log = natural logarithm (logarithm to the base e). a Exponentiated coefficient (with 95% CI) derived using linear regression on the log-transformed length of stay (LOS), unadjusted. b Exponentiated coefficient (with 95% CI) derived using multiple linear regression on the log-transformed LOS, adjusted for Acute Physiology and Chronic Health II score and presence of sepsis. c Right ventricular dilatation with the right-to-left ventricular size ratio 1:1 in end-diastole at the papillary muscle level and interventricular septal straightening/ paradoxical motion. d Visually estimated left ventricular ejection fraction 50% or greater. e p < f Inferior vena cava (IVC) variability was present if the IVC had a collapsibility index of > 40% for patients who were spontaneously breathing or who were on continuous positive airway pressure via face mask or a distensibility index of > 18% for patients on noninvasive or invasive mechanical ventilation. n = 293 due to 25 missing values from inadequate acoustic windows. g Normal study requires absent pericardial effusion, nondilated right ventricle, normal left ventricular function, and absent mitral regurgitation. univariate analysis, with the other findings not indicating any trends with length of stay. A prior study showed that one third of medical ICU patients had occult cardiac abnormalities, which were associated with increased ICU and hospital length of stay (19). However, the latter study required complete transthoracic echocardiography with Doppler scanning by full-time academic echocardiographers with 5 15 years experience, while ours used BCCE done by first-year trainees. We believe that the description of our program will encourage noncardiologists to train in BCCE. Apart from the time taken by the supervisors to teach, we did not invest in additional training materials or echocardiography simulators. This was possible by using existing free open-access medical education resources and by integrating regular training within our practice of screening critically ill patients with echocardiography. Although we only had about 10 hours of theoretical Critical Care Medicine

8 See et al self-instruction and 2.5 hours of directly observed practical instruction per trainee, a short training period may be enough to achieve basic competency (4, 8, 13, 20, 21). As such, our program and probably our results can be readily replicated by others. Another strength is that the training program required trainees to scan various ICU patients with and without mechanical ventilation, which provided a rich source of practical experience, as opposed to using echocardiography simulators for training. Of 343 scans, 154 (45.0%) were done when patients were spontaneously breathing (no positive pressure applied). We would expect better echocardiographic views in such patients than in those receiving positive pressure ventilation. However, the proportion of adequate views in the former cases compared with the latter had no statistically significant difference (81.2% vs 72.5%, p = 0.073). Such hands-on training was also shown to be effective in teaching general ultrasonography and echocardiography skills to ICU residents (4, 5, 20). Our study has to be interpreted with certain caveats. First, our findings are only relevant for BCCE, since we did not study more advanced techniques (e.g., assessment for diastolic dysfunction, pulmonary artery systolic pressure estimation, tissue Doppler measurements, and contrast echocardiography). As such, intensivists should use BCCE to supplement clinical examination and continue to request cardiology review for more comprehensive studies on selected patients. Second, we used a conventional though still portable ultrasound system, and simpler handheld devices may yield poorer results (22). Third, we are uncertain of the learning trajectory beyond 40 BCCEs, although there is no current consensus on a minimum number for competency (3). Fourth, we do not claim that our program would produce the best possible outcomes but only that useful training can emanate from a simple training system. A more sophisticated curriculum involving cardiologist-taught simulator training, greater bedside instruction time, and computer-based testing has admittedly yielded faster trainee scanning times (7). Fifth, although we may have missed some findings, a normal study was not associated with increased mortality and length of hospitalization. Sixth, BCCE views may be inadequate because of patients morphology rather than learners competency. As supervisors could not consistently repeat the BCCE due to clinical or logistic reasons, we opted to be strict with the scoring criteria whereby all seven views had to be obtained by the fellows. This would potentially underestimate the competency of trainees. Seventh, potential bias in scoring could occur as our assessors were not blinded. To minimize bias, we adopted strict a priori scoring criteria for individual BCCE domains. Nonetheless, major bias was unlikely, as our trainees learning trajectory (at least for LVEF estimation) was consistent with prior data (13), and the results from our supervisor-adjudicated BCCE assessment had some prognostic impact. Finally, we have not studied whether our approach actually affected management. By showing the achievement of reasonably high levels of trainee competence even in the first year of training, in domains that are clinically significant, we hope to encourage others who face similar resource challenges to set up BCCE programs. Trainee-performed scans can be reasonably accurate for several domains of BCCE and may yield important clinical benefits. Our approach of indirect supervision allowed us to provide ongoing feedback to trainees, while committing minimal teaching time. We believe that the current standard of BCCE by non cardiology trainees can add value to diagnosis and hence management. This can be especially useful when urgent bedside assessment of hemodynamic status is required for critically ill patients and when such a need occurs after office hours. Our results support the recommendation that competency in BCCE requires at least 30 transthoracic studies (3). Further monitoring of the learning trajectory can perhaps inform the suitability of this low-cost approach for certification, accreditation, or privileging processes. In conclusion, we demonstrated that BCCE training using minimal resources is feasible and new trainees can achieve clinically meaningful competencies in most BCCE domains after performing about 30 examinations within the first year. The relatively high prevalence of BCCE abnormalities and the significant association of acute cor pulmonale with ICU mortality support the need for BCCE training, which should be included in the curriculum of all intensivists (3). However, continued supervision is required, particularly to improve the accuracy of visual LVEF assessment. ACKNOWLEDGMENT We thank Dr. Chan Yiong Huak, Yong Loo Lin School of Medicine, National University of Singapore, for providing independent statistical review of this article. REFERENCES 1. Beaulieu Y: Specific skill set and goals of focused echocardiography for critical care clinicians. Crit Care Med 2007; 35:S144 S Kaplan A, Mayo PH: Echocardiography performed by the pulmonary/ critical care medicine physician. Chest 2009; 135: Expert Round Table on Ultrasound in ICU: International expert statement on training standards for critical care ultrasonography. Intensive Care Med 2011; 37: Vignon P, Dugard A, Abraham J, et al: Focused training for goaloriented hand-held echocardiography performed by noncardiologist residents in the intensive care unit. Intensive Care Med 2007; 33: Vignon P, Mücke F, Bellec F, et al: Basic critical care echocardiography: Validation of a curriculum dedicated to noncardiologist residents. Crit Care Med 2011; 39: Oren-Grinberg A, Talmor D, Brown SM: Focused critical care echocardiography. Crit Care Med 2013; 41: Beraud AS, Rizk NW, Pearl RG, et al: Focused transthoracic echocardiography during critical care medicine training: Curriculum implementation and evaluation of proficiency. Crit Care Med 2013; 41:e179 e Manasia AR, Nagaraj HM, Kodali RB, et al: Feasibility and potential clinical utility of goal-directed transthoracic echocardiography performed by noncardiologist intensivists using a small hand-carried device (SonoHeart) in critically ill patients. J Cardiothorac Vasc Anesth 2005; 19: Slasky BS, Auerbach D, Skolnick ML: Value of portable real-time ultrasound in the ICU. Crit Care Med 1983; 11: Manno E, Navarra M, Faccio L, et al: Deep impact of ultrasound in the intensive care unit: The ICU-sound protocol. Anesthesiology 2012; 117: October 2014 Volume 42 Number 10

9 Feature Articles 11. Jardin F, Dubourg O, Bourdarias JP: Echocardiographic pattern of acute cor pulmonale. Chest 1997; 111: Vieillard-Baron A, Prin S, Chergui K, et al: Echo-Doppler demonstration of acute cor pulmonale at the bedside in the medical intensive care unit. Am J Respir Crit Care Med 2002; 166: Melamed R, Sprenkle MD, Ulstad VK, et al: Assessment of left ventricular function by intensivists using hand-held echocardiography. Chest 2009; 135: Gudmundsson P, Rydberg E, Winter R, et al: Visually estimated left ventricular ejection fraction by echocardiography is closely correlated with formal quantitative methods. Int J Cardiol 2005; 101: Picard MH, Adams D, Bierig SM, et al; American Society of Echocardiography: American Society of Echocardiography recommendations for quality echocardiography laboratory operations. J Am Soc Echocardiogr 2011; 24: Muller L, Bobbia X, Toumi M, et al; the AzuRea Group: Respiratory variations of inferior vena cava diameter to predict fluid responsiveness in spontaneously breathing patients with acute circulatory failure: Need for a cautious use. Crit Care 2012; 16:R Barbier C, Loubières Y, Schmit C, et al: Respiratory changes in inferior vena cava diameter are helpful in predicting fluid responsiveness in ventilated septic patients. Intensive Care Med 2004; 30: Mayo PH, Beaulieu Y, Doelken P, et al: American College of Chest Physicians/La Société de Réanimation de Langue Française statement on competence in critical care ultrasonography. Chest 2009; 135: Bossone E, DiGiovine B, Watts S, et al: Range and prevalence of cardiac abnormalities in patients hospitalized in a medical ICU. Chest 2002; 122: Chalumeau-Lemoine L, Baudel JL, Das V, et al: Results of short-term training of naïve physicians in focused general ultrasonography in an intensive-care unit. Intensive Care Med 2009; 35: Alexander JH, Peterson ED, Chen AY, et al: Feasibility of point-of-care echocardiography by internal medicine house staff. Am Heart J 2004; 147: Goodkin GM, Spevack DM, Tunick PA, et al: How useful is handcarried bedside echocardiography in critically ill patients? J Am Coll Cardiol 2001; 37: Critical Care Medicine

10 本文献由 学霸图书馆 - 文献云下载 收集自网络, 仅供学习交流使用 学霸图书馆 ( 是一个 整合众多图书馆数据库资源, 提供一站式文献检索和下载服务 的 24 小时在线不限 IP 图书馆 图书馆致力于便利 促进学习与科研, 提供最强文献下载服务 图书馆导航 : 图书馆首页文献云下载图书馆入口外文数据库大全疑难文献辅助工具