The Utility of Handheld Echocardiography for Early Diagnosis of Rheumatic Heart Disease

The Utility of Handheld Echocardiography for Early Diagnosis of Rheumatic Heart Disease Andrea Beaton, MD, Twalib Aliku, MD, Emmy Okello, MD, Sulaiman Lubega, MD, Robert McCarter, ScD, Peter Lwabi, MD, and Craig Sable, MD, Washington, District of Columbia; Kampala, Uganda

Background: Rheumatic heart disease (RHD) remains endemic in most of the developing world. Echocardiography has proved highly sensitive for early detection of RHD, but it remains too costly for most low-income settings. Handheld ultrasound machines used to perform handheld echocardiography (HAND) are both less expensive and more portable, possibly making them ideal screening tools. HAND has never been tested for the early diagnosis of RHD. The aim of this study was to evaluate the performance of focused HAND compared with focused standard portable echocardiography for the diagnosis of subclinical RHD. Methods: HAND and standard portable echocardiography were performed on 125 Ugandan children, 41 with borderline or definite RHD, and 84 healthy controls. Images were blindly reviewed according to the 2012 World Heart Federation guidelines. Results: HAND was highly sensitive (90.2%) and specific (92.9%) for distinguishing between normal patients and those with RHD, but it performed best with definite RHD. HAND overestimated mitral valve morphologic valve abnormalities, being only 66.7% specific for anterior leaflet thickness > 3 mm and 79.0% specific for restricted leaflet motion. False-negative results (n = 4) were due primarily to underestimation of mitral regurgitation length. Conclusions: In this population, HAND was highly sensitive and specific for early detection of RHD. HAND functions best as a screening tool with confirmation of positive screening results by fully functional echocardiography machines. Technical advances may enable one-step RHD screening using HAND. The performance of HAND should be studied across diverse populations and in field tests before recommending it for widespread screening. (J Am Soc Echocardiogr 2014;27:42-9.) Keywords: Echocardiography, Handheld echocardiography, Rheumatic heart disease, Screening

Rheumatic heart disease (RHD) causes significant disability and premature death in developing nations, despite its virtual elimination in most of the developed world.1 RHD results from cumulative exposure to streptococcal infections, which can lead to cardiac inflammation, scarring, and dysfunction.2 The disease is usually progressive, starting in midchildhood. However, data from developing nations show that most patients present as young adults, once symptoms become severe enough to result in cardiovascular limitations.3 Furthermore, up to 40% of these symptomatic young adults cannot

From the Division of Cardiology, Children’s National Medical Center, Washington, Distric of Columbia (A.B., R.M., C.S.); and Uganda Heart Institute, Mulago Hospital Complex, Kampala, Uganda (T.A., E.O., S.L., P.L.). This project was supported by Award Number UL1TR000075 from the National Institutes of Health National Center for Advancing Translational Sciences. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Center for Advancing Translational Sciences or the National Institutes of Health. Reprint requests: Andrea Beaton, MD, Children’s National Medical Center, Division of Cardiology, 111 Michigan Avenue, Washington, DC 20010 (E-mail: [email protected]). 0894-7317/$36.00 Copyright 2014 by the American Society of Echocardiography. http://dx.doi.org/10.1016/j.echo.2013.09.013

42

recall an episode of acute rheumatic fever,4 casting doubt on the notion that improved primary clinical recognition can significantly reduce the global burden of disease. Symptomatic presentation of RHD in resource-poor settings has dismal outcomes.5 Cardiac surgery or catheterization is often not available, and even when offered, it is often cost prohibitive. Data from Africa suggest particularly poor outcomes, with an average age of RHD-related death of 25.89 years.6 Outcomes are even more tragic for pregnant mothers, 34% of whom will die in the peripartum period.7 By contrast, patients who present with early RHD generally have good outcomes. Secondary prophylaxis, in the form of penicillin injections every 3 to 4 weeks to prevent recurrent streptococcal infection, in these patients is of proven benefit. The best prognosis is for those with mild disease, the majority of whom, with regular secondary prophylaxis, will have no detectable cardiac disease after 10 years.2,8 The challenge, then, is how to effectively find patients with early RHD. Over the past decade, echocardiography has proved to be the most sensitive tool for early detection of RHD.9 The World Health Organization now supports early detection of RHD through echocardiography in high-prevalence regions.10 For the first time, the World Heart Federation (WHF) has provided evidence-based guidelines for the echocardiographic diagnosis of RHD.11 These guidelines are specifically designed to define the ‘‘minimum echocardiographic criteria for the diagnosis of RHD’’11 and are meant for use with

Journal of the American Society of Echocardiography Volume 27 Number 1

patients who live in endemic areas, have no symptoms of FCU = Focused cardiac RHD, and have no histories of ultrasound acute rheumatic fever. The WHF guidelines provide instrucHAND = Handheld tions for the systematic assessechocardiography ment of the morphology and RHD = Rheumatic heart function of the mitral and aortic disease valves through limited echocarSTAND = Standard portable diographic evaluation. They echocardiography were designed for use with a standard portable echocardioWHF = World Heart graphic machine, capable of twoFederation dimensional and color image acquisition as well as continuous-wave Doppler imaging. Despite increasingly affordable standard portable echocardiographic machines, however, the price remains too high for practitioners in most RHD-endemic countries. Newly available handheld ultrasound machines could both reduce RHD screening costs and increase the reach of such initiatives to even the most remote settings. These ultraportable handheld machines have proved effective in many settings, including routine outpatient echocardiographic evaluation,12 inpatient cardiac consultation,13 and assessing left ventricular function in the acute care setting.14 They are simple to use and are able to produce two-dimensional and color images but do not currently have the spectral Doppler capabilities needed to fully implement the WHF criteria. Their sensitivity and specificity for RHD detection compared with standard portable echocardiography (STAND) have not been evaluated. The objective of this study was to determine the sensitivity and specificity of a focused echocardiographic evaluation performed with handheld echocardiography (HAND) compared with STAND for the detection of RHD in Uganda, a known high-prevalence area.15 This is the first systematic evaluation of HAND as a tool for early RHD detection. Abbreviations

METHODS Study Population This prospective observational study included a sample of paired echocardiograms in 125 patients. Participants included 60 patients presenting for follow-up as part of the Ugandan RHD registry project (Uganda Heart Institute, Mulago Hospital Complex, Kampala, Uganda) and 65 asymptomatic Ugandan schoolchildren who took part in an echocardiography-based screening program at their school.15 Each participant underwent a focused echocardiographic examination with STAND. Additionally, each patient underwent a focused echocardiographic examination with HAND. All studies were recorded in a 2-week period in September 2012. Institutional review board approval was obtained from Makerere University and the Children’s National Medical Center. Echocardiographic Protocol One pediatric cardiologist (A.B.) performed all echocardiographic examinations. Handheld echocardiographic equipment (Vscan; GE Medical Systems, Milwaukee, WI) used a 1.7-MHz to 3.4-MHz transducer. This device provides two-dimensional and color Doppler imaging on an integrated 3.5-inch display. Frame rates ranged from 25 to 30 Hz for black-and-white imaging and from 12

Beaton et al 43

to 16 Hz for color Doppler. The device offers an image-based ‘‘Auto-Cycle’’ function for the automatic detection of a full heart cycle beginning with end-diastole. Standard portable echocardiographic equipment (Vivid-I; GE Medical Systems) used a 1.5-MHz to 3.6-MHz transducer. Frame rates ranged from 25 to 35 Hz for black-and-white imaging and from 12-18 Hz for color Doppler. For STAND, electrocardiography was used to detect two heartbeat loops. STAND was used as the gold standard. Grayscale and color Doppler recordings (parasternal long-axis and short-axis and apical fourchamber views) were acquired and stored digitally for later offline assessment. In addition, STAND included continuous-wave Doppler recordings at the mitral and aortic valves (not available on the handheld echocardiographic device). Protocol for Blinding Each imaging study was assigned a unique research identification number, which could be used to link the imaging study to the research participant. A single expert reviewer (C.S.) interpreted all handheld and standard portable echocardiographic studies. The reviewer was blinded to the corresponding echocardiographic examination (either HAND or STAND), but it was not possible to blind the reviewer to the type of study (HAND or STAND), because interpretation occurred on different platforms (Figure 1 shows the handheld echocardiographic platform), each device has a characteristic standardized screen setup, and images were interpreted using different software packages. Image Scoring and Interpretation For analysis, images from the handheld echocardiographic studies were interpreted using the dedicated Vscan Gateway software installed on a standard personal computer. The size of the image was slightly larger on the computer than on the handheld echocardiographic device, but the resolution of the images was identical. Images from the standard portable echocardiographic studies were transferred to our institution’s echocardiographic picture archiving and communication system for interpretation. Studies were analyzed in random order. A score for image quality (ranging from 1 to 5, with 5 being the best) was assigned to describe the quality of the overall image, the quality of the mitral valve images, and the quality of the aortic valve images for all studies. Echocardiographic interpretation of the standard portable echocardiographic studies was conducted according to the 2012 WHF guidelines (Table 1).11 Because of the lack of continuous-wave Doppler capabilities in the handheld echocardiographic studies, modified 2012 WHF criteria were used to determine pathologic regurgitation (Table 2). Overall categorization of disease presence (normal, borderline RHD, or definite RHD) was recorded, as well as the individual morphologic and functional components used to arrive at this categorization. Statistical Analysis Study data were collected and managed using the REDCap electronic data-capture tools hosted at the Children’s National Medical Center.16 Groups were compared using a one-way analysis of variance for continuous data and c2 tests for categorical data. Sensitivity and specificity were calculated for overall RHD detection (combined definite RHD and borderline RHD), as well as separately for definite and borderline RHD. Individual components composing the 2012 WHF criteria were also examined.

44 Beaton et al

Journal of the American Society of Echocardiography January 2014

Figure 1 Deidentified handheld echocardiographic parasternal long-axis images being analyzed with the Vscan Gateway software package on a personal computer. Black-and-white images show measurement of the thickness of the anterior mitral leaflet (limited to 1-cm intervals), and color Doppler images show measurement of the mitral regurgitation jet length. Table 1 WHF 2012 guidelines for echocardiographic diagnosis of RHD (patients aged < 20 years)

Table 2 WHF 2012 guidelines for diagnosis of pathologic regurgitation

Criterion

Pathologic MR

Definite RHD A. Pathologic MR and at least two morphologic features of RHD of the MV B. MS mean gradient > 4 mm Hg* C. Pathologic AR and at least two morphologic features of RHD of the AV† D. Borderline disease of both the AV and MV‡ Borderline RHD A. At least two morphologic features of RHD of the MV without pathologic MR or MS B. Pathologic MR C. Pathologic AR

2012 WHF criteria Seen in two views In at least one view, jet length > 2 cm* Velocity > 3 m/sec for one complete envelope Pansystolic jet in at least one envelope

AR, Aortic regurgitation; AV, aortic valve; MR, mitral regurgitation; MS, mitral stenosis; MV, mitral valve. *Congenital MV anomalies must be excluded. † Bicuspid AV and dilated aortic root must be excluded. ‡ Combined AR and MR in high-prevalence regions and in the absence of congenital heart disease is regarded as rheumatic.

Modified criteria†,‡ Seen in two views In at least one view, jet length > 2 cm* Pansystolic jet (by color Doppler)

Pathologic AR

Seen in two views In at least one view, jet length > 1 cm* Velocity > 3 m/sec in early diastole Pandiastolic jet in at least one envelope Seen in two views In at least one view, jet length > 1 cm* Pandiastolic jet (by color Doppler)

RESULTS

AR, Aortic regurgitation; MR, mitral regurgitation. *Regurgitant jet length should be measured from the vena contracta to the last pixel of regurgitant color (blue or red). † Exclusion of continuous-wave Doppler evaluation for determination of pathologic versus nonpathologic valvular regurgitation. ‡ In this study, pathologic MR or AR was considered present in patients meeting the above criteria.

Paired echocardiograms were available for interpretation on all 125 Ugandan children. Of these children, 84 (67%) were found to be normal, 16 (13%) were found to have borderline RHD, and 25 (20%) were found to have definite RHD by STAND. No children were found to have congenital heart disease. Descriptive characteristics of these three populations are located in Table 3. The majority of patients with borderline and definite RHD had isolated mitral

valve disease, with seven patients having mixed aortic and mitral valve disease (28%) and only one patient (4%) having isolated aortic valve disease. In addition, 12 patients (14%) found to be normal had nonpathologic mitral regurgitation, while none had nonpathologic aortic regurgitation.

Beaton et al 45

Journal of the American Society of Echocardiography Volume 27 Number 1

Table 3 Demographic and echocardiographic characteristics by diagnosis (STAND) Normal

Borderline RHD

Definite RHD

(n = 84)

(n = 16)

(n = 25)

P

Mean age (y) 10.7 Women 52.4% MR (cm) 12 1.1–1.5 cm 6 1.6–1.9 6 2.0–2.5 0 >2.5 0 MS 0 Morphologic MV abnormalities 7 Anterior mitral leaflet 4 thickness > 3 mm Chordal thickening 0 Restricted leaflet motion 3 Excessive leaflet motion 0 AR 0 >1 cm 0 Morphologic AV abnormalities 0 Thickening 0 Coaptation defect 0 Restricted leaflet motion 0 Prolapse 0

11.1 62.5% 15 0 1 13 1 0 7 7

11.1 60% 24 0 3 3 18 5 24 24

.74 .66

Variable

1 2 0 0 0 0 0 0 0 0

20 14 10 8 8 6 6 5 3 4

AR, Aortic regurgitation; AV, aortic valve; MR, mitral regurgitation; MS, mitral stenosis; MV, mitral valve.

Figures 2 and 3 (definite and borderline RHD, respectively) compare parasternal long-axis views obtained in two patients with RHD. Table 4 compares disease categorization by STAND and HAND. Table 5 lists the sensitivity and specificity of HAND in overall RHD diagnosis (definite and borderline RHD). For the overall distinction between normal patients and those with RHD (borderline and definite), HAND performed well, with excellent sensitivity (90.2%) and specificity (92.9%). After elimination of those classified as having definite RHD by STAND, HAND was less sensitive for discriminating between borderline and normal, at only 75%. To understand the reasons for disagreement between machines, the sensitivity and specificity of the individual components of diagnosis were examined (Table 5). Of the 84 patients graded as normal by STAND, HAND incorrectly classified six as having borderline RHD. Only one of these patients was diagnosed with borderline B disease (isolated significant mitral regurgitation), with the length of the mitral regurgitation jet found to be 1.7 cm on STAND and 2.2 cm on HAND (with all other features of pathologic regurgitation true on both machines). The remaining five false-positive results were incorrectly classified as borderline A (at least two morphologic changes without significant mitral regurgitation). Handheld echocardiographic images had poor specificity for both mitral valve thickness (66.7%) and restricted leaflet motion (79.0%); handheld echocardiographic images were interpreted to show morphologic changes not seen on STAND. The prevalence of excessive mitral valve anterior leaflet tip motion was too low to accurately interpret sensitivity and specificity and did not contribute to any change in diagnosis.

Of the 16 patients graded as having borderline RHD by STAND, HAND determined that four were normal, eight had borderline RHD, and four had definite RHD. All four false-negative results resulted from handheld echocardiographic interpretation of nonsignificant mitral regurgitation. In four studies with false-negative results, the length of the mitral regurgitation jet was found to be <2 cm, and the duration of regurgitation was judged not to be pansystolic. These four borderline cases were among the mildest in the STAND group, with mitral regurgitation jet lengths between 2 and 2.2 cm (0.6–1.8 cm on HAND). Underdiagnosis of pathologic mitral regurgitation was also seen in two of eight patients (25%) correctly identified by HAND as having borderline RHD. For these two patients, borderline B standard portable echocardiographic studies were identified as borderline A by HAND, with underestimation of mitral regurgitation length and (again) overestimation of morphologic abnormalities. Overdiagnosis of morphologic abnormalities on handheld echocardiographic images was also seen in the borderline group, four of whose patients were diagnosed with definite RHD. Functional mitral valve changes were detected by HAND with more precision than were morphologic abnormalities. Pathologic regurgitant length (>2 cm) was detected with sensitivity of 83.3% and specificity of 100%. Lack of continuous-wave Doppler necessitated subjective grading of the timing of mitral regurgitation on the basis of color Doppler images. Sensitivity and specificity suffered significantly from this method, being only 76.9% and 68.8%, respectively, for the detection of pansystolic regurgitation. STAND was slightly better than HAND in both overall image quality and in quality of imaging of the mitral valve (3.85 vs 3.7, P = .05, and 3.9 vs 3.8, P = .05). No difference between machines was noted for the quality of aortic valve imaging (3.7 vs 3.6, P = .29).

DISCUSSION In this study, a critical first step in the validation of HAND for RHD screening, we report the results of 125 evaluations using both standard portable and handheld echocardiographic machines. Compared with a field screening study, we used a population with an artificially large number of RHD-positive patients (borderline RHD and definite RHD), reducing the needed sample size (which is in the range of 5,000 patients, assuming an RHD prevalence of 2%). In addition, we strategically chose one blinded expert reviewer for all studies, eliminating the confounder of multiple reviewers and focusing our evaluation solely on current handheld echocardiographic capabilities. In this first study of using HAND for RHD detection, we show that HAND is highly sensitive and specific for the detection of RHD. However, HAND performs better for some components of the 2012 WHF criteria compared with others and may be best suited as a screening tool rather than for full-field implementation of the WHF guidelines. The aim of screening is to identify RHD in apparently healthy individuals.17 A useful screening tool ideally should have high sensitivity to limit the number of false-negatives while maintaining high specificity to limit the number of false-positives. Both are important in applying echocardiography to RHD screening in resourceconstrained settings. False-negatives will result in missed opportunities to prevent disease progression, costing not only lives but also valuable health care funds devoted to treating advanced RHD down the road. Conversely, false-positives impose on individuals the cost and burden of unnecessary long-term penicillin prophylaxis and the potential stigma of a chronic disease diagnosis.

46 Beaton et al

Journal of the American Society of Echocardiography January 2014

Figure 2 (A) Two-dimensional parasternal long-axis view of the mitral valve by STAND (left) and HAND (right) in the same patient with definite RHD. Arrow indicates the thickened anterior mitral valve leaflet, demonstrating excessive leaflet motion. (B) The same patient and devices in color Doppler mode. Arrow indicates the origin and termination of the mitral regurgitation jet in both images. LA, Left atrium; LV, left ventricle. And for society, false-positives may flood limited subspecialist resources with unneeded secondary evaluations.9 Given these considerations, we believe that HAND would need to be both highly sensitive and specific compared with STAND to justify its use in widespread screening. In our study, HAND met this goal. It proved 90.2% sensitive and 92.9% specific for differentiating normal patients and those with disease. It performed even better for those with definite disease: 100% sensitive and 96% specific. These results are consistent with the only reported use of HAND for the evaluation of mitral valve disease; that study reported high sensitivity (96%) and specificity (96%) for detecting more than moderate mitral regurgitation.18 However, 75% to 90% of RHD detected through echocardiographic screening meets criteria for borderline RHD,15,19-22 with trivial to mild mitral regurgitation, for which HAND was only 75% sensitive and 92.9% specific. There are many studies in the literature that highlight the use of focused cardiac ultrasound (FCU) for the detection of chamber enlargement, ventricular dysfunction, and pericardial fluid. Although it is tempting to consider our study as another example of FCU, the recent American Society of Echocardiography guidelines make a clear distinction between the methodology in our study and FCU.23 FCU is defined as being an adjunct to physical examination in specific clinical settings. Applying this terminology to our study (or any study based on WHF echocardiographic guidelines for RHD), which did not involve ultrasound as an adjunct to physical examination, would be inconsistent with the 2013 American Society of Echocardiography definition of FCU.

The only other study to feature HAND for RHD detection used a first-generation machine (Optigo; Philips Medical Systems, Andover, MA). The study assessed the ability of briefly trained operators to diagnose RHD, compared with diagnoses made with previous echocardiograms obtained on standard, fully functional machines. Three final-year medical students, after 8 hours of focused training, each examined 45 patients, 14 of whom had rheumatic mitral valve disease and 31 of whom were normal. The man age in this cohort range from 55.8 to 59.5 years, and all 14 patients with RHD had at least moderate mitral stenosis (n = 13), at least moderate mitral regurgitation (n = 8), or both (number not reported). Sensitivity for ultrasound diagnosis of mitral regurgitation and mitral stenosis was very poor (averages of 21% and 33% across students). Pooled sensitivity (81%) and specificity (95%) for diagnosing rheumatic valve disease was significantly better, which the authors attributed to a training program that focused on normal and rheumatic mitral morphology.24 Multiple confounding variables, including technical image acquisition, the inherent image quality of the machine, and the experience of the students, make it difficult to draw many conclusions from this study. In addition, the advanced age and disease severity of the population make the conclusions unlikely to apply to a screening population. The current generation of handheld ultrasound devices has many limitations that are important to consider. Interpretation of handheld echocardiographic images tended to overdiagnose morphologic mitral valve changes. Measurement of the thickness of the anterior mitral valve leaflet, in particular, performed poorly. Some of this error likely resulted from functional software limitations. Currently, the

Beaton et al 47

Journal of the American Society of Echocardiography Volume 27 Number 1

Figure 3 (A) Two-dimensional parasternal long-axis view of the mitral valve by STAND (left) and HAND (right) in the same patient with borderline RHD. Arrow indicates the thickened anterior mitral valve leaflet and restricted leaflet motion. (B) The same patient and devices in color Doppler mode. Arrow points to the origin and termination of the mitral regurgitation jet in both images, which does not meet the cutoff of 2 cm in length. LA, Left atrium; LV, left ventricle. Table 4 Comparison of HAND and STAND for disease categorization STAND Category

HAND Normal (n = 82) Borderline (n = 14) Definite (n = 29)

Normal (n = 84) Borderline (n = 16) Definite (n = 25)

78 6 0

4 8 4

0 0 25

handheld ultrasound machines used in this study are capable of only 1mm measurement increments. Practically, this results in measurements of 2.5 to 2.9 mm (normal according to the 2012 WHF criteria) being measured at 3 mm, which meets the criteria for thickened. Additionally, in this study, measurements for handheld echocardiographic studies were made using offline software to analyze stillframe images. Although it is possible to complete these measurements in real time on the device, we feel that the additional blur that can come from live images would only exacerbate this limitation. Alone, this single overstatement would not cause the valve to be labeled as morphologically abnormal; handheld echocardiographic interpretations, however, also tended to overdiagnose chordal thickening and restricted leaflet mobility. In theory, suboptimal frame rates can cause an overestimation of valve and chordal thickness if the frame demonstrating maximal valve excursion is not obtained. Conversely, though not seen in our study, slower frame rates may

also underestimate morphologic changes, as subtle tethering of the valve leaflet can also be difficult to detect. Together, overestimation of morphologic abnormalities decreased the machine’s specificity, resulting in the diagnosis of more disease, and more severe disease, compared with standard portable echocardiographic studies. Used as a screening tool, not as a final diagnostic instrument, these overestimations could be corrected during secondary evaluation, reducing the number of initially false-positive patients who unnecessarily receive prophylaxis. Frame rate is unlikely to be a contributing factor to differences in our study, because the frame rates were only slightly higher for STAND. However, slower frame rates are even more important to consider when comparing color Doppler evaluation. Low color Doppler frame rates can result in underestimation of color Doppler jet length if the frame of maximal color jet length is not obtained. Additionally, lower frame rates could make subjective determination of pansystolic regurgitation (needed in the absence of continuous-wave Doppler and electrocardiographic tracing) less accurate. This could have been compounded in our study, in which the standard portable echocardiographic studies contained two heartbeat loops, whereas most of the handheld studies contained one heartbeat loop. Functional assessment of the mitral valve also proved to have some challenges when interpreting handheld echocardiographic imaging. The cutoff of 2 cm for the length of mitral regurgitation had adequate sensitivity and specificity (83.3% and 100%, respectively). However, technical parameters on the ultrasound machine can also influence the length of the regurgitant jet. These include gain, carrier frequency, color scale, filter settings, and frame rate, which, if low, may miss the

48 Beaton et al

Journal of the American Society of Echocardiography January 2014

Table 5 Sensitivity and specificity of HAND for overall RHD diagnosis and for individual components of RHD diagnosis, using STAND as the gold standard Diagnosis

Prevalence (%)

Sensitivity (%)

Specificity (%)

RHD Positive Definite RHD Borderline RHD MV morphology Anterior mitral leaflet thickness > 3 mm Chordal thickening Restricted leaflet motion Excessive leaflet tip motion MV function MR seen in two views* Jet length > 2 cm Pansystolic*

32.8 20.0 12.8

90.2 (75.9–96.8) 100 (83.4–100) 75.0 (47.4–91.7)

92.9 (84.5–97.0) 96.0 (89.5–98.7) 92.9 (84.5–94.0)

59.2 16.8 16.0 8.0

71.6 (59.8–67.8) 95.2 (74.12–99.9) 85.0 (61.1–96.0) 60.0 (27.4–86.3)

66.7 (59.8–81.2) 90.4 (82.6–95.0) 79.0 (69.8–86.1) 99.1 (94.5–99.9)

87.3 28.8 70.9

81.3 (66.9–90.6) 83.3 (66.5–93.0) 76.9 (56.9–89.0)

28.6 (5.1–69.7) 100 (94.8–100) 68.8 (41.5–87.9)

MR, Mitral regurgitation; MV, mitral valve. *Prevalence, sensitivity, and specificity only among those studies in which MR was present.

time point of maximal jet length, contributing to poor accuracy, reproducibility, and relevance of this parameter. Although the WHF guidelines give clear guidance on optimizing machine settings, the current handheld machine only allows control of gain, limiting the user’s ability to optimize image acquisition. Additionally, more severe regurgitant jets can be difficult to measure, because the flow is disturbed and the distal end of the regurgitant jet is irregular in shape (Figure 2B). However, this difficultly is limited to more severe regurgitant jets, and we do not feel that it is an important contributor to determining if a regurgitant length is physiologic or pathologic (>2 or <2 cm). Finally, lack of continuous-wave Doppler capabilities led us to modify the WHF criteria, eliminating the criterion of velocity > 3 m/sec needed to diagnose significant mitral or aortic regurgitation. In this study, elimination of this criterion from the handheld echocardiographic examinations did not independently alter any diagnostic classifications. In addition to some limitations in the current handheld echocardiographic technology, there are also some limitations to the WHF guidelines.11 Although currently, we believe these guidelines to be the best available for early detection of RHD, they represent a compilation of the best available evidence, as well as expert consensus opinion. These criteria are essential for the standardization of worldwide protocols for the diagnosis of latent RHD, but they are likely far from perfect. It is only through time and careful data collection and analysis that these guidelines will be able to be refined, improving their specificity for clinically significant RHD. Currently, the WHF guidelines call for frame rates between 30 and 60 Hz, on the basis of prior studies. The impact of using lower frame rates on accurate image interpretation has not been studied. However, high frame rates are less of an issue in patients with lower heart rates. Most of the patients in our study (and likely most school-aged children) had heart rates < 80 beats/min. Nonetheless, optimization of frame rates is an important consideration in applying the WHF criteria. We also remain cautious when interpreting the WHF recommendation for pathologic mitral regurgitation jet length of 2 cm, because regurgitant length is determined more by afterload than by regurgitant volumes. Although the velocity of aortic or mitral regurgitation certainly does not contribute to assessment of regurgitant severity, it is included in the WHF criteria to differentiate true mitral or aortic regurgitation from artifacts or closing volumes. Using jet length to

distinguish pathologic regurgitation from artifacts or closing volumes for screening seems reasonable, but without defining loading conditions (preload, systolic blood pressure), it is less likely to be an accurate way to track disease progression or improvement. Similarly, we wish to highlight that the velocity of the regurgitation jet does not reflect the severity of the regurgitation, only the pressure drop across the incompetent valve. The requirement of regurgitant flow velocity > 3 m/sec is meant to differentiate true mitral or aortic regurgitation from artifacts or closing volumes and should not be used to grade the severity of disease or to track change in disease severity over time. In turn, in our study, the elimination of this criterion from the 2012 WHF guidelines for handheld echocardiographic interpretation (given no continuous-wave Doppler capability on the handheld machines) did not independently affect the sensitivity or specificity of this device. Finally, the WHF criteria do not provide guidance for how to measure nonlinear regurgitant jets. Often, posteriorly directed mitral regurgitation (most commonly seen with RHD) will track along the atrial wall, changing alignment and often color as the flow axis relative to the ultrasound beam changes from ‘‘toward’’ to ‘‘away.’’ Color flow jets that are directed centrally into the left atrium also tend to appear larger because they entrain red blood cells on all sides of the jet. In contrast, eccentric jets that hug the left atrial wall cannot entrain blood on all sides and tend to appear smaller than central jets of similar severity; the color jet assumes the shape of a spoon in three dimensions.25 Although most commonly seen with longer jets, reducing its importance in screening, it is also possible that shorter jets could be underestimated when they are eccentric and follow the curve of the atrial wall. Standardization of technique to measure these jets will be important for consistent reporting. Our study had several limitations. In our cohort, RHD was limited to the mitral valve in all but a few patients, leaving us without a sufficient basis to accurately determine the sensitivity and specificity of functional and morphologic changes in the aortic valve. Although this bias is representative of disease pattern in Ugandan children, it may vary across settings, making these results less generalizable. The percentage of borderline patients in our cohort was small, resulting in a wide confidence interval for sensitivity for borderline disease. Larger numbers will be needed to more precisely define HAND’s ability to detect borderline disease, because these children

Journal of the American Society of Echocardiography Volume 27 Number 1

represent the largest group of those found through screening. Additionally, although the reviewer was blinded to all clinical information, the physician performing the screens did so consecutively and thus was not blinded to the clinical condition of the patients before handheld echocardiographic screening. Theoretically, although standardized views were obtained, this may have introduced bias to image acquisition. Finally, the population screened was not representative of the general population, as almost half of the patients came from an RHD follow-up clinic. The reviewer was aware of the high prevalence of RHD in this population, which may have caused bias toward reading more abnormalities. CONCLUSIONS HAND holds great promise as a tool in the global effort to control RHD. With time, handheld ultrasound devices likely will increase in functional capabilities, making full implementation of the 2012 WHF guidelines possible. We emphasize that given current specifications (1-mm measurement increments, lack of continuous-wave Doppler), HAND is more reasonably a tool for screening rather than for full implementation of the 2012 WHF guidelines. However, for a small trade-off in specificity, HAND could be applied to diagnosis in settings where secondary evaluation is resource limited. Additionally, future studies should examine how to maximize the sensitivity and specificity of HAND, which may include modification of the WHF criteria. It is likely that handheld ultrasound technology will continue to evolve. The second-generation Vscan has a threefold longer battery time than the first-generation system. Future developments might include even longer battery life, solar-powered systems, wireless (including 3G and 4G) transfer of images, wider angle of imaging, improved imaging quality and frame rate, high-frequency transducers, and fuller functionality, including spectral Doppler. If these technological advances can be made while maintaining affordability, HAND offers the promise of a highly portable, sensitive, and specific tool for widespread RHD screening. This study represents a crucial first step in assessing the utility of HAND for RHD screening. More research is needed to fully understand the performance of HAND for borderline RHD and among patients with aortic valve involvement. Importantly, our data provide clear evidence that the next step, a large-scale cross-sectional field study comparing HAND with STAND, is justified. Additionally, if HAND is ultimately validated for RHD screening, it will be important to develop and test educational modules for training primary health care workers in the use of HAND technology, as it is these in-country nonexperts who will likely compose the workforce worldwide. REFERENCES 1. Carapetis JR, Steer AC, Mulholland EK, Weber M. The global burden of group A streptococcal diseases. Lancet Infect Dis 2005;5:685-94. 2. Majeed HA, Batnager S, Yousof AM, Khuffash F, Yusuf AR. Acute rheumatic fever and the evolution of rheumatic heart disease: a prospective 12 year follow-up report. J Clin Epidemiol 1992;45:871-5. 3. Okello E, Kakande B, Sebatta E, Kayima J, Kuteesa M, Mutatina B, et al. Socioeconomic and environmental risk factors among rheumatic heart disease patients in Uganda. PLoS One 2012;7:e43917. 4. Bland EF, Duckett Jones T. Rheumatic fever and rheumatic heart disease; a twenty year report on 1000 patients followed since childhood. Circulation 1951;4:836-43.

Beaton et al 49

5. Gunther G, Asmera J, Parry E. Death from rheumatic heart disease in rural Ethiopia. Lancet 2006;367:391. 6. Oli K, Asmera J. Rheumatic heart disease in Ethiopia: could it be more malignant? Ethiop Med J 2004;42:1-8. 7. Diao M, Kane A, Ndiaye MB, Mbaye A, Bodian M, Dia MM, et al. Pregnancy in women with heart disease in sub-Saharan Africa. Arch Cardiovasc Dis 2011;104:370-4. 8. Feinstein AR, Stern EK, Spagnuolo M. The prognosis of acute rheumatic fever. Am Heart J 1964;68:817-34. 9. Roberts K, Colquhoun S, Steer A, Remenyi B, Carapetis J. Screening for rheumatic heart disease: current approaches and controversies. Nat Rev Cardiol 2013;10:49-58. 10. World Health Organization. Rheumatic fever and rheumatic heart disease: report of a WHO expert consultation. WHO technical report series. Geneva, Switzerland: World Health Organization; 2004. 11. Remenyi B, Wilson N, Steer A, Ferreira B, Kado J, Kumar K, et al. World Heart Federation criteria for echocardiographic diagnosis of rheumatic heart disease—an evidence-based guideline. Nat Rev Cardiol 2012;9: 297-309. 12. Prinz C, Voigt JU. Diagnostic accuracy of a hand-held ultrasound scanner in routine patients referred for echocardiography. J Am Soc Echocardiogr 2011;24:111-6. 13. Cardim N, Fernandez Golfin C, Ferreira D, Aubele A, Toste J, Cobos MA, et al. Usefulness of a new miniaturized echocardiographic system in outpatient cardiology consultations as an extension of physical examination. J Am Soc Echocardiogr 2011;24:117-24. 14. Razi R, Estrada JR, Doll J, Spencer KT. Bedside hand-carried ultrasound by internal medicine residents versus traditional clinical assessment for the identification of systolic dysfunction in patients admitted with decompensated heart failure. J Am Soc Echocardiogr 2011;24:1319-24. 15. Beaton A, Okello E, Lwabi P, Mondo C, McCarter R, Sable C. Echocardiography screening for rheumatic heart disease in Ugandan schoolchildren. Circulation 2012;125:3127-32. 16. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research Electronic Data Capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform 2009;42:377-81. 17. Wilson JMG, Jungner G. Principles and practice of screening for disease. Geneva, Switzerland: World Health Organization; 1968. 18. Kono Y, Fukuda S, Shimada K, Oe H, Maeda K, Kawasaki T, et al. Pocketsized echo for evaluation of mitral and tricuspid regurgitation. JACC Cardiovasc Imaging 2011;4:921. 19. Marijon E, Ou P, Celermajer DS, Ferreira B, Mocumbi AO, Jani D, et al. Prevalence of rheumatic heart disease detected by echocardiographic screening. N Engl J Med 2007;357:470-6. 20. Carapetis JR, Hardy M, Fakakovikaetau T, Taib R, Wilkinson L, Penny DJ, et al. Evaluation of a screening protocol using auscultation and portable echocardiography to detect asymptomatic rheumatic heart disease in Tongan schoolchildren. Nat Clin Pract Cardiovasc Med 2008;5:411-7. 21. Bhaya M, Panwar S, Beniwal R, Panwar RB. High prevalence of rheumatic heart disease detected by echocardiography in school children. Echocardiography 2010;27:448-53. 22. Saxena A, Ramakrishnan S, Roy A, Seth S, Krishnan A, Misra P, et al. Prevalence and outcome of subclinical rheumatic heart disease in India: the RHEUMATIC (Rheumatic Heart Echo Utilisation and Monitoring Actuarial Trends in Indian Children) study. Heart 2011;97:2018-22. 23. Spencer KT, Kimura BJ, Korcarz CE, Pellikka PA, Rahko PS, Siegel RJ. Focused cardiac ultrasound: recommendations from the American Society of Echocardiography. J Am Soc Echocardiogr 2013;26:567-81. 24. Shmueli H, Burstein Y, Sagy I, Perry ZH, Ilia R, Henkin Y, et al. Briefly trained medical students can effectively identify rheumatic mitral valve injury using a hand-carried ultrasound. Echocardiography 2013;30:621-6. 25. Enriquez-Sarano M, Tajik AJ, Bailey KR, Seward JB. Color flow imaging compared with quantitative Doppler assessment of severity of mitral regurgitation: influence of eccentricity of jet and mechanism of regurgitation. J Am Coll Cardiol 1993;21:1211-9.