Rheumatic heart disease (RHD) is a common and neglected health problem in the developing world and in indigenous regions of some high-income countries [1, 2]. Greater than 33 million people have RHD and nearly 320,000 die from the disease each year [3‐5]. In sub-Saharan Africa, studies from multiple countries report that approximately 0.5–3% of school-age children have echocardiographic signs of definite or borderline disease according to World Heart Federation (WHF) criteria [6‐9].

RHD is a preventable sequela of streptococcal pharyngitis, commonly known as “strep throat.” A small percentage of patients with untreated streptococcal pharyngitis will develop autoimmune-mediated acute rheumatic fever (ARF) and then RHD, which is characterized by heart valve damage and progressive cardiac failure [10]. No specific treatment for RHD exists but, if detected early, continuous antibiotic prophylaxis (also known as “secondary prevention” of RHD) can decrease ARF attack rates and mitigate progression of heart disease [11‐13]. Many patients will eventually require complex pharmaceutical regimens and surgical intervention [14]. Too often, RHD leads to chronic morbidity and premature death [15].

RHD has been poorly addressed in Zambia. Prevalence was estimated at 1.3% when measured nearly three decades ago by auscultation [16]. In 2001, a review of 200 case notes for pediatric cardiac patients at Lusaka University Teaching Hospital (UTH), Zambia’s main referral hospital, revealed that 70% were admitted for RHD (J Musuku, unpublished data). Currently, UTH provides services for approximately ten RHD outpatients each week and on average five inpatients daily. There are no specialized services for RHD available in the country outside of Lusaka.

After receiving specialty training in paediatric cardiology in South Africa, a UTH-based pediatrician mobilized academic colleagues, the Ministry of Health, local professional societies, and a pharmaceutical company in a multi-faceted effort to eliminate RHD in Zambia [17, 18]. Priorities of this program were structured according to the “A.S.A.P.” framework (Awareness-raising, Surveillance, Advocacy, and Prevention) established by the Pan-African Society of Cardiology as a strategy to reduce morbidity and mortality associated with RHD in sub-Saharan Africa [19, 20].

As part of this effort, the prevalence of RHD was measured in Zambian school children using handheld echocardiography [21‐23] and a recently validated simplified screening protocol based on the WHF criteria for echocardiographic diagnosis of RHD [24, 25]. In order to assure effective follow-up care for students identified to have RHD, selected health clinics in Lusaka were enrolled into a novel RHD control program that consisted of training, ongoing supportive supervision, and maintenance of adequate medicine stores [26, 27].

The opportunity of this study was also used to build local echocardiographic capabilities that could be applied to subsequent clinical and research activities in Zambia [28]. Data from this investigation were anticipated to inform health programs combating RHD and to support advocacy for RHD patients.

The sample population consisted of 2276 students in 15 schools that were eligible for participation. Consent was attempted for 2116 students that were in attendance in school on the day that consent forms were distributed (160 students were absent on that day). A total of 1102 students ultimately advanced through the consent process and underwent echocardiographic screening as part of the study (Fig. 3).

Sociodemographic characteristics of students included in the study revealed a slight predominance of female participants, and a majority of students from lower socioeconomic strata as determined by the neighborhood in which they lived (Table 1).

Most (86%) students were aged 13–18 years. Ten students were identified by study nurse-coordinators to have pharyngitis at the time of screening and were referred for medical follow-up. Fifteen students were found to have other medical problems during screening (including dental caries and rashes) and were also referred for medical evaluation. One student was found by screening echocardiography to have congenital heart disease.

Fifty-three students screened positive for RHD by portable echocardiography at schools and were referred for comprehensive echocardiography at UTH. Of those, 42 students followed-up and received confirmatory echocardiography. There were no gender differences in the cohort of children that screened positive. On confirmatory echocardiography, it was found that 3 students met WHF criteria for definite RHD, 10 students met criteria for borderline RHD, and 29 students did not meet criteria for any category of RHD. The confirmed prevalences of definite, borderline, and total RHD were therefore 2.7 per 1000 (95% CI 0.6–8.0), 9.1 per 1000 (95% CI 3.8–15.6), and 11.8 per 1000 (95% CI 5.7–15.6), respectively. Despite extensive and prolonged efforts by the study coordinators to follow-up students that screened positive, eleven students were lost to follow-up between the time they screened positive at schools and the time of a comprehensive echocardiographic assessment at UTH.

Demographic and clinical characteristics of the students that were diagnosed with definite and borderline RHD are shown in Table 2.

Each of these students had abnormalities of the mitral valve, with the most commonly observed abnormal morphological features being anterior mitral valve leaflet thickening and chordal thickening. The aortic valve was not affected in this group of students.

Prior to the main study initiation a formal assessment of the capabilities of the local radiographers to successfully implement the abridged echocardiographic screening protocol was conducted. The echocardiograms of 471 students screened by the local radiographers were reviewed by the VirtualScopics quality assurance (VSQA) specialist. The pairwise kappa test showed fair agreement between the local radiographers and the VSQA quality assurance radiographer (agreement, 97%; ϰ = 0.54; SE = 0.11). The results are listed in Table 3.

Of the 53 students referred for comprehensive echocardiography at UTH, 24 had screened negative at the school screening but the VSQA review indicated mitral regurgitation with jet length ≥ 2 cm in at least one view. From this cohort, 18 students followed-up with a confirmatory echocardiogram resulting in the diagnosis of one definite RHD and five borderline RHD cases. There was concordance between the local radiographers and the VSQA review for 28 students that screened positive, and 23 of these students followed-up with a confirmatory echocardiogram that resulted in the diagnosis of two definite RHD and five borderline RHD cases. One student screened positive at school and the VSQA review was negative; the confirmatory echo diagnosis was normal.

This study found a RHD prevalence of 11.8 per 1000 in Zambian school children. In addition, this study demonstrated the practicality of successfully training radiographers in echocardiography-based screening. The main finding that 1.2% of school children have evidence of borderline or definite RHD is consistent with investigations using similar methodologies in urban settings of other sub-Saharan African countries (Table 4).

The finding is also similar to the single previous estimate of RHD in Zambia that was obtained several decades ago albeit that study used different screening (auscultation alone) and diagnostic criteria (not utilizing current World Heart Federation criteria) [16]. It is notable that this research did not take place in isolation; rather, from the start it was incorporated into a broader health systems strengthening program that included the majority of government health clinics in Lusaka and aims to provide tailored support for the prevention and treatment of RHD in Zambia.

We trained radiography technicians that had little or no previous experience with echocardiography to perform school-based screening. This was necessary because health workers in Lusaka are not skilled in cardiac echocardiography except for a few health professionals at UTH. The feasibility of task-shifting simple echocardiographic screening protocols to non-expert health workers has been demonstrated in similar settings [38‐41]. In Zambia, this approach necessitated greater resources to conduct training and assess their results, and it also delayed the study until the quality of the screening echocardiograms could be assured. However, the study team felt strongly about this approach as the investment in time and resources was viewed as a method for building local echocardiographic capabilities to support subsequent clinical and research activities in Zambia. Hence the screening for each student involved in the study was conducted by a local radiographer. Ultimately, the performance of the local radiographers in the study was characterized by significant inconsistencies when compared with the VSQA quality assurance review. Without the VSQA review, the school screening process would have missed nearly half of the students diagnosed with definite or borderline RHD. The challenges that were faced in providing local radiographers with the skills required to confidently and independently conduct screenings for RHD will be an important consideration in related future research in Zambia and may have implications for the design of similar research programs in other countries.

Various simplified echocardiography protocols have been utilized to facilitate population-based screening for asymptomatic RHD [42, 43], and these are now being used in conjunction with ultra-portable handheld echocardiography devices in an effort to further streamline the screening process [44, 45]. In this study we used the recently validated FOCUS screening protocol, which was previously shown in a controlled investigation to be 92% sensitive and 100% specific for definite RHD [24]. This protocol, which utilizes a single view and measurement, was particularly advantageous in the current study in which screening was conducted by general radiographers who lacked echocardiography experience. Of the 42 students that screened positive and underwent comprehensive follow-up echocardiography, 31% (n = 13) were diagnosed with definite or borderline RHD according to WHF criteria. We attribute the decreased specificity to “real-world” factors that include uncertainty with some measurements obtained in the “field” (i.e., at schools) and the motivation by first-time study radiographers and coordinators to be conservative in their screening assessments of children to minimize the risk of missing a student who may have RHD. The number of referred patients for confirmatory echocardiography did not impose an excessive burden on the study team.

Just over half of students that received a consent form and questionnaire were ultimately screened (Fig. 3). Of the consent forms that were distributed to school children to take home to their parents, 36% were not returned or were returned blank. Nine percent of families opted out of participation. The consent form included a questionnaire with 11 questions to obtain demographic information and the child’s health history, with a focus on the incidence and treatment of sore throat. We hypothesize that completing the questionnaire may have been perceived as burdensome to some parents/guardians. Also, students may have forgotten to give the form to their parents, or failed to bring a completed form back to school. Absenteeism was also a factor on screening day (5.7% of students).

A novel education and training program was, for the first time in Zambia, implemented in Lusaka clinics by the study team in order to assure appropriate follow-up clinical care for participants identified through the study to have RHD [26]. The program comprised an introductory on-site training workshop, dissemination of educational materials for staff and patients, ongoing supportive supervisory visits by UTH staff, and assessment of antibiotic stocks with replenishment of benzathine penicillin when necessary. Key elements of the program included guidance on when and how to administer secondary antibiotic prophylaxis, how to manage the unlikely event of penicillin allergy, and how to deliver primary prevention of RHD (i.e., detection and proper treatment of bacterial pharyngitis). At the time of the investigation eight clinics were enrolled that were selected specifically due to their geographic proximity to study schools. Since then the program has grown to include all 29 government clinics in the Lusaka municipality, and activities are underway to expand the RHD control program to other provinces beyond Lusaka. Future work will involve the exploration of innovative ways to scale up the RHD control program (such as the use of electronic training modules), and to provide broad access to educational materials through the Pan African Society of Cardiology (PASCAR).

Borderline RHD is an echocardiography-based diagnostic category established in recent years by the World Heart Federation to improve the sensitivity of early detection of RHD in individuals ≤ 20 years [25]. In theory, early diagnosis paves the way for clinical interventions that prevent or mitigate disease progression, although this has not been proven in practice [46]. Because this category is new and its natural history is not yet fully understood, a major research priority has been to determine optimal management of individuals with borderline RHD. In the current study participants with borderline RHD were offered antibiotic prophylaxis along with routine clinical follow-up and periodic echocardiography, an approach that was taken in some previously described programs [6, 8] and not others [7, 47]. A recent prospective observational study of 34 school children in South Africa diagnosed with borderline RHD were followed up after 5 years [48]; 16% continued to have borderline RHD and 16% progressed to definite RHD, which suggests that initiation of antibiotic prophylaxis in children with borderline disease may have clinical benefit if antibiotic prophylaxis prevents progression of latent RHD. A similar conclusion was drawn in a prospective cohort study of indigenous Australian children with borderline RHD who were found to be at increased risk of progression to definite RHD [34]. At this time the approach in Zambia is to initiate antibiotic prophylaxis in individuals with borderline RHD and to discontinue prophylaxis on a case-by-case basis according to each patient’s clinical evolution. We will refine these tactics as new data and recommendations from leading health authorities become available.

There were some limitations in this study. Eleven participants who screened positive for RHD at schools were lost to follow-up. In most cases, parents or guardians were successfully contacted but ultimately could not be persuaded to bring their children for a confirmatory echocardiogram at UTH. This was despite offers of travel reimbursement and numerous attempts over three months or more by study staff to facilitate follow-up, including repeated telephone calls and in-person visits back to schools to discuss with families and school staff. In many cases, appointments were made for follow-up visits but not kept. If future screening studies are conducted in Zambia, we aim to incorporate on-site confirmatory echocardiography such that definitive diagnoses can be made at the time of screening and obviate the need for subsequent visits. Another limitation of the study is that the population screened was limited to school-going children in the urban setting of Zambia’s capital; this group may not be representative of the entire country. The study was limited to Lusaka mainly because of resource considerations, including clinical staff that would be available to conduct the study. Access to healthcare is generally accepted to be greater in Lusaka than in rural and remote regions of the country, which may result in improved treatment of streptococcal pharyngitis and prevention of RF and RHD. Indeed, more patients with RHD are reportedly referred to UTH from communities that are located outside of Lusaka, although there is not yet a national patient registry or surveillance system in place. Finally, while the screening protocol was associated with a validated sensitivity above 90%, there remains the possibility that some participants with potential RHD erroneously screened negative which would have led to an underestimation of the disease prevalence in the population studied.

The prevalence of RHD in Zambia is similar to that reported in neighboring countries and other low-resource settings globally. This information is being used to inform ongoing advocacy and education efforts designed for local health authorities, health workers, and the public. The abbreviated echocardiography screening protocol performed well in practice and may be considered for use in future prevalence studies. The conduct of this investigation was viewed as an opportunity to incorporate capacity-building for the local radiographers and to launch primary and secondary prevention programs in local clinics. The authors recommend prioritizing this approach whenever possible, even if it requires greater investment in time and other resources.