One-year Outcomes and Factors Associated with Mortality Following Acute Myocardial Infarction in Northern Tanzania

Abstract

Background

Little is known about long-term outcomes and uptake of secondary preventative therapies following acute myocardial infarction (AMI) in sub-Saharan Africa.

Methods

Consecutive patients presenting with AMI (as defined by the Fourth Universal Definition of AMI Criteria) to a northern Tanzanian referral hospital were enrolled in this prospective observational study. Follow-up surveys assessing mortality, medication use, and rehospitalization were administered at 3, 6, 9, and 12 months following initial presentation, by telephone or in person. Multivariate logistic regression was performed to identify baseline clinical and sociodemographic factors associated with one-year mortality.

Results

Of 152 enrolled patients with AMI, 5 were lost to one-year follow-up (96.7% retention rate). Mortality rates were 34.9% (53 of 152 participants) during the initial hospitalization, 48.7% (73 of 150 patients) at 3 months, 52.7% (78 of 148 patients) at 6 months, 55.4% (82 of 148 patients) at 9 months, and 59.9% (88 of 147 patients) at one year. Of 59 patients surviving to one-year follow-up, 43 (72.9%) reported persistent anginal symptoms, 5 (8.5%) were taking an antiplatelet, 8 (13.6%) were taking an antihypertensive, 30 (50.8%) had been rehospitalized, and 7 (11.9%) had ever undergone cardiac catheterization. On multivariate analysis, one-year mortality was associated with lack of secondary education (OR 0.26, 95% CI 0.11–0.58, p=0.001), lower BMI (OR 0.90, 95% CI 0.82–0.98, p=0.015), and higher initial troponin (OR 1.30, 95% CI 1.05–1.80, p=0.052).

Conclusions

In northern Tanzania, AMI is associated with high all-cause one-year mortality, and use of evidence-based secondary preventative therapies among AMI survivors is low. Interventions are needed to improve AMI care and outcomes.

Introduction

As sub-Saharan Africa (SSA) proceeds through the epidemiologic transition,^1,2 the burden of acute myocardial infarction (AMI) across the region is increasing.^3,4 Despite this, there are little existing data describing AMI outcomes in SSA.⁵ Understanding AMI outcomes is essential for clinicians to identify interventions to improve AMI care and for policy makers as they allocate scarce healthcare resources.

Limited existing data regarding AMI outcomes in SSA comes primarily from in-hospital mortality data from retrospective studies at hospitals with percutaneous coronary intervention (PCI) capacity.^6–8 However, there are nearly no published data describing long-term post-hospital outcomes including mortality. To our knowledge, only one site, has reported prospective one-year AMI outcomes in all of SSA: Abidjan Heart Institute in the Ivory Coast, which has a robust cardiology service and PCI capacity.^9,10 Beyond this single site, however, there are no long-term prospective AMI data published from any country in the region, including in East Africa. Moreover, outcomes at advanced cardiac centers likely do not reflect outcomes across the region, where access to PCI and coronary artery bypass surgery remains extremely limited.^11–13 In Tanzania, a country with population exceeding 56 million, there is currently only one PCI-equipped hospital, which is located on the eastern coast.¹³

Moreover, there are few published data regarding use of evidence-based secondary prevention therapies following AMI in SSA. International guidelines emphasize long-term treatment with medications such as anti-platelet agents, beta blockers, ACE inhibitors, and statins, which are known to reduce morbidity and mortality among AMI survivors.^14–16 These guidelines also call for evidence-based non-pharmaceutical interventions such as smoking cessation, heart-healthy diets, and exercise.^14–16 Although there has been extensive study of adherence to secondary prevention guidelines among AMI patients in high-income countries,^17–19 to our knowledge there are no published reports describing long-term adherence to these therapies in SSA. Such data would allow for understanding gaps in evidence-based AMI care to identify targets for improvement.

There is an urgent need to describe long-term AMI outcomes in SSA, particularly in settings with limited access to advanced cardiac care and where rates of chronic disease are increasing. In Tanzania, prior study has demonstrated high mortality and frequent re-hospitalization within thirty days of initial AMI,²⁰ but long-term AMI outcome data in Tanzania is currently lacking. The purpose of this study was to examine one-year outcomes of mortality and secondary preventative therapy uptake among a cohort of AMI patients presenting to a hospital in northern Tanzania where coronary revascularization is currently unavailable.

Methods

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Setting

This study was conducted at Kilimanjaro Christian Medical Centre (KCMC), a tertiary care hospital whose catchment area comprises most of northern Tanzania. Community prevalence of hypertension and diabetes among adults living in the districts surrounding KCMC are 28% and 6%, respectively.^21,22 At the time of this study, KCMC was equipped with thrombolytics, but did not have PCI capacity and did not have any trained cardiologists on staff. The closest PCI-capable facility to KCMC at the time of this study was located in Dar es Salaam, Tanzania’s economic capital, approximately 10 hours away by road.¹³

Participant recruitment

Participants with AMI were consecutively enrolled via a prospective AMI screening study in the KCMC emergency department, with detailed methods published previously.⁴ Briefly, consecutive adults presenting to the emergency department with acute chest pain or shortness of breath were screened for AMI with an electrocardiogram (ECG) and point-of-care troponin-I testing (Abbott Point-of-Care, Princeton, NJ). Presence of AMI was defined using Fourth Universal Definition of AMI criteria:²³ Those with pathologic ST elevation in consecutive leads or pathologic troponin elevation were considered to have AMI. For study definition purposes, ECGs were adjudicated by two independent physician judges; a third physician served as a tiebreaker in cases of disagreement. All participants who met criteria for AMI were enrolled into the present cohort study.

Study procedures

Upon initial presentation to the emergency department, trained research assistants collected baseline data regarding sociodemographics, medical history, and presenting symptoms from each participant, and height, weight, and blood pressure were measured. Enrollment and initial surveys were conducted when patients were awaiting test results or admission; clinical care was not delayed by study procedures. Following enrollment, trained research assistants conducted a follow-up survey with participants 3, 6, 9 and 12 months after initial presentation. Follow-up surveys were administered via telephone; when participants were unreachable, a research assistant visited the participant’s home to administer the follow-up survey. If a participant died in the interim, the follow-up survey was administered to a relative. The follow-up survey comprised questions about symptom status, medication use, rehospitalizations, and barriers to care. To enhance the accuracy of medication use data, participants were asked to read the names of the medications directly from medication container; when in-home visits were conducted, research assistants requested to directly examine medications that were being taken by participants.

Study Definitions

Body mass index (BMI) was calculated directly from measured height and weight at enrollment. Baseline medical history, substance use, and medication use were self-reported. Participants who self-reported a history of attending secondary school or university were characterized as having post-primary education. Sedentary lifestyle was defined by self-report of less than 150 minutes of moderately intense exercise per week, as per World Health Organization recommendations.²⁴ Anginal severity at time of follow-up was graded by Canadian Cardiovascular Society guidelines.²⁵ Presence of heart failure symptoms was defined as a self-report of bilateral leg swelling and orthopnea. At follow-up, follow-up appointment attendance, dietary changes, regular exercise, tobacco use, re-hospitalization, and cardiac catheterization were defined by self-report.

Statistical Analyses

Descriptive statistics, including mean (standard deviation) and frequencies were used to characterize baseline participants and their outcomes. For mortality rate calculations, participants lost to follow-up were censored at time of loss (3, 6, 9, or 12 months). To identify factors associated with one-year mortality, univariate and multivariate analyses were performed on the following variables: participant age, sex, BMI, tobacco use, alcohol use, level of education (a surrogate for socioeconomic status^26,27), antiplatelet use at baseline, antihypertensive use at baseline, BMI, symptom duration prior to initial presentation, and initial serum troponin level. These variables were selected among available variables based on a priori assumptions about putative causal relationships with post-MI mortality and data from high-income settings.^27–29 Univariate associations between baseline participant characteristics and one-year mortality were assessed using Pearson’s chi-squared or Welch’s t-test. Multivariate logistic regression was then performed to assess factors associated with of one-year mortality. Any variable demonstrating evidence of univariate association with one-year mortality (p<0.10) was included in the model; age and sex were also included in the model a priori. Participants lost to follow-up were excluded from regression analyses. All statistical analyses were performed in the R Suite. An odds ratio with 95% confidence interval not including 1.0 or a p-value < 0.05 were considered statistically significant.

Ethical Approval

All study participants provided written informed consent prior to enrollment. This study was approved by ethical review committees at the Tanzania National Institute for Medical Research, Kilimanjaro Christian Medical Center, and Duke Health.

Results

Of 681 enrolled participants presenting with chest pain or shortness of breath, 152 participants had confirmed AMI, including 61 with ST-elevation myocardial infarction (STEMI) and 91 with non-STEMI. The baseline characteristics of participants with AMI are presented in Table 1. The mean (sd) age of participants was 61.2 (18.5) years, and 61 (40.1%) were female. The mean (sd) BMI among participants was 23.9 (4.4) kg/m², and 54 (35.5%) had completed post-primary education. At initial presentation, 18 (11.8%) participants were taking a daily aspirin and 42 (27.6%) were taking a daily anti-hypertensive medication. As previously reported,⁴ none of the participants received thrombolysis at initial hospital presentation, likely due to prolonged delays between symptom onset and hospital presentation.

Table 1.

Baseline characteristics of patients presenting to hospital with acute myocardial infarction, northern Tanzania, 2019 (N=152)

	Number of participants	(%)
Female sex	61	(40.1)
Age, years, mean (sd)	61.2 (18.5)
Post-primary education	54	(35.5)
Ongoing tobacco use	16	(10.5)
Ongoing alcohol use	54	(35.5)
Sedentary lifestyle	91	(59.9)
Body mass index, kg/m2, mean (sd)	23.9 (4.4)
Medication use at baseline
Aspirin	18	(11.8)
Anti-hypertensive	42	(27.6)
Statin	3	(2.0)
Antihyperglycemic	11	(7.2)
Initial serum troponin at baseline, ng/ml, mean (sd)*	0.97 (3.49)
Duration of symptoms prior to hospital presentation, days, mean (sd)	6.6 (12.2)
Duration of initial hospitalization, days, mean (sd)	5.3 (6.1)

^*Manufacturer-recommended 99^th percentile cut-off value 0.08 ng/ml

Over the course of the one-year follow-up period, 5 (3.3%) participants with AMI were lost to follow-up. As previously reported,^4,20 the in-hospital mortality rate among participants during the initial hospitalization was 34.9% (53 of 152 participants). Among participants completing follow-up, the mortality rate was 48.7% (73 of 150 participants) at three months, 52.7% (78 of 148 participants) at six months, 55.4% (82 of 148 participants) at nine months, and 59.9% (88 of 147 participants) at twelve months (Figure 1). Table 2 presents detailed outcomes among surviving participants at each follow-up. Throughout the follow-up period, a majority of surviving participants reported persistent anginal symptoms; at one year, 43 (72.9%) reported ongoing angina. At twelve months, 15 (25.4%) participants reported heart failure symptoms.

Figure 1.

All-cause mortality following acute myocardial infarction, northern Tanzania, 2019–2020 (N=152)^a

^a In-hospital mortality data previously published^4,20

Table 2.

Clinical status of surviving participants following acute myocardial infarction, northern Tanzania, 2019

	Participants surviving to 3 months (N=77)	Participants surviving to 6 months (N=70)	Participants surviving to 9 months (N=66)	Participants surviving to 12 months (N=59)
Symptoms
Persistent anginal symptoms	62 (80.5%)	55 (78.6%)	57 (86.4%)	43 (72.9%)
Severity of anginal symptoms
CCS* class I	25 (32.5%)	20 (28.6%)	13 (19.7%)	14 (23.7%)
CCS* class II	8 (10.4%)	17 (24.3%)	17 (25.8%)	5 (8.5%)
CCS* class III	19 (24.7%)	9 (12.9%)	16 (24.2%)	11 (18.6%)
CCS* class IV	10 (13.0%)	9 (12.9%)	11 (16.7%)	13 (22.0%)
Heart failure symptoms	26 (33.8%)	18 (25.7%)	21 (31.8%)	15 (25.4%)
Medications
Taking aspirin or other antiplatelet	6 (7.8%)	2 (2.9%)	4 (6.1%)	5 (8.5%)
Taking any antihypertensive	10 (13.0%)	7 (10.0%)	6 (9.1%)	8 (13.6%)
Taking any statin	2 (2.6%)	2 (2.9%)	1 (1.5%)	1 (1.7%)
Health behaviors
Currently using tobacco	3 (3.9%)	0 (0%)	1 (1.5%)	0 (0%)
Exercising regularly	47 (61.0%)	46 (65.7%)	44 (66.7%)	38 (64.4%)
Limiting salt intake	70 (90.9%)	64 (91.4%)	63 (95.5%)	53 (89.8%)
Limiting fatty foods	68 (88.3%)	60 (77.9%)	57 (86.4%)	52 (88.1%)
Health literacy
Feels they understand their diagnosis	43 (55.8%)	45 (64.3%)	48 (72.7%)	46 (78.0%)
Feels they understand their treatment	66 (85.7%)	61 (87.1%)	53 (80.3%)	53 (89.8%)
Healthcare
Received follow-up appointment in preceding 3 months	62 (80.5%)	41 (58.6%)	34 (51.5%)	28 (47.5%)
Ever re-hospitalized (cumulative)	26 (33.8%)	28 (40.0%)	29 (43.9%)	30 (50.8%)
Ever undergone cardiac catherization (cumulative)	5 (6.5%)	6 (8.6%)	6 (9.1%)	7 (11.9%)

^*CCS: Canadian cardiovascular society²⁵

Post-MI Preventative Therapy

Throughout the follow-up period, uptake of post-MI preventative therapies was low (Table 2). Among surviving participants, fewer than one in ten were taking an antiplatelet agent at any point over the one-year follow up period: 6 (8%) participants reported taking an antiplatelet medication at three months, 2 (3%) participants at six months, 4 (6%) participants at nine months, and 5 (9%) participants at one year. Similarly, use of any class of antihypertensive medication among surviving participants throughout the follow-up period was rare: 10 (13%) participants reported taking any antihypertensive at three months, 7 (10%) participants at six months, 6 (9%) participants at nine months, and 8 (14%) at one year. Almost no surviving participants reported taking a statin in the year following their MI: 2 (3%) were taking a statin at three months and only 1 (2%) was taking a statin at one year. By twelve months, none of the surviving participants reported tobacco use. Throughout the follow-up period, the majority of surviving participants reported exercising regularly, limiting salt intake, and limiting consumption of fatty foods (Table 2). Among participants surviving to one year, 30 (50.8%) had been rehospitalized and 7 (11.9%) had ever undergone cardiac catheterization.

Table 3 presents univariate associations between baseline participant characteristics and one-year mortality. Factors associated with one-year mortality on univariate analysis included post-primary education (OR 0.27, 95% CI 0.13–0.55, p<0.001), older age (p=0.039), lower BMI (p=0.019), and higher initial troponin (p=0.057). In multivariate analysis (Table 3), factors associated with one-year mortality were lack of post-primary education (OR 0.26, 95% CI 0.11–0.58, p=0.001), lower BMI (OR 0.90, 95% CI 0.82–0.98, p=0.015), and higher initial troponin (OR 1.30, 95% CI 1.05–1.80, p=0.052).

Table 3.

Univariate associations between baseline characteristics and one-year mortality among AMI patients, northern Tanzania, 2019 (N=147)

Baseline participant characteristic	Patients alive at one year (N=59), n (%) or mean (sd)	Patients dead at one year (N=88), n (%) or mean (sd)	Univariate OR (95% CI)	Univariate p	Multivariate OR (95% CI)	Multivariate p
Male sex	36 (61.0%)	51 (60.0%)	0.88 (0.45–1.73)	0.711	1.01 (0.46, 2.17)	0.989
Post-primary education	31 (52.5%)	20 (22.7%)	0.27 (0.13–0.55)	<0.001*	0.26 (0.11, 0.58)	0.001*
Tobacco use	7 (11.9%)	8 (9.1%)	0.74 (0.25–2.28)	0.586	--
Alcohol use	19 (32.2%)	32 (36.4%)	1.20 (0.60–2.45)	0.603	--
Sedentary lifestyle	37 (62.7%)	53 (60.2%)	0.90 (0.45–1.78)	0.762	--
Aspirin use at baseline	10 (16.9%)	8 (9.1%)	0.49 (0.17–1.35)	0.154	--
Anti-hypertensive use at baseline	19 (32.2%)	23 (26.1%)	0.75 (0.36–1.56)	0.425	--
Age, years	58.1 (15.3)	64.2 (20.0)	--	0.039*	1.01 (0.99, 1.03)	0.453
BMI, kg/m2	24.8 (4.1)	23.1 (4.4)	--	0.019*	0.90 (0.82, 0.98)	0.015*
Initial serum troponin at presentation, ng/ml	0.42 (1.26)	1.39 (4.44)	--	0.057	1.30 (1.05, 1.80)	0.052*
Symptom duration prior to initial hospital presentation, days	7.0 (14.6)	6.4 (10.6)	--	0.787	--

^*p<0.05 or odds ratio 95% confidence interval not including 1

AMI: Acute myocardial infarction

OR: Odds ratio

BMI: Body mass index

Discussion

In this observational cohort study, we present one of the first reports of long-term outcomes among AMI patients in SSA. At a Tanzanian referral hospital without PCI capacity, AMI is associated with high one-year mortality, use of evidence-based preventative therapies is infrequent, and re-hospitalization is common. These findings call attention to the need for improvements in AMI care in northern Tanzania, in particular with regard to access to PCI and uptake of secondary preventative therapies.

One-year mortality in this AMI cohort was an alarming 60%--among the highest one-year AMI mortality rates ever reported globally. Recent studies from the United States, Israel, Spain, Austria, Poland, China, and Brazil reported all-cause one-year AMI mortality rates ranging from 3 to 19%.^30–37 Thus the mortality rate observed in our cohort is far in excess of what is experienced in high-income settings outside of SSA. Caution is warranted, however, when making comparisons between mortality rates in Tanzania and high-income countries, as general life expectancy in Tanzania (65 years in 2019)³⁸ is lower than these comparator countries, which may explain some of the differences in observed mortality. Moreover, there may have been substantial case mix differences between our study population and study cohorts in other countries. To our knowledge, only one prior study has examined one-year mortality following AMI in SSA; in that study, conducted at a PCI-capable specialized cardiac hospital in the Ivory Coast, a 10% mortality rate was reported among AMI patients surviving to hospital discharge.⁹ The substantially higher mortality rate in our study may be due to lack of PCI access, lack of cardiology subspecialty care, lower use of evidence-based therapies, differences in patient populations, or inclusion of all patients regardless of survival to hospital discharge. Additional study is needed to explore the reasons for the particularly high mortality rate observed in northern Tanzania; additional study is also needed across SSA to determine whether similarly high rates of AMI-associated mortality exist in other settings without PCI access. Although studies in other settings have found sex-based differences in AMI mortality particularly in high-income settings,³⁹ we did not observe any association between sex and post-AMI mortality in Tanzania.

In addition to high mortality, we found low usage rates of evidence-based secondary prevention therapies in our cohort. International guidelines call for long-term treatment with medications such as aspirin, beta blockers, ACE inhibitors, and statins following AMI—guidelines supported by evidence showing morbidity and mortality benefits.^14–16 In northern Tanzania, however, fewer than 15% of AMI patients were taking any of these medications at any point in the year following their initial MI. By comparison, uptake of these evidence-based therapies is much higher in settings outside of SSA: a recent multicenter study examining post-discharge AMI outcomes in Europe, Asia, and Latin America found that substantial geographic variation in AMI care existed, but that approximately half of AMI survivors were taking dual antiplatelet therapy two years after hospital discharge, and two-year mortality was less than 8% in all settings.^40,41 Additional research is needed to explore the reasons for the low uptake of preventative therapies in Tanzania; possible explanations include lack of provider training, poor follow-up, cost of medications, inadequate patient education, medication availability, and low adherence, among others. Interventions are needed to address these and other barriers to increase use of these life-saving medications. Uptake of non-pharmacologic preventative interventions was substantially better than uptake of pharmacologic therapies in our cohort; very few patients reported tobacco use following their AMI, and the majority of patients reported exercising regularly and limiting salt and fat intake. Rehospitalization rates in our cohort were high, with more than half of surviving AMI patients rehospitalized within one year. Only 12% of the surviving AMI patients in our cohort underwent cardiac catheterization in the year following their AMI—an unsurprisingly low proportion given the lack of local cardiac catheterization lab. The low rates of cardiac catheterization likely explain in part why the majority of participants reported persistent anginal symptoms over the course of the year following their AMI. These findings call attention to the need for expansion of PCI access in Tanzania and across SSA—a need which has been well-documented.^11–13 In settings outside SSA, standardized quality indicators are used to monitor the quality of AMI care and identify areas for improvement.⁴² In SSA, a similar mechanism for establishing and monitoring contextually-appropriate quality indicators for AMI care may assist ministries of health to identify opportunities for improvement.

Factors associated with one-year mortality in our study included lack of secondary education, lower BMI, and higher initial troponin. The protective effect of post-primary education is perhaps not surprising, as there is a strong association between wealth and educational status in Tanzania.²⁶ In the United States, more years of education is associated with lower mortality following AMI, presumably due to socioeconomic disparities for which education is a surrogate.²⁷ Lower BMI was predictive of higher mortality in our cohort. The apparent protective effect of higher BMI on post-AMI mortality—dubbed “the BMI paradox”—has been noted in a wide-variety of high-income settings,^28,43,44 and remains poorly understood. In Tanzania, in a cohort where obesity was relatively uncommon, we found the same protective effect of higher BMI. Finally, higher initial troponin was predictive of higher risk of mortality in our cohort, consistent with what has long been observed in high-income settings where troponin level is considered a surrogate for infarct size.²⁹ Additional study is needed across SSA to identify other predictors of AMI mortality and develop interventions to mitigate their impact.

This study had several limitations. First, the study only included patients who survived to presentation at a referral hospital; thus patients from more isolated, rural settings were likely under-represented. Exclusion of patients who did not survive to hospital or did not have the capacity to travel to a referral hospital likely resulted in under-estimation of the mortality rate in the general population. Second, we relied on patient self-report for medication use, which may have resulted in an under-estimation of the use of evidence-based therapies. Similarly, we relied on patient self-report regarding tobacco use, exercise, and diet; social desirability bias may have led to an over-estimation of the proportion of patients engaging in heart-healthy lifestyle modifications. Additionally, we did not have data on some commonly-reported AMI outcomes, such as ejection fraction; such data may have allowed us to identify additional predictors of mortality. Furthermore, we assessed all-cause mortality and were unable to determine whether or not individual deaths were directly or indirectly related to the participant’s initial AMI. Finally, this was a single-center study, limiting generalizability.

In conclusion, we found that, in northern Tanzania AMI is associated with high one-year mortality and that use of evidence-based secondary preventative therapies over the year following AMI is uncommon, suggesting potentially modifiable barriers to improving AMI outcomes. Additional research is needed to understand the reasons for these poor clinical outcomes and to develop interventions to reduce AMI-associated morbidity and mortality.

What is Known

• Long-term clinical outcomes after acute myocardial infarction have been improving in high-income settings over recent decades.

• Little is known about long-term myocardial infarction outcomes in sub-Saharan Africa, where access to advanced cardiac care is limited.

What this study adds

• In northern Tanzania, 6 in 10 patients with acute myocardial infarction died within twelve months.

• In the year following acute myocardial infarction, few surviving patients underwent cardiac catheterization or regularly took secondary preventative therapies such as aspirin or antihypertensives.

• There are multiple opportunities to improve post-myocardial infarction care and outcomes in northern Tanzania.

Non-standard Abbreviations and Acronyms

SSA

sub-Saharan Africa

AMI

acute myocardial infarction

PCI

percutaneous coronary intervention

KCMC

Kilimanjaro Christian Medical Centre

ECG

electrocardiogram

BMI

body mass index

STEMI

ST-elevation myocardial infarction