Development of an inclusive and cost-effective Afrocentric predictive model for stroke: A novel approach

doi:10.21203/rs.3.rs-3319744/v2

Download PDF

Research Article

Development of an inclusive and cost-effective Afrocentric predictive model for stroke: A novel approach

https://doi.org/10.21203/rs.3.rs-3319744/v2

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background: Prediction of stroke is indispensable for the initiation of the necessary preventive and prophylactic measures. The conspicuous omission of African data in most of the validated prediction models questions their appropriateness for the region. This study therefore aimed to develop an inclusive and cost-effective Afrocentric predictive model for stroke (CAPMS). Methods and Analyses: We employed an evidence synthesis approach namely meta-analysis structured as per the Preferred Reporting Item for Systematic Reviews and Meta-analyses (PRISMA) guidelines. We searched PubMed, Scopus, African Journals, Medline, Cochrane Library, Web of Science and Cumulative Index for Nursing and Allied Health Literature, from inception to date. Case-control and cohort studies that reported the risk factors of stroke and corresponding risk estimates were included. Screening of the titles and abstracts was undertaken by two independent reviewers. We conducted meta-analyses using Comprehensive Meta-analysis (CMA) version 3. Result: Majority of the eligible studies (>50%) investigated both ischemic and haemorrhagic stroke. In Africa, more than 20 risk factors of stroke were identified however only 18 met the eligibility for meta-analysis. Our findings shows that homocysteine (Rw = 13.9, Ri = 0.67), hypertension (Rw = 5.6, Ri =0.94) and cardiac events (Rw =3.1, Ri = 0.8) were the three most powerful independent predictors of stroke in Africa. Low consumption of green vegetables (Rw =2.4, Ri = 1.0), stress (Rw = 1.76, Ri =1.0) and hypertension were the most clinically responsive risk factors of stroke. All biomarkers except homocysteine were cost-effective having an investigative price of less than $13.6. Critical risk point was fitted at the 90th percentile being 12.7. The cumulative Rw and investigative costs of CAPMS 1 (15.8 and $2.4) and CAPMS 2 (14.9 and $7.2) showed good performance index and cost-effectiveness. Conclusion: Targeted screening with the CAPMS 1 & CAPMS 2 model represents cost-effective innovation in practice of stroke screening in African clinics and communities. We recommend immediate validation of CAPMS to ascertain its performance, feasibility and acceptability in the region.

Stroke

cardiovascular disease

prediction

secondary prevention

Africa

Cerebrovascular accident also known as stroke is a major public health concern and the leading cause of adult-onset disability, and dementia [1]. Annually, 15 million people worldwide suffer a stroke, of which an estimated 5 million die and an additional 5 million are disabled [2]. There is a prominent geographical inequality in age-standardized stroke incidence, mortality, prevalence and disability-adjusted life year rates across the globe. The highest rates are recorded in low and middle-income countries (LMIC), particularly in sub-Saharan Africa [3, 4]. Of note, approximately 80% of all global stroke mortalities occur in sub-Saharan Africa, and this burden is expected to increase in the coming decades [3, 4]. Comparably, in the preceding four decades the incidence of stroke decreased by 42% in high-income countries (HICs) but increased to more than 100% in LMICs [5, 6]. This disparity in incidence rates is indicative of the dissimilarities in stroke pathogenesis and risk factors, preventive care, and funding [7]. In HICs, research funding, healthcare insurance, and preventive care including lifestyle modification are more accessible [8], in contrast to Africa where curative care remains the mainstay [9]. Data emanating from the Eastern North Carolina study which reported a 12% higher stroke mortality than the rest of the state, confirms a reduction in stroke prevalence and mortality following the completion of 4900 community outreach risk factor screenings between 2007 and 2011[10], thereby accentuating the advantages of preventive strategies.

We anticipate a major shift from curative to preventive approaches and health promotion as the first line of action in the region, which is consistent with the sustainable development goal target 3.4 [11]. However, effective preventive measures must be predicated upon adequate characterization and stratification of traditional and emerging risk factors [4, 12]. Globally, several stroke risk factors such as atrial fibrillation, hypertension, dyslipidaemia, and diabetes mellitus, have been identified [13] and risk assessment models including the Framingham risk score have been developed to predict stroke [14]. Noteworthy, of all the current validated risk assessment models available, only one Afrocentric model was built using African data thus questioning their appropriateness for the region. Moreover, major limitations including the lack of systematic risk factor stratification, global insensitivity and cost-ineffectiveness emerged in all models, resulting in the assimilation of clinically insignificant risk factors, and limited external validity and accessibility [15, 12]. More specifically, the Afrocentric model was built on Nigerian and Ghanaian data, with only seven factors included in the model [4]. This study, therefore, aims to develop an Afrocentric predictive model for stroke using a novel evidence synthesis approach in an attempt to ensure inclusiveness and cost-effectiveness.

PECOT criteria

The population of the review included individuals without stroke followed up until the stroke occurred as in a prospective cohort, as well as individuals with or without stroke whose case files were retrospectively examined as in a retrospective cohort of individuals with or without stroke who were examined prospectively as in a case-control study [16]. Primary outcomes included stroke factors namely hypertension, obesity, diabetes, age, etc., as well as risk estimates and risk of bias score. Additional primary outcomes were derived risk estimates namely risk stability index (Ri), and risk weight (RW) [12]. The review covered the time between the inception of the oldest database (1961) and 10th May 2023.

Protocol and registration

This protocol is registered with the International Prospective Register of Systematic Reviews (PROSPERO; ID: CRD42023430437). The study utilized a systematic review with a meta-analysis approach to develop the Afrocentric predictive model and a prospective observational design to validate the emerging model. The protocol was structured using the Preferred Reporting Item for Systematic Reviews and Meta-analyses (PRISMA) checklist [17] and adhered to the recommendations of Meta-analysis of Observational Studies in Epidemiology (MOOSE) [18].

Eligibility Criteria

Inclusion criteria

Studies were considered if they met the following inclusion criteria: (1) must be peer-reviewed African studies; (2) examined risk factors of stroke patients; (3) used case-control (CS); cohort design (C) or systematic review of CS and/or CS; (4) were of low or moderate risk of bias; (5) reported the risk factor and corresponding risk estimates (Odds ratio) or provide sufficient information for the risk estimate to be calculated and transformed into odds ratios; (6) were written and published in English or French and (7); the risk factor had to be reported by at least three studies to be included in the meta-analysis. There was no restriction regarding publication country, race, and gender.

Exclusion criteria

Studies were excluded if it focused on (1) cross-sectional studies, (2) demonstrated high risk of bias, and (3) published in other languages other than English and French, and (4) studies that omitted the risk estimate (i.e., odds ratio or its equivalent) despite reporting the prevalence of risk factors of stroke.

Outcome measures

Outcome measures may include laboratory and non-laboratory measures. The outcomes included stroke risk factors and corresponding risk estimates. Stroke risk factors may include hypertension, obesity, atrial fibrillation, low physical activity, etc.

Sources of Information

In line with Medline, PubMed. Cochrane Library, CINAHL, Web of Science, Scopus and African Journal (SABINET) from 1966 to 10th May 2023. A draft MEDLINE search strategy developed by the subject librarian and the primary investigator is provided in Appendix 1.

Search strategy development

An information specialist (SBM) together with the NM developed, tested, and refined the search strategy. Search terms were identified from the key concepts in the review model. All searches were executed using a range of terms from the medical subject categories. A PubMed pilot search was carried out to ascertain the appropriateness of the search strings (Appendix 1). The search terms were then modified to fit the syntax and subject headings of the other databases, which included Medline, and PubMed. Cochrane Library, CINAHL, Web of Science, Scopus and African Journal (SABINET). Apart from MeSH terms, keywords were searched in the title, abstract and keyword fields. Wildcards and truncation for the keywords were also implemented. A combination of index terms and keywords as well as Boolean operators, (AND, OR and NOT) were implemented to assist in retrieving the most comprehensive results.

Data Management

The results retrieved for the literature search were exported to EndNote 20 and duplicates were eliminated. After removing duplicates, articles were subjected to title and abstract screening using EndNote 20. Included and excluded articles were exported to and organised in EndNote 20 for the generation of the PRISMA flow chart and in-text citations.

Study selection and data extraction

The principal investigator (MN) and co-author (PO) undertook the initial title and abstract screening. Training of PO was undertaken by MN. Discussions were used to resolve divergent points of view. Data was then extracted by PO and verified by MN. None of the corresponding authors of the full-text articles were contacted as the required information was found within the texts. The PRISMA diagram displays the details regarding the flow of studies throughout the selection process, as well as the rationale for exclusion.

Data Items

Primary basic data were risk factors and their corresponding estimates (odds ratio or proportion of events for computation of OR) as well as risk of bias score; and subsequently included risk responsiveness, risk weight and critical risk point. Secondary data included study characteristics such as type of stroke, setting, study design, sample size, etc.

Risk of Bias Assessment

We undertook a risk of bias using the Joanna Briggs Institute (JBI) risk of a bias assessment instrument for case-control and cohort designs [19]. Both MN and PO independently appraised 20% of the studies and established a good-inter-rater agreement. We denoted inter-rater agreement as the ratio of the total number of articles correctly assessed by both raters to the total number of articles assessed. The remaining publications (80%) were evaluated by PO.

Summary Measures

Mean and standard deviation was used to describe the quantitative characteristics of the participants in the included studies. The participants' educational achievements and gender were summarized in percentage. The critical risk point was calculated in line with Nweke et al. [12] and Nwagha & Nweke [20].

Data synthesis and analysis

For ease of comparison and where appropriate, all risk estimates were converted to odds ratio. Sociodemographic characteristics constitute the secondary data in line with previous reports [21, 22]. We converted OR to 1/OR when OR < 1 to place all effects in a common frame [23]. Similarly, to frame the confidence interval for OR < 1, we used a multiplier effect [1/OR²]. Using the random-effect meta-analysis model, the pooled odds ratio and measures of heterogeneity including I² and publication bias were estimated per risk factor. Afterwards, relative risk weight and critical risk point were located in the fourth quartile. Comprehensive Meta-Analysis (CMA, version 4) software and the statistical package for the social sciences (SPSS) were used to conduct the meta-analysis. The level of significance was set at α = 0.05.

Determining cost-effectiveness

A transaction is cost-effective when the greatest benefit is gained for a comparatively low price [24]. The computation of the critical risk point allows the model to be built from any prudent combinations of clinically relevant and responsive risk factors/biomarkers, provided that such combinations meet the critical point [20]. In our study, we contextualized prudence as cost-effectiveness and high model performance. The definition of cost-effectiveness was based on the average investigative cost associated with the putative model (clinically relevant and responsive) factors and/or biomarkers obtained in representative African regions (West Africa, Southern Africa, East Africa, and North Africa) [20]. Data on the cost of risk factors of interest, namely, HIV testing, blood glucose, lipid profile, ECG and homocysteine was obtained from selected African laboratories in Nigeria, South Africa, Morocco, and Uganda. Data were sampled from the selected from the established laboratories directly or by proxy. Prices were obtained and converted into US dollars for ease of comparability [20]. A risk factor otherwise known as the model component was said to be cost-effective if the associated investigative cost was within the average cost. Similarly, a model was said to be cost-effective if the associated investigative cost was within the average cost.

Study selection and characteristics

We initially identified 4517 records. Following de-duplication, title and abstract screening, only 61 records were left for full-text review. Of the 61 full text publications, 29 were excluded. Our final review included 33 articles (27 case-control, 5 cohorts and one systematic review), and encompassed 38544 African participants (Fig. 1). Of the 33 studies, 13 (39.4%) were conducted in West Africa, twelve (36.4%) in North Africa, three (9%) in East Africa, one involved selected countries from the sub-Saharan African countries [25], two included participants from all regions of Africa [26, 27], and two involved participants from East and Southern African [28, 29]. Noteworthy, 17 (51.5%) investigated both ischemic and haemorrhagic stroke, 14 (42.4%) investigated only ischemic stroke while two (6%) examined only haemorrhagic stroke (Table 1). We included 26 articles in the meta-analysis of 18 risk factors (Fig. 1) and excluded risk factors which were reported once or twice. Risk factors thus excluded from the meta-analyses were regular sugar consumption (2) history of stroke (2), unemployment (2), alleles (7 different alleles each reported once) and depression (1) (Appendix 2).

Table 1

Characteristics of the studies included in the review
Authors	Study Design	Method of data collection	Stroke type	Sample size	Country	Region
Akpalu et al. [4]	Case-control	Retrospective	Ischaemic and intracerebral haemorrhage	3553	Nigeria and Ghana	West Africa
Elagib et al. [30]	Case-control	Retrospective	Thrombotic CVA	400	Sudan	North Africa
O'Donnell et al. [26]	Case-control	Prospective	Ischaemic and intracerebral haemorrhage	1949	Africa (Mozambique, Nigeria, South Africa, Sudan, and Uganda),	All region
Owolabi [31]	Case-control	Retrospective	Ischaemic and haemorrhagic	252	Nigeria	West Africa
Owolabi [32]	Cohort	Retrospective	Ischaemic	702	Nigeria	West Africa
Chehaibi et al. [33]	Case-control	Retrospective	Ischaemic	388	Tunisia	North Africa
de Mast et al. [34]	Case-control	Retrospective	Ischaemic	527	Tanzania	East Africa
Fekih-Mrissa et al. [35]	Case-control	Retrospective	Ischaemic	184	Tunisia	North Africa
Saidi et al. [36]	Case-control	Retrospective case-control	Ischaemic and haemorrhagic	773	Tunisia	North Africa
Saidi et al. [37]	Design missing	Retrospective	Ischaemic	498	Tunisia	North Africa
Saidi et al. [38]	Case-control	Retrospective	Ischaemic and haemorrhagic	253	Tunisia	North Africa
Akpa et al [39]	Case-control	Retrospective	Ischaemic and haemorrhagic	7066	Ghana and Nigeria	West Africa
Akpalu et al. [40]	Case-control	Retrospective	Ischaemic and haemorrhagic	1998	Ghana	West Africa
Amu et al. [41]	Case-control	Prospective case-control	Ischaemic, intracerebral and subarachnoid haemorrhage	160	Nigeria	West Africa
Diakite et al. [42]	Case-control	Retrospective	Ischaemic	381	Morocco	North Africa
Diakite et al. [43]	Case-control	Retrospective	Ischaemic	376	Morocco	North Africa
Fekadu et al. [44]	Cohort	Prospective	Ischaemic and haemorrhagic	116	Ethiopia	East Africa
O’Donnell et al. [27]	Case-control	Prospective	Ischaemic and haemorrhagic	646	African (Mozambique, Nigeria, South Africa, Sudan, and Uganda).	All regions
Saidi et al. [45]	Case-control	Retrospective	Ischaemic stroke	551	Tunisia	North Africa
Okubadejo et al. [46]	Case-control	Unamed	Ischaemic stroke	115	Nigeria	West Africa
Chang et al. [29]	Case-control	Prospective	Ischaemic and haemorrhagic	772	Africa (Kenya, Zambia, Zimbabwe)	Southern Africa
Poulter et al. [28]	Case-control	Prospective	Ischaemic	56	Africa (Kenya, Zambia, Zimbabwe)	Southern Africa
Mbonde et al. [47]	Cohort	Prospective	Intracerebral haemorrhage	73	Uganda	East Africa
Namale et al. [25]	Systematic review	Systematic review	Ischaemic and haemorrhagic	2003	Angola, Benin, Botswana, Burkina Faso, Burundi, Cameroon, Cape Verde, Central Africa Republic, Chad etc	Sub-Saharan Africa
Saidi et al. [48]	Case-control	Retrospective	Ischaemic stroke	773	Tunisia	North Africa
Salem-Berrabah et al. [49]	Case-control	Retrospective	Ischaemic stroke	147	Tunisia	North Africa
Sarfo et al. [50]	Cohort	Prospective	Ischaemic and haemorrhagic	3296	Ghana	West Africa
Sarfo et al. [51]	Case-control	retrospective	Haemorrhagic and ischaemic stroke	2118	Nigeria and Ghana	West Africa
Sarfo et al. [52]	Case-control	Prospective	Haemorrhagic and ischaemic stroke	1080	Nigeria and Ghana	West Africa
Sarfo et al. [53]	Cohort	Prospective	Ischaemic and haemorrhagic stroke	1365	Ghana	West Africa
Sarfo et al. [54]	Case-control	Retrospective	Intra cerebral haemorrhage	2944	Ghana and Nigeria	West Africa
Sarfo et al. [55]	Case-control	Retrospective	Ischaemic stroke	2431	Ghana and Nigeria	West Africa
Walker et al. [56]	Case-control	Prospective	Ischemic and haemorrhagic	598	Tanzania	North Africa

Risk factors and categories of stroke

Eighteen studies investigated hypertension as a risk factor for stroke: 12 examined patients with ischaemic and haemorrhagic stroke, five investigated ischaemic stroke only, and one studied haemorrhagic stroke only. Of the seventeen studies that investigated diabetes as a risk factor for stroke, a trend similar to hypertension as a risk factor was observed, viz., 10 investigated patients with ischaemic and haemorrhagic stroke, six with ischaemic stroke, and one with haemorrhagic stroke. Likewise, seventeen studies investigated dyslipidaemia as a risk factor for stroke, revealing nine investigating patients with ischaemic and haemorrhagic stroke, seven with ischaemic stroke and one with haemorrhagic stroke. Smoking status as a risk factor for stroke was investigated in 14 studies, of which 11 included patients with ischaemic and haemorrhagic stroke, two with ischemic stroke only and one with haemorrhagic stroke. Of note, 10 studies reported cardiac causes (i.e., atrial fibrillation, myocardial infarction, left ventricular hypertrophy etc.) as a risk factor: eight evaluated patients with both ischaemic and haemorrhagic stroke and two examined patients with ischaemic stroke only. Of the seven studies that investigated alcohol consumption as a risk factor, only one examined ischemic stroke whereas 6 involved patients with ischaemic and haemorrhagic stroke. Physical activity was identified as a risk factor of stroke in six studies, which all involved ischaemic and haemorrhagic stroke. In the ten studies that investigated obesity as a risk factor for stroke, eight involved patients with ischaemic and haemorrhagic stroke in contrast to 2 that involved patients with ischaemic stroke only. Six studies reported stress as a risk factor: of which four involved patients with ischaemic and haemorrhagic stroke, one investigated ischaemic stroke and one haemorrhagic stroke only. Eight studies reported age as a risk factor, out of which six involved patients with ischaemic and haemorrhagic stroke, and two included patients with only ischaemic stroke. The distribution of various stroke categories for the remaining eight risk factors are detailed in Appendix 2.

Risk Stratification and the critical risk point

Of the 18 factors sampled for meta-analysis, age (Ri = (0.25), sex (0.12) and education (0.25) were considered clinically non-responsive, having Ri values less than 0.5. Homocysteine (Rw = 13.9, Ri = 0.67) and hypertension (Rw = 5.6, Ri = 0.94) were the most important risk factor for stroke. Low consumption of green vegetables and hypertension were the most clinically responsive factors with a perfect and near-perfect risk responsiveness of Ri = 1.0 and Ri = 0.94, respectively. Other clinically relevant risk factors include cardiac disease (atrial fibrillation, myocardial infarction & other heart diseases: Rw = 3.1, Ri = 0.8), diabetes (Rw = 2.5, Ri = 0.76), Low consumption of green vegetables (Rw = 2.4, Ri = 1.0), dyslipidaemia (Rw = 2.4, Ri = 0.83), and high level of stress (Rw = 1.76, Ri = 1.0), HIV (Rw = 1.70, Ri = 0.67). Low HDL (Rw = 3.3) and high total cholesterol (1.70) were identified as the most important risk factors as a cause of dyslipidaemia (Table 2). The critical point for the detection of stroke lies within the last (4th ) quartile. Based on the distribution of risk weight across risk factors, the critical point was calculated as the last fourth quartile between the 76th and 100th percentile (12.7) of the upper quartile class. Hence, a cumulative risk weight of ≥ 12.7 may possess the highest diagnostic accuracy for stroke and could mark the critical point at which pre-emptive interventions should be instituted (Table 2).

Table 2

Risk factors and/or biomarkers of stroke
Factors	Reference Category	Ri	OR	95% CI	p-value	I²	Eggers’ test t- value p		N-factor	Risk category
Dyslipidaemia	TC ≥ 5.2mmol/L, HDL ≥ 1.03mmol/L, LDL ≥ 3.4mmol/L or TAG ≥ 1.7mmol/L	0.83	2.84	1.84–4.38	< 0.001	85.24	0.503	0.642	2.36	Upper middle class
Hypertension	History of HTN or sustained BP ≥ 140/90mmHg for 72hrs	0.94	5.94	3.93–8.98	< 0.001	95.59	2.167	0.046	5.58	Upper class
Diabetes	History of diabetes or HbA1C ≥ 6.5%	0.76	3.23	2.63–3.97	< 0.001	73.55	0.713	0.487	2.45	Upper class
Alcohol consumption	Current alcohol consumption especially high/heavy episodic intake (seven or more drinks per week)	0.57	1.76	1.39–2.24	< 0.001	47.71	0.150	0.887	1.00	Low middle class
Smoking status	Previous/current smoker	0.64	2.22	1.69–2.92	< 0.001	54.196	2.323	0.039	1.42	Low middle Class
Physical activity	Physical inactivity (Less than 4hrs moderate intensity/strenuous leisure activity per week)	0.67	1.78	1.60–1.99	< 0.001	1.975	1.493	0.210	1.19	Low middle Class
Obesity	BMI ≥ 30Kg/m²	0.38	1.65	1.13–2.39	0.009	88.931	0.746	0.484	0.63	Lower class
Salt intake	Table added salt	0.33	1.62	1.13–2.31	0.009	< 0.001	SE= 0	≥ 1.00	0.53	Lower class
Meat consumption	Regular meat consumption	0.75	1.78	1.51–2.10	0.874	< 0.001	1.400	0.296	1.34	Low middle class
Stress	High level of stress (combined measures of stress at home/work (e.g. irritability, anxiety or sleeping difficulties) and life events, experienced in the 2 weeks preceding)	1	1.76	1.48–2.08	< 0.001	31.121	0.272	0.799	1.76	Upper middle class
Monthly income	Monthly income < 100USD	0.6	1.60	1.38–1.85	< 0.001	< 0.001	0.598	0.592	0.96	Lower middle class
Age	NSF	0.25	1.08	0.98–1.18	0.132	70.741	2.648	0.038	0.27	NSF
Cardiac causes (Atrial fibrillation, myocardial infarction, prosthetic heart valve, rheumatic valvular heart disease etc)	Previous history, review of baseline electrocardiograph results	0.8	3.91	2.51–6.08	< 0.001	76.751	1.022	0.336	3.13	Upper class
HIV	HIV⁺	0.67	2.53	1.25–5.13	0.010	69.727	0.132	0.916	1.70	Low middle class
Total cholesterol	≥ 5.2mmol/L	0.57	2.98	1.33–6.69	0.008	95.485	1.512	0.191	1.70	Low middle class
Gender	Male gender	0.12	1.25	1.05–1.49	0.011	< 0.001	1.449	0.221	0.15	Lower class
Triglycerides	TAG ≥ 1.7mmol/L	0.33	1.56	1.16–2.08	0.003	< 0.001	0.029	0.982	0.51	Lower class
Education	No formal education	0.25	1.32	1.01–1.62	0.009	< 0.001	1.148	0.370	0.33	Lower class
Consumption of green vegetables	Low consumption of green leafy vegetable	1	2.44	1.87–3.19	< 0.001	49.280	0.450	0.697	2.44	Upper middle class
Family history of CVD (including stroke)?	YES	0.8	1.88	1.41–2.51	< 0.001	44.559	1.210	0.313	1.50	Upper middle class
HDL	HDL ≥ 1.03mmol/L	1	2.48	0.72–8.55	0.151	84.113	0.381	0.768	3.30	Upper class
Homocysteine	Abnormally elevated Hcy i.e. tHcy > 15 micromol/L	0.67	20.80	1.05-41	0.047	95.047	0.056	0.964	13.94	Upper class
CVD: cardiovascular disease; Critical point (95th percentile): 12.7; First quartile: ≤0.61; Second quartile: 0.62–1.46; Third quartile: 1.47–2.44; Fourth quartile: ≥2.5

Cost of investigation relevant risk factors (putative model components) in USD

To substantiate cost-effectiveness, we sourced the investigative cost of the selected biomarkers from four African countries namely Nigeria, South Africa, Morocco and Uganda, and subsequently calculated the average cost. We defined a biomarker (i.e., specialised tests) as cost-effective if the cost of investigation falls within the average $13.6. A biomarker or model cost was considered ineffective if the associated investigative cost was ≥ $13.6; all biomarkers except homocysteine were identifed as being cost-effective, with associated prices ≤$13.6 (Table 3).

Table 3

Cost of investigation of relevant risk factors (putative model components) in USD
Biomarker (test)	Nigeria(¹ I$ = 770 as at 05/07/2023	South Africa (R) Rate: I$ = R17.87 as at 22/07/2023	The average cost in dollars
Blood glucose	700 (0.9)¶	R70 (3.9)^Ɨ	$2.4
HIV	1000 (1.3)¶	R220 (12.3) ^Ɨ	$6.8
Lipid profile	5000 (6.5) ¶	R380 (21.3) ^Ɨ	$13.9
ECG	3000 (3.9)¶	R350 (19.6) ^Ɨ	$11.75
Homocysteine	37,800 (48.8)	R310 (17.3) ^Ɨ	$ 33.1
Average	9500 (12.3)	R266 (14.9)	$13.6
IUDS = R17.87 as at 22/07/2023; 1USD: 770 as at 05/07/2023

¶: Department of Medical Laboratory Enugu State University Teaching Hospital Parklane Enugu

*: https://healthtracka.com/panels/homocysteine-133/

Ɨ: Lancet Labs, South Africa. Durban. 52 Ismail C Meer St, Durban Central, Durban, 4001, South Africa

Assembling models and model components

An assembly of the emerging models constructed are illustrated in Table 4. All models met the CRP but only two viz., cost-effective stroke screening model (CAPMS) 1 & CAPMS 2, were considered cost-effective, as they demonstrated an associated investigative cost of < $12.3. Homocysteine was identified as the only univariate model (Table 4).

Table 4

assembling models and model components
Potential models	Components	Rw vis-à-vis CRP = 12.7	Remarks
Standalone Homocysteine model	Homocysteine	13.9	Univariate & cost-ineffective
CHH model	Cardiac causes, hypertension & homocysteine	22.8	Multivariate & cost-ineffective
CaLHVeF model	Cardiac causes, dyslipidemia, high levels of stress, low consumption of leafy vegetables, hypertension	14.8	Multivariate & relatively cost-effective
CAPMS 1	High levels of stress, hypertension, diabetes, dyslipidemia, high level of stress, low consumption of leafy vegetable, smoking	15.8	Multivariate & cost-effective
CAPMS 2	High level of stress, hypertension, diabetes, high level of stress, low consumption of leafy vegetables, Family history of CVD, smoking	14.9	Multivariate & cost-effective

Examining pre-existing models/tools vis-à-vis our study findings

Examining the pre-existing risk assessment tools in the light of our study findings, we identified two major flaws namely suboptimal predictive power (CRP < 12.7) and cost-ineffectiveness vis-à-vis African countries where an average person lives below the poverty line. Risk assessment tools demonstrating suboptimal predictive power include the CUORE, SCORE, non-laboratory NHANES and MUCA models respectively. In contrast, the Framingham, PROCAM, Reynolds Risk Score, QRISK, The Cardiovascular Health Study Cohort (CHSCO), ARIC and SMART-REACH models, satisfy the CRP criteria but are costly and ineffective tools as they incur a procedural cost that may be unaffordable by patients in LMICs. Noteworthy, the ASSIGN, Self-Reported Stroke Risk Stratification scale, Zhao’s Risk Stratification Tool for Ischemic Stroke, and NHANES laboratory model all met the criteria for the CRP, are cost-effective but show no evidence of non-African studies. The Akpa’s Afrocentric model is West African and met the CRP, and is relatively cost-effective but time-consuming (Table 5).

Table 5

Examining the flaws in pre-existing stroke risk assessment models/tools
Risk assessment tool	Component factors	Critics/flaws	The implication of flaws vis-à-vis our study findings
Framingham model [57]	Systolic BP, Hypertensive therapy, diabetes mellitus, cigarette smoking, history of coronary heart disease, presence of atrial fibrillation & left ventricular hypertrophy	It is based on the small US largely white middle-income community [58] It predicts only future coronary heart diseases and does not predict future total cardiovascular events [58] 3. The Framingham risk score could overestimate or underestimate risk in populations [59]	Met the CRP but cost-ineffective as ECG test alone is not sufficient to suggest left ventricular hypertrophy; it will require further tests. African data is not involved hence may not be appropriate for the region
PROCAM model [60]	LDL cholesterol, Smoking, HDL cholesterol, Systolic blood pressure, Family history of premature Myocardial infarction, diabetes mellitus & triglycerides	It is based on a sample of industrial German employees [61] Only estimates the risk for fatal or nonfatal myocardial infarction [61] It may be considered somewhat underpowered for risk estimation for women [61]	Met the CRP but cost-ineffective Limited application to the African populace
Reynolds Risk Score [62]	Systolic BP, Hypertensive therapy, diabetes mellitus, cigarette smoking, history of coronary heart disease, presence of atrial fibrillation & left ventricular hypertrophy, parental history of premature coronary heart disease & high sensitivity C- reactive protein	1. It is derived from only two cohort studies hence its external validity is questionable [62] It is unclear whether high sensitivity c-reactive protein is a risk factor for CVD i.e. lack of stratification of risk factors [62]	Met the CRP but not cost-effective 2. Limited application to the African populace
MUCA model [63]	Weight, Height, 3 consecutive blood pressure readings, 12-hour fasting blood sample, and blood lipids to measure total cholesterol	Its sample size is limited to Chinese individuals living in China [64]	Cumulative risk weight less than the CRP (Suboptimal predictive potential) Limited application to the African populace
ASSIGN [65]	Sex, age, total cholesterol, HDL cholesterol, systolic blood pressure, Cigarette smoking status and deprivation, diabetes, family history of coronary heart disease	Random samples from the general population in Scotland only hence the external validity is questionable [65] The basis for the exclusion of obesity or BMI is not understood i.e. lack of stratification of risk factors Woodward et al. [65]	Met the CRP but we consider age and sex clinically insignificant factors in Africa Limited application to the African populace
CUORE [66]	Age, systolic blood pressure, total cholesterol, smoking habit, diabetes, & hypertension therapy	Not applicable outside Italy and hence lacks external validation for global application [66]	Cumulative risk point less than CRP hence may perform sub-optimally in Africa Limited application to the African populace
SCORE [67]	Patients with established disease, asymptomatic individuals at high risk of CVD mortality, first-degree relatives of patients with premature CVD& other individuals encountered during clinical practice	The scoring system only predicts the risk of CV death- does not take into account non-fatal CVD events [68] The score function is only recommended for use in the 40 to 65-year age range [67] May overestimate risk in countries with decreasing CVD mortality and underestimate risk in countries with increasing CVD mortality [69]	Cumulative risk point less than CRP hence may perform sub-optimally in Africa Limited application to the African populace
QRISK [61]	Age, sex, smoking, systolic blood pressure, the ratio of total serum cholesterol to HDL, BMI, family history of coronary heart disease aged less than 60 years, are a measure of deprivation,& existing treatment with the antihypertensive agent, ethnic origin, type 2 diabetes, rheumatoid arthritis, atrial fibrillation and chronic renal disease.	The data is only from England and Wales [61] The data set very incomplete for some risk factors; therefore imputation and statistical modelling are needed to reduce biased estimates [61]	Met the CRP but not cost-effective Limited application to the African populace
Self-Reported Stroke Risk Stratification (SRSRSF): The Regards Study) [70]	Age, race, sex, self-report of a physician diagnosis of hypertension, diabetes mellitus, atrial fibrillation, & heart disease status, current cigarette smoking, education, general self-reported health, any history of stroke symptoms as assessed by questionnaire	1. The major limitation is that REGARDS included only black and white participants, so it is not clear whether these findings can be generalized to other race-ethnic groups [70]	Met the CRP and cost-effective Limited application to the African populace
Risk Stratification Tool for Ischemic Stroke: A Risk Assessment Model Based on Traditional Risk Factors Combined With White Matter Lesions and Retinal Vascular Caliber [71]	Number of eyes, age, BMI, Sex, height, weight, history of hypertension, hyperlipidaemia, diabetes, coronary heart disease, cigarette smoking habits and consumption levels of alcohol	The included cases were collected from a single centre and the small sample limits the ability to predict stroke [71] The stroke group was derived from hospitalized patients and two subtypes of ischemic stroke, other stroke types were not included [71]	Met the CRP & cost-effective Limited application to the African populace
The Cardiovascular Health Study Cohort (Ch0 [72]	Age, aspirin usage, diabetes, impaired glucose tolerance, higher systolic blood pressure, increase time needed to walk 15ft, frequent falls, elevated creatinine levels, abnormal left ventricular wall motion and increased LV mass on echocardiography, ultrasound-defined carotid stenosis, and atrial fibrillation	The CHS contains more men and African American participants [72] The external validity is not fully assessed [72]	Met the CRP but not cost-effective Limited application to the African populace
Prediction of Ischemic Stroke Risk in the Atherosclerosis Risk in Communities Study (ARIC Study) [73]	Age, race, smoking status, diabetes mellitus, left ventricular hypertrophy, previous coronary heart disease, use of antihypertensive medication, systolic blood pressure.	This study included an African-American cohort from just one city [73] Other ethnic groups were too sparsely represented in the ARIC study, therefore it is possible that the ARIC score will not apply to US populations [73]	Met the CRP, not cost-effective and inclusion of clinically insignificant may affect its performance in Africa Limited application to the African populace
NHANES I Follow-Up Study Cohort [74]	Laboratory-based model: Age, sex, systolic blood pressure, smoking status, total cholesterol, reported diabetes status, current treatment for hypertension Non-laboratory-based model: Age, sex, systolic blood pressure, smoking status, BMI, reported diabetes status, current treatment for hypertension	This study represents a population from the USA, and the proposed risk score may not be as useful in developing countries [74] Only a few low-income countries have cohort data to make estimates [74] The basis for the exclusion of HDL is not understood i.e. lack of stratification of risk factors [74]	Met the CRP, cost-effective but the inclusion of clinically insignificant may affect its performance in Africa Limited application to the African populace Cumulative risk point less than CRP hence may perform sub-optimally in Africa Limited application to the African populace
SMART-REACH [75]	Age, sex, BMI, current smoking, diabetes mellitus, history of heart failure, history of heart failure, history of atrial fibrillation, systolic blood pressure, serum creatinine concentration, number of locations of CVD cerebrovascular, coronary, and peripheral artery disease) and total and low-density lipoprotein cholesterol)	Missing data for clinical Measurements at the 3-month follow-up and excluded oldest patients [75] Did not account for changes in risk factor levels over time i.e. lack of stratification of risk factors [75]	Met the CRP but cost-ineffective Limited application to the African populace
NOVEL AFROCENTRIC MODEL [39]	Baseline age, income < USD 100, No formal education, Hypertension, dyslipidaemia, diabetes mellitus, cardiac disease, family history of CVD, Obesity, stress, added salt, low consumption of leafy green vegetables, Regular sugar consumption, meat consumption, & physical inactivity	The model is Afrocentric but non-inclusive; only data from West Africa (Nigeria & Ghana) were included. Multi-component makes model administration hectic	Met CRP but cost-effective and time-consuming

A detailed examination of different risk assessment models for survival in breast cancer patients reveals a triad of key challenges namely limited sample frame with conspicuous omission of African data, lack of stratification based on responsiveness of determinants and/or cost-ineffectiveness vis-à-vis affordability in African settings. In this study, we took care of the triad and assembled a cost-effective Afrocentric models⸻ CAPMS 1 and CAPMS 2. In comparison to values reported with pre-existing models, we anticipate a decrease in the false negative and false positive rates once the model has been validated. Finding multiple biomarker combinations that satisfy the threshold enables individuality and cost-effectiveness, particularly in Africa, where the average person may find it difficult to afford the high cost of investigating biomarker inclusion in some of the existing models. Although more research is needed to validate the study's findings, the findings from this study may necessitate recalibration of several existing models. Being able to accurately predict the onset of stroke would enhance the successful implementation m of preventative healthcare measures as well as include and exclude patients from randomized trials [76, 77]. This study offers a thorough list of relevant stroke predictors in a hypothetical clinical relevance-order which has potential to further inform therapeutic targets.

Of the sampled risk factors and/or biomarkers of stroke in this study, homocysteine was identified as a significant risk factor for stroke in Africa, demonstrating a high-risk weight of 13.9 and risk responsiveness of 0.67. Similarly, non-African studies have highlighted the importance of homocysteine in the pathogenesis of stroke [78, 79]. Our findings indicate that individuals who have abnormally elevated homocysteine levels (tHcy > 15 micromol/L) appear to be approximately 21 times more likely to suffer from a stroke compared with healthy individuals. A recent meta-analysis by the Homocysteine Studies Collaboration confirmed homocysteine as a risk factor for the first events of stroke and coronary heart disease [79]. Apart from being a direct risk factor for stroke, homocysteine has been linked to hypertension and dyslipidaemia, and is confirmed as primary risk factor for haemorrhagic stroke [80] and ischemic stroke [81]. Homocysteine (Hcy) is a toxic, sulfur-containing intermediate of methionine metabolism [82]. Hyperhomocysteinemia (hHcy), as a consequence of impaired Hcy metabolism or defects in crucial co-factors that participate in its recycling, is considered as an independent human stroke risk factor [82]. Neural cells are sensitive to prolonged hHcy treatment because Hcy cannot be metabolized by the transsulfuration pathway or the folate/vitamin B12 independent remethylation pathway [82]. The adverse consequences following ischemia-induced damage encompass the build-up of reactive oxygen species (ROS) and the posttranslational alterations of proteins through homocysteinylation and thiolation. Ischemic preconditioning (IPC) is an adaptive response of the CNS to sub-lethal ischemia, which elevates tissue tolerance to subsequent ischemia [82]. Despite its predictive potential, the associated high cost of investigation may inform its exclusion in most stroke screening models. The shortcomings associated with its risk responsiveness adds to the limitations of homocysteine as a risk factor. This is suggestive that the occurrence of some types of stroke in Africa may be unrelated to homocysteine, hence building a stroke predictive mode based exclusively on homocysteine is impractical and unscientific. It is also possible that the lower risk responsiveness of homocysteine reported in our study may be associated with the review of only a few African studies that evaluated the association between homocysteine and stroke and thus may not be unconnected to the associated high investigative cost [35, 40, 46]. Nonetheless, in Africa, homocysteine is an emerging biomarker and its clinical and economic importance is warranted as this will aid the search for a devise-based screening tool for predicting the onset of stroke in the region.

Our findings also demonstrates the significance of hypertension as a risk factor for stroke, with a near-perfect risk responsiveness of 0.94 and a risk weight of 5.94, thus confirming its consideration when modelling stroke in Africa. Furthermore, hypertension was implicated in 94% of all stroke cases reported in the region, which is consistent with previous studies [27, 83]. It is evident that hypertension is essential in predicting stroke [84], with negligible investigative costs incurred thus recommending its inclusion in most of the pre-existing stroke models. Whilst no reasonable predictive and/or screening model for stroke should omit hypertension, it is important to reiterate that hypertension alone may not always predict stroke based on the risk weight being below the critical risk point. This is suggestive that in Africa, the role of hypertension in the pathogenesis of stroke may be accentuated by factors such as homocysteinemia or its proxy ⸻poor folate intake [85], nutritional and dietary issues including regular meat consumption, low consumption of green vegetables and high salt (sodium intake) (ref). This is substantiated by the adoptiong of the Dietary Approaches to Stopping Hypertension (DASH) eating plan, which has potential to reduce blood pressure by 8–14 mmHg [86]. The DASH diet comprises vegetables, whole grains, fruits, low-fat dairy products, poultry, fish and nuts with less of red meat, sweets and sugar [86]. Hence, stroke strategies targeting risk mitigation amongst hypertensive patients should target blood pressure control as well as nutritional and lifestyle factors.

Our study also highlights ‘cardiac causes’ including atrial fibrillation, myocardial infarction, prosthetic heart valve, rheumatic valvular heart disease and left ventricular failure, as crucial independent predictors of stroke, especially ischaemic stroke, based on their risk stability index (0.8) and risk weight (3.1). This is consistent with global reports that heart cardiac causes namely heart failure and atrial fibrillation are associated with a 2–5 folds increase in the risk of stroke [87–91]. In high-income countries, the investigative cost associated with cardiac causes is affordable and together with its prognostic importance, informs its inclusion in almost all the pre-existing stroke predictive models. However, in Africa, the investigative cost associated with cardiac causes is affordable only if it’s the only laboratory-based factor included in the model (ref). However, the greater affinity of cardiac causes towards ischaemic stroke [92] may be required to balance the greater affinity of hypertension towards haemorrhagic stroke [93] especially when a generic stroke model is sought. In Africa, the role of “cardiac causes” in the pathogenesis of stroke in Africa could be accentuated by nutritional and dietary factors including poor folate intake [94], regular meat consumption [95], alcohol consumption [96], low consumption of green vegetable and high salt (sodium intake) [94]. The associated cost of investigation and the inefficiency associated with testing procedures may limit the inclusion of cardiac causes in stroke screening models, especially amongst individuals with unknown cardiac histories, and in settings where cardiac causes are not public health concerns. Albeit, in regions of Africa where 20–35% of the population is affected by cardiovascular disease (CVD) [97], cardiac causes are a candidate putative factor for stroke screening models. Given the challenges associated with its testing in our setting, a history of CVD with an excellent risk stability index of 0.8 and odds ratio of 1.88 might serve as an alternative to field testing of cardiac causes since no investigative cost will be required. However, we recognize that the history taken will be most appropriate in communities where routine health checks are carried out, unlike in African settings where patients present at late disease stages.

Our study further demonstrates dyslipidaemia as a major risk factor of stroke in Africa, with a risk weight of 2.36. Dyslipidaemia, defined as cholesterol ≥ 5.2mmol/L, HDL ≥ 1.03mmol/Land LDL ≥ 3.4mmol/L or TAG ≥ 1.7mmol/L (ref). Even though dyslipidaemia may be an independent predictor of stroke, it is unable to function as a standalone predictor and/or biomarker. Our result shows that it possesses a risk stability index of 0.83, indicative of its involvement in 83% of stroke cases, especially ischaemic stroke. Therefore, it is a candidate biomarker for stroke predictive models, however, its inclusion and utility as a putative model component may be limited by relatively high associated investigated costs. Nevertheless, its interaction with homocysteinemia [81], hypertension [98] and heart diseases [99], supports it universal effectiveness as an ideal biomarker for models intended for use among low-risk groups. Interestingly, our study confirms a differential predictive potential of various causes of dyslipidaemia. HDL was identified as the most important cause of dyslipidaemia with a perfect risk responsiveness and risk weight (3.3), which potentially informed its inclusion in the pre-existing PROCAM and ASSIGN models. Furthermore, we require a further weighting factor when assigning a score to dyslipidaemia, and such factor must be based on the cause, with the highest weight attributed to the most important factor. Similar to hypertension and “cardiac causes”, the role of dyslipidaemia in the pathogenesis of stroke in Africa could be accentuated by nutritional and dietary factors including poor folate intake [94], regular meat consumption [95], alcohol consumption [96], low consumption of green vegetable and high salt (sodium intake) [94], which represent important targets of primary stroke preventive strategies.

Our findings also highlights diabetes as an independent predictor of stroke, based on its risk weight of 2.45 and a moderate risk stability index of 0.76. Notably, individuals who have a history of diabetes or an HbA1C ≥ 6.5% appear to be 3 times more liable to get a stroke. Our findings are consistent with the notion that stroke may result from direct damage to the cerebral blood vessels associated with pathologic alterations caused by diabetes [100]. Atherosclerosis emanating from diabetes (high blood sugar levels) induced long-term vascular damage is a major risk factor for stroke [101]. Additionally, high levels of the "bad" cholesterol LDL and low levels of the “good" cholesterol HDL, may result from diabetes [102]. The likelihood of stroke is further raised by these abnormalities in lipids [102]. Diabetes can also disrupt the blood clotting processes and raise the risk of thrombotic events like ischemic strokes [103] as well as increase the risk of developing peripheral artery disease, which reduces blood flow to the limbs [104]. Its clinical relevance in stroke prediction and prognosis could be justified by its inclusion in several pre-existing stroke models including the Framingham, PROCAM, Reynolds Risk Score, ASSIGN, CUORE, QRISK and Self-Reported Stroke Risk Stratification (SRSRSF) models. The relative low cost associated with diabetes diagnosis makes a suitable risk factor for predictive stroke models, however, diabetes alone does not meet the critical risk point. Hence, its relevance in stroke prediction is manifest only when combined with other relevant risk factors in a multivariate model.

Noteworthy, Our findings also highlights HIV as a risk factor for stroke, with a risk stability of 0.67 and a risk weight of 1.70, suggestive that 67% of any given situation, HIV has the potential to predict the occurrence of stroke. Specifically, HIV infected individuals were 2.5 times more likely to develop stroke compared to normal healthy individuals. However, its predictive relevance manifests only in combination with other important risk factors, as previously reported [56], whereby HIV status is an independent stroke risk factor within an antiretroviral naïve population. Factors responsible for increased risk of stroke among HIV patients include inflammation, coagulatory issues, and side effects of antiretroviral therapy on cardiovascular health [105–107]. Surprisingly, no pre-existing prediction models has included HIV status as a component of a predictive stroke model thus far, despite the relatively low investigative cost.

Limitations

The significant degree of heterogeneity found in the risk estimates for seven of the twelve biomarkers examined in our study constitutes a limitation. The risk estimates for two factors viz: smoking and age suffered publication bias. The application of discretionary/subjective judgment in establishing the critical risk is a limitation. However, the subjectivity simply represents a trade-off between sensitivity and specificity. The emerging models requires validation to confirm the best critical risk point and determine the diagnostic performance, feasibility and acceptability of the models.

Targeted screening with the CAPMS 1 and CAPMS 2 model have the potential to represent cost-effective innovation in practice of stroke screening in African clinics and communities. However, future studies are warranted to validate the models in order to establish the optimum CRP for the prediction of stroke in Africa.

Ethics approval and consent to participate: Not applicable

Consent for publication: Not applicable

Availability of data and materials

Review data is available from the corresponding author upon reasonable request.

Competing interest

No potential conflict of interest to this article was reported

Funding

The primary investigator is a postdoctoral fellow with the Department of Physiotherapy, University of Pretoria, South Africa.

Authors' contributions

NM and PO conceived the study. NM and MSB conducted the searches. OP downloaded the studies and conducted initial screening under the supervision of MN. MN, GN, NP and IS constituted the content experts. NM and OP ran the analysis. PO and UM conducted quality appraisal. The manuscript was drafted by NM, PO and UM. All authors read through and approved the final draft of the manuscript.

Acknowledgements

The authors wish to acknowledge the authors of all the original studies to be included in the meta-analysis. We are grateful for the shoulders of giants (the primary study authors) upon which we now stand, awaiting the validation of this product.

Salvadori E, Papi G, Insalata G, Rinnoci V, Donnini I, Martini M, et al. Comparison between ischemic and hemorrhagic strokes in functional outcome at discharge from an intensive Rehabilitation Hospital. Diagnostics (Basel). 2020;11(1):38.
World Health Organization. World Health Organization [online]. ; 2022 [cited Dec 30, 2022]. Available from: https://www.who.int.
Krishnamurthi RV, Ikeda T, Feigin VL. Global, regional and country-specific burden of ischaemic stroke, intracerebral haemorrhage and subarachnoid haemorrhage: A systematic analysis of the global burden of disease study 2017. Neuroepidemiology. 2020;54(2):171–9.
Akpalu A, Sarfo FS, Ovbiagele B, Akinyemi R, Gebregziabher M, Obiako R, et al. Phenotyping stroke in sub-Saharan Africa: stroke investigative research and education network (SIREN) phenomics protocol. Neuroepidemiology. 2015;45(2):73–82.
Feigin VL, Norrving B. A new paradigm for primary prevention strategy in people with an elevated risk of stroke. Int J Stroke. 2014;9(5):624–6.
Akinyemi RO, Ovbiagele B, Adeniji OA, Sarfo FS, Abd-Allah F, Adoukonou T, et al. Stroke in Africa: profile, progress, prospects and priorities. Nat Rev Neurol. 2021;17(10):634–56.
Cruz-Flores S, Rabinstein A, Biller J, Elkind MSV, Griffith P, Gorelick PB, et al. Racial-ethnic disparities in stroke care: the American experience: A statement for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. 2011;42(7):2091–116.
Pandian JD, Kalkonde Y, Sebastian IA, Felix C, Urimubenshi G, Bosch J. Stroke systems of care in low-income and middle-income countries: challenges and opportunities. Lancet. 2020;396(10260):1443–51.
Ofoli JNT, Ashau-Oladipo T, Hati SS, Ati L, Ede V. Preventive healthcare uptake in private hospitals in Nigeria: A cross-sectional survey (Nisa Premier Hospital). BMC Health Serv Res. 2020;20(1):273.
Congleton TM, Small CW, Freeman SD. Abstract WP345: Stroke risk factors screening and education: A regional strategy to address stroke prevalence and mortality in Eastern North Carolina. Stroke. 2013;44;Suppl 1:Abstract WP345:. doi: 10.1161/str.44.suppl_1.AWP345(suppl_1) (2 Meeting Abstracts): AWP345:Abstract WP345:. 10.1161/str.44.suppl_1.AWP345.
United Nations. Transforming our world: the 2030 agenda for sustainable development; 2020 [cited Aug 19, 2020]. Available from: https://sdgs.un.org/sites/default/files/publications/21252030%20Agenda%20for%20Sustainable%20Development%20web.pdf.
Nweke M, Ukwuoma M, Adiuku-Brown AC, Ugwu P, Nseka E. Characterization and stratification of the correlates of postpartum depression in sub-Saharan Africa: A systematic review with meta-analysis. Womens Health (Lond). 2022;18:17455057221118773.
Yong HMD, Foody JMD, Linong JMD, Dong ZMD, Wang YMD, Ma L, et al. A systematic literature review of risk factors for stroke in China. Cardiol Rev Mar/Apr. 2013;21(2):77–93.
Bosomworth NJ. Practical use of the Framingham risk score in primary prevention: a Canadian perspective. Can Fam Phys Med Fam Can. 2011;57(4):417–23.
Cooney MT, Dudina AL, Graham IM. Value and limitations of existing scores for the assessment of cardiovascular risk: a review for clinicians. J Am Coll Cardiol. 2009;54(14):1209–27.
Song JW, Chung KC. Observational studies: cohort and case-control studies. Plast Reconstr Surg. 2010;126(6):2234–42.
Moher D, Shamseer L, Clarke M, Ghersi D, Liberati A, Petticrew M, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev. 2015;4(1):1.
Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. Obs Stud Epidemiol (Moose). Group JAMA. 2000;283(15):2008–12.
Moola S, Munn Z, Tufanaru C, Aromataris E, Sears K, Sfetc R, et al. Chapter 7. Systematic reviews of aetiology and risk. In: Joanna Briggs Institute reviewer’s manual; 2017. pp. 20–5.
Nwagha TU, Nweke M. Stratification of risk factors of lung cancer-associated venous thromboembolism and determining the critical point for preemptive intervention: A systematic review with meta-analysis. Clin Med Insights Oncol. 2023;17:11795549231175221.
Borenstein M, Hedges LV, Higgins JPT, et al. Introduction to meta-analysis. John Wiley & Sons; 2009.
Lenhard W, Lenhard A. Calculation of effect sizes. Psychometrika. 2015.
Chen H, Cohen P, Chen S. How bid is a bid odds ratio? Interpreting the Magnitude of Odds Ratios in Epidemiological studies. Commun Stat Simul Comput. 2010;39(4):860–4. 10.1080/03610911003650383.
Center for Disease Control and Prevention. Cost-effectiveness analysis. Centers for Disease Control and Prevention, Office of Policy, Performance, and Evaluation [cited 5/6/2023]. Available from: https://www.cdc.gov/policy/polaris/economics/cost-effectiveness/index.html.
Namale G, Kamacooko O, Kinengyere A, Yperzeele L, Cras P, Ddumba E et al. Risk factors for hemorrhagic and ischemic stroke in sub-Saharan Africa. J Trop Med. 2018; 2018:4650851.
O’Donnell MJ, Chin SL, Rangarajan S, Xavier D, Liu L, Zhang H, et al. Global and regional effects of potentially modifiable risk factors associated with acute stroke in 32 countries (INTERSTROKE): a case-control study. Lancet. 2016;388:761–75.
O’Donnell MJ, Xavier D, Liu L, Zhang H, Chin SL, Rao-Melacini P, et al. Risk factors for ischaemic and intracerebral haemorrhagic stroke in 22 countries (the INTERSTROKE study): a case-control study. Lancet. 2010;376(9735):112–23.
Poulter NR, Chang CL, Farley TM, Marmot MG. Reliability of data from proxy respondents in an international case-control study of cardiovascular disease and oral contraceptives. World Health Organization collaborative study of cardiovascular disease and steroid hormone contraception. J Epidemiol Community Health. 1996;50(6):674–80.
Chang C-L, Marmot MG, Farley TMM, Poulter NR. The influence of economic development on the association between education and the risk of acute myocardial infarction and stroke. J Clin Epidemiol. 2002;55(8):741–7.
Elagib AH, Ahmed AE, Hussein A, Musa AM, Khalil EA, El-Hassan AM. Possible predisposing factors for thrombotic cerebral accidents in Sudanese patients. Saudi Med J. 2008;29(2):304–6.
Owolabi MO, Agunloye AM. Which risk factors are more associated with ischemic rather than hemorrhagic stroke in black Africans? Clin Neurol Neurosurg. 2013;115(10):2069–74.
Owolabi MO, Agunloye AM. Risk factors for stroke among patients with hypertension: a case-control study. J Neurol Sci. 2013;325(1–2):51–6.
Chehaibi K, Hrira MY, Trabelsi I, Escolà-Gil JC, Slimane MN. Gene variant and level of IL-1β in ischemic stroke patients with and without type 2 diabetes mellitus. J Mol Neurosci. 2015;57(3):404–9.
de Mast Q, Molhoek JE, van der Ven AJ, Gray WK, de Groot PG, Jusabani A, et al. Antiphospholipid antibodies and the risk of stroke in urban and rural Tanzania A community-based case-control study. Stroke. 2016;47(10):2589–95.
Fekih-Mrissa N, Mrad M, Klai S, Mansour M, Nsiri B, Gritli N, et al. Methylenetetrahydrofolate reductase (C677T and A1298C) polymorphisms, hyperhomocysteinemia, and ischemic stroke in Tunisian patients. J Stroke Cerebrovasc Dis. 2013;22(4):465–9.
Saidi S, Mahjoub T, Almawi WY. Aldosterone synthase gene (CYP11B2) promoter polymorphism as a risk factor for ischaemic stroke in Tunisian Arabs. J Renin Angiotensin Aldosterone Syst. 2010;11(3):180–6.
Saidi S, Slamia LB, Ammou SB, Mahjoub T, Almawi WY. Association of apolipoprotein E gene polymorphism with ischemic stroke involving large-vessel disease and its relation to serum lipid levels. J Stroke Cerebrovasc Dis. 2007;16(4):160–6.
Saidi S, Slamia LB, Mahjoub T, Ammou SB, Almawi WY. Association of PAI-1 4G/5G and – 844G/A gene polymorphism and changes in PAI-1/tPA levels in stroke: a case-control study. J Stroke Cerebrovasc Dis. 2007;16(4):153–9.
Akpalu A, Sarfo FS, Akinyemi J, Wahab K, Komolafe M, Obiako R, et al. Frequency & factors associated with recurrent stroke in Ghana and Nigeria. J Neurol Sci. 2022;439:120303.
Akpalu A, Nyame P. Plasma homocysteine as a risk factor for strokes in Ghanaian adults. Ghana Med J. 2009;43(4):157–63.
Amu E, Ogunrin O, Danesi M. RE – appraisal of risk factors for stroke in Nigerian Africans – a prospective case – control study. Afr J Neurol Sci. 2005;24(2).
Diakite B, Hamzi K, Hmimech W, Nadifi S. First study of C2491T FV mutation with ischaemic stroke risk in Morocco. J Genet. 2015;94(2):313–5.
Diakite B, Hamzi K, Hmimech W, Nadifi S, GMRAVC. Genetic polymorphisms of T-1131C APOA5 and ALOX5AP SG13S114 with the susceptibility of ischaemic stroke in Morocco. J Genet. 2016;95(2):303–9.
Fekadu G. Current challenges and strategies in management and prevention of stroke. J Neurol Stroke. 2019;9(3):149–53.
SAIDI S, MAHJOUB T, ALMAWI WY. Lupus anticoagulants and anti-phospholipid antibodies as risk factors for a first episode of ischemic stroke. J Thromb Haemost. 2009;7(7):1075–80.
Okubadejo NU, Oladipo OO, Adeyomoye AA, Awosanya GO, Danesi MA. Exploratory study of plasma total homocysteine and its relationship to short-term outcome in acute ischaemic stroke in Nigerians. BMC Neurol. 2008;8:26.
Mbonde AA, Chang J, Musubire A, Okello S, Kayanja A, Acan M, et al. An analysis of stroke risk factors by HIV serostatus in Uganda: implications for stroke prevention in sub-Saharan Africa. J Stroke Cerebrovasc Dis. 2022;31(7):106449.
Saidi S, Mallat SG, Almawi WY, Mahjoub T. Association between renin-angiotensin-aldosterone system genotypes and haplotypes and risk of ischemic stroke of atherosclerotic etiology. Acta Neurol Scand. 2009;119(6):356–63.
Salem-Berrabah OB, Mrissa R, Machghoul S, Hamida AB, N’Siri B, Mazigh C, et al. Hyperhomocysteinemia, C677T MTHFR polymorphism and ischemic stroke in Tunisian patients. Tunis Med. 2010;88(9):655–9.
Sarfo FS, Mobula LM, Plange-Rhule J, Ansong D, Ofori-Adjei D. Incident stroke among Ghanaians with hypertension and diabetes: A multicenter, prospective cohort study. J Neurol Sci. 2018;395:17–24.
Sarfo FS, Opare-Sem O, Agyei M, Akassi J, Owusu D, Owolabi M, et al. Risk factors for stroke occurrence in a low HIV endemic West African country: A case-control study. J Neurol Sci. 2018;395:8–16.
Sarfo FS, Ovbiagele B, Gebregziabher M, Wahab K, Akinyemi R, Akpalu A, et al. Stroke among Young West Africans: evidence from the SIREN (stroke investigative research and educational network) large multisite case-control study. Stroke. 2018;49(5):1116–22.
Sarfo FS, Mobula LM, Adade T, Commodore-Mensah Y, Agyei M, Kokuro C, Adu-Gyamfi R, Duah C, Ovbiagele B. Low blood pressure levels & incident stroke risk among elderly Ghanaians with hypertension. J Neurol Sci. 2020;413:116770.
Sarfo FS, Ovbiagele B, Gebregziabher M, Akpa O, Akpalu A, Wahab K, et al. Unraveling the risk factors for spontaneous intracerebral hemorrhage among West Africans. Neurology. 2020;94(10):e998–e1012.
Sarfo FS, Ovbiagele B, Akpa O, Akpalu A, Wahab K, Obiako R, et al. Risk factor characterization of ischemic stroke subtypes among West Africans. Stroke. 2022;53(1):134–44.
Walker RW, Jusabani A, Aris E, Gray WK, Unwin N, Swai M, et al. Stroke risk factors in an incident population in urban and rural Tanzania: a prospective, community-based, case-control study. Lancet Glob Health. 2013;1(5):e282–8.
Wolf PA, D’Agostino RB, Belanger AJ, Kannel WB. Probability of stroke: a risk profile from the Framingham Study. Stroke. 1991;22(3):312–8.
Sacco RL, Khatri M, Rundek T, Xu Q, Gardener H, Boden-Albala B, et al. Improving global vascular risk prediction with behavioral and anthropometric factors. The multiethnic NOMAS (Northern Manhattan Cohort Study). J Am Coll Cardiol. 2009;54(24):2303–11.
Brindle P. Predictive accuracy of the Framingham coronary risk score in British men: prospective cohort study. BMJ (PMC Free article). 2003;327(7426):1267.
Assmann G. Simple scoring scheme for calculating the risk of acute coronary events based on the 10-year follow-up of the prospective cardiovascular Munster (PROCAM) study. Circulation. 2002;3:105.
Hippisley-Cox J. Predicting cardiovascular risk in England and Wales: prospective derivation and validation of QRISK2. BMJ. 2008;7659:1475–82.
Ridker PM, Paynter NP, Rifai N, Gaziano JM, Cook NR. Creactive protein and parental history improve global cardiovascular risk prediction: the Reynolds risk score for men. Circulation. 2008;118:2243–51.
Xing X, Yang X, Liu F, Li J, Chen J, Liu X, et al. Predicting 10-year and lifetime stroke risk in Chinese population. Stroke. 2019;50(9):2371–8.
The Collaborative Study Group on Trends of Cardiovascular Disease in China and Preventive Strategy. Current status of major cardiovascular risk factors in Chinese populations and their trends in the past two decades [in Chinese]. Zhonghua xin xue guan bing za zhi. 2001;29:74–9.
Woodward M. Adding social deprivation and family history to cardiovascular risk assessment: the ASSIGN score from the Scottish Heart Health Extended Cohort (SHHEC). Heart. 2007;93(2):172–21.
Ferrario M. Prediction of coronary events in a low incidence population. Assessing accuracy of the CUORE Cohort Study prediction equation. Int J Epidemiol. 2005;2:413–5.
Perk J et al. DBGGH. European Guidelines on cardiovascular disease prevention in clinical practice version 2012. The fifth joint task force of the European Society of Cardiology and other societies on cardiovascular disease preventio; 2012.
Conroy RM, Pyörälä K, Fitzgerald AP, Sans S, Menotti A, De Backer G, et al. Estimation of ten-year risk of fatal cardiovascular disease in Europe: the SCORE project. Eur Heart J. 2003;24(11):987–1003.
Brindle P, Beswick A, Fahey T, Ebrahim S. Accuracy and impact of risk assessment in the primary prevention of cardiovascular disease: a systematic review. Heart. 2006;92(12):1752–9.
Howard G, McClure LA, Moy CS, Howard VJ, Judd SE, Yuan Y, et al. Self-reported stroke risk stratification: reasons for geographic and racial differences in stroke study. Stroke. 2017;48(7):1737–43.
Weiming W, Tingting Z, Kang G, Gang Y, Yue C, Youhua X. Smoking and the pathophysiology of peripheral artery disease. Front Cardiovasc Med. 2021;8. 10.3389/fcvm.2021.704106.
Manolio TA, Kronmal RA, Burke GL, O’Leary DH, Price TR. Short-term predictors of incident stroke in older adults. The cardiovascular health study. Stroke. 1996;27(9):1479–86.
Chambless LE, Heiss G, Shahar E, Earp MJ, Toole J. Prediction of ischemic stroke risk in the Atherosclerosis Risk in Communities Study. Am J Epidemiol. 2004;160(3):259 – 69. [published correction appears in Am J Epidemiol. 2004;160(9):927.
Gaziano TA, Young CR, Fitzmaurice G, Atwood S, Gaziano JM. Laboratory-based versus non-laboratory-based method for assessment of cardiovascular disease risk: the NHANES I Follow-up Study cohort. Lancet. 2008;371(9616):923–31.
Gynnild MN, Hageman SHJ, Dorresteijn JAN, Spigset O, Lydersen S, Wethal T, et al. Risk stratification in patients with ischemic stroke and residual cardiovascular risk with current secondary prevention. Clin Epidemiol. 2021;13:813–23.
Borenstein M, Hedges LV, Higgins JPT, Rothstein HR. Converting among effect sizes. Introduction Meta-Anal. 2009:45–9.
Teoh D. Towards stroke prediction using electronic health records. BMC Med Inform Decis Mak. 2018;18:127. https://doi.org/10.1186/s12911-018-0702-y.
Rabelo NN, Telles JPM, Pipek LZ, Farias Vidigal Nascimento R, Gusmão RC, Teixeira MJ, et al. Homocysteine is associated with higher risks of ischemic stroke: A systematic review and meta-analysis. PLoS ONE. 2022;17(10):e0276087.
Zhao M, Wang X, He M, Qin X, Tang G, Huo Y, et al. Homocysteine and stroke risk: modifying effect of methylenetetrahydrofolate reductase C677T polymorphism and folic acid intervention. Stroke. 2017;48(5):1183–90.
Wu Q, Liu J, Wang Y, Cheng Y, Liu M. Higher serum homocysteine levels are associated with an increased risk of hemorrhagic transformation in patients with acute ischemic stroke. BMC Neurol. 2023;23(1):103.
Arina CA, Amir D, Siregar Y, Sembiring RJ. Correlation between homocysteine and dyslipidemia in ischaemic stroke patients with and without hypertension. IOP Conf Ser Earth Environ Sci. 2018;130:012005. 10.1088/1755-1315/130/1/012005.
Ján L, Barbara T, Maria K, Dušan D, Anna B, Dagmar K, et al. Role of homocysteine in the ischemic stroke and development of ischemic tolerance. Front Neurosci. 2016;10:538.
Li C, Engström G, Hedblad B, Berglund G, Janzon L. Blood pressure control and risk of stroke: a population-based prospective cohort study. Stroke. 2005;36(4):725–30.
Anker D, Santos-Eggimann B, Santschi V, Del Giovane C, Wolfson C, Streit S, et al. Screening and treatment of hypertension in older adults: less is more? Public Health Rev. 2018;39:26.
Xiong Y, Huang J, Amoah AN, Liu B, Bo Y, Lyu Q. Folate, vitamin B6, and vitamin B12 intakes are negatively associated with the prevalence of hypertension: A national population-based study. Nutr Res. 2023;112:46–54.
World Health Organization. World Health Day 2013. Diet, nutrition and hypertension. Available from: https://www.emro.who.int/world-health-days/2013/nutrition-hypertension-factsheet-whd html#:~:text = Following%20such%20a%20diet%20reduces,sweets%20and%20sugar%2Dcontaining%20beverages.&text = Manage%20stress.
Adelborg K, Szépligeti S, Sundbøll J, Horváth-Puhó E, Henderson VW, Ording A, et al. Risk of stroke in patients with heart failure: A population-based 30-year cohort study. Stroke. 2017;48(5):1161–8.
Tai YH, Chang CC, Yeh CC, Sung LC, Hu CJ, Cherng YG, et al. Long-term risk of stroke and poststroke outcomes in patients with heart failure: two nationwide studies. Clin Epidemiol. 2020;12:1235–44.
Hamatani Y, Nagai T, Nakai M, Nishimura K, Honda Y, Nakano H, et al. Elevated plasma D-dimer level is associated with short-term risk of ischemic stroke in patients with acute heart failure. Stroke. 2018;49(7):1737–40.
Alberts VP, Bos MJ, Koudstaal PJ, Hofman A, Witteman JC, Stricker BH, et al. Heart failure and the risk of stroke: the Rotterdam Study. Eur J Epidemiol. 2010;25(11):807–12.
Kozdag G, Ciftci E, Vural A, Selekler M, Sahin T, Ural D, et al. Silent cerebral infarction in patients with dilated cardiomyopathy: echocardiographic correlates. Int J Cardiol. 2006;107(3):376–81.
Vahedi K, Amarenco P. Cardiac causes of stroke. Curr Treat Options Neurol. 2000;2(4):305–18.
Kitagawa K. Blood pressure management for secondary stroke prevention. Hypertens Res. 2022;45(6):936–43.
Foroughi M, Akhavanzanjani M, Maghsoudi Z, Ghiasvand R, Khorvash F, Askari G. Stroke and nutrition: a review of studies. Int J Prev Med. 2013;4(Suppl 2):165–79.
Kaluza J, Wolk A, Larsson SC. Red meat consumption and risk of stroke: a meta-analysis of prospective studies. Stroke. 2012;43(10):2556–60.
Smyth A, O’Donnell M, Rangarajan S, Hankey GJ, Oveisgharan S, Canavan M, et al. Alcohol Intake as a Risk Factor for Acute Stroke: The INTERSTROKE Study. Neurol Jan. 2023;100(2):e142–53.
Keates AK, Mocumbi AO, Ntsekhe M, Sliwa K, Stewart S. Cardiovascular disease in Africa: epidemiological profile and challenges. Nat Rev Cardiol. 2017;14(5):273–93.
Tang N, Ma J, Tao R, Chen Z, Yang Y, He Q, et al. The effects of the interaction between BMI and dyslipidemia on hypertension in adults. Sci Rep. 2022;12(1):927.
Hedayatnia M, Asadi Z, Zare-Feyzabadi R, Yaghooti-Khorasani M, Ghazizadeh H, Ghaffarian-Zirak R, et al. Dyslipidemia and cardiovascular disease risk among the MASHAD study population. Lipids Health Dis. 2020;19(1):42.
Chen R, Ovbiagele B, Feng W. Diabetes and stroke: epidemiology, pathophysiology, pharmaceuticals and outcomes. Am J Med Sci. 2016;351(4):380–6.
Centers for Disease Control and Prevention. 2022. Diabetes and your heart. https://www.cdc.gov/diabetes/library/features/diabetes-and-heart.html. Accessed August 22, 2023.
American Heart Association., 2021. Cholesterol and Diabetes. https://www.heart.org/en/health-topics/diabetes/diabetes-complications-and-risks/cholesterol-abnormalities--diabetes. Accessed August 22, 2023.
Vaidya AR, Wolska N, Vara D, Mailer RK, Schröder K, Pula G. Diabetes and Thrombosis: A Central Role for Vascular Oxidative Stress. Antioxid (Basel). 2021;10(5):706.
Thiruvoipati T, Kielhorn CE, Armstrong EJ. Peripheral artery disease in patients with diabetes: Epidemiology, mechanisms, and outcomes. World J Diabetes. 2015;6(7):961–9.
Benjamin LA, Bryer A, Emsley HC, Khoo S, Solomon T, Connor MD. HIV infection and stroke: current perspectives and future directions. Lancet Neurol. 2012;11(10):878–90.
Ismael S, Moshahid Khan M, Kumar P, Kodidela S, Mirzahosseini G, Kumar S, et al. 3HIV associated risk factors for ischemic stroke and future perspectives. Int J Mol Sci. 2020;21(15):5306.
Kuate LM, Tchuisseu LAK, Jingi AM, Kouanfack C, Endomba FT, et al. Cardiovascular risk and stroke mortality in persons living with HIV: a longitudinal study in a hospital in Yaounde. Pan Afr Med J. 2021;40:8.

No competing interests reported.

Download PDF

Editorial decision: Revision requested
10 Apr, 2024
Reviews received at journal
31 Mar, 2024
Reviews received at journal
12 Oct, 2023
Reviewers agreed at journal
03 Oct, 2023
Reviewers agreed at journal
30 Sep, 2023
Reviewers agreed at journal
30 Sep, 2023
Reviewers invited by journal
30 Sep, 2023
Editor invited by journal
13 Sep, 2023
Editor assigned by journal
13 Sep, 2023
Submission checks completed at journal
10 Sep, 2023
First submitted to journal
05 Sep, 2023

You are reading this latest preprint version

Development of an inclusive and cost-effective Afrocentric predictive model for stroke: A novel approach

Status:

Version 2

Abstract

Figures

Introduction

PECOT criteria

Methods

Protocol and registration

Eligibility Criteria

Inclusion criteria

Exclusion criteria

Outcome measures

Sources of Information

Search strategy development

Data Management

Study selection and data extraction

Data Items

Risk of Bias Assessment

Summary Measures

Data synthesis and analysis

Determining cost-effectiveness

Results

Study selection and characteristics

Risk factors and categories of stroke

Risk Stratification and the critical risk point

Cost of investigation relevant risk factors (putative model components) in USD

Assembling models and model components

Examining pre-existing models/tools vis-à-vis our study findings

Discussion

Limitations

Conclusions

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 2