Stroke represents a major cause of death, disability and dementia worldwide (1). It has been known for decades that the incidence and prognosis of stroke are worse in low and middle income countries relative to high income countries, a difference that has been increasing over time (1). Furthermore, within high income countries, there are also major differences in incidence and prognosis of stroke between the more socially disadvantaged groups relative to less disadvantaged (2). Although it is difficult to disentangle ethnicity and social factors, it has been suspected that ethnicities of African ancestry were more at risk for stroke; they also had a poorer prognosis. Whereas, some studies suggested that 39% of the excess mortality in Blacks was due to lower socioeconomic status, others, controlling for risk factors and health insurance level, found that the differences between blacks and non-blacks were in fact explained by the higher prevalence of standard cardiovascular risk factors in blacks relative to non-blacks (3).

French Guiana is a multiethnic Amazonian territory with populations from African, European, and Asian ancestry. It is also a territory of great social contrast: It has the highest GDP per capita in Latin America, while 29% of the total population, and nearly half of adults, consists of foreigners with great social vulnerabilities (4). The incidence of patients hospitalized for ischemic stroke and hemorrhagic strokes in French Guiana is the highest of all French territories (5, 6). The overall standardized incidence of stroke was 297 per 100 000 in 2017 (7). We have previously shown that social inequalities of health were present for nearly all major pathologies –HIV, Cancer, Preterm birth, terminal renal failure- and that precarious populations often renounced care (8–13). Despite being a French territory, French Guiana has a shortage of medical and paramedical staff, and a lack of specialized structures. Hence in 2018, there were an estimated 652 doctors and 2.4 neurologists per 100,000 inhabitants (14). The social differences in incidence, severity, and prognosis of strokes are possibly the result of greater frequency of risk factors, problems in accessing care, and the quality of emergency care and rehabilitation therapy. At a time when progress on the control of risk factors and improvement of emergency care of strokes have led to significant progress, the most socially vulnerable segments of the population may not benefit equally from these advances. Between 2011 and 2015 INDIA a prospective study was conducted in French Guiana and in Dijon, France. Because of the sociocultural complexity of French Guiana (4, 15) and its lack of specialized care (16), we hypothesized that precariousness would be frequent in our cohort, and that it may influence presentation and outcomes. Our main objectives were to compare the epidemiology of ischemic strokes and their outcomes between French Guiana and mainland France, and to compare strokes between precarious and non-precarious patients within French Guiana; the main outcomes studied were risk factors, case fatality, stroke mechanism, treatment of stroke and outcomes were both compared between French Guiana and Dijon, and within French Guiana between the precarious and non-precarious.

The comparison of ischemic stroke patient characteristics and outcomes between 2 contrasted French territories is important to pinpoint specific areas of intervention. Indeed, the characteristics of patients included in French Guiana were very different from those included in Dijon. Patients included in French Guiana were more precarious (EPICES score >30) than patients included in Dijon. Patients in French Guiana had a more precarious family situation; they were less often followed medically by general practitioners, specialists, and dentists. They were also less likely to have health insurance than in Dijon. The use of thrombolysis appeared less frequent in the patients included in French Guiana than in the patients included in Dijon but the difference disappeared after adjusting for patient characteristics.

Within French Guiana, there were also significant health inequalities which were more salient when comparing foreign patients with French patients than when comparing precarious patients using the EPICES score. Among stroke patients in French Guiana, Foreign patients were about 4 years younger than French patients. There was no significant difference between precarious and non-precarious patients regarding the delay between symptoms and admission, but among those who benefited from thrombolysis, French patients had a slightly shorter delay between admission and thrombolysis. Non-precarious patients were also more likely to benefit from first intention MRI than precarious patients. A counterintuitive finding was the lower 1-month mortality in socially precarious patients than in non-precarious patients, however the trend was no longer significant after adjusting for age and initial severity. This is different of what was observed in the UK where foreign patients generally have had a consistently higher mortality than UK citizens (19, 20). One of the hypotheses to explain this lower case-fatality is that socially precarious and foreign patients were younger and had fewer comorbidities, which improved survival rates despite apparently more severe strokes. Recently, it was shown that socially precarious patients had increased mortality which was delayed between 3 and 12 months; but here again this was not observed, perhaps within French Guiana the total number of deaths was insufficient for robust statistical analyses (21).

The study design did not include all patients with a stroke and thus did not allow calculating incidence. Over half of patients were foreign, which seems high and would suggest a higher incidence among foreign patients, but this corresponds to the demographic situation among adults in the general population of French Guiana (15). Over 2/3rds of the patients were considered socially precarious whereas in Dijon France it was 39.4%. Foreign patients were more frequently socially precarious than French-born patients. In French Guiana, 53% of persons live under the French National poverty level (4) and a study in 2011 showed that 73.3% of hospital patients in Cayenne were socially precarious (EPICES score>30). The proportion of socially precarious persons among stroke patients was high in French Guiana but it is difficult to affirm that the incidence of stroke was higher in precarious persons than in non-precarious persons; however the younger age at diagnosis would definitely suggest so. An important potential limitation in the present study is that the strict inclusion criteria requiring the presence of kin for patients with severe strokes and the requirement of reachability after discharge may have led to the exclusion of very precarious patients coming from the most remote areas of French Guiana, often late and alone. Not including these patients may have introduced a bias minimizing the representation of severe precarious patients consulting late. Future prospective studies in French Guiana should therefore have broader inclusion criteria to limit this bias. Patient follow-up by phone may have biased estimations as the availability of a phone may be less frequent in the precarious. Another limitation is that the Dijon Comprehensive Stroke Center may concentrate large artery occlusions that are referred from other smaller hospitals, which may partly explain the observed differences in mechanisms (more cardio-embolic etiology) and severity (greater NIHSS score on admission). Another potential limitation is that the comparisons within French Guiana may have required more statistical power to unmask differences between groups. Nevertheless, the present study was prospective and gives the first rigorous comparison of ischemic stroke between French Guiana and Dijon in France and the first study of health inequalities and stroke within French Guiana.

We have shown that among precarious populations, and notably among migrants, the picture for many diseases is often similar: a later presentation with more advanced disease and often more deaths (8–10, 22). Socially vulnerable populations will renounce care for reasons pertaining to cost or transport (11, 12, 16). For reasons that are unclear, the more socially vulnerable patients and foreign-born seemed less likely to benefit from MRI both in 1st intention and for control. Foreign-born patients were less likely to benefit from supra-aortic and transcranial doppler and ultrasonography. Foreign-born patients also had longer intervals between admission and thrombolysis but not between stroke and thrombolysis. The interval between stroke and admission seemed longer but this was not statistically significant. Previous studies in other countries have also shown some social inequalities in the provision of care notably regarding delays, imagery, rehabilitation treatment and prevention, but there was no consistent pattern of inequality by socioeconomic status (2). Here, one plausible explanation could be that uncertainties linked to the frequent linguistic barriers among foreign patients require longer intervals to ensure patients do not have contraindications for MRI or thrombolysis.

Overall, foreign and more socially vulnerable patients were younger and had more severe lesions but they were less likely to have a past vascular history. Despite some elements suggesting decreased access to care and diagnostics they had increased survival and were generally oriented in the same manner. The overall picture is thus nuanced. There seem to be social inequalities of health regarding stroke, but not quite to a level observed elsewhere, because patient survival was higher among the most socially vulnerable, despite more severe lesions, presumably because they were younger. This difference with observations from elsewhere (2) may reflect particularities in the demographics of immigrants in French Guiana; it could also show that hospital care is relatively accessible for the most vulnerable (11, 12) in the universal health care system in French Guiana. In line with theoretical work on the strength of weak ties and structural holes, ego network studies suggested that social networks topology also had an influence on delays to consult (23). Here we did not find any differences in pre-hospital delays between precarious and non-precarious groups, perhaps the small size of the cities and limited alternatives makes the health system easier to grasp in case of emergency. However, for prevention, screening, non-urgent care and chronic diseases such as hypertension or diabetes social vulnerability may definitely lead to dramatic outcomes. Three quarters of stroke risk factors are behavioral and thus amenable to prevention (1). Proactive interventions have demonstrated potential for rapid impact, for instance the US Hypertension Detection and Follow-up Project 1976–1979 rapidly lowered cardiovascular deaths by 26% (24). In the French territory with the highest incidence of stroke (6), this suggests that health promotion and screening efforts should be made to reach the population, and notably the most vulnerable groups, with rapid potential gains in terms of stroke-related morbidity and mortality. Beyond social inequalities in identifying the proximal determinants of stroke, studies have shown that the buildup toward these inequalities starts early in life with childhood circumstances, nutritional deficiencies during pregnancy and infancy (2, 25). In the territory with the highest fertility rate in Latin America, and one of the highest preterm delivery rate (9), where 30% of the population is below the poverty level, improvements in mother and child care may lead to additional benefits with regards to stroke. Moreover, lead, which has been listed among the risk factors for stroke (1) is at concentrations above 50 micrograms among 25% pregnant women (26) and the awareness of the magnitude of this problem, which predominantly affects poorer populations, is only dawning on health authorities. What impact this could have on the risk of stroke among the more socially vulnerable remains to be determined.

In conclusion, within French Guiana, over 2/3rds of patients were socially precarious and over half were foreigners; the differences between socially precarious and non-precarious patients were not as important as described elsewhere; the main difference was that foreigners were younger than French patients at the time of their stroke. Ischemic stroke in French Guiana was strikingly different from ischemic stroke in Dijon: patients were substantially younger and poorer in French Guiana than in Dijon; strokes were predominantly lacunar in French Guianaand cardio-embolic in Dijon but this may have reflected referrals of patients with large artery obstruction to the large reference stroke center in Dijon. Although thrombolysis was substantially more frequent in Dijon than in French Guiana in the crude analysis, taking into account patient differences showed no difference in the proportion of thrombolysis between French Guiana and Dijon. Within French Guiana, the risk factors, mechanisms, and outcomes seemed relatively homogenous across socioeconomic strata. Therefore, the marked differences between 2 contrasted French territories suggest that beyond health inequalities there may be a different epidemiology of cardiovascular risk factors between French Guiana and mainland France.

The raw data supporting the conclusions of this article will be made available by the authors, upon reasonable request.

The cohort was approved by the French Regulatory authorities CNIL (Commission Nationale de l'Informatique et des Libertés, autorisation n°911040) and the CPP Est 1/AFSSAPS (Comité de protection des personnes Est 1, Agence National de Sécurité du médicament et des Produits de Santé, autorisation n°2010/40 – AFSSAPS n°2010-A00940-39). All patients or legal representatives gave consent to participate. The patients/participants provided their written informed consent to participate in this study.

DR, MN, IF, and YB: conception. DR, EM, CM, MP-P, IF, YB, and ED: study conduct and inclusions. DR and IF: data analysis. MN: first draft writing. DR, CM, BdT, NS, IF, and MN: review and editing. All authors contributed to the article and approved the submitted version.

The study was funded by the PHRC (Programme Hospitalier de Recherche Clinique), French Ministry of Health.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.