A systematic review of interventional acute ischaemic stroke treatments in Latin American countries has found that mechanical thrombectomy usage in these regions is “comparable to international levels”. Writing in the Journal of Stroke and Cerebrovascular Diseases, Alejandro Gonzalez-Aquines (University of Bradford, Bradford, UK) and colleagues note that this provides evidence of the recent advances Latin America has made with regard to stroke care.
“The data regarding safety outcomes and the effectiveness of mechanical thrombectomy were consistent across the studies we examined, not to mention supported by the available literature,” they further state. “However, mortality rates did remain high. The poor quality of evidence and the limited number of eligible countries suggests that further research is necessary. Future studies on mechanical thrombectomy outcomes in both public and private healthcare settings throughout Latin America would be greatly beneficial, but they must be conducted using standard time window criteria, and reduce the risk of bias by including representative samples and comparison groups.”
In an attempt to describe the use, effectiveness and safety of endovascular stroke thrombectomy procedures in Latin American countries, Gonzalez-Aquines et al identified relevant studies across multiple databases (CINAHL, MEDLINE, Web of Science, SciELO, EMBASE and LILACS), and synthesised them according to effectiveness—as per recanalisation rates—and safety measures—via mortality and functional independence at 90 days.
Some 17 studies were included in the researchers’ review after their identification and screening process, and only six countries were represented in these articles: Argentina, Brazil, Chile, Ecuador, Mexico and Puerto Rico. According to Gonzalez-Aquines et al, most of the studies were conducted in public or university hospitals; the average duration of studies was 51.4 months; and just one was a randomised trial, with the remainder being either case series (nine) or cohort studies (seven).
“The included studies had great methodological heterogeneity due to differences in study design, the mechanical thrombectomy time window, and stroke location,” the authors concede, acknowledging the limitations and potential risk of bias in their review. All but two of the studies (87.5%) were classified as poor quality, and most failed to recruit a representative population sample, while only two included comparison groups, they add.
Nevertheless, across the 17 studies they included, stroke thrombectomy utilisation rates ranged from 2.6% to 50.1%. Recanalisation rates following mechanical thrombectomy treatments varied between 63.7% and 95%, while the majority of studies reported a modified Rankin Scale (mRS) score of 0–2 in less than 40% of followed-up patients. Gonzalez-Aquines et al point out that the single randomised trial included in their analysis reported a positive odds ratio (OR) of 90-day mRS of 2.28—indicating favourability of thrombectomy over standard care.
“The rate of intracranial haemorrhage (ICH) also differed in the included studies,” the authors add. “Some reported asymptomatic haemorrhage, while others had symptomatic haemorrhage. Among those reporting asymptomatic haemorrhage, rates varied between 3.7% and 45.5%. However, the rates of symptomatic haemorrhage remained below 10% in all studies with this outcome, ranging from 4.1% to 8.7%.”
Discussing the final safety measure of interest, Gonzalez-Aquines et al write that mortality rates 90 days after the thrombectomy procedure remained below 30% in most studies, with only three reporting rates above 30%. However, these three studies all involved patients with strokes caused by posterior circulation occlusions.
“The findings on the review’s three main outcomes (mechanical thrombectomy rate, effectiveness and safety) were consistent among the included studies despite the heterogeneity of the study design,” the authors state. “The lack of a representative sample and comparison groups compromised the quality of the studies and increased the risk of bias—with the exception of the only randomised trial, which showed an overall low risk of bias.”
A further limitation they highlight is the fact that, although they originally considered conducting a meta-analysis, this was not possible due to methodological differences. The majority of studies also lacked a representative sample, constituting “a considerable limitation in generalising their findings”.
“Nonetheless, there is consistency in the outcomes of the studies in our review, including wider comparisons with available international literature,” Gonzalez-Aquines et al add.