This descriptive study of ischemic stroke in young patients is the biggest one done in Colombia so far. It was performed in a University Hospital. Most patients were female, which differs to previous case series,3,7,10,15−22 although concordant with most recent global data.4
Contrary to previous observations,7,8,19,21,23 no predominance of female over males under 30 years was observed, which is similar to other studies.24 However, over 30 years, female predominance was observed, contrary to most series. A recent cohort study found a higher ischemic stroke incidence on female patients of all age groups, particularly between 18–44 years;25 this merits further research, and could be explained by an improval on healthcare access, biases on previous descriptive studies, among other things.
A high frequency of traditional risk factors, specially arterial hypertension and smoking, was observed, which is similar to the evidence previously published.3,7,8,10,13,15,16,18–20,22,24,26−32 Nonetheless, dyslipidemia wasn’t identified as one of the main risk factors, unlike previous studies.3,7,8,10,15,19,20,22,24,27,29−31 In some series, alcohol consumption has been described as one of the main risk factors,18–20, 23,30 similar to what was observed in our study. The age-related increase of such traditional risk factors may explain, in part, a similar tendency on the atherosclerotic and lacunar stroke; nevertheless, the influence of these traditional risk factors is unclear for the larger proportion of patients with undetermined etiology and other determined etiologies observed at younger age.
A remarkable finding in the present study is the high rate of illicit substance consumption, compared to previous studies,7,13,19,24 which is directly related to the determined etiologies of stroke in this series.
The frequency of migraine was discretely lower than the general population,33 which differs from previous studies that consider migraine an important risk factor, primarily on women.7,13,22,24,26,31,36 The low frequency may be explained by underdiagnosis, and/or a lower relevance on this population. Other possible explanations may be considered.
Unlike previous reports, with high frequency of oral contraceptives use on younger women which could explain a female predominance at younger ages,3,7,18,19,22,24,26,31,32 this wasn’t observed in the present study.
The results according to TOAST classification from this study are similar to those obtained on the largest series (to our knowledge) which included 3,331 patients in multiple centers from Europe.34
The proportion of large-vessel atherosclerosis (6.8%) was similar to previously published series in Europe and South America (6.7–9.3%),7,17,19,24,34 and particularly to the previous Colombian study (6.5%);13 Mexico has reported lower rates.23
Small-vessel disease in this study (7.6%) was lower than what was previously reported (12.2–42.5%),7,17,19,20,24,27,34,35 however, it triples the previous Colombian study which found a 2.6% rate.13 It was similar to the Italian and Swiss series (5–9%).3,18 A subject was diagnosed with CADASIL, a condition of interest in our country, as an extensive family group with this condition has been described.36,37
The frequency of cardioembolic etiologies (17.7%) was similar to the European publications (15.8–20.1%),7,18,19,24,27,34 although lower than previous reports in Latin-America (23.6–28.3%).13,17,23 It is remarkable the higher frequency of rheumatic heart disease in our study (10%) and other developing countries17,23 compared to European countries.7,18 On the other hand, the frequency of cardioembolism was higher in the present study compared to a french one,15 probably due to low-moderate risk cardioembolic sources were classified as undetermined causes.
Cardioembolic rates were higher at older ages, mostly over 30 years, as has been previously described.8,17,22,23,31
The rate of other determined etiologies (25.7%) was similar to those reported by European case series between 19–29%.3,7,8,15,18,19,22,24,26,27,34,35 Conversely, Latin American case series have reported a higher rate of this category, between 33.5–39.6%.13,17,23 The most common determined etiology was craniocervical arterial dissection, in 11.4% of cases, similar to the 12.8% found in the largest study,34 and in the same direction as previous studies which report values up to 24%.3,7,8,13,15,22–24,26,35
A worrying point is the high rate of substance abuse-related stroke (2.9%) compared to European series which have reported a 0.3–1.5% rate7,18,24,34 which is double or even 10 times the case number. Additionally, substance abuse participated on the genesis of the vascular event with other causes, hence being classified as undetermined etiology, but it remarks the importance of this matter. In general, the most commonly implied substance was cocaine-induced vasculopathy. In Latin-american case series, this cause was rare.13,17,23
It’s worthwhile mentioning that a vasculitis case of unconfirmed etiology was observed, although with high suspicion for neurocysticercosis. This etiology has been observed in Brazil and Mexico case series, with frequencies of 1.9% and 4.6%, respectively.17,23 A Colombian case of ischemic stroke due to neurocysticercosis related vasculitis has been reported.38 Another remarkable findings was two cases of tuberculous vasculitis in our series, an etiology not reported in previous series; hence, the proposal of considering it as a potential etiology in undetermined cases, particularly in countries where Mycobacterium tuberculosis infection is endemic. Additionally, meningovascular syphilis was another neuro-infection causing stroke.
Most studies found a low rate of migrainous infarction on 0.2–2.8%3,7,13,17,22,24,26,27,31,34,35 while other studies found a slightly higher rate (3.3–4.8%).18,23 The present study found no migrainous infarction cases similar to an Italian case series.8
Patients with more than one potential etiology of stroke, and those with no etiology identified despite extensive/insufficient work-up, were included in the undetermined etiology category. This category obtained the higher proportion in our study (42.4%), similar to previous studies (32–44%).3,7,18,19,23,27,34,39 One of the previous series presented a higher rate (62.4%) of undetermined cases,15 which may be explained due to potential cardioembolic sources (such as PFO or ASA) being classified as undetermined.
The high frequency of undetermined TOAST in the present study may be explained by various factors. Nonetheless, some patients underwent different diagnostic studies at different institutions, hence, the result was unknown; in other cases, it wasn’t possible to access the chart data or diagnostic studies results in older charts, which made it difficult to determine possible etiologies of stroke.
The vascular territory was determined with neuroimaging results. The most commonly involved was the anterior one, similar to most previous studies.3,7,8,13,16,18, 20–22,24,30,32 The distribution of involved territories was very similar to previous Colombian series.13 Despite the frequent use of cerebral magnetic resonance imaging (MRI), no high rate of posterior circulation involvement was observed unlike some previous studies.7,22,24
Among the strengths of the present study include: that it was performed on a University Hospital which takes care of a great volume of patients from all economic strata, which could potentially be reflected on the wide variety of etiologies. The biggest series on Colombia and the world were developed on University Hospital, just as the present study; however, a larger number of determined etiologies was observed, compared to the previous Colombian series, despite being developed in similar times.13 Another strength was the description of demographic characteristics, which may support further studies to explore social determinants of health in these patients.
There are some limitations in our study. One being the high frequency of the undetermined etiology as was previously developed. The observational design has intrinsic biases. Variable definition depending on chart report may be different from previous studies, which implies some difficulties for comparison. On substance consumption, it wasn’t possible to determine the frequency, hence it was defined in a dichotomous way. No interaction analysis between risk factors (contraceptives-thrombophilia, migraine-smokin, migraine-contraceptives, contraceptives-hypertension, etc) which may show a differential performance for stroke etiologies. Finally, NIHSS (National Institute of Health Stroke Scale) score wasn’t registered, nor was disability through modified Rankin score.