An accurate and timely minimum diagnostic work-up in early-onset ischemic stroke is necessary to target treatment properly, prevent recurrent stroke, and facilitate recovery.18
Imaging of brain parenchyma and vasculature supplying the brain
The first step in etiological diagnosis of ischemic stroke is imaging of the brain and intra- and extra-cranial arteries, including the aortic arch. In clinical practice and especially in a hyperacute setting, the first imaging is most often performed with CT and, if needed, with a CTA. The radiation dose of a head CT is comparable to approximately 3 months of background radiation.155 However, with its best sensitivity and spatial resolution, MRI offers improved differential diagnosis,156 and it should include at least diffusion-weighted imaging (DWI), fluid-attenuated inversion-recovery (FLAIR), and susceptibility-weighted imaging (SWI). It was also shown in systematic reviews that contrast-enhanced magnetic resonance angiography (MRA) is the most specific and sensitive non-invasive modality to exclude intracranial or extra-cranial arterial abnormalities, such as intramural hematoma, in dissection with axial T1 fat-saturated sequences of the neck.157,158 As described above, in some cases, e.g. with strong suspicion of RCVS, arterial imaging should be repeated after the first week of symptom onset and also DSA might be necessary if the diagnosis remains unclear. Furthermore, in brain vasculitis, especially in primary angiitis of the central nervous system, biopsy of the leptomeninges and the cortex is still considered the gold standard since with suspected vasculitis high-resolution contrast-enhanced vessel wall imaging is recommended.159 However, MRI findings might not be specific, and angiography has a low sensitivity and specificity.160
Clinical examination and laboratory testing
In addition to imaging, patients should be interviewed for full clinical and family history and examined for cardiovascular status, inspecting the skin, extremities, nails, eyes, mouth, and mucous membranes. Stature and deformities should be registered.111 Diagnostic work-up in early-onset stroke should include routine blood samples and hypercoagulability testing to exclude disorders such as prothrombin gene mutation and the presence of antiphospholipid antibodies.161 When rarer etiologies as suspected, such as vasculitis, a wider spectrum of blood
30 and cerebrospinal fluid samples should be considered. A thorough work-up is necessary to exclude also stroke mimics such as migraine (with aura) or symptoms caused by inflammatory demyelinating diseases and central nervous system infections.
Cardiac examinations
Cardiac diagnostic work-up of young ischemic stroke patients is necessary in the absence of an obvious other cause.162 A 12-lead ECG is the first-line investigation that can reveal occult AF and suggest other high-risk sources of embolism. For example, a Finnish study showed that P-terminal force in lead V1 was independently associated with CE from a high-risk source in young patients.163
Both TTE and TEE are recommended for young patients since detection of PFO in TEE could lead to PFO closure in selected patients with high-risk PFO features and/or high RoPE score.73 However, ESUS criteria necessitates only TTE,43 albeit Katsanos and colleagues recently showed the benefit of TEE particularly in younger ESUS patients in changing treatment strategy.164 Transcranial Doppler ultrasound with bubble study (TCD-BS) belongs to the current cardiac work-up of young ischemic stroke patients and provides a tool for detecting right-to-left shunts. According to the meta-analysis of Mojadidi and colleagues, TCD-BS had a high weighted mean sensitivity and specificity (97% and 93%, respectively), with TEE as the reference method.165 TCD-BS is performed using a contrast, either agitated 10 ml saline-blood mixture or commercial ultrasound contrast agent, simultaneously detecting potential micro embolic signals (MES), e.g. in the middle cerebral artery via a transtemporal bone window.166 At least two tests should be done, first at rest and then repeated with Valsalva maneuver.
However, if there is already evidence of large right-to-left shunt at rest, Valsalva maneuver is usually not required. The size of the right-to-left shunt in unilateral TCD monitoring can be classified as low grade (1-10 MES), medium grade (over 10 but no curtain of MES), or high grade, i.e. a curtain. In bilateral monitoring, the corresponding definitions are 1-20 MES, over 20 MES but no curtain, and a curtain where a single MES cannot be discriminated with the TCD spectra.167
Cardiac CT should be used as first-line diagnostics in e.g. valve patients with suspected valve embolus and other cardioembolic causes of ischemic stroke.162 Cardiac MRI can be used as an ancillary study to exclude e.g. amyloidosis, sarcoidosis, heart failure, or hypertrophic cardiomyopathy when TTE and TEE do not find an embolic source or TEE is
31 contraindicated.168 Chest CT can be performed to detect pulmonary arteriovenous fistula or pulmonary embolism. Notably, right-to-left shunts can be caused also by e.g. esophageal variceal-pulmonary venous fistula or a direct right pulmonary artery to left atrium fistula.169,170 Furthermore, prolonged ECG monitoring can detect occult AF more frequently in older ESUS patients. One study with 157 young ischemic stroke patients (mean age 43 years) reported that AF might actually be more common than expected, with 14 subjects presenting with AF. AF was also associated with higher NIHSS scores.54 Another study reported that prevalence of AF was as high as 10.2% in 98 young patients and that using long-term ECG monitoring improved the detection of AF.171
Ruling out venous thrombosis
Ultrasound of the lower extremities to exclude venous thrombosis as a source of paradoxical embolism and screening for coagulopathies should be considered.18 Incidence rates for deep venous thrombosis alone in young patients vary between 0.2 and 0.9 per 1000 person-years and seem to be higher in women of childbearing age than in same-aged men.172,173 A recent study suggested a trending association between pelvic thrombi and high RoPE Score in stroke patients with PFO.174 Furthermore, Lapergue and colleagues proposed that patients with cryptogenic stroke and PFO might benefit from combined CT venography and pulmonary angiography in detecting deep vein thrombosis and pulmonary embolism (total combined prevalence 11% in 114 patients (95% CI 5.5-17.7).175 Stroke can also be caused by a cerebral venous thrombosis, but the topic is beyond the scope of this review focusing on ischemic stroke.
Testing endothelial function
Measuring endothelial function in early-onset ischemic stroke patients is still not part of the routine diagnostic work-up, and it has not been examined or validated in early-onset cryptogenic ischemic stroke. However, testing endothelial function might provide useful information regarding stroke mechanisms and for tailoring secondary prevention. In regulation of vascular tone, endothelial cells release vasoactive substances such as prostacyclin, nitric oxide, endothelin, and angiotensinogen. Furthermore, the endothelium plays an essential role in thrombosis, platelet activation, and leukocyte adhesion.176 Several risk factors, including
32 increasing age, male sex, higher body mass, smoking, and diabetes, have been shown to be associated with impaired endothelial function causing disturbances in regulation of vascular tone, platelet aggregation, and modulation of inflammation.177,178
Endothelial function can be measured either with flow-mediated dilatation (FMD) test or indirectly with peripheral arterial tonometry (PAT), the latter being less invasive.177 Other methods include venous occlusion pletysmography and evaluation of circulatory markers.
Debate continues regarding which of these methods is the most appropriate, i.e. the “gold standard”, for assessing endothelial function. Some of these, such as PAT, measure response to hemodynamic provocation tests or pharmacological agents.176 Only a few studies have suggested an association of these factors and endothelial dysfunction in younger ischemic stroke patients. One included young patients with spontaneous cervical artery dissection and showed that cervical artery dissection patients had significantly lower FMD index than patients with cryptogenic stroke (5.0 ± 9.3% in vertebral artery dissection, 5.7 ± 6.2% in internal carotid artery dissection, and 13.2 ± 6.5% in cryptogenic stroke; P<0.0005). Another study in middle-aged patients compared those with ischemic stroke with patients with hemorrhagic stroke and reported a significant difference in FMD index (3.8 ± 0.9% vs. 6.0 ± 1.6%).179,180
Taken together, more studies are needed to assess factors affecting endothelial function and the association between endothelial dysfunction and early-onset cryptogenic ischemic stroke.
Furthermore, no data exist regarding whether improving endothelial function also decreases patients’ risk of recurrent strokes and long-term outcome.