Petechial hemorrhages accompanying lobar hemorrhage

A significant proportion of spontaneous lobar hemorrhage in the elderly is due to cerebral amyloid angiopathy (CAA). ^[1] CAA can cause petechial as well as large lobar hemorrhages. ^[2] Like the large lobar bleeds, the petechial hemorrhages are typically situated in cortical or corticosubcortical (C/CS) regions. They are generally not resolved on CT, but were detected by MRI in a few published cases of pathologically confirmed CAA. ^[3-5] The frequency of radiographically detectable petechial hemorrhages in patients with lobar hemorrhage, however, is unknown.

Petechial hemorrhages, if detectable in a significant proportion of patients, might aid in the diagnosis and staging of CAA during life. Although petechial bleeds can accompany hypertensive as well as CAA-related hemorrhage, ^[6] multiple bleeds restricted to C/CS territories are not characteristic of hypertensive disease and are therefore supportive of a clinical diagnosis of probable CAA. ^[7] Petechial hemorrhages may also be responsible for some clinical manifestations of CAA, such as recurrent episodes of spreading neurologic symptoms. ^[5]

We utilized the technique of gradient-echo MRI for detection of petechial hemorrhages. Gradient-echo MRI enhances the magnetic susceptibility (and resultant signal dropout) produced by chronic blood products, thus increasing sensitivity for hemorrhage. ^[8] We include a report of one patient in whom gradient-echo MRI detected a small bleed apparently causing recurrent neurologic spells.

Methods.

Between July 1994 and June 1995, 23 patients over age 60 years presented to the Massachusetts General Hospital with primary lobar hemorrhage on CT. (This figure excludes patients with known causes of hemorrhage such as trauma, cerebral tumor, coagulopathy, or vascular malformation.) Of the 23 patients, 13 were prospectively evaluated using T₂-weighted and gradient-echo MRI. The 10 patients not evaluated by MRI consisted of 5 with terminal clinical state (1 with CAA at postmortem examination), 2 with diagnoses previously established by hematoma resection (both showing CAA in the pathologic tissue), and 3 whose treating physicians chose not to obtain MRI. An additional 2 patients (patients 1 and 10) were evaluated by MRI during this time period for hemorrhages that had occurred 1 to 2 years previously and were added to the series. Mean age of evaluated patients was 75.9 years (range, 65 to 90) and sex distribution was eight men and seven women. MRI evaluation was generally performed within 3 days of the lobar hemorrhage, except as noted. Controls consisted of 10 consecutive patients (7 men, 3 women) over the age of 60 years (mean 72.4, range 62 to 83) undergoing MRI for other indications.

CTs were performed without intravenous contrast on General Electric 9800 scanners. MRI was performed on a 1.5-T superconductive magnet (General Electric). Fast spin-echo T₂-weighted (TR 4,000-4,200/TE 96-104/NEX 1) and multiplanar gradient-echo (TR 749-750/TE 50/NEX 2/flip angle 10/time 3:41) images were obtained in the axial plane.

Large lobar hemorrhage was defined as a hematoma measuring at least 1 cm in all dimensions occurring in the C/CS region. Petechial hemorrhage consisted of smaller lesions showing increased density on CT or decreased signal intensity on T₂-weighted or gradient-echo MRI. All studies were reviewed by two of the authors (S.M.G. and P.W.S.).

Results.

We prospectively obtained MRIs on 15 elderly patients with primary lobar hemorrhage demonstrated on CT. MRI examination included axial gradient-echo as well as routine T₂-weighted sequences.

Nine of the 15 patients with large lobar bleeds (60%) demonstrated accompanying petechial hemorrhages restricted to the C/CS region Table 1, Figure 1. The petechial bleeds were seen as decreased signal on T₂-weighted or gradient-echo MRI without surrounding increased T₂ signal to suggest hemorrhage into a larger lesion Figure 1. One of these 9 patients (patient 4) ultimately demonstrated CAA on pathologic examination of resected hematoma. Another 3 of the 15 patients evaluated, 2 with C/CS hemorrhages and one without, had one or more petechial bleeds at non-C/CS loci such as putamen, thalamus, or brainstem. Overall, 12 of the 15 patients with lobar bleed (80%) had detectable petechial hemorrhages, the number ranging from 0 to 14 per patient (median 3).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Neuroimaging studies in patient 5. Shown are CT (left), T₂-weighted MRI (center), and gradient-echo MRI (right) 3 days following lobar hemorrhage. In addition to the acute bleed seen on all studies, the gradient-echo sequence shows multiple areas of decreased signal in the corticosubcortical region (arrows) consistent with chronic hemorrhages.

Gradient-echo MRI, as expected, had the greatest sensitivity for petechial hemorrhage. This technique demonstrated all the bleeds seen by the other techniques as well as additional bleeds in 11 of the 12 patients with detectable petechial hemorrhages. CT alone detected definite petechial C/CS bleeds in 3 of 15 patients, and T₂-weighted MRI detected these lesions in 5 of 15 patients.

Gradient-echo MRI in 10 consecutive elderly control patients undergoing MRI for other indications revealed no similar C/CS petechial bleeds. Two control patients demonstrated T₂-bright lesions associated with a degree of hemorrhage, suggestive of bleeding into a previous ischemic infarction or traumatic injury rather than primary hemorrhage.

Gradient-echo MRI in one of the patients (patient 1) demonstrated a possible cause of recurrent neurologic symptoms. The patient was an 82-year-old ambidextrous man with history of a right parietal lobar hemorrhage in May 1993. He presented again in November 1994 with spells of tingling in the right fingers spreading smoothly to the right palm or elbow over 2 to 4 minutes and lasting 10 to 30 minutes. He had three such spells over 3 days, one accompanied by clumsiness of the right hand. Physical examination demonstrated a normal examination on the right. Carotid noninvasive testing showed only mild atherosclerotic disease of the carotid arteries. Gradient-echo MRI Figure 2 revealed multiple C/CS foci of hemorrhage not visible on CT or T₂-weighted MRI. One such lesion was located in the left parietal cortex near the postcentral gyrus, a position corresponding to the patient's spells of right hand paresthesias and clumsiness.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. MRI in patient 1. T₂-weighted (left) and gradient-echo (right) MRI were performed 19 months after right parietal hemorrhage for recurrent episodes of spreading symptoms in the right hand. Petechial hemorrhages seen on gradient-echo (arrows) included lesions in the right posterior parietal and left postcentral gyrus corticosubcortical regions.

Discussion.

Radiographically detectable petechial hemorrhages accompanying large lobar hemorrhages were surprisingly common in this prospective series. In 9 of 15 patients, petechial bleeds were present and restricted to a C/CS distribution, thus meeting our suggested criteria for probable CAA. ^[7] Although not entirely specific, the observed MRI pattern in the proper setting can be highly suggestive of CAA. Other possible causes of petechial hemorrhages, such as multiple cavernous angiomas, ^[9] hypertensive encephalopathy, ^[10] or diffuse axonal injury, are clinically less likely or excluded in these cases. The absence of T₂ abnormalities associated with the chronic petechial bleeds in these patients also helps exclude structural lesions that can cause acute lobar hemorrhage such as arteriovenous malformation or hemorrhagic tumor.

We demonstrated petechial bleeds in the putamen, thalamus, or brainstem in three patients. Although we await further pathologic correlation, we believe the likelihood of CAA is less certain. CAA-related hemorrhages in these locations can occur, ^[1] but are unusual. Another possibility is hypertensive hemorrhage, which more characteristically involves the deep territories; indeed, each of these three patients was hypertensive.

The sensitivity and widespread availability of gradient-echo MRI make it a valuable addition to the evaluation of lobar hemorrhage. This technique requires no additional software or hardware relative to routine MRI. It typically takes about 4 minutes of scanning time and generally does not add to the cost of the study.

We suggested previously ^[5] and in the current study that petechial C/CS hemorrhage can cause specific neurologic symptoms such as spreading weakness, numbness, or paresthesias. As patients with these symptoms are frequently treated with anticoagulant or antiplatelet agents for presumed transient ischemic attacks, the diagnosis of hemorrhage assumes the greatest practical importance.