Thrombotic thrombocytopenic purpura: Brain CT and MRI findings in 12 patients

Methods.

We present 12 consecutive TTP patients seen between 1990 and 1997. TTP diagnosis required microangiopathic hemolytic anemia (Coombs’ negative), thrombocytopenia, and multi-organ dysfunction without other causes.4 Patient 5, who had TTP that was thought to be postinfectious, was also taking cyclosporine. His clinical recovery occurred with TTP treatment while continuing the same dose of cyclosporine. His cyclosporine levels were therapeutic. Therefore, this patient most likely had TTP rather than cyclosporine toxicity. Although Patient 12 had postpartum TTP, there was no evidence of toxemia during pregnancy. Patients had neurologic and neuroimaging evaluations during acute neurologic symptoms. Fifteen CTs were performed on 11 patients (all except Patient 12), and 12 MRIs were performed on 8 patients (Patients 3 through 6, 8, and 10 through 12). MRI was performed at 1.5 T using conventional spin-echo noncontrast T1-weighted images (T1WI), T2-weighted images (T2WI), and proton-density images (PDWI). Postcontrast MRI was performed in three patients (4, 8, and 11). Both CT and MRI were performed in seven patients (3 through 6, 8, 10, and 11).

Results.

Clinical and laboratory features are summarized in the table. Seven patients had elevated blood pressure at the time of neurologic symptoms, with moderate hypertension at worst. The hypertension was seen in combination with acute renal failure. Neurologic manifestations were varied (see table). CSF analysis (n = 3) was normal except for mild protein elevation in one patient and minimal leukocytosis in another. EEG showed generalized slowing (n = 6) or normal findings (n = 1). Autopsies (Patients 7 and 9) showed microthrombi in small cerebral vessels. Patients were treated with plasmapheresis (n = 11) or fresh frozen plasma (Patient 10).

Neuroimaging studies demonstrated acute brain lesions in nine patients (75%) (see table and figures 1 and 2⇓). Patients 1 and 2, studied with CT only, and Patient 11, studied with CT and pre/postcontrast MRI, all had normal neuroimaging. MRI was sensitive in detecting TTP lesions, showing abnormal findings in 88% of patients, whereas CT was abnormal in 45% of patients (see figure 1). Three lesion types were noted: reversible bilateral cerebral edema (n = 7), ischemic strokes (n = 3), and frank hematomas (n = 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Patient 5. CT versus MRI in the detection of brain lesions associated with thrombotic thrombocytopenic purpura. (A) Noncontrast CT performed on the same day of neurologic symptom onset. This axial slice shows no striking abnormality, although vague hypodensities in occipital and temporal regions are present. (B) MRI (T2-weighted, TR/TE 2,000/80 msec) performed on the same day as the CT shows hyperintense lesions (arrows) involving bilateral posterior temporal and occipital lobes (both gray and white matter) and basis pontis on the left. Lesions are seen more clearly than they are by CT and are consistent with reversible posterior leukoencephalopathy syndrome and hypertensive encephalopathy.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Patient 12. Serial MRI shows reversible cerebral edema (reversible posterior leukoencephalopathy) associated with thrombotic thrombocytopenic purpura. (A–C) Initial MRI (T2-weighted, TR/TE 2,300/80 msec) performed on the same day of neurologic symptom onset shows bilateral hyperintense lesions (arrows) involving posterolateral temporal, posterior frontal, and occipitoparietal lobes (white matter more involved than gray matter). (D–F) Repeat MRI (T2-weighted, TR/TE 2,300/80 msec) performed 12 weeks later shows resolution of all lesions.

Reversible cerebral edema lesions (see figures 1 and 2⇑) were mildly hypointense on T1WI and hyperintense on PDWI/T2WI. This edema was typically posterior—in occipitoparietal lobes, posterior frontal lobes, brainstem, and cerebellum (see table and figures 1 and 2⇑). Lesions usually involved white matter and sometimes extended to the gray-white junction and the cerebral cortex (see figure 2). These edematous lesions resembled reversible posterior leukoencephalopathy syndrome (RPLS),5 including cases that we have reported.6 Petechial hemorrhage was uncommon (n = 2). These lesions showed minimal enhancement (Patient 4) or no enhancement (Patient 8). Serial studies showed edema resolution (see figure 2). One patient had marked neurologic recovery suggesting reversal of lesions that was not confirmed by imaging (Patient 5). Reversible edema was associated with renal failure and hypertension because most patients with this edema had both renal failure and hypertension (n = 5/7). However, elevated blood pressure and acute renal dysfunction existed in two patients without cerebral edema. Patients with edema alone (no strokes or hematomas) had a favorable neurologic outcome (4 excellent, 1 fair) (see table).

Strokes involved posterior cerebral artery (n = 3), middle cerebral artery (n = 1), lenticulostriate (n = 1), and cerebellar (n = 1) territories, indicating that large- or small-vessel involvement could occur. Many infarctions were hemorrhagic (n = 3). One patient had a pontine hematoma. Patients with infarction or hematoma had an unfavorable outcome (see table).

Discussion.

This study shows that TTP is often associated with clinical and neuroimaging features of RPLS. Similar to RPLS due to other causes, TTP-associated RPLS lesions resolved on follow-up imaging studies and led to neurologic improvement. TTP was also associated with strokes and brain hemorrhages, which led to a poorer outcome. Therefore, in addition to coagulopathy, cerebrovascular autoregulatory dysfunction seems to contribute to the pathophysiology of TTP.

RPLS may be triggered by hypertension, toxemia, and immunosuppressive agents.5,6 RPLS is most likely due to impaired cerebrovascular autoregulation leading to edema.5 Renal dysfunction appears to predispose the brain to RPLS.6 Therefore, RPLS in TTP may relate to multiple factors, including endothelial injury, hypertension, and renal failure.1 Consistent with this hypothesis, RPLS in our TTP patients was associated with acute hypertension and renal failure. We have reported that RPLS may develop with only moderate acute elevations in blood pressure, especially in patients with renal failure.6 Therefore, TTP is an important cause of hypertensive encephalopathy and should be added to the expanding spectrum of RPLS. Furthermore, moderate hypertension in TTP patients should not be ignored.

Previous neuroimaging correlations of neurologic dysfunction in TTP are sparse, consisting primarily of case reports. These previous CT/MRI studies of TTP most commonly reported ischemic strokes,2,7,8 hemorrhages,2,9 or normal findings.2,3 Although some authors have described disappearing abnormalities that may have represented edema,8,10 these lesions were not well characterized. The only previous MRI series of TTP showed normal findings in five patients; however, these patients were studied during the chronic phase (long after neurologic illness.)3 Therefore, ours is the first series showing acute MRI manifestations of TTP. Documenting characteristic multiple neuroimaging lesions, including RPLS with or without hemorrhagic strokes, may aid in the recognition of TTP. In addition to improving the diagnosis of TTP, neuroimaging may be useful in assessing prognosis. Previous CT work indicates that normal findings portend a favorable neurologic outcome, whereas infarctions or hemorrhages indicate more severe disease.2 The present study adds that MRI demonstration of RPLS-like lesions alone may predict a favorable neurologic outcome. Our study suggests that MRI is more sensitive than CT in demonstrating brain lesions in TTP. For example, three patients had abnormal MRI when CT was normal. However, caution must be exercised because of the small number of patients and the possibility that MRI may have been performed at a time in the evolution of edema more likely to show changes than when CT was performed. Our hypotheses regarding MRI diagnostic sensitivity and prognostic utility in TTP will be strengthened by larger prospective clinical-neuroimaging studies.

The demonstration of reversible edema, infarctions, and hemorrhages in our series may be explained by the multi-factorial pathogenesis of TTP.1 The salient pathophysiologic factor in TTP is endothelial injury followed by the release of a variety of mediators that promote platelet activation and widespread hyaline microthrombi formation. A consumptive coagulopathy results in thrombocytopenia and predisposition to bleeding. TTP is a multi-organ disease, with a propensity to affect the brain, gut, and kidneys. Therefore, the myriad of neurologic complications in TTP is explained by the variety of vascular insults that may affect the brain such as endothelial injury, thrombus formation, bleeding diathesis, and hypertensive encephalopathy.

TTP often results in multiple cerebral lesions of several types that are commonly demonstrated acutely by MRI. RPLS, especially hypertensive encephalopathy, may occur in TTP, and infarcts and hemorrhages may also occur. These multiple lesion types may occur in the same patient. MRI may aid in the diagnosis and prognostication in TTP and may be more sensitive than CT in demonstrating brain lesions. Despite the extensive MRI abnormalities, prompt treatment of TTP-associated RPLS often results in neurologic recovery. TTP should be added to underlying causes to be searched for when RPLS is suspected.