Clinical Spectrum of Reversible Posterior Leukoencephalopathy Syndrome FREE

Reversible posterior leukoencephalopathy syndrome (RPLS), also referred to as posterior reversible encephalopathy syndrome, is characterized by neuroimaging findings of reversible vasogenic subcortical edema without infarction. Clinical presentation typically involves global encephalopathy, seizures, headache, or visual symptoms.¹ This clinicoradiographic syndrome can be triggered by eclampsia,² hypertensive emergency,³ or exposure to immunosuppression (most notably the calcineurin inhibitors).^{4 - 7} The clinical presentation of RPLS is often nonspecific, and therefore the diagnosis of RPLS has come to increasingly rely on magnetic resonance imaging (MRI).

With Mayo Clinic institutional review board approval, we retrospectively identified patients with RPLS clinically diagnosed at the Mayo Clinic between January 1, 1999, and December 31, 2006. We used a text-retrieval system (the Mayo Clinic Life Sciences System) that searches the final diagnosis in electronic clinic notes for coded text. Included patients had a clinical presentation and neuroimaging abnormalities consistent with RPLS with documented recovery clinically and on repeated neuroimaging. We collected data on demographics, predisposing conditions, presenting symptoms, documented blood pressure measurements the day of presentation, time to neuroimaging, time to clinical recovery, and time to repeated imaging. In addition, we gathered the results of laboratory studies, including cerebrospinal fluid analysis and electroencephalography.

All neuroimages (including brain computed tomographs and MRIs) were reviewed by study neurologists (V.H.L. and A.A.R.) for identification of atypical RPLS features, including significant frontal involvement, gray matter/cortical lesions, unilateral lesions, presence of hemorrhage, confluent RPLS lesions, brainstem/cerebellar involvement, corpus callosum lesions, recurrent episodes, and foci of permanent injury on repeated neuroimaging. The second neurologist (A.A.R.) was blinded to the clinical aspects of the cases. Permanent injury was operationally defined as lesions that were still present on brain imaging studies 2 weeks after initial MRI.

We identified 38 episodes of RPLS in 36 patients (20 females and 16 males). Mean age at presentation was 44.7 years (range, 14-78 years) (Table). The RPLS etiologies included hypertension (n = 26; 68%), eclampsia (n = 4; 11%), calcineurin inhibitor use (n = 4; 11%), and other (n = 4; 11%). Comorbid conditions were common in this cohort and included hypertension in 20 (53%), kidney disease in 17 (45%), dialysis dependency in 8 (21%), and transplantation (4 bone marrow and 5 solid organ) in 9 (24%). Four patients (11%) were nonambulatory at baseline, and 6 (16%) were alcoholics. Malignancy was present in 32% of the episodes and included squamous cell cancer (n = 2), metastatic sarcoma (n = 2), multiple myeloma (n = 3), lymphoma (n = 4), and cholangiocarcinoma (n = 1).

Clinical seizures occurred in 33 episodes (87%), of which there was focal onset in 10 (26%) and multiple seizures in 14 (37%). One patient presented with status epilepticus (episode 31). Encephalopathy was a presenting symptom in 35 episodes (92%), visual symptoms in 15 (39%), and headache in 20 (53%). Blood pressure data were available in 36 occurrences within the day before presentation, and the mean peak systolic blood pressure was 187 mm Hg (range, 80-240 mm Hg). The mean duration of hospitalization was 20 days (range, 1-140 days). Intubation was performed in 14 episodes (39%), and these patients remained intubated for a mean of 3.1 days (range, 1-11 days). Clinical recovery occurred within a mean of 5.3 days (range, 0-32 days). Of the 36 patients, 5 died and 3 had no follow-up data available. The remaining 28 patients had follow-up data available for a mean of 1.8 years (range, 19 days to 6.8 years), and none had recurrent seizures.

Cerebrospinal fluid data were available for 18 patients and demonstrated a mean protein level of 92 mg/dL (0.092 g/dL) (reference range, 10-455 mg/dL) [0.010-0.455 g/dL]; (to convert grams per deciliter to grams per liter, multiply by 10.0) and a mean white blood cell count of 1.6/μL (range, 0-5/μL) (to convert to ×10⁹ per liter, multiply by 0.001). Electroencephalographic data were available in 28 patients and showed focal sharp waves in 3 (8%), slowing in 22 (58%), and normal findings in 3 (8%).

The mean delay from the onset of symptoms to neuroimaging was 3 days (range, 0-17 days; median, 2 days), and MRI was performed in 36 episodes (95%). Gadolinium enhancement was seen in 5 of the 15 patients who received contrast. All the patients had significant improvement on repeated neuroimaging (Figure 1). The repeated neuroimaging technique was MRI in 35 episodes (92%). Repeated neuroimaging occurred with a mean delay of 138 days (range, 4-1321 days). Most MRIs (80%) were repeated within 6 months after presentation. Complete neuroimaging resolution of RPLS occurred in 25 episodes (66%), and the earliest complete resolution documented was at 5 days.

Atypical neuroimaging features included significant frontal involvement in 22 episodes (58%), gray matter/cortical lesions in 16 (42%), unilateral lesions in 2 (5%), presence of hemorrhage in 2 (5%), and confluent RPLS in 2 (5%) (Figure 2). Corpus callosum involvement in RPLS occurred in 5 episodes (13%). Brainstem and cerebellar neuroimaging involvement were present in more than half of the patients (58%). Recurrent separate episodes of RPLS occurred in 2 patients (episodes 14 and 15 and episodes 25 and 26) (Figure 3). Permanent injury, typically in the form of small lesions, was seen on repeated MRI in 10 episodes (26%).

As the name implies, RPLS is classically associated with the features of subcortical vasogenic edema, patchy symmetrical bilateral involvement with preferential involvement of the posterior head regions, and complete clinical and radiographic resolution.¹ It is postulated to have a pathogenesis similar to hypertensive encephalopathy,^{8 - 9} although correlation with elevated blood pressure has not been demonstrated. Schwartz⁹ suggested that the vertebrobasilar territory, owing to its relatively sparse sympathetic innervation, may experience preferential disruption of autoregulatory mechanisms, leading to increased perfusion and edema. Lesions of RPLS have been associated with increased perfusion on single-photon emission computed tomography, supporting a vasodilatory mechanism.^{10 - 11} Encephalopathy and seizures remain the major presenting symptoms in RPLS and can include status epilepticus, as demonstrated in this series.¹² Whereas treatment with antiepileptic medications is the standard of care for seizures associated with acute episodes of RPLS, this does not necessitate long-term antiepileptic drug therapy. In this limited series, no patients had recurrence of seizures on follow-up, suggesting that most patients with RPLS, including those who present early with multiple seizures, do not routinely progress to chronic epilepsy.

When an easily identifiable trigger is present (eclampsia, a hypertensive episode, or medication use), RPLS is more readily diagnosed than in patients without obvious precipitants. A common feature in the patients in this series was the presence of comorbid medical conditions in nearly all of the patients, including renal disease, malignancy, alcohol use, pregnancy, inflammatory disease, nonambulatory status, and significant infection. As medical advances escalate the use of chemotherapeutic agents, immunosuppressive medications, and cytotoxic treatments for malignancy, the risk and frequency of RPLS can be expected to increase.

Neuroimaging of RPLS is typically associated with high signal intensity on T2-weighted images predominantly in the posterior regions, which is caused by subcortical white matter vasogenic edema.^{1 ,13} Abnormalities are more conspicuous on fluid-attenuated inversion recovery imaging, which increases the ability to detect subtle lesions in RPLS.¹⁴ Supplemental diffusion-weighted imaging and apparent diffusion coefficient (ADC) map images are helpful in distinguishing vasogenic edema, the predominant abnormality in RPLS, from cytotoxic edema, which can also occur and may represent foci of irreversible ischemia.^{14 - 20} Because vasogenic rather than cytotoxic edema is the principal basis for the lesions, regions demonstrating high signal intensity on T2-weighted images correspond to slightly increased or isointense signal intensity on diffusion-weighted images.^{13 ,21} In ADC mapping, the hyperintense lesions on T2-weighted images correlate with increased ADC values (appear brighter).^{13 ,21} Most RPLS lesions do not enhance on T1-weighted images. Contrast enhancement in the areas of T2 signal abnormality occurs in a few patients and may represent a dynamic feature related to the timing of imaging after the onset of symptoms and severity of lesions.^{22 - 23} Gadolinium enhancement in areas of T2 abnormality is inconsistent and typically not a striking finding²² and in this series was seen in only 5 patients. Complete reversibility is generally regarded as a defining feature of RPLS. The ideal timing of repeated brain imaging to document recovery is unclear. In this series, the earliest neuroimaging resolution occurred in 5 days (patient 10). Of interest, 4 patients with improved but incomplete MRI resolution at 3 to 7 days went on to demonstrate complete resolution with further serial imaging weeks to years later (patients 18, 20, 32, and 37). Thus, resolution of RPLS neuroimaging abnormalities probably occurs in the range of several days to weeks.

Although classic neuroimaging features of RPLS with involvement of the posterior head regions are easily recognized, features that may generally be regarded as atypical were often present in our patients, such as significant anterior involvement, cortical lesions, recurrent RPLS episodes, foci of permanent injury, hemorrhage into lesions, and unilaterality (Figure 1). High signal intensity on T2-weighted image lesions can occur in regions other than the parieto-occipital areas, frequently involving the frontal lobes, basal ganglia, thalami, or brainstem.¹³ The RPLS lesions typically do not exclude the anterior head regions. Frontal lobe involvement is a common feature seen in most of these patients and is consistent with previous studies.²² Similarly, RPLS lesions do not have to be restricted to the white matter. Cortical lesions or gray matter lesions occurred in 42% of these episodes and have been reported at an even higher frequency in other series.¹⁴ Although white matter is predominantly involved, gray matter lesions, anterior lesions, and isolated brainstem involvement are becoming increasingly appreciated.^{24 - 26}

Recurrent episodes of RPLS have been reported in the same patient.²⁷ Recurrent discrete episodes of RPLS occurred in 2 different patients in this series (episodes 14 and 15 and episodes 25 and 26). Hemorrhage is recognized as a potential complication of hypertensive encephalopathy, particularly in patients with clotting abnormalities or coagulopathy,^{9 ,11 ,28} and hemorrhage into the RPLS lesion was demonstrated in 2 patients in the present series.

Although RPLS is typically reversible, lesions can progress to irreversible damage and leukomalacia.^{16 ,28} Residual foci of abnormalities on follow-up neuroimaging after 2 weeks are suggestive of small residual infarcts and were found in approximately one-fourth of our patients. It is conceivable that some of these lesions may take longer than 2 weeks to resolve. Use of ADC mapping to distinguish cytotoxic edema (indicating acute infarction) from vasogenic edema can be equivocal with small foci. An alternative consideration is that some of these foci of hyperintense lesions on repeated neuroimaging may represent preexisting ischemic disease. This series shares the intrinsic limitations of all retrospective studies, including lack of uniformity in timing of follow-up imaging. The strength of this study arises from the fact that all the patients had repeated neuroimaging documenting recovery, ensuring the diagnosis of RPLS and excluding mimickers.

Correspondence: Vivien H. Lee, MD, Department of Neurological Sciences, Rush University Medical Center, 1725 W Harrison St, Ste 1121, Chicago, IL 60612 (vivien_lee@rush.edu).

Accepted for Publication: March 19, 2007.

Author Contributions: All authors had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Lee, Wijdicks, and Rabinstein. Analysis and interpretation of data: Lee, Wijdicks, Manno, and Rabinstein. Drafting of the manuscript: Lee and Wijdicks. Critical revision of the manuscript for important intellectual content: Wijdicks, Manno, and Rabinstein. Administrative, technical, and material support: Lee. Study supervision: Wijdicks and Manno.

Financial Disclosure: None reported.