Pontine Reversible Edema: A Newly Recognized Imaging Variant of Hypertensive Encephalopathy?

Although the typical imaging findings of hypertensive encephalopathy have been well described in the medical literature, the diagnosis can occasionally present a challenge. This may be related to the infrequency of patients presenting with hypertensive encephalopathy, relatively vague clinical symptoms (headache, visual disturbances, and seizures), and failure to communicate the patient's elevated systemic blood pressure to the radiologist. Imaging findings in mild cases of hypertensive encephalopathy are those of edema, usually within the cortex and subcortical white matter of the parietal, occipital, temporal, and to a lesser degree, the posterior frontal lobes, typically with bilaterality, although not with perfect symmetry. More severe cases tend to have flagrant involvement of the subcortical white matter and may extend to the frontal, posterior temporal, cingulate, and central sylvian regions, as well as to the cerebellar white matter. The most severe cases can have various degrees of thalamic, insular, and pontine involvement.

In most cases, the changes of hypertensive encephalopathy appear to represent reversible vasogenic edema, which is seen on T2-weighted images and can sometimes be shown by diffusion-weighted imaging (DWI) and apparent diffusion coefficient (ADC) maps. Most investigators believe that hypertensive encephalopathy begins with progressive hypertension, eventually leading to failure of autoregulation and cerebral hyperperfusion. This in turn results in breakdown of the blood-brain barrier with vasogenic edema, which begins in the cerebral cortex and later accumulates in the subcortical white matter.

The symptoms and imaging findings of hypertensive encephalopathy have been found to be remarkably similar, if not identical, to changes associated with a number of other acute illnesses, including eclampsia/pre-eclampsia, cyclosporin-A (CSA) neurotoxicity, and other more obscure illnesses (1). It is now believed there is a distinct clinico-neuroradiologic syndrome that encompasses these various conditions, reported recently in the New England Journal of Medicine by Hinchey et al (1), wherein the name “reversible posterior leukoencephalopathy syndrome” was suggested. With the increasing numbers of organ and marrow transplant patients on immunosuppressive agents, this syndrome may be seen quite frequently in some tertiary transplant centers. Animal models and our recent experience with fluid-attenuated inversion-recovery imaging suggest that the early changes in this syndrome occur within the cerebral cortex, rather than in the white matter. Thus, we prefer the term “posterior reversible encephalopathy (or edema) syndrome (PRES),” as a more appropriate descriptor of the syndrome. In recent years, this syndrome has been known by several names, including “hypertensive encephalopathy,” “hyperperfusion encephalopathy,” “reversible leukoencephalopathy,” “occipito-parietal encephalopathy,” and “reversible posterior cerebral edema syndrome.” Additional risk factors include renal failure, systemic lupus erythematosis, and numerous drugs including other immunosuppressants, such as FK506 and high-dose corticosteroids, as well as various chemotherapeutic agents (1). Often patients have a combination of risk factors and may have been placed on one or more of these medications, putting them at risk for PRES. Most patients have some hypertension, although many of the patients, especially children and those with CSA neurotoxicity, do not have systemic blood pressure levels as high as are typically encountered with “pure” hypertensive encephalopathy. It is likely that, outside the category of hypertensive encephalopathy, there are various superimposed pathophysiologic mechanisms leading to the common imaging appearance of PRES among its different entities. The physiologic changes of PRES are dynamic, and pathologic correlation is fortunately rarely obtained. In past reports, pathologic analysis has yielded only evidence of fibrinoid necrosis within the arteriole walls, interstitial edema, and petechial microhemorrhages, but little or no evidence of infarction (1).

At least two unusual variants of PRES may exist. The first is a syndrome of cerebral autoregulation disturbance that occurs as an uncommon complication in chronically hypoperfused portions of the brain following revascularization procedures. After carotid endarterectomy or stenting, this phenomenon is termed the “postcarotid endarterectomy hyperperfusion syndrome” and results in reversible edema unilaterally in the affected cerebral hemisphere on the same side as the endarterectomy.

It appears that de Seze et al, in this issue of the AJNR (page 391), report a second imaging variation of PRES. They describe two patients with hypertensive encephalopathy manifested as reversible increased signal on T2-weighted images essentially isolated to the pons. The MR findings described by de Seze et al in these two cases are indeed unusual, but may not be as rare as one might think. The differential diagnosis for such pontine T2 hyperintensity includes pontine glioma, ischemic and radiation changes (generally irreversible conditions), as well as central pontine myelinolysis (CPM), often a devastating condition except in mild cases. The absence of abnormal serum sodium levels, the clinical recovery of their patients, and the resolution of the MR abnormalities make the diagnosis of CPM less likely. It is interesting to note that there is an association of CSA neurotoxicity, one of the better known etiologies of PRES, with CPM in patients after liver transplantation (2). Some of these patients have undergone severe serum sodium flux and truly have CPM. Most of these patients, however, are reported to have good clinical recovery. In reality, there is no pathologic proof that all of these cases represent reversible CPM, as opposed to a central variant of PRES, or a combination of the two conditions. Perhaps some neuroradiologists may be able to recall a case with similar pontine imaging findings to those shown by de Seze et al, but in which a diagnosis of CPM was unlikely from the absence of the typical clinical course and the absence of serum sodium alterations. Some of these cases may, in retrospect, have been secondary to this newly described imaging variation of hypertensive encephalopathy or PRES, perhaps associated with less than obvious elevations of systemic blood pressure. For instance, we recently encountered the following case in point.

A 44-year-old man presented with a decreased level of consciousness and a history of recent cocaine use and was admitted to our emergency department. CT revealed a punctate hemorrhage in the right thalamus, but the diagnosis of hypertensive hemorrhage was complicated by the presence of bilateral thalamic and pontine hypoattenuation. MR imaging was performed, verifying a small intraparenchymal hemorrhage within the right thalamus, but also showing extensive edema within the thalami bilaterally, the pons, and the midbrain. Upon inquiry, it was learned that the patient's systemic blood pressure was extremely elevated. Nevertheless, none of the typical posterior cerebral cortical and subcortical T2 hyperintensities of PRES were evident on the MR images. DWI and ADC maps in this patient showed increased ADC in the thalami, midbrain, and pons, therefore excluding an acute ischemic phenomenon and indicating, instead, a process of reversible vasogenic edema. After institution of antihypertensive therapy, the patient had clinical resolution of his symptoms. The imaging changes, although not completely resolved, were greatly improved on follow-up MR imaging 3 days later. This case would thus appear to be a similar example of brain stem involvement of hypertensive encephalopathy in the absence of peripheral cerebral lesions.

In patients with such pontine T2 hyperintensities, DWI is useful for two reasons. First, it is helpful in refining the differential diagnosis because hyperintensity on DWIs might favor other processes, such as acute ischemia or acute CPM (we have seen two such autopsy-proven cases of CPM), although it may still be possible in severe cases of PRES. Although in the majority of cases the lesions of PRES can be assumed to be reversible with treatment, we have seen a small number of severe cases with ischemic complications heralded by restriction of fluid movement on ADC maps. The use of DWI in these cases also provides prognostic information regarding the likelihood of reversibility.

As we learn more about PRES, we recognize the wider spectrum of imaging appearances of this condition. This spectrum has already been reported in the uremic encephalopathies, a group of conditions including hemolytic-uremic syndrome, hepatorenal syndrome, and thrombotic thrombocytopenic purpura. Some reports have demonstrated bilateral lesions in the parieto-occipital regions identical in appearance to those seen in PRES, whereas other reports have demonstrated predominant involvement of relatively central structures, such as the basal ganglia or the brain stem. It now appears that the uremic encephalopathies represent additional etiologies of PRES, which, for reasons unknown, have a greater tendency for central distribution.

Future research on PRES should be directed at perfusion imaging. To date, there have been a number of conflicting reports using perfusion imaging in these entities, with some reporting hyperperfusion and others reporting hypoperfusion. It is likely that these results depend on the time of imaging relative to the onset of therapy in patients with PRES. Present data tends to favor the theory that the condition begins with hyperperfusion, resulting in failure of autoregulation, and breakthrough accumulation of vasogenic edema. We believe that overly aggressive antihypertensive therapy, in the setting of disturbed cerebral blood flow autoregulation, can result in hypoperfusion, even with apparently normal blood pressures. In some severe cases, this can lead to infarction predominantly in the posterior border zones. Ischemia may also result from status epilepticus and hypoxic complications. Over the years, this mix of transient hyperperfusion and the infrequent ischemic complication has created much confusion in our attempt to elucidate the pathophysiologic mechanisms contributing to the various etiologies of PRES. Given the dynamic nature of brain perfusion in PRES, it may be useful to perform perfusion imaging in selected patients who are responding poorly to therapy as a guide for a more “personalized” titration of antihypertensive therapy.

In conclusion, it is important for neuroradiologists to become aware of the spectrum of imaging findings in the acute presentation of PRES. The diagnosis of PRES is one of the more satisfying diagnoses made in our practice, as it is often unsuspected by clinicians, and relatively dramatic changes on MR imaging can be predicted to be predominantly, if not completely, reversible. The neuroradiologist will often be the first physician to suggest the appropriate diagnosis in the hope of averting unnecessary biopsies and initiating appropriate therapy. Clinicians also need to become more familiar with this syndrome and the treatment issues in order to minimize underlying risk factors and to avoid potential ischemic complications.