Article Text


Posterior reversible encephalopathy syndrome: long-term follow-up
  1. Christian Roth,
  2. Andreas Ferbert
  1. Department of Neurology, Klinikum Kassel, Kassel, Germany
  1. Correspondence to Dr Christian Roth, Department of Neurology, Klinikum Kassel, Moenchebergstraße 41–43, Kassel D-34125, Germany; roth99{at}


Background Posterior reversible encephalopathy syndrome (PRES) has been known for more than 10 years. The long-term prognosis of this condition remains unknown.

Patients and methods In 2006, the authors screened retrospectively the medical records of our department between 1993 and 2006 for PRES. The authors identified 13 patients. Since 2006, another 12 patients have been included prospectively. Since then, follow-up has been performed yearly for all patients. They were investigated in the outpatient clinic or, if they declined to attend, were interviewed by telephone.

Results The authors identified 25 patients with 27 episodes of PRES. Eighty-four per cent of the patients had generalised seizures. Their mean blood pressure was 167/100 mm Hg. Follow-up was performed for all patients over a mean period of 2250 days (range 59–9396; median 1699). Symptoms resolved, on average, after 7.5 days. Restitution of imaging abnormalities could be shown in 72% of cases. All others showed a clear improvement, but without complete restitution, after a mean duration of 41 days. Recurrence of PRES was observed in two patients (8%), 3 years after complete recovery from their first episode.

Conclusion These data show that PRES has a good short-term and long-term prognosis. Recurrence is infrequent, even though trigger factors for PRES were repeatedly experienced by the patients. Resolution of MRI lesions is slower than clinical recovery.

  • PRES
  • reversible
  • recurrent
  • follow-up
  • long-term
  • prognosis
  • encephalopathy
  • leucoencephalopathy
  • cerebral oedema
  • cerebrovascular disease
  • clinical neurology
  • epilepsy
  • headache

Statistics from

Posterior reversible encephalopathy syndrome (PRES) was first described in 19961 and has attracted increasing interest over the last few years. While Hinchey et al described 15 cases in their original paper, there have been many single case studies or publications that have dealt primarily with the radiological features of the disease.2–4 Clinical studies with more than a few cases are rare. Follow-up data on patients with PRES are scarce in the literature. As yet, no publication exists that focuses on the long-term prognosis. If one postulates that hypertension, for example, is a cause of the disease, then one would expect that recurrent episodes of increased blood pressure should be accompanied by recurrent episodes of PRES. It was the aim of this paper to study the long-term prognosis of PRES with particular emphasis on possible recurrence.

Patients and methods

In 2006, we reviewed the medical records of our department between 1993 and 2006 for diagnoses of encephalopathy. These were screened for the inclusion criteria of PRES. We defined PRES patients as those with an acute neurological syndrome, including seizures, disturbed vision, altered mental function, headache or a combination of these syndromes with CT or MRI findings that appeared in parallel with the clinical symptoms. We identified 13 patients retrospectively. Since 2006, all patients with the diagnoses of PRES have been recorded prospectively. Until February 2009, we were able to include a further 12 patients. Data concerning past medical history, symptoms and paraclinical investigations were collected. The peak systolic and diastolic blood pressures attained within the first hours following admission were noted. All patients had a CT or MRI scan within 24 h of the onset of the clinical syndrome. All had undergone at least two imaging sessions. MRI investigations (1.5 T) were made in 20 patients, including T1-, T2- and FLAIR-weighted images, 17 of them receiving an additional diffusion-weighted MRI. Eleven patients received contrast medium. The affected brain structures were noted. Five patients underwent only CT scans with contrast medium.

From 2006 on, follow-up investigations were performed yearly for all patients, up until July 2009. They were invited to attend our department on an outpatient basis. If they did not attend, the follow-up was made by telephone interview with the patient, or with the patient's general practitioner. Follow-up investigations included a history, neurological status and, in selected cases, a new MRI. Special care was taken concerning risk factors such as hypertension, pregnancy and chemotherapy. We analysed these data using Microsoft Excel (Microsoft, Seattle, Washington) and the SPSS 15.0 software package (SPSS, Chicago, Illinois).


Overall, we studied 25 patients with 27 episodes of PRES. Their mean age was 35±23 (range 9–76) years. None of these patients had previous episodes of PRES in their past medical history. Women were affected more often (males/females=1/2.6). There were 10 patients in the peripartal period.

Clinical syndrome

The clinical symptoms and signs are summarised in table 1.

Table 1

Symptoms of 25 patients with posterior reversible encephalopathy syndrome

Generalised seizures occurred in 88% of the patients. In one-third of these patients, the general seizure had started from a focal seizure, or the patients had had additional focal seizures. Ten of the 22 patients had a single seizure, eight patients had two seizures, and four patients had a series of generalised seizures. There were no patients with status epilepticus. Disturbed vision, altered mental function and headache were also frequent symptoms. Three patients had additional brainstem symptoms.

The mean systolic blood pressure for all patients was 170±37 (median 167) mm Hg, and the mean value of the diastolic blood pressure was 98±29 (median 100) mm Hg. The presumed causes of PRES and the related systolic blood pressures are listed in table 2.

Table 2

Supposed trigger factors for 25 patients with posterior reversible encephalopathy syndrome

There were three cases with severely elevated blood pressure with systolic values >170 mm Hg with no evidence of other underlying causes or risk factors for PRES. In one patient with normal blood pressure values, the cause of PRES remained unclear. Alterations in laboratory studies were seen in nine patients with elevated liver enzymes (3× alcoholism, 2× pancreatitis, 2× haemolysis, elevated liver enzymes and low platelets, haemochromatosis, toxic liver damage after chemotherapy) and two patients with elevated serum creatinine (1.4 and 2.8 mg/dl). Lumbar punctures were performed in 15 cases. There were no pathological findings, except, in a few cases, for a mild protein elevation. Sixty-eight per cent of all PRES patients were treated on an intermediate or intensive care unit for a mean duration of 4 (range 1–9) days. Only one patient required intubation and artificial ventilation for more than 12 h because of a series of generalised seizures.

Long-term follow-up

Twenty-three patients had a significant amelioration of their symptoms during their stay in hospital. Clinical symptoms disappeared after a period of between 1 and 53 (median 2) days. In 25% of the patients, the symptoms had disappeared within 1 day. The 75% quartile for disappearance of symptoms was 10.5 days.

We obtained follow-up data for all patients over a mean period of 2250 days (range 59–9346; median 1699) after the acute event. In three patients, symptoms did not improve during their stay in hospital: one 71-year-old male died from septicaemia 59 days after the initial syndrome. Another 78-year-old female patient had dementia and could not be re-examined. A control MRI had shown a complete remission 67 days after the initial symptoms. A 67-year-old female had residual symptoms after 76 days in the form of mild disturbed vision. Forty-three days after the initial event, an MRI had shown an almost complete remission of the white-matter lesions. New intracerebral bleeding, however, was found in the area of the posterior lobe.

During follow-up, we found 12 patients with hypertensive disease, whereas only four of them had been affected before their first episode of PRES. All three patients who developed PRES during therapy with chemotherapeutic drugs received further chemotherapy after the initial episodes of PRES (cyclophosphamid/cytarabin, methotrexat/cyclophosphamid/cytarabin and 6-mercaptopurin). Four of the 10 women in the peripartal period delivered a second baby, 1, 2, 5 and 10 years after the first birth, respectively, without further occurrence of PRES.

We found an incidence for recurrence of PRES of 8% (n=2). One 35-year-old patient had recurrent PRES 3 years after the initial episode. Both events started with epileptic seizures, with no additional neurological deficits. Her blood pressure during the first episode was 200/130 mm Hg and 156/117 mm Hg during the second. She was treated with metoprolol. There was also some irregular alcohol intake. Both episodes resolved completely. MRI revealed bilateral parieto-occipital lesions in the white matter on FLAIR-weighted imaging, with no lesion visible in the diffusion-weighted image. Two months after the first episode, the MRI findings had resolved completely (figure 1). Four years after the second episode, she was free of epileptic seizures and of further PRES episodes. A 70-year-old patient had two episodes of PRES within 3 years. In the first episode, he presented with symptoms of lost vision and nausea. During both episodes, he had increased blood pressure (188/118 mm Hg and 160/66 mm Hg). At the time of the first episode, the patient suffered from a mild acute pancreatitis. Fourteen days after the first symptoms, an MRI showed complete remission of the lesions. After 3 years without any symptoms, the patient was admitted again because of a generalised seizure. MRI again showed hyperintense lesions in the occipital region, which again declined completely. During the second admission, we diagnosed an acute renal failure due to benign prostatic hyperplasia as a possible trigger factor. Four months after this second episode, the patient was still free of symptoms (figure 2).

Figure 1

Thirty-five-year-old female with recurrence of posterior reversible encephalopathy syndrome. MRI (fluid-attenuated inversion recovery) during the first episode shows hyperintense lesions of the posterior, parietal and frontal cortex/subcortical white matter (a). On a control MRI 2.5 months later, the lesions have resolved completely (not shown here). During the second episode, 3 years later, MRI reveals minor lesions in the parietal cortex. In addition, there is some hyperintensity in the right nucleus caudatus (b).

Figure 2

Seventy-year-old patient with two episodes of posterior reversible encephalopathy syndrome. (A) MRI during the first episode. Fourteen days later, control MRI shows a complete remission of the lesions (not shown here). After 3 years without any symptoms, the patient was admitted again because of a generalised seizure. (B) MRI showing hyperintense lesions occipital and fronto-parietal on both sides.

Imaging findings

Twenty of the 25 patients underwent MRI investigation. Eighteen of these cases had an additional CT before the MRI scan. Whereas MRI showed clear abnormalities in all patients, CT showed alterations only in nine of these patients. Five patients were examined only with CT scans and contrast medium. In all cases, both hemispheres where affected, sometimes asymmetrically. The main MRI lesions were located particularly in the occipital or parieto-occipital regions (65% and 25%, respectively). The affected brain structures seen in MRI are summarised in table 3.

Table 3

Affected brain structures in 20 patients with MRI investigations

An isolated involvement of brainstem or cerebellum could not be detected. However, lesions in the brainstem or lesions of the cerebellum in addition to hemispheric lesion were found in 30% and 20% of cases, respectively. Twenty-three per cent of the patients who underwent MRI with diffusion-weighted imaging (DWI) showed some minor lesions. These were always much smaller than in FLAIR-weighted imaging. Contrast enhancement with gadolinium was seen in only one MRI. All patients showed an improvement in the white-matter lesions in imaging sessions. In 72% of imaging sessions, we found a complete recovery of the imaging abnormalities. All other patients showed a clear improvement in their white-matter lesions after a median of 41 days. Haemorrhage was observed in only one patient, in the occipital lobe (see above).


In our study, women were predominantly affected. This can be explained by the fact that the largest group of patients were young women with eclampsia/pre-eclampsia. There were various other suggested trigger factors for PRES. In accordance with other studies, PRES occurred, for example, in connection with elevated blood pressure or during chemotherapy.1 5 6 Both children and older patients were affected. A rapid rise in blood pressure seems to be a cofactor in the pathogenesis of PRES. We found a mean systolic blood pressure of 170 mm Hg. Epileptic seizures were the most frequent symptom. They showed a similar prevalence at a rate comparable with other studies.1 5 We observed focal seizures in one-third of the cases. This might well be explained by the fact that PRES is often a rather localised syndrome as seen on MRI, and often includes the cortex. We found bilateral involvement, mainly occipital or parieto-occipital, on FLAIR-weighted MRI, with only minor asymmetry. These findings are in accordance with other studies.1 7 8 Brainstem lesions have been described in the literature, sometimes even in isolation.9–11 We found lesions in the brainstem in 30% of our patients, but never in isolation.

The prevalence of PRES remains unknown. It is also unclear whether recurrent attacks of PRES may occur. Patients undergoing allogeneic bone marrow transplantation and immune suppressive therapy with ciclosporin are reported to have PRES in 4–8% of cases.12 13 Six to eight per cent of pregnant women develop pre-eclampsia, and 0.3% develop eclampsia.14 15 The risk of having repeated pre-eclampsia or eclampsia is about 22%.16 Although many studies of PRES have been published since its original description in 1996,1 it was only in 2002 that a single case of recurring PRES during long-term therapy with tacrolimus was reported.17 Other single-case studies have described recurrence under chemotherapy with tacrolimus or cysplatin and 5-flurouracil.18 19 In Thailand, a case was reported of a 20-year-old woman with a systemic lupus erythematosus who had multiple episodes of PRES (at least four) over a time period of 2½ years, occurring in connection with recurrent glomerulonephritis. However, no information was provided on the clinical details of this patient's glomerulonephritis, or of its further course.20 In addition, there are two case reports of recurrent PRES episodes21 22 and, finally, a report of 18 children with kidney disease and PRES in which two of the children each developed three recurrent episodes of PRES within 21 months. Elevated blood pressure was measured in five of these episodes; blood pressure was therefore regarded as the major trigger factor in these patients.23

If one postulates that a particular vulnerability of the brain and certain trigger factors are prerequisites for the recurrence of PRES, one would expect recurring attacks of PRES to occur much more frequently. Trigger factors such as acute elevation of blood pressure occur frequently in patients with hypertension and should, therefore, lead to recurrent PRES attacks in those patients who are particular susceptible. The same holds for other trigger factors such as a second or third pregnancy, recurrent chemotherapy, etc. The fact that recurrent PRES has so seldom been described may be due to the fact that few studies have focused on long-term follow-up. One study described 36 patients with 38 episodes of PRES, but with a mean follow-up time of only 1.8 years.5 Further details of the two recurrences and the corresponding risk factors are sparse in this publication. Our study is the first to document long-term prognosis after PRES. The mean follow-up period for our patients was almost 6 years. Four of the 10 pregnant patients had a second pregnancy. Of the three patients who had PRES after chemotherapy, all underwent further cycles of chemotherapy. In addition, 11 patients had a sustained arterial hypertension during the follow-up without developing recurrent PRES. Two of our 25 patients had a second episode of PRES after the initial event. Both showed elevated blood pressure on both occasions, but only one developed hypertensive disease after the first episode. This represents an incidence of recurrence of 8%. There has been a recent report with an incidence of 3.8% retrospectively.24 In this latter study, patients had sickle cell anaemia, antibody positive autoimmune disease or an acute myeloid leukaemia with allogeneic bone marrow transplantation and ciclosporin therapy. The authors concluded that infection and inflammation were trigger factors in these cases, as well as multiple organ dysfunction and lesions of the endothelia.

Most of our patients were treated on an intensive care unit for a mean duration of 3.9 days. This shows that PRES occurs as a severe and acute disease. While intubation was necessary in 39% of the PRES patients in the study by Lee,5 intubation and artificial ventilation were needed in only one of our patients, due to a series of generalised seizures. There have been repeated reports of patients with a poor neurological outcome.2 25 26 It would seem that in these cases, vasogenic oedema had converted into irreversible cytotoxic oedema. This was demonstrated by hyperintensities seen in DWI-weighted images.27 A recent publication postulated an incidence of haemorrhage in PRES of 15%.28 In our study, only one patient failed to show a clinical improvement, due to a haemorrhage in the occipital lobe. The course in two further patients was complicated by multimorbidity. All other patients had an excellent outcome of their PRES, with respect to their clinical symptoms and the imaging findings. The time interval between the acute stage and the complete remission of MRI abnormalities was much longer than that for the recession of the clinical symptoms.

In summary, this study demonstrates that even if PRES occurs as a severe and acute disease, the short-term and especially long-term prognosis is very good. In parallel to the excellent clinical recovery, there was a good recovery from MRI abnormalities. However, these MRI lesions resolve somewhat slower than the clinical recovery. Recurrent PRES attacks are rare. Further research is necessary to better define the factors which predispose a few patients to develop PRES, in contrast to the majority of patients experiencing the same trigger factors, such as hypertension, chemotherapy and eclampsia, who do not develop PRES.


We would like to thank T Dimpfl, Department of Gynecology, and Obstectrics and FP Kuhn, Department of Radiology, for their cooperation.


View Abstract


  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.