The effect of presymptomatic hypertension in posterior reversible encephalopathy syndrome

Abstract Objective The effect of blood pressure (BP) on the lesion distribution of posterior reversible encephalopathy syndrome (PRES) is controversial. The aim of this study was to identify the relationship between brain lesion distribution patterns and BP. Methods Sixty‐five patients with PRES were selected from the database. Data regarding brain MRI findings, clinical symptoms, medical conditions, and BP at the presymptomatic period (24 hr before the symptom onset) and at the symptom onset were collected. The brain lesion distribution degree was numerically calculated (lesion scoring point [LSP]) and compared with BP and medical conditions. Results Mean onset‐MAP was higher than mean pre‐MAP. Pre‐MAP correlated with onset‐MAP. The LSP was significantly correlated with pre‐MAP (p = 0.009, correlation coefficient [cc] = 0.323), whereas no significant correlation was found between LSP and onset‐MAP (p = 0.159, cc = 0.177). Similarly, when patients were grouped by mean MAP values, LSP was significantly higher in the patients with high MAP at the presymptomatic period (p = 0.004), whereas no difference was found in the LSP value between patients with low MAP and high MAP at the symptom onset (p = 0.272). Conclusion The patient with PRES who has relatively higher BP in the presymptomatic period would be more likely to have wider lesion distribution than the patient with lower BP. BP elevation during presymptomatic period may be a heralding sign of impending PRES and a factor affecting the severity of PRES although BP was not investigated at earlier time points.

of auto-regulation caused by severe hypertension or direct toxic insult to endothelium is the most popular theory for pathophysiology (Bartynski, 2008b;Dinsdale, 1983;Schwartz et al., 1995;Strandgaard, Olesen, Skinhoj, & Lassen, 1973). One of the unsettled questions about PRES is the correlation between the brain lesion distribution patterns and blood pressure (BP). The aim of this study is to identify the relationship between brain lesion distribution patterns and BP.

| Data collection
We investigated patients with PRES who were admitted to Severance Hospital between January 2001 and December 2014. All patients had been admitted already at least 24 hr before the symptom onset.
Diagnosis of PRES was based on clinical features (predisposing conditions, headache, seizures, mental alteration, and visual disturbances); multifocal lesions on MRI, mainly suggesting vasogenic edema; clinical recovery; and when available, reversibility of MRI lesions. Clinical information such as past medical histories, comorbid illnesses, brain magnetic resonance imaging (MRI), BP measured 1 day before the symptom onset (pre-BP), and BP checked (onset-BP) at symptom onset (mental alteration, seizure, severe headache, or visual disturbances) was collected. BP had been measured every at least 8 hr. BP measured between 24 and 32 hr before the symptom onset was regarded as BP 1 day before the symptom onset.
Cases with unclear symptom onset, without BP documentation, without brain MRI, with pre-existing neurologic deficits, or with a history of seizure disorder were excluded. As a result, 65 patients were selected. This study was approved by the Institutional Review Board of Severance Hospital.

| Brain imaging and brain lesion distribution
Brain MRI using either 1.5-T (Signa Horizon, GE Medical System, Milwaukee, WI, or Gyroscan Intera, Philips Medical Systems, Best, The Netherlands) or 3.0-T scanner (Achieva, Philips Medical Systems, Best, The Netherlands), including conventional spin-echo T 1 -weighted axial, T 2 -weighted axial, and fluid-attenuated inversion recovery axial sequences, were commonly reviewed in all the patients. The mean time to the MRI evaluation was 1.6 days (range, 0-6 days).
The lesion distribution of the PRES was determined by one experienced neuro-radiologist (SK Lee). Affected brain regions were tabulated as frontal, parietal, occipital, temporal lobes, basal ganglia (BG), thalamus, brain stem (BS), and cerebellum (Cbll). Each brain region was scored as one point, and then the scores from the brain regions were added up (lesion scoring point, LSP) in each patient ( Figure 1).

| Statistical analysis
Statistical analysis was performed using SPSS (version 20) for Windows (IBM Corp., Armonk, NY, USA). The Chi-square, Mann-Whitney U, Spearman's correlation, and Kruskal-Wallis tests were used to determine the statistical significance (p < 0.05).
F I G U R E 1 The lesion scoring method. Two sample cases are illustrated. O, occipital lobe; T, temporal lobe; BS, brain stem; F, frontal lobe; P, parietal lobe; Cbll, cerebellum; N, no; Y, yes; LSP, lesion scoring point | 3 of 5 LEE Et aL.

| Brain lesion distribution
Five of 10 patients with preeclampsia-eclampsia had lesions in BG (p = 0.010).

| D ISCUSS I ON
In this study, we assumed an evolution period before the toxic symptom onset. We collected the BP profiles in the presymptomatic and onset periods, and compared lesion distribution with BPs. As a result, LSP correlated with pre-MAP although pre-MAP did with onset-MAP. Patients with higher BP profiles appeared to have increased chance of wider lesion distribution than patients with lower BP profiles in the presymptomatic period.
There have been several studies concerning the relationship between BP at symptom onset, MRI findings, and clinical characteristics. However, their results were incongruent. Bartynski and Boardman found four types of cerebral edema pattern which was classified by main edema lesion location and symmetricity between two cerebral hemispheres. No significant association between MAP at toxicity and imaging patterns, and MAP and underlying medical conditions was found (Bartynski & Boardman, 2007). Later, they assessed the extent and severity of hemispheric cortex-white matter edema by visual grading method. Hemispheric edema was greater in normotensive patients than in those with moderate to severe hypertension (Bartynski & Boardman, 2008). In a study reported in 2009, authors also used a visual grading method to measure the extent of the abnormal signal. In the comparison of the hypertensive group and normotensive group at symptom onset, no significant differences in the extent of disease and the number of affected brain regions between two groups were found. The number of affected brain regions was significantly higher in patients with eclampsia, and the basal ganglia region was more frequently involved in these pa- Although their study did not directly investigate the relationship between BP at the presymptomatic period and at symptom onset

| CON CLUS ION
This study may suggest that the patients with PRES who had higher BP in the presymptomatic period might be more likely to show wider spatial distribution of edematous lesions.

ACK N OWLED G M ENTS
None.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interests.