Discourse Type Clustering using POS n-gram Profiles and High-Dimensional Embeddings

Abstract: To cluster textual sequence types (discourse types/modes) in French texts, K-means algorithm with high-dimensional embeddings and fuzzy clustering algorithm were applied on clauses whose POS (part-ofspeech) n-gram profiles were previously extracted. Uni-, bi- and trigrams were used on four 19th century French short stories by Maupassant. For high-dimensional embeddings, power transformations on the chi-squared distances between clauses were explored. Preliminary results show that highdimensional embeddings improve the quality of clustering, contrasting the use of bi and trigrams whose performance is disappointing, possibly because of feature space sparsity.

[112-POS] : Increased post-partum prevalence of hypertension and renal dysfunction after preeclampsia

OBJECTIVES: Women with a history of preeclampsia (PE) are at increased risk of long term cardiovascular and end-stage renal diseases. However, follow up of preeclamptic women is often omitted, mainly due to a weakness of knowledge of maternal caregivers and lack of comprehensive guidelines. The aim of this study was to define the prevalence of albuminuria, high blood pressure, and renal dysfunction 6 weeks after a preeclampsia. METHODS: This is a prospective case-control study comparing women presenting with preeclampsia to an unmatched control group of women with no hypertensive disorders of pregnancy. A complete medical assessment was performed at 6 weeks post-partum. Recruitment started in June 2010. RESULTS: 324 women were included in the PE group and 50 in the control one. Characteristics of both groups and results of the medical work-up at 6 weeks post-partum are presented in Table 1. Women with preeclampsia presented with a higher BMI, higher prevalence of office high blood pressure, pathological albuminuria and renal hyper-filtration than women in the control group. CONCLUSIONS: Prevalence of post-partum hypertension, and renal dysfunction is higher in women with PE than in uncomplicated pregnancies. Systematic assessment of renal risk factors 6 weeks after preeclampsia allows identification of high-risk women and early implementation of preventive and therapeutic strategies. DISCLOSURES: A. Ditisheim: None. B. Ponte: None. G. Wuerzner: None. M. Burnier: None. M. Boulvain: None. A. Pechère-Bertschi: None.

Post-partum persistence of abnormal circadian pattern of blood pressure after preeclampsia [109-POS]

OBJECTIVES: Blunted nocturnal dip of blood pressure (BP) and reversed circadian rhythm have been described in preeclampsia (PE). Non-dipper status and preeclampsia are both associated with an increased risk of cardiovascular disease later in life. Complete recovery of BP in PE is reported to occur over a variable period of time. Twenty-four hours-ambulatory blood pressure measurement (ABPM) in the post-partum follow-up after a PE has not been described. The aim of this study was to assess 24h-ambulatory blood pressure pattern after a PE and to determine the prevalence of non-dipper status, nocturnal hypertension, white coat hypertension and masked hypertension. METHODS: This is an observational, prospective study on women who suffered from a preeclampsia. A 24h-ABPM was done 6 weeks post-partum at the Hypertension Unit of the University Hospitals of Geneva, concomitantly with a clinical and biological evaluation. RESULTS: Forty-five women were included in a preliminary analysis. Mean age was 33±6years, 57.3% were Caucasian, mean BMI before pregnancy was 24±5kg/m(2). Office and ambulatory BP are shown in Table 1. Prevalence of nocturnal hypertension was high and half of the women had no nocturnal dipping. The diagnosis of hypertension based on office BP was discordant with the diagnosis based on ABPM in 25% of women. CONCLUSIONS: The prevalence of increased nighttime BP and abnormal BP pattern is high at 6weeks post-partum in preeclamptic women. Early assessment of BP with ABPM after preeclampsia allows an early identification of women with persistent circadian abnormalities who might be at increased risk. It also provides a more accurate assessment than office BP. DISCLOSURES: A. Ditisheim: None. B. Ponte: None. G. Wuerzner: None. M. Burnier: None. M. Boulvain: None. A. Pechère-Bertschi: None.