Statistical analysis of incomplete longitudinal ordinal data - Impact of Multiple Imputation on data distribution

Scientific conference (2013, June 17)

Is there a good way to impute missing data in longitudinal studies with ordinal outcomes ?

Scientific conference (2013)

Cardiac surgery and acute kidney injury: retrospective study

Conference (2012, October 27)

Background: In cardiac surgery, acute kidney injury (AKI) is a severe postoperative complication and associated with increased rates of mortality, morbidity, and length of stay in intensive care units ... [more ▼]

Background: In cardiac surgery, acute kidney injury (AKI) is a severe postoperative complication and associated with increased rates of mortality, morbidity, and length of stay in intensive care units (ICU). It occurs in 5% to 30% of patients depending on the definition used [1] [2] [3]. The aim of this study is to present an overview of AKI following cardiac surgery associated or not with cardiopulmonary bypass, in our center. Methods: This retrospective study includes patients treated by cardiac surgery from April 1st, 2008 to March 31th, 2009 in a single center. We selected patients who underwent on-pump coronary artery bypass surgery (CABG), off-pump CABG (OPCAB), aortic valve replacement, mitral valve repair or replacement and aortic valve replacement combined with CABG. Patients undergoing renal replacement therapy preoperatively were excluded. The RIFLE classification (Risk, Injury, Failure, Loss and End stage kidney disease) allowed stratifying the patients into the 3 grades of AKI severity. The stratification was based on the most pejorative element observed within 7 days after surgery: increased serum creatinine level or decreased urine output, or decreased glomerular filtration rate according to criteria of Bellomo [4]. Occurrence of AKI was studied by type of cardiac surgery as its impact on the length of stay in ICU and in the hospital. Proportions were compared by the Chi-square test and median values by the Mann-Whitney U test. Results were considered significant at p < 0.05. Results: Four hundred and thirty-four patients were included: median (IQR) age 69.0(60.0-76.0) year, 30.2% females, 2.76% urgent/emergent cases. Fifty-eight patients (13.4%) underwent OPCAB, 182(41.9%) on-pump CABG, 104(24.0%) aortic valve replacement, 44(10.1%) mitral valve repair or replacement and 46(10.6%) aortic valve replacement combined with CABG. AKI occurred in 213(49.1%) patients: 79(37.1%) “Risk”, 108(50.7%) “Injury” and 26(12,2%) “Failure”. Distribution of AKI by type of surgery was equal to 36.2% among OPCABG, 44.0% on-pump CABG, 49.0% aortic valve replacement, 52.3% mitral valve surgery and 82.6% aortic valve replacement combined with CABG, respectively. AKI occurrences differed significantly according to the type of surgery (p<0.0001). Lengths of stay in ICU were significantly longer (p<0.0001) in AKI group compared with non AKI group: 3(2-4) days versus 2(2-3) days. However, no difference (p = 0.65) was observed between the two (AKI and NON-AKI) groups in hospital length of stay: 13(10-18) days versus 12(10-16) days. Conclusions: The incidence of AKI is very high in this population as compared to the literature. This may be due to the fact that the three elements of the RIFLE classification for all the population studied have been used. This study emphasizes the need for clear definition of AKI in order to compare different studies adequately. AKI after cardiac surgery with cardiopulmonary bypass would be further studied in order to develop more appropriate preventive measures. [less ▲]

Simulation-based comparative performance of multiple imputation methods for incomplete longitudinal ordinal datasets

Poster (2012, October 24)

Cardiac surgery and acute kidney injury: retrospective study

Conference (2012, September 29)

Background: In cardiac surgery, acute kidney injury (AKI) is a severe postoperative complication and associated with increased rates of mortality, morbidity, and length of stay in intensive care units ... [more ▼]

Background: In cardiac surgery, acute kidney injury (AKI) is a severe postoperative complication and associated with increased rates of mortality, morbidity, and length of stay in intensive care units (ICU). It occurs in 5% to 30% of patients depending on the definition used [1] [2] [3]. The aim of this study is to present an overview of AKI following cardiac surgery associated or not with cardiopulmonary bypass. Methods: This retrospective study includes patients treated by cardiac surgery from April 1st, 2008 to March 31th, 2009 in a single center. We selected patients who underwent on-pump coronary artery bypass surgery (CABG), off-pump CABG (OPCAB), aortic valve replacement, mitral valve repair or replacement and aortic valve replacement combined with CABG. Patients undergoing renal replacement therapy preoperatively were excluded. The RIFLE classification (Risk, Injury, Failure, Loss and End stage kidney disease) allowed to stratify the patients into the 3 grades of AKI severity. The stratification was based on the most pejorative element observed within 7 days after surgery: increased serum creatinine level or decreased urine output, or decreased glomerular filtration rate according to criteria of Bellomo [4]. Occurrence of AKI was studied by type of cardiac surgery as its impact on the length of stay in ICU and in the hospital. Proportions were compared by the Chi-square test and median values by the Kruskal-Wallis. Results were considered significant at p < 0.05. Results: Four hundred and thirty four patients were included: median (IQR) age 69.0(60.0-76.0) year, 30.2% females, 2.76% urgent/emergent cases. Fifty-eight patients (13.4%) underwent OPCAB, 182(41.9%) on-pump CABG, 104(24.0%) aortic valve replacement, 44(10.1%) mitral valve repair or replacement and 46(10.6%) aortic valve replacement combined with CABG. AKI occurred in 213(49.1%) patients: 79(37.1%) “Risk”, 108(50.7%) “Injury” and 26(12,2%) “Failure”. Occurrence of AKI after OPCAB was 21(9.86%), CABG 80(37.6%), aortic valve replacement 51(23.9%), mitral valve surgery 23(10.8%) and aortic valve replacement combined with CABG 38(17.8%). AKI occurrences differed significantly between the different groups of surgery (p<0.0001). Lengths of stay in ICU were significantly longer (p<0.0001) in AKI group compared with non AKI group: 3(2-4) days versus 2(2-3) days. No difference (p = 0.65) was observed between the two (AKI and NON-AKI) groups in hospital length of stay: 13(10-18) days versus 12(10-16) days. Conclusions: The incidence of AKI is very high in this population as compared to the literature. This may be due to the fact that the three elements of the RIFLE classification for all the population studied have been used: serum creatinine level, urine output and glomerular filtration rate. This study emphasizes the need for clear definition of AKI in order to compare adequately different studies. AKI after cardiac surgery with cardiopulmonary bypass would be further studied in order to develop more appropriate preventive measures. [less ▲]

Multiple imputation methods for incomplete longitudinal ordinal data: a simulation study

Conference (2012, September 14)

On the potential increase of the oxidative stress status in patients with abdominal aortic aneurysm.

in Redox Report : Communications in Free Radical Research (2012), 17(4), 139-44

BACKGROUND: Abdominal aortic aneurysm (AAA) is a major cause of preventable deaths in older patients. Oxidative stress has been suggested to play a key role in the pathogenesis of AAA. However, only few ... [more ▼]

BACKGROUND: Abdominal aortic aneurysm (AAA) is a major cause of preventable deaths in older patients. Oxidative stress has been suggested to play a key role in the pathogenesis of AAA. However, only few studies have been conducted to evaluate the blood oxidative stress status of AAA patients. METHODS AND RESULTS: Twenty seven AAA patients (mean age of 70 years) divided into two groups according to AAA size (</= 50 or > 50 mm) were compared with an age-matched group of 18 healthy subjects. Antioxidants (vitamins C and E, beta-carotene, glutathione, thiols, and ubiquinone), trace elements (selenium, copper, zinc, and copper/zinc ratio) and markers of oxidative damage to lipids (lipid peroxides, antibodies against oxidized patients, and isoprostanes) were measured in each subject. The comparison of the three groups by ordinal logistic regression showed a significant decrease of the plasma levels of vitamin C (P = 0.011), alpha-tocopherol (P = 0.016) but not when corrected for cholesterol values, beta-carotene (P = 0.0096), ubiquinone (P = 0.014), zinc (P = 0.0035), and of selenium (P = 0.0038), as AAA size increased. By contrast, specific markers of lipid peroxidation such as the Cu/Zn ratio (P = 0.046) and to a lesser extent isoprostanes (P = 0.052) increased. CONCLUSION: The present study emphasizes the potential role of the oxidative stress in AAA disease and suggests that an antioxidant therapy could be of interest to delay AAA progression. [less ▲]

Population compliance with national dietary recommendations and its determinants: findings from the ORISCAV-LUX study.

in British Journal of Nutrition (2012)

The objective of the present study was to determine the proportion of adults meeting national recommendations for food and nutrient intake and to identify the demographic, socio-economic and behavioural ... [more ▼]

The objective of the present study was to determine the proportion of adults meeting national recommendations for food and nutrient intake and to identify the demographic, socio-economic and behavioural factors that may contribute to weaken dietary compliance. ORISCAV-LUX is a cross-sectional study that took place in Luxembourg (2007-8). A representative stratified random sample of 1352 adults aged 18-69 years participated in the nationwide cardiovascular health survey. A FFQ was used to estimate food intake. Radar charts were built to compare graphically the compliance of the participants with different key dietary guidelines on the same set of axes. The thirteen food- and nutrient-based recommendations were scored and summed to create a recommendation compliance index (range - 0.5 to 14). Ordinal logistic regression analyses were performed to determine the factors contributing to poor dietary compliance. Several food- and nutrient-based guidelines were insufficiently respected compared with others. The greatest gaps occurred in the adherence to grain and dairy product consumption guidelines, as well as to total fat and notably to SFA recommendations. Age, country of birth, economic status, smoking status and subject's awareness of the importance of balanced meals emerged as independently associated with weak dietary compliance. Obese subjects conformed more to dietary recommendations compared with normal-weight subjects. The findings underscore the need for specific nutrition education messages along with targeted interventions. Efforts should be continued to increase population awareness of the importance of a healthy lifestyle and a balanced diet. [less ▲]

Epidemiology of the metabolic syndrome in Luxembourg findings from ORISCAV-LUX study

Conference (2011)

Prevalence of the metabolic syndrome in Luxembourg according to the Joint Interim Statement definition estimated from the ORISCAV-LUX study.

in BMC Public Health (2011), 11(1), 4

ABSTRACT: BACKGROUND: The prevalence of the metabolic syndrome (MS) has been determined in many countries worldwide but never in Luxembourg. This research aimed to 1) establish the gender- and age ... [more ▼]

ABSTRACT: BACKGROUND: The prevalence of the metabolic syndrome (MS) has been determined in many countries worldwide but never in Luxembourg. This research aimed to 1) establish the gender- and age-specific prevalence of MS and its components in the general adult population of Luxembourg, according to the most recent Joint Interim Statement (JIS) definition, by using both the high and low cut-off points to define abdominal obesity, and 2) compare and assess the degree of agreement with the Revised National Cholesterol Education Programme-Adult Treatment Panel III (R-ATPIII) and the International Diabetes Federation (IDF) definitions. METHODS: A representative stratified random sample of 1349 European subjects, aged 18-69 years, participated to ORISCAV-LUX survey. Logistic regression and odds ratios (OR) were used to study MS prevalence with respect to gender and age. The Framingham risk score (FRS) to predict the 10-year coronary heart disease (CHD) risk was calculated to compare the proportion of MS cases below or above 20%, according to both high and low waist circumference (WC) thresholds. Cohen's kappa coefficient (kappa) was utilized to measure the degree of agreement between MS definitions. RESULTS: The prevalence of the MS defined by the JIS was 28.0% and 24.7% when using the low (94/80) and the high (102/88) WC cut-off points, respectively. The prevalence was significantly higher in men than in women (OR = 2.6 and 2.3 for the low and high WC thresholds), as were all components of the MS except abdominal obesity measured by both thresholds. It also increased with age (OR values in age categories ranging from 2.7 to 28 when compared to the younger subjects for low WC and from 3.3 to 31 for the high WC cut-offs). The 10-year predicted risk of CHD by FRS did not depend on the threshold used. Globally, excellent agreement was observed between the three definitions of MS (kappa= 0.89), in particular between JIS and IDF (kappa = 0.93). Agreement was significantly higher in women than in men, and differed between age groups. CONCLUSION: Regardless of the definition used, the adult population of Luxembourg reveals a high MS prevalence. Our findings contribute to build evidence regarding the definitive construct of the MS, to help selecting the waist circumference thresholds for Europid populations, and to support the need to revise the guidelines for abdominal obesity levels. [less ▲]