Interact CardioVasc Thorac Surg 2009;9:484-490. doi:10.1510/icvts.2009.203836
© 2009 European Association of Cardio-Thoracic Surgery
Negative results - Cardiac general |
Microalbuminuria and short-term prognosis in patients undergoing cardiac surgery
Martin Majlund Mikkelsena,b,e,*,
Niels Holmark Andersenb,c,
Thomas Decker Christensena,b,
Troels Krarup Hansenb,d,
Hans Eiskjaerb,c,
Carl Erik Mogensenb,d,
Vibeke Elisabeth Hjortdala,b and
Søren Paaske Johnsenb,e
a Department of Cardiothoracic and Vascular Surgery T, Aarhus University Hospital, Skejby, Brendstrupgaardsvej 100, 8200 Aarhus N, Denmark
b Institute of Clinical Medicine, Aarhus University Hospital, Skejby, Brendstrupgaardsvej 100, 8200 Aarhus N, Denmark
c Department of Cardiology, Aarhus University Hospital, Skejby, Brendstrupgaardsvej 100, 8200 Aarhus N, Denmark
d Department of Medicine, Aarhus University Hospital, Nørrebrogade, 8000 Aarhus C, Denmark
e Department of Clinical Epidemiology, Aarhus University Hospital, Oluf Palmes Allé, 8200 Aarhus N, Denmark
Received 27 January 2009;
received in revised form 22 May 2009;
accepted 25 May 2009
 The study was financially supported by The Institute of Clinical Medicine, Beckett Fonden, Jens Anker Andersens Fond, Karl G. Andersens Fond, Snedkermester Sophus Jacobsen og hustru Astrid Jacobsens Fond, Ragnhild Ibsens Legat for Medicinsk Forskning, Linex Fonden, Civilingenør Frode V Nyegaards Fond, Korning Fonden, Søster og Verner Lipperts Fond, A.P. Møllers Fond. The work was independent of the funders.
*Corresponding author. Department of Clinical Epidemiology, Aarhus University Hospital, Oluf Palmes Allé, 8200 Aarhus N, Denmark. Tel.: +4589424811; fax: +4589424801.
E-mail address: majlund{at}ki.au.dk (M.M. Mikkelsen).
 |
Abstract
|
|---|
Objective: To examine if preoperative microalbuminuria (MA) is associated with in increased risk of adverse outcomes in patients undergoing elective cardiothoracic surgery, and if adding information on MA could improve the accuracy of the additive EuroSCORE. Methods: In a follow-up study we included 962 patients undergoing elective cardiothoracic surgery from 1 April 2005 to 30 September 2007 at our department. MA (urine albumin/creatinine ratio between 2.5–25 mg/mmol) was assessed in a morning spot-urine sample. We used population-based medical registries for 30-day follow-up and compared the length of stay and adverse outcomes including (i) all-cause death, myocardial infarction, stroke, or atrial fibrillation, (ii) surgical reintervention, renal insufficiency, sternal wound infection, or septicaemia among patients with and without MA. Results: MA was found in 180 (18.7%) patients. The risk of both combined outcomes (adjusted odds ratios (ORs): 1.00 (95% confidence interval (CI): 0.77–1.30) and 1.18 (95% CI: 0.79–1.75), respectively) and most individual outcomes did not differ between the micro- and normoalbuminuric patients. The patients with MA and an additive EuroSCORE of 5 had a significantly prolonged median length of intensive care unit (ICU) stay (0.15 days [95% CI: 0.04–0.26]) and total hospital stay (0.5 days [95% CI: 0.04–0.96]). Patients with MA had a higher risk of postoperative septicaemia (OR: 12.1 [95% CI: 3.2–45.9]). Area under receiver operating characteristics curves of the EuroSCORE with regard to 30-day mortality was 0.86 both with and without MA. Conclusions: Preoperative MA in patients undergoing elective cardiothoracic surgery was not associated with most early adverse outcomes. However, risk of septicaemia was higher and patients with MA also had a marginally longer length of ICU and hospital stay. Information on preoperative MA did not improve the accuracy of the additive EuroSCORE.
Key Words: Adult; Cardiothoracic surgery; Comorbidity; Kidney; Outcomes
|
1. Introduction
|
|---|
Microalbuminuria (MA) is a well-known cardiovascular marker of risk in both diabetic and non-diabetic patients [1–3]. Increased urinary albumin excretion is a marker of endothelial dysfunction and severity of vascular disease [4], and for primary prevention of cardiovascular adverse events the early detection of MA is important [5]. The association between preoperative MA and postoperative outcomes in patients undergoing cardiothoracic surgery has only been sparsely examined and is as yet not clarified. In a relatively small prospectively studied population of 257 diabetic patients undergoing coronary artery bypass surgery, MA with a 30-mg/day cut-off was associated with a longer intensive care unit (ICU) stay but not significantly with a higher mortality [6]. Echahidi and colleagues retrospectively analyzed data from 5304 consecutive patients that had undergone coronary artery bypass surgery and reported that MA was not associated with operative mortality [7]. However, MA is closely related to the metabolic syndrome, a clustering of several cardiovascular risk factors, which was found to be a powerful risk factor of adverse outcome in patients undergoing coronary bypass surgery [7, 8]. Predictive models, such as the EuroSCORE are widely used in routine clinical practice for risk-prediction of patients undergoing cardiothoracic surgery [9]. These models are of major importance both for clinical decision making and for analysis and comparison of patient outcomes (e.g. between different types of interventions or across institutions). New markers of disease or adverse outcomes are constantly proposed and some of these markers may improve the accuracy of the established risk scoring systems. No existing studies have to our knowledge evaluated the effect of adding information on MA to the models. The aim of this study was to evaluate whether preoperative MA was associated with adverse outcomes at 30-day follow-up and whether MA could improve preoperative risk-prediction with EuroSCORE in patients undergoing elective cardiothoracic surgery.
|
2. Materials and methods
|
|---|
2.1. Design and setting
We conducted a prospective follow-up study in a region of Western Denmark with a mixed rural-urban population of  1.2 million. From 1 April 2005 to 30 September 2007 we included patients undergoing elective cardiothoracic surgery at the Department of Cardiothoracic and Vascular Surgery at Aarhus University Hospital, Skejby, Denmark. All patients gave informed consent prior to inclusion. The study protocol was approved by the Regional Ethics Committee and the Danish Data Protection Agency.
2.2. Study population
Inclusion criteria were (i) age older than 18 years, (ii) elective cardiothoracic surgery (surgery performed more than two days after planning of the procedure) – including CABG or OPCAB surgery, valve surgery, thoracic aortic surgery, pulmonary thromboendarterectomy, grown up congenital heart disease procedures, or combined procedures. Exclusion criteria were (i) severe renal disease defined as a serum creatinine above 200 mmol/l or (ii) macroalbuminuria defined as a urine albumin creatinine excretion ratio (UACR) above 25 mg/mmol and (iii) previous heart or renal transplant surgery.
During the study period a total of 2216 patients underwent cardiothoracic surgery at the department. Patient screening and recruitment was done with assistance of a project nurse working half-time. Approximately 50% (n=1193) of the potential candidates for the study could therefore be screened consecutively. Seventy-eight were excluded because of acute surgery, 54 had known renal disease with serum creatinine >200 mmol/l. Out of the remaining, 1061 patients were randomly invited to participate. Only 12 patients did not accept participation and 1049 patients were included. Thereafter, a preoperative in-hospital baseline serum creatinine analyses revealed another 11 patients with increased levels above our exclusion criteria (Fig. 1). The UACR assessment resulted in exclusion of 22 patients with macroalbuminuria, and another 54 patients were excluded due to failure of UACR analyses, leaving 962 patients available for follow-up (Fig. 1).
2.3. Laboratory analyses
A preoperative fasting blood sample and a morning spot urine sample were collected and analyzed at the Department of Clinical Biochemistry, Aarhus University Hospital, Skejby, Denmark.
We defined MA as a UACR from 2.5 to 25 mg/mmol. Using a preoperative urine sample, the urinary albumin (mg/l) was assessed quantitatively by immunoturbidimetry and the urinary creatinine (mmol/l) was estimated by an enzymatic colorimetric test. The sensitivity level for urinary albumin was 7 mg/l. Since the UACR was constructed by dividing urinary albumin with urinary creatinine, albumin values below 7 mg/l could not contribute to an exact UACR (mg/mmol). We defined the UACR to be 0.1 mg/mmol in these patients.
2.4. Study outcomes
The study outcomes were a composite of (i) all-cause mortality, myocardial infarction, stroke, and atrial fibrillation or flutter, and (ii) surgical reintervention, deep sternal wound infection, septicaemia (defined as a positive bacteriaemic blood sample and/or clinical sepsis), and renal insufficiency (defined as more than a 100% increase of s-creatinine from baseline and/or use of dialysis). We also examined the individual elements of the composite outcomes as well as the length of stay in the ICU, and the total length of hospital stay.
Since 1968, all Danish residents have been assigned a unique civil registration number that allows unambiguous record linkage between the Danish health databases. The following Danish health databases were used for follow-up: the Danish Registry of Patients, the Prescription Database of Central Denmark Region, the Western Denmark Heart Registry and the National Register of Causes of Death.
The Danish National Registry of Patients was established in 1977 and holds data on all hospitalizations from somatic Danish hospitals, including dates of admission and discharge, procedure(s) performed, and up to 20 discharge diagnoses coded by physicians according to the International Classification of Diseases [8th revision (ICD-8) until the end of 1993, end 10th revision (ICD-10) thereafter]. Discharge diagnoses are determined exclusively by the physician who discharges the patient and cannot be altered later for any other purposes. The Prescription Database contains data on all redeemed prescriptions at all pharmacies in the region since 1998. The main variables are the name of the drug, ATC code, package identifier, date of refill and codes identifying the prescribing physician. The Western Denmark Heart Registry, established in 1999, is a regional clinical register including detailed patient baseline characteristics, data for all cardiothoracic procedures performed, and per- and postoperative outcomes. The National Register of Causes of Death is a complete registry of dates and all causes of death in Denmark since 1973.
2.5. Covariates
Baseline characteristics and in-hospital preoperative data were collected from a preoperative interview, patient medical records, the Prescription Database, and the National Patient Register. For each patient a case-report-form was used.
Baseline data included age, gender, smoking habits, body mass index, blood pressure, prior ischaemic peripheral-, cerebro-, or cardiovascular disease, history of arrhythmias, diabetes and dyslipidaemia, cardiac ejection fraction, EuroSCORE, Charlson Comorbidity Index, serum levels of creatinine, electrolytes, albumin, glucose, white and red blood cell counts, thrombocytes and the UACR. We obtained information on use of medication up to 180 days preoperatively (e.g. ACE inhibitors, AT-II antagonists, beta blockers, calcium channel blockers, diuretics, lipid lowering drugs and antidiabetics).
The EuroSCORE was obtained from linkage to the Western Denmark Heart Registry. Missing values of the EuroSCORE were computed from information in the case-report-form and the medical records.
The Charlson Comorbidity Index was constructed by combining data from the case-report-form with data from the National Registry of Patients. The Charlson Comorbidity Index is a method of classifying comorbidity and in longitudinal studies predicts both short- and long-term mortality [10].
The pre- and postoperative covariates included type of operation, cardiopulmonary bypass time, aortic cross-clamp time, number and type of grafts, use of inotropics and antibiotics, and mechanical ventilation.
2.6. Statistical analyses
Baseline characteristics and pre- and postoperative variables were compared between groups with and without MA using Fisher's Exact test, independent t-tests and the Mann–Whitney test. The associations between MA and the composite and individual outcomes were then adjusted for both the additive EuroSCORE and other baseline potential confounding factors using multivariate logistic regression analyses.
Data on the length of ICU and hospital stay were analyzed on a logarithmic scale using linear regression analyses. Thereafter, we transformed the regression estimate and estimated the absolute difference in median length of stay between groups at different levels of the EuroSCORE. The standard error was calculated using the delta method. Kaplan–Meier survival curves were also constructed.
Discrimination and calibration analyses were done in order to examine whether MA in combination with EuroSCORE would predict risk more accurately than EuroSCORE alone. The association between the EuroSCORE and the combined as well as the single outcomes was estimated using logistic regression. Receiver operating characteristic curves were constructed to perform discrimination analyses. Calibration of the predicted risk was performed by Hosmer–Lemeshow goodness of fit analyses. The analyses were repeated using the logistic EuroSCORE. We also examined the association between the level of UACR and risk of various outcomes with spline regression in order to identify any non-linear patterns. Finally, separate subgroup analyses were done for the 877 patients undergoing only bypass surgery, valve surgery, thoracic aortic surgery, or combined procedures.
A two-tailed P-value <0.05 was considered statistically significant. Analyses were performed using the Stata® 10.0 package (StataCorp LP, Texas, US).
|
3. Results
|
|---|
3.1. Study cohort and surgical characteristics
Baseline patient and preoperative characteristics according to presence of MA are displayed in Table 1. The overall prevalence of MA in the cohort was 18.7% (n=180). The EuroSCORE was available for all included patients and the median EuroSCORE of the study cohort was 5.3 (range: 0–18).
On average patients with MA were 3.5 years older and had a higher EuroSCORE than the normoalbuminuric patients. Likewise, there were minor differences in smoking habits, type II diabetes, previous stroke, atrial fibrillation, ejection fraction, and use of antihypertensive medication].
Patients with MA had less solitary bypass procedures and more combined procedures performed, but there were no differences between groups with regard to use of extra corporal circulation and aortic cross-clamp time. The number of grafted coronary vessels in patients undergoing on- or off-pump coronary artery bypass surgery was slightly lower in the MA group.
3.2. The association between MA and postoperative adverse outcomes
In both the crude analyses and when adjusted for EuroSCORE length of ICU stay and total hospital stay were significantly prolonged in patients with MA (Fig. 2 and Fig. 3). A patient with an additive EuroSCORE equal to 5 had a 0.15 days (95% confidence interval (CI): 0.04–0.26) prolonged ICU stay and a 0.50 days (95% CI: 0.04–0.96) prolonged total hospital stay. Adjustment for additional covariates did not change these findings.
The crude and adjusted associations between MA and study outcomes are displayed in Table 2. Overall, the first predefined combined adverse outcome occurred in 369 (38%) patients, whereas the second combined outcome occurred in 140 (14.6%) patients.
MA was not an independent predictor of most individual outcomes when adjusted for EuroSCORE. Still, MA remained a strong independent predictor of developing septicaemia (odds ratio (OR): 12.1 [3.19–45.9]). In the crude analysis MA was significantly associated with the second combined outcome (reintervention, sternal wound infection, septicaemia, and renal insufficiency) but not the first combined outcome (all-cause mortality, myocardial infarction, stroke, and atrial fibrillation or flutter). However, when adjusted for EuroSCORE this association did not remain significant. Further adjustment for other covariates did not change the risk estimates substantially.
Analyzing the UACR as a continuous spline function did not reveal any threshold values in the association with postoperative outcomes. Furthermore, neither repeated analyses using the logistic EuroSCORE nor restricting all the analyses only to patients undergoing bypass surgery, valve surgery, thoracic aortic surgery, or combined procedures had significant impact on the risk estimates.
3.3. MA and the accuracy of the EuroSCORE
Table 3 shows the associations between the additive EuroSCORE and the postoperative outcomes. The ORs represent the increase in risk for each increment of 1 in the additive EuroSCORE. The EuroSCORE was an independent predictor of operative mortality (OR: 1.4 [1.20–1.63]) and was significantly associated with most other postoperative outcomes. Adding MA to the model had no impact on any of the relative risk estimates. The MA adjusted EuroSCORE was not statistical significantly associated with myocardial infarction, sternal wound infection, or septicaemia. The Hosmer–Lemeshow test for calibration showed good agreement for both the EuroSCORE (P-value=0.22,  2=9.4) and the EuroSCORE+MA (P-value=0.44, 2=8.0) model between the observed and expected number of deaths within 30 days after surgery. Except from ‘surgical reintervention’ (P-value=0.03, 2=15.3) the calibration tests of both the combined (P-value=0.86, 2=3.3) and the individual outcomes (P-values >0.51, 2<6.2) also showed statistical acceptable agreements, which were not substantially changed when information about MA wad added to the EuroSCORE (P-values >0.32, 2<9.3). The results of the discriminative statistics displayed in Table 4 show the areas under curves, which were not changed when the receiver operating characteristics of the EuroSCORE alone and the EuroSCORE+MA were compared.
|
4. Discussion
|
|---|
MA was found in almost one out of five patients in this prospective cohort of elective patients undergoing cardio- thoracic surgery. Patients with MA had marginally longer ICU and total hospital stays and a higher risk of postoperative septicaemia. However, MA was not independently associated with other postoperative outcomes, including mortality, and adding information on MA did not improve the accuracy of the additive EuroSCORE.
The prevalence of MA depends on the underlying source population. Our study included patients with an a priori high cardiovascular risk-profile. The prevalence of MA was low when compared to the study by Yorgancioglu et al. [6], who found an overall prevalence of 34.6%. However, in contrast to their study, we included primarily non-diabetic patients.
The longer ICU stay in the group with MA corresponds well to earlier findings regarding MA and length of ICU stay [6]. The absolute differences in ICU and total hospital stay were small and may only be of little clinical relevance. However, the prolonged stay may represent complicated admissions in the group with MA. A longer ICU stay could potentially also through more intense care reduce the incidence of adverse outcomes in the patients with MA and thereby weaken the risk associated with MA.
MA is considered a marker of endothelial disease severity. Gosling and colleagues found a transient increase in albumin excretion rate in patients admitted with acute myocardial infarction [11]. The pathophysiologic cause of this increased leakage of albumin is not fully understood, but an acute inflammatory response is thought to participate in a worsening of endothelial dysfunction. In other settings it has been reported that the level of MA is related to adverse outcome after surgery or in acute diseases [11–15]. The surgical stress may accentuate this preoperative condition, primarily attributed to a systemic and renal increase in vascular permeability, as part of an early induced inflammatory process. We found a 12-fold increase in the risk of developing septicaemia postoperatively in patients with MA. This has not been reported before in patients undergoing cardiothoracic surgery. However, the result needs to be interpreted with caution, since the number of septicaemic events was small and the diagnosis problematic as it is dependent on either a positive bacteriaemic blood sample or clinical sepsis in a patient population that has a systemic inflammatory response due to the surgery performed and extra corporal circulation.
Our study has both strengths and limitations: we studied a well-defined cohort that was representative of the patient population undergoing cardiothoracic surgery at our department. We had a complete follow-up on all included patients, since our design relied on national health registries with complete coverage. Recruitment of patients to be invited to participate in the study was random and the level of UACR was not known at the time of invitation. The usability of hospital discharge and other disease registries as data sources is dependent on the validity of the registered data and it is well known that the validity of routine hospital discharge diagnoses may vary. However, the predictive values have previously been reported to be high (80–99%) for several of the outcomes in our study including myocardial infarction, stroke and atrial fibrillation [16–19]. Any misclassification would in any case most likely be independent of the presence of MA and would bias the findings toward the null hypothesis.
We calculated the UACR from a morning spot-urine sample, which is considered a well-validated sampling approach with regard to UACR [20]. This is an inexpensive and feasible approach. The one-time measurement of UACR introduces limitation to the study because of intra-individual variation in the urinary albumin excretion with time. This variation is mainly because of the state of hydration and physical activity, and three consecutive samples over three days would be optimal. However, such a setup does not seem feasible [21]. A preoperative morning spot-urine tends to minimize the effect of exercise and the UACR adjusts for part of both the individual and physical albumin excretion variation by correcting for urinary creatinine concentration, and by including a large number of patients the risk of bias due to variation in albumin excretion was reduced. However, we cannot rule out some extent of bias because of hydration and different muscle masses between individuals.
MA assessed in a morning spot-urine did not substantially contribute with more knowledge about adverse 30-day postoperative outcomes in elective cardiothoracic surgery than did the EuroSCORE itself. The EuroSCORE is a sensitive predictor of 30-day postoperative mortality, but it has been shown to overestimate mortality in low-risk patients and to underestimate mortality in high-risk patients [22]. Therefore, also the EuroSCORE should be continually used and refined in risk prediction of patients undergoing cardiothoracic surgery. The present results do not seem to advocate for the general assessment of preoperative UACR as a supplement to EuroSCORE in elective cardiac surgery. Whether MA independently predicts long-term postoperative outcome in this patient group remains to be assessed.
We conclude that preoperative presence of MA in patients undergoing elective cardiothoracic surgery was not associated with a broad range of early postoperative outcomes, and that adding information on MA will not improve the accuracy of the additive EuroSCORE. However, postoperative septicaemia was significantly more frequent and the length of ICU and total hospital stay was longer in patients with MA than patients without MA.
|
References
|
|---|
- Arnlov J, Evans JC, Meigs JB, Wang TJ, Fox CS, Levy D, Benjamin EJ, D'Agostino RB, Vasan RS. Low-grade albuminuria and incidence of cardiovascular disease events in nonhypertensive and nondiabetic individuals: the Framingham Heart Study. Circulation 2005;112:969–975.[Abstract/Free Full Text]
- Damsgaard EM, Froland A, Jorgensen OD, Mogensen CE. Microalbuminuria as predictor of increased mortality in elderly people. Br Med J 1990;300:297–300.[Abstract/Free Full Text]
- Klausen KP, Scharling H, Jensen JS. Very low level of microalbuminuria is associated with increased risk of death in subjects with cardiovascular or cerebrovascular diseases. J Intern Med 2006;260:231–237.[CrossRef][Medline]
- Wang Y, Li Y. [The relationship between microalbuminuria and coronary artery lesions]. Zhonghua Nei Ke Za Zhi 2007;46:981–983.[Medline]
- Farbom P, Wahlstrand B, Almgren P, Skrtic S, Lanke J, Weiss L, Kjeldsen S, Hedner T, Melander O. Interaction between renal function and microalbuminuria for cardiovascular risk in hypertension: the Nordic diltiazem study. Hypertension 2008;52:115–122.[Abstract/Free Full Text]
- Yorgancioglu C, Tokmakoglu H, Suzer K, Zorlutuna Y. Does microalbuminuria in diabetic patients affect the postoperative course after coronary artery bypass surgery? Eur J Cardiothorac Surg 2002;21:395–400.[Abstract/Free Full Text]
- Echahidi N, Pibarot P, Despres JP, Daigle JM, Mohty D, Voisine P, Baillot R, Mathieu P. Metabolic syndrome increases operative mortality in patients undergoing coronary artery bypass grafting surgery. J Am Coll Cardiol 2007;50:843–851.[Abstract/Free Full Text]
- Klausen KP, Parving HH, Scharling H, Jensen JS. The association between metabolic syndrome, microalbuminuria and impaired renal function in the general population: impact on cardiovascular disease and mortality. J Intern Med 2007;262:470–478.[CrossRef][Medline]
- Gogbashian A, Sedrakyan A, Treasure T. EuroSCORE: a systematic review of international performance. Eur J Cardiothorac Surg 2004;25:695–700.[Abstract/Free Full Text]
- de Groot V, Beckerman H, Lankhorst GJ, Bouter LM. How to measure comorbidity. A critical review of available methods. J Clin Epidemiol 2003;56:221–229.[CrossRef][Medline]
- Gosling P, Brudney S, McGrath L, Riseboro S, Manji M. Mortality prediction at admission to intensive care: a comparison of microalbuminuria with acute physiology scores after 24 hours. Crit Care Med 2003;31:98–103.[CrossRef][Medline]
- Gosling P, Hughes EA, Reynolds TM, Fox JP. Microalbuminuria is an early response following acute myocardial infarction. Eur Heart J 1991;12:508–513.[Abstract/Free Full Text]
- Koulouris S, Lekatsas I, Karabinos I, Ioannidis G, Katostaras T, Kranidis A, Triantafillou K, Thalassinos N, Anthopoulos L. Microalbuminuria: a strong predictor of 3-year adverse prognosis in nondiabetic patients with acute myocardial infarction. Am Heart J 2005;149:840–845.[CrossRef][Medline]
- Slowik A, Turaj W, Iskra T, Strojny J, Szczudlik A. Microalbuminuria in nondiabetic patients with acute ischemic stroke: prevalence, clinical correlates, and prognostic significance. Cerebrovasc Dis 2002;14:15–21.[CrossRef][Medline]
- Thorevska N, Sabahi R, Upadya A, Manthous C, moateng-Adjepong Y. Microalbuminuria in critically ill medical patients: prevalence, predictors, and prognostic significance. Crit Care Med 2003;31:1075–1081.[CrossRef][Medline]
- Andersen TF, Madsen M, Jorgensen J, Mellemkjoer L, Olsen JH. The Danish National Hospital Register. A valuable source of data for modern health sciences. Dan Med Bull 1999;46:263–268.[Medline]
- Frost L, Andersen LV, Johnsen SP, Mortensen LS. Lost life years attributable to stroke among patients with nonvalvular atrial fibrillation: a nationwide population-based follow-up study. Neuroepidemiology 2007;29:59–65.[CrossRef][Medline]
- Johnsen SP, Overvad K, Sorensen HT, Tjonneland A, Husted SE. Predictive value of stroke and transient ischemic attack discharge diagnoses in The Danish National Registry of Patients. J Clin Epidemiol 2002;55:602–607.[CrossRef][Medline]
- Joensen AM, Jensen MK, Overvad K, Dethlefsen C, Schmidt E, Rasmussen L, Tjonneland A, Johnsen S. Predictive values of acute coronary syndrome discharge diagnoses differed in the Danish National Patient Registry. J Clin Epidemiol 2009;62:188–194.[CrossRef][Medline]
- Bakker AJ. Detection of microalbuminuria. Receiver operating characteristic curve analysis favors albumin-to-creatinine ratio over albumin concentration. Diabetes Care 1999;22:307–313.[Abstract/Free Full Text]
- Mogensen CE, Vestbo E, Poulsen PL, Christiansen C, Damsgaard EM, Eiskjaer H, Froland A, Hansen KW, Nielsen S, Pedersen MM. Microalbuminuria and potential confounders. A review and some observations on variability of urinary albumin excretion. Diabetes Care 1995;18:572–581.[Medline]
- Choong CK, Sergeant P, Nashef SA, Smith JA, Bridgewater B. The EuroSCORE risk stratification system in the current era: how accurate is it and what should be done if it is inaccurate. Eur J Cardiothorac Surg 2009;35:59–61.[Free Full Text]
|
|
Top
Abstract
1. Introduction
