La lecture critique de la littérature médicale dans le - CClin Sud-Est
Transcription
La lecture critique de la littérature médicale dans le - CClin Sud-Est
La lecture critique de la littérature médicale dans le groupe de lecture DIALIN L. Ayzac, CCLIN Sud Est VI° journée annuelle du Réseau DIALIN – 22 septembre 2009 Les critères de validité des publications médicales • Où est publié l’article? – Notion de comité de lecture – Notion d’index de citations : Impact factor • Qui sont les auteurs? – Autres publications – Notoriété • La lecture critique 1 Objectifs • Où se situe cette étape ? – La recherche bibliographique a été réalisée et de qualité – Les articles ont été obtenus • Quels sont les objectifs ? – Lire les documents – Juger de la qualité de leur contenu – « Séparer le bon grain de l’ivraie » – Sélection + évaluation C’est le concept de lecture critique. Juger la valeur des publications • Quelle est la crédibilité de la publication ? ((validité interne)) – En fonction du type d’étude, identifier les différentes étapes du protocole de l’étude et évaluer le niveau de crédibilité de chaque étape – Les conclusions sont-elles valides ? • Quelle est l’applicabilité des informations contenues dans la publication ? (validité externe) – Les conclusions sont-elles applicables à la pratique médicale du praticien ou au projet de recherche du réseau DIALIN? 2 Une méthode (1) Deux livres • 1. Salmi LR: Lecture critique et rédaction médicale édi l scientifique. i tifi C Commentt lilire, rédiger et publier une étude clinique ou épidémiologique. Paris, Elsevier, 1998. • 2. Landrivon G, Delahaye F: La recherche clinique De l'idée l idée à la publication. Paris, Masson, 1995. Une méthode (2) Utiliser une grille de lecture 3 Une méthode (3) 8 étapes du protocole d’étude Ö 8 critères d’évaluation • Pour chaque critère : – Est-il possible de trouver l’information sur le critère ? – L’abord du critère est-il correcte ? – Si l’abord est incorrect, cela menace-t-il la validité de l’étude l étude ? Une méthode (4) 8 étapes du protocole d’étude Ö 8 critères d’évaluation Situation du problème Î Objectifs Î Hypothèse • Il y a 4 grandes préoccupations possibles : • L’histoire naturelle d’une maladie, son évolution et son pronostic. • L’utilisation, L’ tilisation la performance et l’intérêt d’ d’un n test diagnostic. •L’impact d’une intervention thérapeutique, de dépistage, de prévention ou d’éducation (utile/inutile/dangereux) •La détermination d’une étiologie ou d’une causalité • Replacer l’article dans un de ces 4 cadres Ö une idée claire sur les objectifs • Y a-t-il une hypothèse à vérifier par rapport à l’objectif ? 4 Une méthode (5) 8 étapes du protocole d’étude Ö 8 critères d’évaluation Situation du problème Î Objectifs Î Hypothèse • Il existe une hiérarchie parmi les 5 modèles possibles de plan d’études dans le niveau de preuve et dans le coût. Plan de l’étude Niveau de preuve 0 Coût Une méthode (6) 8 étapes du protocole d’étude Ö 8 critères d’évaluation Situation du problème Î Objectifs Î Hypothèse • L’étude rapporte un cas intéressant et inhabituel ou une série de cas Plan de l’étude Niveau de preuve Rapport de cas 0 Coût 5 Une méthode (7) 8 étapes du protocole d’étude Ö 8 critères d’évaluation Situation du problème Î Objectifs Î Hypothèse • L’étude transversale est la description de fréquence d’une maladie, de ses facteurs de risque dans une population déterminée à un temps déterminé. Plan de l’étude Niveau de preuve Etude transversale Rapport de cas 0 Coût Une méthode (8) 8 étapes du protocole d’étude Ö 8 critères d’évaluation Situation du problème Î Objectifs Î Hypothèse Plan de l’étude • L’étude cas-témoin est une étude d’observation, rétrospective, comparant les caractéristiques de patients atteints d’une maladie (les cas) avec celles d’individus indemnes de la maladie (les témoins). Niveau de preuve Étude cas témoin Etude transversale Rapport de cas 0 Coût 6 Une méthode (9) 8 étapes du protocole d’étude Ö 8 critères d’évaluation Situation du problème Î Objectifs Î Hypothèse Plan de l’étude • L’étude de cohorte est une étude d’observation, prospective, suivant un groupe d’individus exposés à des facteurs de risque d’une maladie pendant une période de temps donnée. Le taux d’incidence de la maladie peut être comparé à celui d’un groupe témoin non exposé. Niveau de preuve Étude de cohorte Étude cas témoin Etude transversale Rapport de cas 0 Coût Une méthode (10) 8 étapes du protocole d’étude Ö 8 critères d’évaluation Situation du problème Î Objectifs Î Hypothèse Plan de l’étude • L’essai contrôlé est une étude expérimentale, comparant le résultat d’une intervention sur un groupe de sujets recevant l’intervention au résultat dans un groupe semblable qui ne reçoit pas l’intervention. Niveau de preuve Essai randomisé Étude de cohorte Étude cas témoin Etude transversale Rapport de cas 0 Coût 7 Une méthode (11) 8 étapes du protocole d’étude Ö 8 critères d’évaluation Situation du problème Î Objectifs Î Hypothèse Plan de l’étude • Le plan d’étude retenu est-il approprié à la question posée ? Une méthode (12) 8 étapes du protocole d’étude Ö 8 critères d’évaluation Situation du problème Î Objectifs Î Hypothèse Plan de l’étude • Quel est le facteur étudié ? • Exposition • ou Intervention Facteur étudié • ou Test diagnostique • Sont-ils bien décrits ? • Comment sont ils mesurés ? • Même méthode de mesure chez tout les sujets ? • Même méthode de mesure dans tous les groupes ? • Méthode de mesure à l’aveugle ? • Pour les test diagnostiques : comparaison indépendante avec l’étalon ? 8 Une méthode (13) 8 étapes du protocole d’étude Ö 8 critères d’évaluation Situation du problème Î Objectifs Î Hypothèse • Quel est le critère de jugement (facteur résultant) ? Plan de l’étude • Comment est il mesuré ? Facteur étudié • Même méthode de mesure chez tout les sujets ? Critère de jugement • Même méthode de mesure dans tous les groupes ? • Méthode de mesure à l’aveugle ? • Tous T les l critères itè de d jugement j t pertinents ti t sont-ils t il évalués ? Une méthode (14) 8 étapes du protocole d’étude Ö 8 critères d’évaluation Situation du problème Î Objectifs Î Hypothèse Plan de l’étude Facteur étudié Critère de jugement (facteur résultant) Population étudiée / échantillon • La sélection est-elle correcte ? • Y a t il randomisation ? • Les groupes diffèrent-ils diffèrent ils par des caractéristiques autres que le facteurs étudiés ? •Quel est la proportion de sujets atteignant la fin du suivi ? g q : y a-t-il un •Pour les tests diagnostiques large éventail de patients ? 9 Une méthode (15) 8 étapes du protocole d’étude Ö 8 critères d’évaluation Situation du problème Î Objectifs Î Hypothèse Plan de l’étude Facteur étudié Critère de jugement (facteur résultant) Population étudiée / échantillon Biais et facteurs de confusion • Un biais est une erreur systématique qui contribue à produire des estimation s systématiquement plus élevées ou plus basses que la valeur réelle des paramètres à estimer • Biais dus à la sélection • Biais dus à la mesure… • Un facteur de confusion est un facteur qui modifie les effets du facteur étudié sur le critère de jugement du fait de son lien à la fois avec le facteur étudié et avec le facteur de jugement. jugement • Sont-ils tous envisagés ? • Sont-ils tous pris en compte ? Une méthode (16) 8 étapes du protocole d’étude Ö 8 critères d’évaluation Situation du problème Î Objectifs Î Hypothèse Plan de l’étude Facteur étudié • Résultats statistiquement significatif • Tests statistiques, place du p • Intervalles de confiance • Direction de l’association Critère de jugement (facteur résultant) • Force de l’association Population étudiée / échantillon • Taille de l’échantillon suffisant ? Biais et facteurs de confusion • Cliniquement intéressant ? Résultats • Puissance du test / taille de l’échantillon ? 10 Une méthode (17) 8 étapes du protocole d’étude Ö 8 critères d’évaluation Situation du problème Î Objectifs Î Hypothèse • Y a t il toutes les réponses aux questions posées ? Plan de l’étude • L’hypothèse est-t-elle vérifiée ? Facteur étudié • L’objectif est-il atteint ? Critère de jugement (facteur résultant) Population étudiée / échantillon Biais et facteurs de confusion • Les résultats sont-ils acceptables appliqués li é à la l population l ti source d de lla publication ? (Validité) Résultats Synthèse • Les résultats sont-ils acceptables pour votre propre pratique ? (Applicabilité) Un exercice corrigé (1) Situation du problème Î Objectifs Î Hypothèse 11 Un exercice corrigé (2) Plan de l’étude Facteur étudié Le facteur étudié (l (l’exposition exposition ou ll’intervention intervention qui est supposé avoir des conséquences sur un problème de santé, une maladie ou un état clinique) est la présence de bactériémie à Staphylococcus aureus (SAB) chez des patients hémodialysés chroniques. Elle est mesurée dans ses conséquences en terme d’hospitalisation, de pronostique – dissémination de l’infection , décès -et de coûts. Un exercice corrigé (3) Critère de jugement (facteur résultant) Le critère de jugement est la nature du site d’accès (cathéter, fistule native, fistule greffe). 12 Un exercice corrigé (4) Population étudiée / échantillon Biais et facteurs de confusion Les facteurs de confusion et les biais Les données sont collectées au moment de l’hospitalisation du patient et de sa saisie dans un dossier informatisé. Qu’en est il si le patient n’est pas hospitalisé ou hospitalisé dans un autre établissement? (Biais de sélection). La mortalité est observée à la sortie du patient et 12 semaines après la survenue de la bactériémie (pour éliminer un biais d’observation). Pour l’étude de la mortalité et des coûts d’autres facteurs que le site d’accès rentre en ligne de compte : ce sont les facteurs de confusions. Ils sont recueillis systématiquement : facteurs liés au patient (gravité – APACHE II) ou à la prise en charge (antibiotiques, corticothérapie…) Un exercice corrigé (5) Résultats Les analyses statistiques (régression logistique multiple) ont consisté en une estimation du risque relatif (Odds Ratio OR) et de son intervalle de confiance. Les intervalles contiennent 1. Il n’y a donc pas d’association significative entre type de site d’accès et mortalité/coûts 13 Un exercice corrigé (6) Résultats Un exercice corrigé (7) Synthèse 14 Au total Situation du problème Î Objectifs Î Hypothèse Plan de l’étude Facteur étudié Critère de jugement (facteur résultant) Population étudiée / échantillon Biais et facteurs de confusion Résultats Synthèse 15 Fiche de lecture critique d’une étude clinique ou épidémiologique DIALIN (V3 08/03/2007) (d’après SALMI L.R. Lecture critique et rédaction médicale scientifique. Comment lire, rédiger et publier une étude clinique ou épidémiologique. Paris :Elsevier ; 1995) Référence : N° ..................... Auteurs : .............................................................................................................................................................. Titre : .................................................................................................................................................................... Revue : ................................................................................................................................................................ Année : ............................ Volume (n°) : .............. ( ....... ) Pages : ..................................................... Résumé : Objectif et justification : Schéma d’étude ; groupes comparés : Population étudiée et nombre de sujets : Conduite de l’étude : Principaux résultats : Cocher la case correspondant au respect du critère : 0 = oui, N = non, I = incomplet ; NA = ne s’applique pas, NSP = ne sait pas. Une réponse cochée N à un critère en italique = étude inacceptable. Critères Schéma d’étude Formulation claire de l’objectif Hypothèse faite à priori Schéma adapté à l’objectif Procédures de sélection Critères d’inclusion décrits et adéquats Critères d’exclusion décrits et adéquats Indication du nombre de refus avant l’étude Procédures identiques dans tous les groupes Taille d’échantillon adaptée Règles éthiques respectées Conduite de l’étude Retraits indiqués, expliqués et raisonnables Retraits équilibrés entre les groupes Mesures principales fiables et valides Analyse des résultats Méthodes adaptées à la question et au schéma Prise en compte des variables importantes Résultats vérifiables des données brutes Prise en compte des comparaisons multiples Conclusion : Qualité Exceptionnelle Très bonne Plutôt bonne Faible mais acceptable Inacceptable Pas qualifié pour juger O I N NA NSP Discussion, forces et faiblesses : Cocher Faits avérés ou réfutés : Décision Mettre un seul X devant un des deux mots clefs INC A inclure dans la biblio EXC A ne pas inclure dans la biblio Justifications de l’inclusion Justifications de l’exclusion Mettre au moins un X devant un des mots clefs Mettre au moins un X devant un des mots clefs IAV Infection de l’accès vasculaire HOS Hors sujet BAC Bactériémie MAQ Mauvaise qualité HCV Hépatite virale C AUT Autre HIV SIDA En clair : VIR Autres viroses TIC Taux d’incidence TPR Taux de prévalence FAC Facteurs de risque (de confusion) REC Recommandations SUR Surveillance STA Standardisation des résultats INO Innovation EXC Exceptionnel, publication princeps Autres mots clefs réservés à la gestion des références bibliographiques SEL En cours de sélection NOL Non lu (non retenu par la sélection) LEC En cours de lecture (retenu par la sélection) Lecteur : ................................................... Date : ......... / ........./ ......... Commentaires Relationship between Clinical Outcomes and Vascular Access Type among Hemodialysis Patients with Staphylococcus aureus Bacteremia Jula K. Inrig,* Shelby D. Reed,† Lynda A. Szczech,* John J. Engemann,‡ Joelle Y. Friedman,† G. Ralph Corey,‡ Kevin A. Schulman,§ L. Barth Reller,‡ and Vance G. Fowler, Jr.‡ *Department of Medicine, Division of Nephrology, ‡Department of Medicine, Division of Infectious Diseases, and § Department of Medicine, Duke University Medical Center, and †Center for Clinical and Genetic Economics, Duke Clinical Research Institute, Durham, North Carolina The association between hemodialysis vascular access type, costs, and outcome of Staphylococcus aureus bacteremia (SAB) among patients with ESRD remains incompletely characterized. This study was undertaken to compare resource utilization, costs, and clinical outcomes among SAB-infected patients with ESRD by hemodialysis access type. Adjusted comparisons of costs and outcomes were based on multivariable linear regression and multivariable logistic regression models, respectively. A total of 143 hospitalized hemodialysis-dependent patients had SAB at Duke University Medical Center between July 1996 and August 2001. A total of 111 (77.6%) patients were hospitalized as a result of suspected bacteremia; 32 (22.4%) were hospitalized for other reasons. Of the 111 patients, 59.5% (n ⴝ 66) had catheters as their primary access type, 36% (n ⴝ 40) had arteriovenous (AV) grafts, and 4.5% (n ⴝ 5) had AV fistulas. Patients with fistulas were excluded from analyses because of small numbers. Patients with catheters were more likely to be white, had shorter dialysis vintage, and had higher Acute Physiology and Chronic Health Evaluation II scores compared with patients with grafts. Unadjusted 12-wk mortality did not significantly differ between patients with catheters compared with patients with grafts (22.7 versus 10.0%; P ⴝ 0.098); neither did 12-wk costs differ by access type ($22,944 ⴞ 18,278 versus $23,969 ⴞ 13,731, catheter versus graft; P > 0.05). In adjusted analyses, there was no difference in 12-wk mortality (odds ratio 1.63; 95% confidence interval 0.29 to 9.02; catheter versus graft) or 12-wk costs (means ratio 0.84; 95% confidence interval 0.60 to 1.17; catheter versus graft) among SAB-infected patients with ESRD on the basis of hemodialysis access type. Twelve-week mortality and costs that are associated with an episode of SAB are high in hemodialysis patients, regardless of vascular access type. Efforts should focus on the prevention of SAB in this high-risk group. Clin J Am Soc Nephrol 1: 518 –524, 2006. doi: 10.2215/CJN.01301005 I nfections are the second leading cause of mortality in dialysis patients, accounting for nearly one fourth of all deaths in patients with ESRD (1). Staphylococcus aureus is a leading cause of bacteremia among hemodialysis-dependent patients (1–11), and rates may be increasing (1). Complications of S. aureus bacteremia (SAB) include sepsis, endocarditis, osteomyelitis, septic arthritis, and metastatic abscesses (12–14). Vascular access is a major risk factor for bacteremia (15,16) and infection-related hospitalization and mortality (17,18) among hemodialysis patients. Evidence suggests that tunneledcuffed catheters might be associated with a higher frequency of infectious complications than other forms of dialysis vascular access. For example, the frequency of hospitalizations for infectious complications among patients with ESRD has increased (1) at the same time that the proportion of patients who have Received October 12, 2005. Accepted February 22, 2006. Published online ahead of print. Publication date available at www.cjasn.org. Address correspondence to: Dr. Jula K. Inrig, Duke University Medical Center, North Pavilion, 2400 Pratt Street, Box 3646, Durham, NC 27705. Phone: 919-6687516; Fax: 919-668-7128; E-mail: [email protected] Copyright © 2006 by the American Society of Nephrology ESRD and have catheters as their primary vascular access has climbed (19 –21). However, the association between vascular access type and clinical outcomes among patients with ESRD and SAB still is unclear. Despite its clinical significance, the economic impact of SAB among patients with ESRD remains incompletely characterized. Previous studies provided insight into the economic burden of S. aureus (22,23). Engemann et al. (24) found that the mean cost of an episode of SAB among a cohort of patients with ESRD was $24,034 per episode, and patients with complicated SAB episodes incurred even higher costs. Similarly, Reed et al. (25) found community-dwelling patients who had ESRD and methicillin-resistant S. aureus (MRSA) bacteremia incurred higher costs compared with patients with methicillin-sensitive S. aureus (MSSA) bacteremia. Despite these observations, however, the potential relationship between intravascular access type, costs, and clinical outcomes among patients with SAB is undefined. Herein, we sought to analyze differences in resource utilization, costs, and clinical outcomes by access type among patients who have ESRD and are admitted to the hospital for suspected SAB. ISSN: 1555-9041/103-0518 Clin J Am Soc Nephrol 1: 518 –524, 2006 Materials and Methods Study Design A prospective cohort of hemodialysis patients with SAB was used to describe and compare resource utilization, costs, and clinical outcomes among patients with various types of hemodialysis access. Characteristics of these patients have been reported previously (24,25). Patients and Study Setting All adult patients who had ESRD, were undergoing hemodialysis, and were admitted to Duke University Medical Center with SAB between July 1, 1996, and August 31, 2001, were identified prospectively and included in this study. Daily reports were received from the microbiology laboratory regarding all patients with at least one positive blood culture for S. aureus, and all patients subsequently were evaluated and consented for enrollment within 36 h of identification of S. aureus. Exclusion criteria were SAB in nonhospitalized patients, age ⬍18 yr, polymicrobial infection, neutropenia (absolute neutrophil count ⬍1.0 ⫻ 109/L), or death before evaluation. For maintaining independence of observations, the analysis was limited to the first episode of SAB for each patient but was inclusive of recurrent infections that occurred during the 12 wk after the date of the first positive blood culture. The institutional review board of Duke University Medical Center approved this study. Data Collection Clinical data were collected at the time of a patient’s hospitalization and entered into an electronic database (Access; Microsoft Corp., Redmond, WA). The data that were collected included patient demographic characteristics, comorbid conditions, antibiotic therapy, complications of infection, duration of ESRD and hemodialysis, type of vascular access, history of kidney transplant, and suspected cause of the patient’s kidney disease. Acute Physiology and Chronic Health Evaluation II (APACHE II) scores were calculated on the date of initial positive blood culture (26). Outcomes (including death or metastatic complications) were assessed during the initial hospitalization and 12 wk after the date of the initial positive blood culture. Data regarding resource utilization (e.g., medical tests, procedures, length of hospitalization) were obtained by medical record abstraction. Cost data information for all patients was retrieved from the Duke University Medical Center hospital accounting system (Transition Systems, Inc., Boca Raton, FL). Definitions Patients were categorized into groups on the basis of type of intravascular access used for hemodialysis at the time of admission: (1) Tunneled, cuffed, long-term catheter access; (2) synthetic arteriovenous (AV) graft; or (3) native AV fistula. Patients with catheter access were categorized further into two subgroups for sensitivity analyses: (1) Patients with catheters as their sole active access type and (2) patients with alternative active access types (maturing or active AV grafts or AV fistulas). Corticosteroid use was defined as either chronic low-dose or acute high-dose therapy used within the previous 30 d. Prevalent congestive heart failure at baseline was ascertained through self-report with medical record verification. Self-report was used to classify injection drug use. Sepsis was defined by standard criteria (27). Central nervous system involvement was defined by a positive culture for S. aureus obtained from cerebrospinal fluid. Infective endocarditis was defined according to Duke criteria (28). Metastatic infections were defined by the presence of an infection remote from the determined primary focus of infection (e.g., vertebral osteomyelitis, distant abscess). Methicillin susceptibility was determined from the oxacillin test using Staphylococcus aureus Bacteremia and Outcomes 519 criteria of the National Committee for Clinical Laboratory Standards (29). Cost Data The cost analysis was performed from the perspective of the health system and includes all direct medical costs that are associated with treatment of SAB and its sequelae during a 12-wk period. Because patients were included in the analysis only when their hospitalization was due to SAB, the entire cost of the admission was attributed to the infection. Cost data were available for all patients who were admitted after July 1, 1998. Hospital costs were obtained from the hospital cost accounting system and were adjusted to 2001 US dollars using the Consumer Price Index for Medical Care (30). Physician fees for inpatient and outpatient services were assigned using the 2001 reimbursement rates from the Centers for Medicare and Medicaid Services for providers in North Carolina. Total inpatient costs for the initial hospitalization included inpatient costs and physician fees. Outpatient costs and rehospitalization costs that occurred within 12 wk after the date of the initial positive blood culture were included only for procedures and admissions related to S. aureus infection. To be consistent with preferred methods (31), costs results were reported as means. Statistical Analyses Because of small numbers, patients with native AV fistulas as their sole access were excluded from further analyses. Descriptive statistics are presented as counts and percentages for discrete variables. Continuous variables are reported as means unless noted otherwise. In unadjusted analyses, continuous variables were compared using t tests when the data approximated a normal distribution and Wilcoxon rank sum tests when the data were non-normally distributed. Categorical variables were compared using Pearson 2 tests. Unadjusted comparisons of costs were based on nonparametric bootstrapping because the data did not approximate either a normal or log-normal distribution and because of nonequal variances (32,33). A bias-corrected method was applied to obtain the 95% confidence interval (CI) of the difference between cost estimates. An unadjusted comparison of clinical outcomes (mortality at discharge and at 12 wk) by access type was performed using Pearson 2 tests. Adjusted comparisons of costs were based on a multivariable generalized linear model (GLM) specified with a ␥ distribution and a log link to account for heteroscedasticity and skewness associated with the cost data (34 –36). Previous investigators have shown through a series of simulated data sets that GLM are less biased than ordinary least squares (OLS) regression on log-transformed cost data (34). GLM are more resilient to various data structures that are often represented with cost data, and the GLM approach typically is more conservative than OLS because the standard errors tend to be greater than those estimated by OLS (34). Cost analyses were performed using SAS PROC GLM. Adjusted comparisons of clinical outcomes were performed using a multivariable logistic regression model. Variables that were entered into the regression model included demographic characteristics, comorbidities, APACHE II scores, infection-related factors (admitted from home, nursing home, or other hospital), and infection with methicillin-sensitive S. aureus or MRSA. All analyses were performed using SAS statistical software, version 8.2 (SAS Institute, Inc., Cary, NC). Results Baseline Characteristics From July 1996 to August 2001, 143 hemodialysis patients who were admitted to Duke University Medical Center had 520 Clinical Journal of the American Society of Nephrology positive blood cultures for S. aureus. Of these, 111 (77.6%) patients were admitted for suspected bacteremia and 32 (22.4%) were admitted for other reasons and therefore were excluded from further analyses. Of the 111 patients, 59.5% (n ⫽ 66) had a catheter as their sole access type for hemodialysis upon admission, 36.0% (n ⫽ 40) had a synthetic AV graft, and 4.5% (n ⫽ 5) had a native AV fistula. Patient Characteristics According to Intravascular Access Type Patients with catheters as their sole access type were more likely to be white and have a shorter dialysis vintage compared with patients with AV grafts (Table 1). Patients with catheters also had higher total APACHE II scores on diagnosis with SAB compared with patients with AV grafts. There were no significant differences between groups with regard to gender, age, admission from a nursing home, underlying cause of renal disease, other comorbidities, or infection with MRSA. Clin J Am Soc Nephrol 1: 518 –524, 2006 Resource Utilization and Outcomes In unadjusted analyses, resource utilization during the initial hospitalization was similar between patients with catheters and patients with AV grafts, with the exception of the number of procedures that were performed related to a foreign device (4.1 versus 2.3 for catheter versus graft; P ⫽ 0.001; Table 2). During the 3-mo follow-up period, 13.8% of patients with catheters were re-hospitalized compared with 23.7% of patients with grafts (P ⫽ 0.21) with a similar number of mean inpatient days during the 12-wk period (12.6 versus 12.4 d for catheter versus graft; P ⫽ 0.33). Metastatic infections occurred in 22 (33.3%) patients with catheters and in 12 (30.0%) patients with AV grafts (P ⫽ 0.72). Metastatic complications included endocarditis (n ⫽ 17), septic emboli (n ⫽ 8), abscess (n ⫽ 6), septic arthritis (n ⫽ 5), meningitis (n ⫽ 3), nonvertebral osteomyelitis (n ⫽ 3), vertebral osteomyelitis (n ⫽ 3), septic thrombophlebitis (n ⫽ 2), cerebro- Table 1. Baseline patient characteristics by access type among ESRD patients hospitalized with SABa Characteristic Tunneled, Cuffed Catheter (n ⫽ 66) AV Graft (n ⫽ 40) Pb Male Race white black other Age (mean ⫾ SD) Duration of hemodialysis (mean ⫾ SD) Admitted from home nursing facility another hospital Cause of renal failure hypertension diabetes glomerulonephritis cystic kidney disease other unknown Comorbidities cancer HIV corticosteroids within 30 d diabetes injection drug use previous infective endocarditis none Total APACHE II score (mean ⫾ SD) Infection with MRSA 32 (48.5%) 17 (42.5%) 0.55 18 (27.3%) 46 (69.7%) 2 (3.0%) 55.3 ⫾ 15.0 1.87 ⫾ 2.24c 2 (5.0%) 38 (95.0%) 0 (0%) 57.7 ⫾ 15.5 4.26 ⫾ 4.01c 0.02 54 (81.2%) 10 (15.2%) 1 (3.6%) 35 (87.2%) 3 (7.5%) 2 (5.0%) 0.46 34 (51.5%) 35 (53.0%) 3 (4.6%) 1 (1.5%) 8 (12.1%) 0 (0%) 25 (62.5%) 16 (40.0%) 4 (10.0%) 0 (0%) 3 (7.5%) 2 (5.1%) 0.27 0.19 0.27 0.43 0.45 0.29 3 (4.6%) 1 (1.5%) 6 (9.1%) 39 (59.1%) 5 (7.6%) 3 (4.5%) 14 (21.2%) 19.8 ⫾ 5.4 24 (36.4%) 1 (2.5%) 0 (0%) 2 (5.0%) 19 (47.5%) 4 (10.0%) 0 (0%) 13 (32.5%) 17.5 ⫾ 4.4 11 (37.5%) 0.59 0.43 0.44 0.24 0.66 0.17 0.20 0.043 0.35 0.42 ⬍0.001 a APACHE II, Acute Physiology and Chronic Health Evaluation II; AV, arteriovenous; MRSA, methicillin-resistant Staphylococcus aureus; SAB, S. aureus bacteremia. b P values for continuous variables are based on t tests or Wilcoxon rank sum tests; P values for categorical variables are based on 2 tests. c Missing data for three patients in the catheter group and three patients in the graft group. Clin J Am Soc Nephrol 1: 518 –524, 2006 Staphylococcus aureus Bacteremia and Outcomes 521 Table 2. Unadjusted resource utilization, outcomes, and costs among patients with ESRD and SAB during their initial hospitalization and at 3-monthsa Tunneled, Cuffed Catheter (n ⫽ 66) AV Graft (n ⫽ 40) Difference with 95% CI or P Resource utilization initial hospitalization inpatient days (mean ⫾ SD) 10.6 ⫾ 10.1 10.3 ⫾ 6.4 0.79 procedures related to (mean ⫾ SD) foreign device 4.1 ⫾ 3.1 2.3 ⫾ 2.5 0.001 nonforeign device 0.3 ⫾ 1.1 0.1 ⫾ 0.3 0.75 12-wk period rehospitalizationb 8 (13.8%) 9 (23.7%) 0.21 inpatient days (mean ⫾ SD) 12.6 ⫾ 11.1 12.4 ⫾ 6.8 0.33 Outcomes initial hospitalization metastatic infection 22 (33.3%) 12 (30.0%) 0.72 died 8 (12.1%) 2 (5.0%) 0.22 12-wk period died 15 (22.7%) 4 (10.0%) 0.098 Costs (n ⫽ 39) (n ⫽ 23) initial hospitalization mean ⫾ SD $16,679 ⫾ 12,063 $17,509 ⫾ 11,267 ⫺$830 (⫺$7,161 to $5,117) median (⫾IQR) $13,776 (⫾82961 to 18,698) $13,663 (⫾10,868 to 23,607) 12-wk period mean ⫾ SD $22,944 ⫾ 18,278 $23,969 ⫾ 13,731 ⫺$1,024 (⫺$8,459 to $6,792) median (⫾IQR) $17,392 (⫾11,945 to 29,425) $18,318 (⫾13,663 to 31,660) a CI, confidence interval; IQR, interquartile range. Patients who died during initial hospitalization were excluded. b vascular event (n ⫽ 2), epidural abscess (n ⫽ 1), and other (n ⫽ 5). More than one metastatic complication could be present in the same patient. Inpatient mortality during the initial hospitalization was 12.1% for patients with catheters and 5.0% for patients with AV grafts (P ⫽ 0.22). Unadjusted 3-mo mortality did not differ significantly on the basis of hemodialysis access type (22.7 versus 10.0% for catheter versus graft; P ⫽ 0.098). In the 15 patients who had catheters and died, cause of death was related to SAB in eight and unrelated in seven. All four deaths in patients with AV grafts were due to SAB or its complications. Cost Analysis Patients with catheters had similar mean costs during the initial hospitalization for SAB ($16,679 ⫾ $12,063) compared with patients with AV grafts ($17,509 ⫾ $11,267; difference ⫺$830; 95% CI ⫺$7,161 to $5,117). Mean total 3-mo costs that were associated with SAB also were similar between patients with catheters ($22,944 ⫾ $18,278) and patients with AV grafts ($23,969 ⫾ $13,731; difference ⫺$1,024; 95% CI ⫺$8,459 to $6,792). Multivariable Analysis After adjustment for patient characteristics, the odds of death at 3 mo did not differ significantly by access type (Table 3). In contrast, an increasing APACHE II score (odds ratio [OR] 1.19; 95% CI 1.02 to 1.39) and bacteremia caused by MRSA (OR 10.33; 95% CI 2.15 to 49.66) were associated with increased mortality at 12 wk. Twelve-week costs did not differ significantly between patients with catheters and patients with AV grafts (means ratio 0.84; 95% CI 0.60 to 1.17) after adjustment for patient characteristics. Subgroup Analyses of Patients with Catheters Among the 66 patients with catheters, 20 had alternative active access (maturing or active AV grafts or fistulas), 18 had failed AV grafts or fistulas, and the remaining 28 had no alternative active or inactive access type. Inpatient and 12-wk mortality did not differ between patients without alternative active access (n ⫽ 46) compared with patients with maturing AV fistulas or grafts (n ⫽ 20; inpatient mortality 13 versus 10% [P ⫽ 0.73]; 12-wk mortality 26 versus 15% [P ⫽ 0.32]; catheter only versus alternative active access, respectively). Among patients who had a catheter but no alternative active access, the mean costs for the initial hospitalization were $18,162 (⫾$11,687) compared with $12,903 (⫾$12,740) among patients with maturing AV grafts or fistulas (difference $5259; 95% CI ⫺$5056 to $12,627). At 12 wk, mean costs for patients who had a catheter but no alternative access were $25,738 (⫾$19,415) compared with $15,831 (⫾$13,198) for patients who had a catheter and a maturing AV graft or fistula (difference $9907; 95% CI ⫺$944 to $20,183). After adjustment for patient characteristics, there was no difference in 12-wk mortality between patients who had a catheter but no alternative access and those with a maturing AV 522 Clinical Journal of the American Society of Nephrology Clin J Am Soc Nephrol 1: 518 –524, 2006 Table 3. Associations between clinical and demographic variables and mortality and costs among patients with ESRD and SABa 12-Wk Mortality 12-Wk Costs Variable Odds Ratio Catheter onlyb Female gender White race Diabetes Hypertension History of kidney transplant CNS involvement Heart failure Sepsis APACHE II MRSA Admit from nursing facility/other hospital a 1.63 0.24 0.90 1.72 3.02 0.29 3.32 2.11 3.93 1.19 10.33 0.69 95% CI 0.29 0.05 0.12 0.34 0.53 0.01 0.48 0.16 0.47 1.02 2.15 0.10 to to to to to to to to to to to to 9.02 1.12 6.90 8.80 17.36 6.32 22.89 27.85 32.76 1.39 49.66 4.65 Means Ratio 0.84 0.89 1.19 1.12 0.86 1.05 1.39 1.54 1.14 0.97 1.09 1.45 95% CI 0.60 0.65 0.77 0.81 0.61 0.61 0.79 0.80 0.67 0.94 0.77 0.87 to to to to to to to to to to to to 1.17 1.23 1.86 1.54 1.21 1.21 2.46 2.98 1.93 1.01 1.55 2.44 CNS, central nervous system. Graft only served as the reference group. b graft or fistula (OR 1.71; 95% CI 0.25 to 11.75). In adjusted analyses, 12-wk costs also did not differ significantly on the basis of the absence or presence of alternative active access (means ratio 1.52; 95% CI 0.98 to 2.35). Discussion SAB is a serious, common complication of intravascular access in hemodialysis-dependent patients (1). Our investigation shows that morbidity, mortality, and costs that are associated with an episode of SAB among hemodialysis-dependent patients are pronounced but do not vary significantly by different forms of intravascular access. The clinical impact of an episode of SAB in hemodialysis patients is considerable and is highlighted by our cohort. Metastatic complications occurred in 34% of patients, including infective endocarditis, osteomyelitis, and abscesses. The high rate of metastatic complications among our patients is consistent with previous reports (10,12) and highlights the virulence of S. aureus, particularly in an immunocompromised host with an intravascular foreign device. It is interesting that the rate of metastatic infections did not differ among patients on the basis of their intravascular access type. This observation suggests that the risk for metastatic complications that are associated with SAB in hemodialysis patients is high regardless of the potential source of infection. Overall mortality also was high in this investigation but did not vary significantly among patients with different vascular access types, even after adjustment for differences in acute and chronic comorbidities. Although vascular access type did not affect length of hospitalization, the number of procedures that were performed during the initial hospitalization was highest among patients with catheters. This reflects the standard practice at our institution of prompt removal of a hemodialysis catheter in any patient with documented SAB whenever clinically possible. Patients who had AV grafts and developed SAB also required extensive medical therapy, because half of these patients required surgical removal or revision of their vascular access, and approximately one quarter required re-hospitalization during the follow-up period (37). The findings from our cost analysis of hemodialysis-dependent patients with SAB are unique and show that overall costs are pronounced (approximately $23,000 during 12 wk). Surprisingly, costs did not differ statistically between patients with catheters and patients with AV grafts, despite the high surgical costs associated with an AV graft revision/removal and replacement (38). As the total costs for treating the 111 hemodialysis patients who had SAB and were reported in this study was $2,440,987, access-related SAB represents a significant economic burden to the health care system. It should be highlighted that these costs are in addition to the already substantial health care costs associated with maintenance hemodialysis and initial dialysis access placement (38 – 40). There are a number of potential reasons that no significant associations were noted between a patient’s hemodialysis access type and his or her ultimate clinical outcome. First, among patients with AV grafts, complications and costs that are associated with an AV graft revision/resection can be significant (37,38,41). In contrast, patients with catheters required a large number of procedures (e.g., catheter removal and replacement, prosthetic device removal, surgery for infective endocarditis), which likely escalated the mortality rate and cost in this cohort. The lack of our ability to detect a difference in costs among patients who had a catheter and who did and did not have alternative active access types may be due to the lack of power within this subgroup analysis. Alternatively, the other active access type may not have been readily usable at the time of admission and therefore was unable to confer a cost savings; in addition, if the maturing AV graft or fistula had been placed recently, then management of postsurgical infectious complications may have increased costs in this subgroup. Clin J Am Soc Nephrol 1: 518 –524, 2006 Our study has a number of limitations. First, our data were derived from an observational design with a modest sample size, thereby limiting our ability to detect statistically significant differences between groups. Second, groups that were based on access type had distinct differences in comorbidities; however, we tried to identify and adjust for all potential confounders between groups in our statistical analysis. Third, although we did not capture costs for patients who were admitted to outside hospitals, follow-up was complete on all patients at 12 wk, thus making the risk for missing an outside hospital admission small. Fourth, we did not include in our cost analysis outpatients who were treated for bacteremia, although this is less relevant as it is standard practice for patients with documented SAB to be treated as inpatients and therefore this should not have affected our results. Finally, referral bias may overestimate the cost and outcome of SAB if only sicker patients were admitted to our institution. This risk is minimized by the fact that the majority of patients in this study are cared for by university-affiliated hemodialysis units that use our institution for all hospital admissions. Despite these limitations, our cohort highlights the significant morbidity, mortality, and costs that are associated with an episode of SAB among hemodialysis patients. Furthermore, our investigation suggests that a patient’s hemodialysis access type is not significantly associated with differential costs or clinical outcomes during an episode of SAB, even if the vascular access is readily removable. Given the significant patient, physician, and cost burden associated with an episode of SAB in patients with ESRD, effective interventions to try to prevent S. aureus infections in this high-risk group are warranted. Acknowledgments This research was supported in part by a research agreement between Duke University and Nabi Biopharmaceuticals. J.K.I. was supported by a clinical research fellowship endowment from Earl Dalbey and by National Institutes of Health grant K12 RR-017630. V.G.F. was supported by National Institutes of Health grant R01-AI059111. This study was presented in abstract form at the European Renal Association-European Dialysis and Transplant Association Annual Congress, Lisbon, Portugal, May 15 to 18, 2004. References 1. US Renal Data Systems: USRDS 2003 Annual Data Report: Atlas of End-Stage Renal Disease in the United States, Bethesda, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, 2003 2. Krishnasami Z, Carlton D, Bimbo L, Taylor ME, Balkovetz DF, Barker J, Allon M: Management of hemodialysis catheter-related bacteremia with an adjunctive antibiotic lock solution. Kidney Int 61: 1136 –1142, 2002 3. Mokrzycki MH, Schoppel B, Von Gersdorff G, Rush H, Zdunek MP, Feingold R: Tunneled-cuffed catheter associated infections in hemodialysis patients who are seropositive for the human immunodeficiency virus. J Am Soc Nephrol 11: 2122–2127, 2000 4. Boelaert JR, Van Landuyt HW, Gordts BZ, De Baere YA, Messer SA, Herwaldt LA: Nasal and cutaneous carriage of Staphylococcus aureus in hemodialysis patients: The effect of Staphylococcus aureus Bacteremia and Outcomes 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 523 nasal mupirocin. Infect Control Hosp Epidemiol 17: 809 – 811, 1996 Chow JW, Yu VL: Staphylococcus aureus nasal carriage in hemodialysis patients: Its role in infection and approaches to prophylaxis. Arch Intern Med 149: 1258 –1262, 1989 Saad TF: Bacteremia associated with tunneled, cuffed hemodialysis catheters. Am J Kidney Dis 34: 1114 –1124, 1999 Dryden MS, Samson A, Ludlam HA, Wing AJ, Phillips I: Infective complications associated with the use of the Quinton ‘Permcath’ for long-term central vascular access in haemodialysis. J Hosp Infect 19: 257–262, 1991 Beathard GA: Management of bacteremia associated with tunneled-cuffed hemodialysis catheters. J Am Soc Nephrol 10: 1045–1049, 1999 Tanriover B, Carlton D, Saddekni S, Hamrick K, Oser R, Westfall AO, Allon M: Bacteremia associated with tunneled dialysis catheters: Comparison of two treatment strategies. Kidney Int 57: 2151–2155, 2000 Poole CV, Carlton D, Bimbo L, Allon M: Treatment of catheter-related bacteremia with an antibiotic lock protocol: Effect of bacterial pathogen. Nephrol Dial Transplant 19: 1227–1244, 2004 Marr KA, Sexton DJ, Conlon PJ, Corey RG, Schwab SJ, Kirkland KB: Catheter-related bacteremia and outcome of attempted salvage in patients undergoing hemodialysis. Ann Intern Med 127: 275–280, 1997 Marr KA, Kong L, Fowler VG, Gopal A, Sexton DJ, Conlon PJ, Corey GR: Incidence and outcome of Staphylococcus aureus bacteremia in hemodialysis patients. Kidney Int 54: 1684 –1689, 1998 Schwab SJ, Beathard GA: The hemodialysis catheter conundrum: Hate living with them, but can’t live without them. Kidney Int 56: 1–17, 1999 Sexton DJ: Vascular access infections in patients undergoing dialysis with special emphasis on the treatment of Staphylococcus aureus. Infect Dis Clin North Am 15: 731–742, 2001 Hoen B, Paul-Dauphin A, Hestin D, Kessler M: EPIBACDIAL: A multicenter prospective study of risk factors for bacteremia in chronic hemodialysis patients. J Am Soc Nephrol 9: 869 – 876, 1998 Powe NR, Jaar B, Furth SL, Hermann J, Briggs W: Septicemia in dialysis patients: Incidence, risk factors, and prognosis. Kidney Int 55: 1081–1090, 1999 Pastan S, Soucie M, McClellan WM: Vascular access and increased risk of death among hemodialysis patients. Kidney Int 62: 620 – 626, 2002 Allon M, Depner TA, Radeva M, Bailey J, Beddhu S, Butterly D, Coyne DW, Gassman JJ, Kaufman AM, Kaysen GA, Lewis JA, Schwab SJ; HEMO Study Group: Impact of dialysis dose and membrane on infection-related hospitalization and death: Results of the HEMO study. J Am Soc Nephrol 14: 1863–1870, 2003 Tokars JI, Miller ER, Alter MJ, Arduino MJ: National surveillance of dialysis-associated diseases in the United States, 2000. Semin Dial 15: 162–171, 2002 Pisoni RL, Young EW, Dykstra DM, Greenwood RN, Hecking E, Gillespie B, Wolfe RA, Goodkin DA, Held PJ: Vascular access use in Europe and in the United States: Results from the DOPPS. Kidney Int 61: 305–316, 2002 National Kidney Foundation: K/DOQI: Clinical practice guidelines for chronic kidney disease: Evaluation, classifi- 524 22. 23. 24. 25. 26. 27. 28. 29. 30. Clinical Journal of the American Society of Nephrology cation, and stratification. Available: http://www.kidney. org. Accessed March 1, 2005 Rubin RJ, Harrington CA, Poon A, Dietrich K, Greene JA, Moiduddin A: The economic impact of Staphylococcus aureus infection in New York City hospitals. Emerg Infect Dis 5: 9 –17, 1999 Abramson MA, Sexton DJ: Nosocomial methicillin-resistant and methicillin-susceptible Staphylococcus aureus primary bacteremia: At what costs? Infect Control Hosp Epidemiol 20: 408 – 411, 1999 Engemann JJ, Friedman JY, Reed SD, Griffiths RI, Szczech LA, Kaye KS, Stryjewski ME, Reller B, Schulman KA, Corey RG, Fowler VF: Clinical outcomes and costs due to Staphylococcus aureus bacteremia among patients on chronic hemodialysis. Infect Control Hosp Epidemiol 36: 534 – 539, 2005 Reed SD, Friedman JY, Engemann JJ, Griffiths RI, Anstrom KJ, Kaye KS, Stryjewski ME, Szczech LA, Reller B, Corey RG, Schulman KA, Fowler VF: Costs and outcomes among hemodialysis-dependent patients with methicillin-resistant or methicillin-susceptible Staphylococcus aureus bacteremia. Infect Control Hosp Epidemiol 26: 175–183, 2005 Knaus WA, Draper EA, Wagner DP, Zimmerman JE: APACHE II: A severity of disease classification system. Crit Care Med 13: 818 – 829, 1985 American College of Chest Physicians/Society of Critical Care Med Consensus Conference: Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit Care Med 20: 864 – 874, 1992 Durack DT, Lukes AS, Bright DK: New criteria for diagnosis of infective endocarditis: Utilization of specific echocardiographic findings. Duke Endocarditis Service. Am J Med 96: 200 –209, 1994 National Committee for Clinical Laboratory Standards: Performance Standards for Antimicrobial Susceptibility Testing: Twelfth Informational Supplement, Wayne, PA, National Committee for Clinical Laboratory Standards, 2002 [Approved standard M100 –S112] Bureau of Labor Statistics Data: US Department of Labor Clin J Am Soc Nephrol 1: 518 –524, 2006 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. Statistics web site. Available at http://www.bls.gov/data. Accessed October 15, 2005 Thompson SG, Barber JA: How should cost data in pragmatic randomised trials be analyzed? BMJ 320: 1197–1200, 2000 Briggs A, Gray A: The distribution of health care costs and their statistical analysis for economic evaluation. J Health Serv Res Policy 3: 233–245, 1998 Zhou XH, Melfi CA, Hui SL: Methods for comparison of cost data. Ann Intern Med 127: 752–756, 1997 Manning WG, Mullahy J: Estimating log models: To transform or not to transform? J Health Econ 20: 461– 494, 2001 Blough DK, Ramsey SD: Using generalized linear models to assess health care costs. Health Services and Outcomes Research Methodology 1: 185–202, 2000 Manning WG, Basu A, Mullahy J: Generalized modeling approaches to risk adjustment of skewed outcomes data. J Health Econ 24: 465– 488, 2005 Minga TE, Flanagan KH, Allon M: Clinical consequences of infected arteriovenous grafts in hemodialysis patients. Am J Kidney Dis 38: 975–978, 2001 Lee H, Manns B, Taub K, Ghali WA, Dean S, Johnson D, Donaldson C: Cost analysis of ongoing care of patients with end-stage renal disease: The impact of dialysis modality and dialysis access. Am J Kidney Dis 40: 611– 622, 2002 Manns B, Tonelli M, Yilmaz S, Lee H, Laupland K, Klarenbach S, Radkevich V, Murphy B: Establishment and maintenance of vascular access in incident hemodialysis patients: A prospective cost analysis. J Am Soc Nephrol 16: 201–209, 2005 United States Renal Data System: The economic cost of ESRD, vascular access procedures, and Medicare spending for alternative modalities of treatment. USRDS. Am J Kidney Dis 30[Suppl 1]: S160 –S177, 1997 Tabbara MR, O’Hara PJ, Hertzer NR, Krajewski LP, Beven EG: Surgical management of infected PTFE hemodialysis grafts: Analysis of a 15-year experience. Ann Vasc Surg 9: 378 –384, 1995 Fiche de lecture critique d’une étude clinique ou épidémiologique DIALIN (V3 08/03/2007) (d’après SALMI L.R. Lecture critique et rédaction médicale scientifique. Comment lire, rédiger et publier une étude clinique ou épidémiologique. Paris :Elsevier ; 1995) Référence : N° BLA2588 Auteurs : Inrig JK, Reed SD, Szczech LA, Engemann JJ, Friedman JY, Corey GR, Schulman KA, Reller LB, Fowler VG, Jr. .................................................................................................................... Titre : Relationship between clinical outcomes and vascular access type among hemodialysis patients with Staphylococcus aureus bacteremia. .............................................. Revue : Clin J Am Soc Nephrol. ............................................................................................................... Année : 2006 .................. Volume (n°) : 1 (3)) ..Pages : 518-524 ....................................................... Résumé : Objectif et justification : Comparer l’utilisation des ressources, les coûts et les résultats cliniques chez les patients hémodialysés atteints d’une bactériémie à Staphylococcus aureus (SAB) selon le type d’accès vasculaire. Schéma d’étude ; groupes comparés : Etude de cohorte prospective avec 3 groupes : cathéters, fistule greffe, FAV. Population étudiée et nombre de sujets : 143 patients hémodialysés hospitalisés entre juillet 1996 et août 2001 dans un hôpital universitaire de Caroline du Nord avec SAB : 111 (77.6%) pour bactériémie suspectée, 40 (36%) pour une autre raison. Conduite de l’étude : Sur les 111 patients, 66 (59.5%) cathéters, 40 (36%) fistule greffe et 5 (4.5%) FAV (ce dernier groupe trop peu nombreux étant exclus des analyses Principaux résultats : Sur les analyses ajustées : Pas de différence significative de mortalité à 12 semaines (OR cathéter/greffe 1.63 [0.29 9.03]95% Pas de différence significative de coûts à 12 semaines (OR cathéter/greffe 0.84 [0.60 1.17]95% Cocher la case correspondant au respect du critère : 0 = oui, N = non, I = incomplet ; NA = ne s’applique pas, NSP = ne sait pas. Une réponse cochée N à un critère en italique = étude inacceptable. Critères Schéma d’étude Formulation claire de l’objectif Hypothèse faite à priori Schéma adapté à l’objectif Procédures de sélection Critères d’inclusion décrits et adéquats Critères d’exclusion décrits et adéquats Indication du nombre de refus avant l’étude Procédures identiques dans tous les groupes Taille d’échantillon adaptée Règles éthiques respectées Conduite de l’étude Retraits indiqués, expliqués et raisonnables Retraits équilibrés entre les groupes Mesures principales fiables et valides Analyse des résultats Méthodes adaptées à la question et au schéma Prise en compte des variables importantes Résultats vérifiables des données brutes Prise en compte des comparaisons multiples Conclusion : Qualité Exceptionnelle Très bonne Plutôt bonne Faible mais acceptable Inacceptable Pas qualifié pour juger O I N NA NSP X X X X X X X X X X X X X X X X Discussion, forces et faiblesses : Echantillon faible Cocher Faits avérés ou réfutés : X Pas de différence significative de mortalité ajustée ou de coûts ajustés selon le type d’accès vasculaire Décision Mettre un seul X devant un des deux mots clefs X INC A inclure dans la biblio EXC A ne pas inclure dans la biblio Justifications de l’inclusion Justifications de l’exclusion Mettre au moins un X devant un des mots clefs Mettre au moins un X devant un des mots clefs IAV Infection de l’accès vasculaire HOS Hors sujet X BAC Bactériémie MAQ Mauvaise qualité HCV Hépatite virale C AUT Autre HIV SIDA En clair : VIR Autres viroses TIC Taux d’incidence TPR Taux de prévalence X FAC Facteurs de risque (de confusion) REC Recommandations SUR Surveillance STA Standardisation des résultats INO Innovation EXC Exceptionnel, publication princeps Autres mots clefs réservés à la gestion des références bibliographiques SEL En cours de sélection NOL Non lu (non retenu par la sélection) LEC En cours de lecture (retenu par la sélection) Lecteur : LA Commentaires Date : 8/1/2009