Assessment by mini clinical evaluation exercise (Mini-CEX) in the intensive care unit.

INTRODUCTION: Mini Clinical Evaluation Exercise (mini-CEX) is one of the assessment tools in medical education. It includes three steps: overview of clinical situation, observation and feedback. AIM: To evaluate the feasibility of mini-CEX as a formative assessment tool for medical trainees in 5th year of medicine in a teaching intensive care unit (ICU). METHODS: Single-center qualitative research conducted in ICU during the 2nd semester of the academic year 2022-2023. Seven core clinical skill assessments were done, and the performance was rated on a 9-point scale. An assessment of the method was conducted with both trainees and clinical educators. RESULTS: We conducted six mini-CEX recorded sessions. All medical students had marks under the average of 4.5. In the first period, the highest mark was obtained for counselling skills (4.5). The best score was obtained for clinical judgement (4) in the second period and for management plan (4) in the third period. Most of medical trainees (11 sur 12) were satisfied with the method and feedback was according to them the most useful step. Ten students agreed fully to introduce this assessment tool in medical educational programs. Two medical educators out of three did not practice this method before. They agreed to include mini-CEX in the program of medical education of the faculty of medicine of Tunis. However, they did not agree to use it as a summative assessment tool. CONCLUSION: Our study demonstrates that we can use the mini-CEX in medical teaching. Both trainees and educators were satisfied with the method.

Early administration of norepinephrine in sepsis: Multicenter randomized clinical trial (EA-NE-S-TUN) study protocol.

One of the most important components of sepsis management is hemodynamic restoration. If the target mean arterial pressure (MAP) is not obtained, the first recommendation is for volume expansion, and the second is for norepinephrine (NE). We describe the methodology of a randomized multicenter trial aiming to assess the hypothesis that low-dose NE given early in adult patients with sepsis will provide better control of shock within 6 hours from therapy starting compared to standard care. This trial includes ICU septic patients in whom MAP decrease below 65 mmHg to be randomized into 2 groups: early NE-group versus standard care-group. The patient's attending clinician will determine how much volume expansion is necessary to meet the target of a MAP > 65 mm Hg. If this target not achieved, after at least 30 ml/kg and guided by the available indices of fluid responsiveness, NE will be used in a usual way. The latter must follow a consensual schedule elaborated by the investigating centers. Parameters to be taken at inclusion and at H6 are: lactates, cardiac ultrasound parameters (stroke volume (SV), cardiac output (CO), E/E' ratio), and P/F ratio. MAP and diuresis are recorded hourly. Our primary outcome is the shock control defined as a composite criterion (MAP > 65 mm Hg for 2 consecutive measurements and urinary output > 0.5 ml/kg/h for 2 consecutive hours) within 6 hours. Secondary outcomes: Decrease in serum lactate> 10% from baseline within 6 hours, the received fluid volume within 6 hours, variation of CO and E/E', and 28 days-Mortality. The study is ongoing and aims to include at least 100 patients per arm. This study is likely to contribute to support the indication of early initiation of NE with the aim to restrict fluid intake in septic patients. (ClinicalTrials.gov ID: NCT05836272).

Resource utilization and preparedness within the COVID-19 pandemic in Tunisian medical intensive care units: A nationwide retrospective multicentre observational study.

BACKGROUND: The worldwide SARS-CoV-2 pandemic represents the most recent global healthcare crisis. While all healthcare systems suffered facing the immense burden of critically-ill COVID-19 patients, the levels of preparedness and adaptability differed highly between countries. AIM: to describe resource mobilization throughout the COVID-19 waves in Tunisian University Medical Intensive Care Units (MICUs) and to identify discrepancies in preparedness between the provided and required resource. METHODS: This is a longitudinal retrospective multicentre observational study conducted between March 2020 and May 2022 analyzing data from eight University MICUs. Data were collected at baseline and at each bed expansion period in relation to the nation's four COVID-19 waves. Data collected included epidemiological, organizational and management trends and outcomes of COVID-19 and non-COVID-19 admissions. RESULTS: MICU-beds increased from 66 to a maximum of 117 beds. This was possible thanks to equipping pre-existing non-functional MICU beds (n = 20) and creating surge ICU-beds in medical wards (n = 24). MICU nurses increased from 53 to 200 of which 99 non-ICU nurses, by deployment from other departments and temporary recruitment. The nurse-to-MICU-bed ratio increased from 1:1 to around 1·8:1. Only 55% of beds were single rooms, 80% were equipped with ICU ventilators. These MICUs managed to admit a total of 3368 critically-ill patients (15% of hospital admissions). 33·2% of COVID-19-related intra-hospital deaths occurred within the MICUs. CONCLUSION: Despite a substantial increase in resource mobilization during the COVID-19 pandemic, the current study identified significant persisting discrepancies between supplied and required resource, at least partially explaining the poor overall prognosis of critically-ill COVID-19 patients.

Case report: Severe brain hypoxic damages after acute methanol poisoning.

Methanol poisoning is a challenging clinical situation with irreversible neurologic complication mainly encountered in developed countries. We report a case of a 50-year-old patient who presented with methanol poisoning, symptomatic of respiratory and neurologic failure. In this context, cerebral magnetic resonance imaging concluded entangled injury mechanisms leading to neurologic failure.

Presumed Protective Role for Anti-Hepatitis B Virus Antibodies Against COVID-19 Severe Cases: A Clinical Study Confirming in silico Hypothesis.

Background: Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is responsible for COVID-19 disease which is known to have a broad clinical spectrum, from asymptomatic to critical presentation leading to death. Many researchers have investigated the factors impacting the course of the disease. Our previous in silico study suggested a possible protective effect of Hepatitis B, Tetanus and Measles vaccines against COVID-19. In continuity, we conducted a cross-sectional clinical study in order to confirm our in silico assumptions regarding the HBs-Ag antibodies. Methods: A representative sex- and age-matched sample of patients with confirmed COVID-19 was selected (n = 340). All clinical presentations were equally represented. Using an ELISA test, each patient benefited of a serology for the detection and measurement of the anti-HBs specific IgG antibodies. The obtained results allowed determining the different correlations between these antibody titers and the disease severity. The R® software and the MedCalc® software served to calculate the Spearman's coefficient of rank correlation (rho) for the obtained titers per severity group as well as the different other calculations and figure representations. Results: A significant positive correlation was found with the anti-HBs titers (rho = 0.107; p = 0.04). High anti-HBs titers were significantly associated with the mild presentation of COVID-19. A significant difference was found between the obtained titers per severity class (chi-2 test, p = 0.03). Discussion/Conclusion: Our findings demonstrated that anti-HBs titers were significantly higher for patients having mild COVID-19 presentations. We presume that being immunized against the HB may play a protective role in the course of the disease. Our study provided more key elements in understanding the disparity of the clinical spectrum among regions.

Antimicrobial Susceptibility among Pathogens Isolated in Early- versus Late-Onset Ventilator-Associated Pneumonia.

Ventilator-associated pneumonia (VAP) is associated with increased hospital stay and high morbidity and mortality in critically ill patients. The aims of this study were to (i) determine the incidence of multidrug-resistant (MDR) pathogens in the first episodes of VAP and to assess potential differences in bacterial profiles of subjects with early- versus late-onset VAP. This was a retrospective cohort study over a period of 18 months including all patients who had a first episode of VAP confirmed by positive bacterial culture. Subjects were distributed into two groups according to the number of intubation days: early-onset VAP (<5 days) or late-onset VAP (≥5 days). The primary endpoint was the nature of causative pathogens and their resistance profiles. Sixty patients were included, 29 men and 31 women, with an average age of 38 ± 16 years. The IGS 2 at admission was 40.5 [32-44] and APACHE was 19 [15-22]. Monomicrobial infections were diagnosed in 77% of patients (n = 46). The most frequently isolated bacteria were A. baumannii, 53% (n = 32); P. aeruginosa in 37% (n = 22); Enterobacterales in 28% (n = 17) and S. aureus in 5% (n = 3). Ninety-seven percent of the bacteria were MDR. The VAP group comprised 36 (60%) episodes of early-onset VAP and 24 (40%) episodes of late-onset VAP. There was no significant difference in the distribution of the bacterial isolates, nor in terms of antibacterial resistances between early- and late-onset VAPs. Our data support recent observations that there is no microbiological difference in the prevalence of potential MDR pathogens or in their resistance profiles associated with early- versus late-onset VAPs, especially in countries with high rates of MDR bacteria.

Toxin-induced cardiac arrest: frequency, causative agents, management and hospital outcome.

BACKGROUND: Cardiac arrest (CA) is a public health problem, with various etiologies and a fatal issue in 90-95% of cases. Toxin-induced cardiac arrests (TICA) are poorly described. Scarcity of national data prompted us to carry-out this study. AIM: To determine TICA frequency in a Tunisian reference center in toxicology and its hospital prognosis, and to describe its clinical and therapeutic aspects Methods : Data were collected retrospectively over an 8-years period. We included patients admitted for post-CA care with highly suspected or confirmed TICA. Clinical and toxicological data were recorded. RESULTS: We recorded 21 cases of TICA, which represented 48.8% of CA. A single toxic agent was incriminated in 90% of cases. Main causative agents identified in our series were pesticides and betablockers: chloralosed (n = 6), carbamate inhibitor of cholinesterase (n = 5), acebutolol (n = 4) and organophosphate (n = 2). One case of opiates and cocaine poisoning was reported. Median duration of "no flow" was 0 minutes. Mean duration of "low flow" was 13.74±9.15 minutes. An initial shockable rhythm was noted only in three patients. Mortality rate was 76% (16/21). Four of the five survivors had a Cerebral Performance Category Scale (CPC) 1, only one patient survived with a CPC 3. Factors associated with mortality were : the duration of "low flow" (p=0.02) and APACHE II score (p=0.014). APACHE II≥29 was the only independent factor (OR=2.0, 95%CI [1.07;3.71]). CONCLUSION: TICA were most frequently provoked by pesticides, mortality was high and was independently predicted by APACHE II score.

Nephrotoxicity of Ciprofloxacin: Five Cases and a Review of the Literature.

Fluoroquinolones are usually well tolerated with a minimum of serious adverse effects; renal toxicity is uncommon. Apart from the renal side effects of ciprofloxacin, we aimed to highlight the renal impact of a ciprofloxacin overdose, and thus conducted a prospective study in the Department of Nephrology at La Rabta Hospital between 2010 and 2015. The cohort database was continually updated until the inclusion of five patients who were subjected to an overdose and who were initially admitted to the medical intensive care unit and then transferred to our department for acute renal failure (ARF) due to ciprofloxacin ingestion requiring urgent hemodialysis. All patients developed ARF after 12-36 h of ingestion. Renal ultrasound was normal in all cases. Twenty-four-hour proteinuria was present but not significant in one case, while microscopic hematuria was present in one case. Treatment consisted of supportive therapy and extrarenal purification by conventional intermittent hemodialysis. Four patients recovered normal renal function within 3 weeks and the remaining patient eventually had chronic kidney failure.

Methanol poisoning in Tunisia: report of 16 cases.

Methanol poisoning continues to be a public health problem in Tunisia in spite of the different legislative measures. We report a series of 16 cases of methanol poisoning admitted to our Intensive Care Unit between December 2003 and April 2004. The patients' median age was 21.5 years (range 16 to 53 years) with a median SAPS II of 14 (range 12 to 84) and an APACHE II of 8 (range 6 to 36). The median latent period was 9.5 hours (range 4 to 24 hours) with a delay to medical consultation of 36 hours (range 6 to 48 hours), and a median serum methanol concentration of 1.4 g/L (range 0.19 to 3.62 g/L). Clinical signs included central nervous system symptoms (69%), gastrointestinal complaints (87%), visual disturbances (69%) and metabolic acidosis (94%). Three patients (19%) required mechanical ventilation because of deep coma or shock and died within 6 hours. Hemodialysis was performed in eleven patients (69%) because of visual disturbances and/or metabolic acidosis. One patient developed irreversible bilateral blindness and another unilateral blindness secondary to optic neuropathy. Statistical significant risk factors for the developing of visual disturbances were found to be the ingested quantity of methanol, the latent period, acidosis and serum methanol concentration on admission.

Acute renal failure by ingestion of Euphorbia paralias.

Euphorbia paralias is known in traditional medicine as an anti-inflammatory agent, a purgative and for its local anesthetic property. To the best our knowledge, renal toxicity of this substance has not been previously reported. In this paper, we report the case of a 29-year-old male who developed renal damage following ingestion of Euphorbia paralias. He had been on follow-up for nephrotic syndrome since 1986, although irregularly, with several relapses but each responding well to steroid therapy. A kidney biopsy had not been performed earlier due to refusal by the patient. He was off steroids since April 2008 because the patient developed osteoporosis. He was admitted with general malaise and oliguria to our department in May 2009, following repeated vomiting and watery diarrhea for three days. On examination, he was edematous but had normal vital signs except for a pulse rate of 120/min. Hemoglobin was only 5.5 g/dL but with normal white cell and platelet counts. Blood biochemistry showed evidence of advanced renal failure with a serum creatinine level of 1835 µmol/L and urea at 44.6 mmol/L, sodium of 132 µmol/L and potassium at 4.3 mmol/L. He had features of nephrotic syndrome with severe hypoproteinamia and 24-h urinary protein of 10.45 g. Ultrasonography revealed enlarged kidneys with a reduced echogenecity of the medulla and the papillae. Subsequently, after hemodialysis with blood transfusion, a kidney biopsy was performed that showed focal segmental glomerulosclerosis associated with an acute tubular injury. On intensive interrogation, the patient gave a history of ingesting boiled Euphorbia paralias as a native treatment for edema, ten days prior to the onset of the current illness. A diagnosis of acute renal failure (ARF) resulting from the possible nephrotoxic effect of Euphorbia paralias poisoning was made. He was treated with intermittent hemodialysis and corticosteroids. Serum creatinine values improved after 48 days. At six months following the intoxication, serum creatinine of the patient was 240 µmol/L. In cases of unexplained ARF, a toxic mechanism should always be considered and acute renal failure caused by Euphorbia paralias should be included as a cause if renal toxicity is suspected in those places where it is being used as a native medicine.

Risk factors related to poor outcome after methanol poisoning and the relation between outcome and antidotes--a multicenter study.

INTRODUCTION. Thorough prognostic and metabolic studies of methanol poisonings are scarce. Our aims were to evaluate the factors associated with sequelae and death from methanol poisoning, to develop a simple risk-assessment chart to evaluate factors associated with sequelae and death from methanol poisoning, and to compare the antidotes ethanol and fomepizole. PATIENTS AND METHODS. We present a retrospective observational case series of methanol-poisoned patients from Norway (1979 and 2002-2005), Estonia (2001) and Tunisia (2003/2004), and patients from two different centers in Iran (Teheran 2004-2009 and Mashhad 2009-2010) who were identified by a positive serum methanol and had a blood acid-base status drawn on admission. The patients were divided into different groups according to their outcome: Survived, survived with sequelae, and died. RESULTS. A total of 320 patients were identified and 117 were excluded. Of the remaining 203 patients, 48 died, and 34 were discharged with neurological sequelae. A pH < 7.00 was found to be the strongest risk factor for poor outcome, along with coma (Glasgow Coma Scale (GCS) < 8) and a pCO(2) ≥ 3.1 kPa in spite of a pH < 7.00. More patients died despite hyperventilation (low pCO(2)) in the ethanol group. CONCLUSIONS. Low pH (pH < 7.00), coma (GCS < 8), and inadequate hyperventilation (pCO(2) ≥ 3.1 kPa in spite of a pH < 7.00) on admission were the strongest predictors of poor outcome after methanol poisoning. A simple flow-chart may help identify the patients associated with a poor outcome.

Compressive pleural effusion after ovarian hyperstimulation syndrome--a case report and review.

OBJECTIVE: To report a case of early onset ovarian hyperstimulation with massive pleural effusion and respiratory failure before IVF. DESIGN: Case report. SETTING: University teaching intensive care unit. PATIENT(S): A 26-year-old healthy woman with an unexplained infertility transferred to the intensive care unit on day 4 after hCG injection for early severe presentation of ovarian hyperstimulation syndrome with massive compressive pleural effusion before she underwent embryo transfer. INTERVENTION(S): Mechanical ventilation, thoracocentesis. MAIN OUTCOME MEASURE(S): Resolution of symptoms/stopping of embryos transfer. RESULT(S): Drainage of 5,300 mL of sterile exudative pleural fluid for a period of 48 hours, which permitted resolution of symptoms and allowed mechanical weaning. The IVF procedure was stopped. CONCLUSION(S): This case described is unusual in that the patient presented with early massive pleural effusion on day 4 after hCG injection and before embryo transfer. This is much earlier than in any case report elsewhere.

Epidemiology and risk factors for colonization and infection by Acinetobacter baumannii in an ICU in Tunisia, where this pathogen is endemic.

Because Acinetobacter baumannii has become an alarming endemic pathogen in our country we decided to conduct this prospective study, from January 2004 to December 2005, in order to determine risk factors and outcomes involved in clinical colonization or infection by A. baumannii in a 16-bed Tunisian intensive care unit (ICU). One hundred and two A. baumannii isolates were obtained from 63 patients, with an infection rate of 45%. The rate of multidrug-resistant (MDR) A. baumannii was 39% during the 2-year study, with an epidemic outbreak in October 2004. This outbreak was followed by closure of all the involved ICU rooms and the selective intestinal decontamination of patients, with polymyxin. During the 12-month post-intervention program (January-December 2005), the infection rate declined. The analysis of risk factors for the spread of A. baumannii showed that only the Simplified Acute Physiological Score (SAPS II) was involved. On the other hand, no risk factor was identified for multidrug resistance in patients either colonized or infected by A. baumannii. There was a statistically significant difference only in crude mortality (67.5% in MDR A. baumannii vs 46.7% in susceptible A. baumannii; P = 0.04).

Influence of activated charcoal on the pharmacokinetics and the clinical features of carbamazepine poisoning.

Carbamazepine (CBZ) poisoning has been associated with cases of severe toxicity and death. Multiple-dose activated charcoal was proposed to enhance the clearance of CBZ elimination, but there are no prospective controlled studies that demonstrated a change in clinical outcome after the use of multiple-dose activated charcoal. The aim of this study was to determine the CBZ elimination kinetics and the evolution of clinical features according to the dose of activated charcoal in acute poisoning patients. It is a prospective study for 6 months, from January to June 2004, including all pure acute CBZ-poisoned patients. Twelve patients were randomized to receive a multiple-dose activated charcoal (G1) or a simple dose of 1 g/kg (G2). Their mean age was 27.6+/-12.2 years; the Simplified Acute Physiology Score (SAPS II), 16.37+/-8.46; and the Acute Physiology and Chronic Health Evaluation (APACHE II), 8+/-3.96. They were 8 men and 4 women. The mean concentration of blood CBZ at hospital admission was of 29.42+/-6.68 mg/L. Each group includes 6 patients. The peak value of blood CBZ was comparable in the 2 groups: 33+/-3.46 mg/L (G1) vs 32.6+/-5.63 (G2) (P=.5); the requirement of mechanical ventilation was similar also (3 in each group). The duration of both coma and mechanical ventilation was significantly decreased in the first group compared with the second: 20.33+/-3.05 vs 29.33+/-4.11 hours for coma (P=.02) and 24.1+/-4.2 vs 36.4+/-3.6 hours for mechanical ventilation (P=.001). The length of stay was also significantly decreased in the first group: 30.3+/-3.4 vs 39.7+/-7.3 hours in the second group (P=.000006). Concurrently, we have noted a significant constant reduction of the half-life of CBZ from serum in the first group: 12.56+/-3.5 hours after multiple dose vs 27.88+/-7.36 hours after a simple dose (P=.0004). This decrease was correlated to the dose of charcoal. In summary, we can conclude that multiple-dose activated charcoal is more efficient than simple-dose; it permits a constant decrease of the half-life of blood CBZ without any rebound effect and could improve the prognosis by reducing the duration of coma and the length of stay.

Acute myocarditis and myocardial infarction induced by Paraphenylenediamine poisoning. Interest of angiocoronarography.

We present 2 cases of myocardial damage induced by Paraphenylenediamine (PPD) poisoning in young patients without any history of cardiovascular disease. ECG findings revealed the presence of localized ST segment elevation in precordial leads (V1-V4) in the first case and (V3-V6) in the second associated to an increase of the serum troponine T level at 23 ng/ml (case 1) and 29.7 ng/ml (case 2). The transthoracic echocardiography was in favour of diffuse myocarditis in the first case and of localized myocarditis or septo-apical myocardial infarction in the second. The angiocoronarography performed for the first time in the second case conclude on a septo-apical hypokinesia of the left ventricular resulting of a coronary spasm of the anterior interventricular artery.

Prognostic value of human erythrocyte acetyl cholinesterase in acute organophosphate poisoning.

Acute organophosphate poisoning (OPP) such as dichlorvos may be monitored by the measurement of the erythrocyte acetyl cholinesterase (EAChE) and the serum cholinesterase (SChE) activities. The aim of this study was to look at correlation between the severity of the OPP judged by certain parameters such as coma, hemodynamic disturbances, respiratory failure, and the decrease of cholinesterases enzymes including EAChE and SChE at admission. Cholinesterase activity was determined upon admission and then on days 3 and 15 in the morning. Clinical effects, EAChE, and SChE activities data were investigated in 42 patients with OPP aged of 29.6 +/- 11.8 years with acute cholinergic crisis in all cases. They were comatose in 29% of cases, presenting both hypotension or shock and hypoxemia in 17% of cases. Fifteen of them (36%) required mechanical ventilation. The mean EAChE activity at admission was 24.3 +/- 11.6 micromol/mL per hour at 37 degrees C; it was 1260 +/- 2204 IU/L for SChE. There were no correlations between the EAChE and the SChE activities. The EAChE was decreased only in comatose patients and those presenting hypotension, hypoxemia, and bradycardia with a cutoff of 23.5 micromol/mL per hour at 37 degrees C. Death was observed in 2 patients with a deep decrease of the EAChE at 5 micromol/mL per hour at 37 degrees C in 1 case and 9 micromol/mL per hour at 37 degrees C in another. The kinetics of improvement of the EAChE activity below the cutoff showed the absence of statistical improvement of the EAChE activity on day 3 (16.6 +/- 9 vs 19.5 +/- 5.7 micromol/mL per hour at 37 degrees C); this improvement was remarkable on day 15 (16.6 +/- 9 vs 27.5 +/- 6.5micromol/mL per hour at 37 degrees C, P = .0004). In summary, the marked decrease of EAChE activity appears in this study as prognostic factor in acute OPP, and coma, respiratory failure, hemodynamic disturbances, and death are associated with a decrease of the EAChE of less than 23.5 micromol/mL per hour at 37 degrees C.

Impact of ceftazidime restriction on gram-negative bacterial resistance in an intensive care unit.

The present study included three periods: (1) a 12-month pre-restriction and control period in 2001; (2) a 12-month restriction period with reduced ceftazidime prescribing in favor of piperacillin-tazobactam (2002); (3) and a 24 month post-restriction period (2003-2004). Note that, for results, P represents the difference between 2002 and 2001; P', the difference between 2003 and 2001; and P'', the difference between 2004 and 2001. No changes in hygiene practices were observed during these three periods. The purpose of this study was to assess the effect of reducing ceftazidime use in an intensive care unit (ICU) upon Gram-negative bacterial resistance, particularly as regards Pseudomonas aeruginosa. During the three periods of the study, patients were similar concerning age, Simplified Acute Physiology Score (SAPSII), the site of nosocomial infection, and the requirements for mechanical ventilation (75% in 2001, 76% in 2002, 74% in 2003, and 85% in 2004). The most commonly isolated pathogens were P. aeruginosa, Acinetobacter baumannii, and Enterobacteriaceae. The use of ceftazidime decreased significantly from 12.6% in 2001 to 9% in 2002, to 3% in 2003 (P' = 0.0009), and 2.6% in 2004 (P'' = 0.0001) in favor of piperacillin-tazobactam (0% 2001 to 3.7% in 2003; P' = 0.002; and 5% in 2004; P'' = 0.0001). Simultaneously, we observed a significant decrease in isolates of P. aeruginosa resistant to piperacillin-tazobactam (P = 0.03; P' = 0.004; P'' = 0.009), and those resistant to imipenem in 2003 (P' = 0.008). We also noted a significant decrease in A. baumannii isolates resistant to ceftazidime (P' = 0.01; P'' = 0.0004) and those resistant to imipenem in both 2002 and 2004 (P = 0.03; P'' = 0.04), and a considerable decrease in isolates of Klebsiella pneumoniae producing expanded spectrum betalactamase (ESBL) in 2003 and 2004 (P' = 0.04; P'' = 6.10(-5)). In contrast, we noted an increase in penicillinase-producing isolates of K. pneumoniae, from 6% in 2001 to 16% in 2002 (p = 0.01), 20% in 2003 (P' = 0.001), and 32% in 2004 (P'' = 10(-6)). We concluded that restriction of ceftazidime use was demonstrated to be efficient in reducing antimicrobial resistance, especially to K. pneumoniae ESBL.