Introduction: Bacterial infectious diseases in the elderly range of patients have a diagnostic problem that contribute to inadequate prescription of antibiotics (over-prescription or under-prescription) which can increase the mortality rate in this population. The C-reactive protein (CRP) and procalcitonin (PCT) have been developed to differentiate bacterial infections from other causes of inflammation and remain the most used parameters in this population. The aim of our work is to determine in this age range the added value of PCT and CRP for the diagnosis of infection requiring antibiotic therapy. Patients and methods: All patients admitted consecutively to the emergency department in the Avicenne Military Hospital in Marrakech and aged over 75 years over a period of 6 months were included (patients receiving antibiotics for more than 24 hours were excluded). On admission, demographic characteristics, comorbidities, and general signs (respiratory rate, temperature, pulse rate, confusion, falls, chills) were recorded and a biological assessment was carried out containing (PCT, CRP, leukocytes, albumin, urea and creatinine) and recorded for each patient. The PCT (measured using an immunoluminometric method) and CRP were found to be positive when they were ≥ 0.5 ng / mL and ≥3 mg / L respectively. Patients with SIRS (inflammatory systemic response syndrome) have been reported. Septic status, severe sepsis and septic shock present on admission have also been reported. The patients were classified as infected or not infected. Results: 153 patients were included in the study, of which 98 were women and 55 were men. The average age was 83.1 years (± 4.2). On interrogation 29% had end-stage chronic renal failure, 30% had diabetes, 49% had hypertension 31% had dementia, 20% were followed for tumor pathology, and 10% were followed for chronic obstructive pulmonary disease (COPD). The average of the results of the assessment was as follows: CRP = 32.9 ± 61.1 mg / L, albumin = 30.9 ± 5.2g / L, the calculated creatinine clearance was 41.1 ± 21 , 1 ml / min. PCT was positive in 15% of cases and CRP in 69.0% of patients. SIRS was present in 50 patients (32%), sepsis in 22 cases (14.4%) and severe sepsis in 14 cases (9%); but no septic shock was found on admission. Fifty nine per cent of patients were hospitalized, 61% of whom were in intensive care. The most frequently used diagnoses are: pneumonia (19.3%), heart failure (9.0%), falls with head trauma (5.5%), fractures (16.6%) and infection of the upper urinary tract (5.5%). A total of 53 cases of infections were diagnosed and distributed as follows: 24 pneumonias (45%), 9 cases of upper urinary tract infections (17%), 5 cases of miliary tuberculosis (10%), 7 cases of cellulitis ( 13%), 2 cases of endocarditis (4%) and 6 cases of deep abscess (11%). PCT had good specificity (96%) but low sensitivity (26%), with a negative predictive value and positive non-discriminatory predictive value (83% and 57%, respectively). Univariate use of logistic regression has shown that the presence of sepsis on admission was the best predictor of infection. However, PCT (> 0.5 ng / ml), the presence of SIRS, of temperature above 38 ° C or below 36 ° C and CRP (3 mg / L) have also been associated with the infection, but in the multivariate analysis, only sepsis and CRP positive were always associated with the infection Conclusion: The PCT can be useful for identifying seriously ill elderly patients but not for distinguishing between infected and non-infected patients.
