Ventilator-associated pneumonia (VAP) is an important problem in Pediatric intensive care units (PICU). The prevalence ranges from 3%-65% in PICUs in the USA [1]. The clinical diagnosis of VAP is usually based on the presence of fever, leukocyte counts, amount and character of tracheal secretions, and appearance of new or persistence of radio-graphic infiltrates. However, these parameters taken separately have limited diagnostic value [2]. The clinical criteria for the diagnosis of VAP have repeatedly been criticized as inappropriate, leading to over diagnosis or under diagnosis particularly in the setting of Adult respiratory distress syndrome (ARDS) [3-5]. Despite these limitations, the clinical criteria remain the starting point of diagnostic evaluation of suspected VAP case. Pugin, et al. [6] combined the above mentioned parameters with oxygenation (PaO₂/FiO₂) and formed the Clinical Pulmonary Infection Score (CPIS) as a diagnostic tool for pneumonia. The CPIS has been used in multiple studies on VAP in adults [3,7-9] but limited data is available on pediatric cases [10]. This prospective study was conducted to validate the use of CPIS to diagnose VAP in pediatric patients.

The Institutional review board approval was obtained for the study and informed consent for the bronchoscopic BAL was taken from the parents. In accord with previous studies of quantitative bacteriology of BAL cultures [1,14-17], a bacterial density of ≥10⁴cfu /mL was considered as positive for VAP and those episodes were referred to as ‘Definite VAP’ episodes. The organisms isolated on blood culture were compared with the organisms isolated from the BAL. The five parameters of CPIS were compared between "Definite" and "No Definite" VAP patients using independent sample t test and Chi-square test. Taking BAL colony counts of ≥10⁴ cfu/mL as the reference standard, the sensitivity, specificity, positive and negative predictive values and accuracy were calculated at various CPIS levels. The receiver operating characteristic curve (ROC) was plotted and area under curve was also obtained. The distribution pattern of the data was tested with D’Agostino- Pearson test for normal distribution. P value less than 0.05 was considered significant. SPSS 15 version statistical package was used for analysis.

Microbiological results: Eighteen blood cultures were positive. The most common organism cultured was Pseudomonas aeruginosa (43.6%). Other organisms obtained were Acinetobacter baumannii, Enterobacter sps, methicillin resistant Staphy-lococcus aures (MRSA) and Candida albicans. Eleven (60%) blood culture results were concordant with the organism yielded from BAL in colony counts of ≥10⁴ cfu/mL. Twenty one out of 30 BAL samples yielded positive cultures. However, only 19 samples yielded organisms with a colony count of ≥10⁴ cfu/ mL. Micro-organisms cultured included P.aeruginosa (9), A. baumannii (5), K. pneumoniae (4), Methicillin resistant S. aureus (2), P. mirabilis (1), and Enterobacter species and Candida sepsis (3 each). In six samples, polymicrobial growths were obtained. In three samples, Candida sps grew along with Pseudomonas aeruginosa, Klebsiella pneumo-naie and Acinetobacter baumannii.

Table II Comparison of Individual Variables of Simplified CPIS in Patients With Definite 

and No Definite Ventilator-associated Pneumonia. 

Simplified clinical pulmonary infection scores: The individual variables of CPIS were compared between patients with definite and no definite VAP (Table II). None of the patient had normal chest X-ray at the time of enrolment. The higher temperature of patient (39.2 ± 0.7 vs 38.5 ± 0.8, P = 0.01) and low PaO₂/FiO₂ ratio (155.5 ± 41.6 vs 234 ± 69, P < 0.001) were significant variables for the diagnosis of VAP. A value of 8 on the CPIS was found to have the best accuracy with a sensitivity and specificity of 80% each (Table III). Area under the ROC curve was 0.812 (P < 0.001) (Fig. 1).

TABLE III  Operative Indices of Simplified Clinical Pulmonary Infection Score for the Definite 

Diagnosis of Ventilator-associated Pneumonia 

Fig. 1 Receiver operating characteristic curve of clinical pulmonary infection score.

In this study, we have evaluated the clinical diagnosis of VAP assessed by simplified CPIS using bronchoscopic BAL culture as the reference standard. In the present study, individual variables used in CPIS were also compared in patients with definite VAP and with no definite VAP. Presence of fever and PaO₂/FiO₂ ratios were significantly different in two groups.

Pugin, et al. [6] found significant difference of temperature in adult patients with or without pulmonary infection, while other studies found presence of fever as a poor diagnostic marker, especially in ARDS patients [3,4,7]. The presence of fever may be explained by mechanisms other than pneumonia in patients on ventilator. Meduri, et al. [18] found that only 42% of patients with fever and pulmonary infiltrates had pneumonia. Leucocytosis and purulent tracheal secretions had good sensitivity (77% and 69%, respectively) but poor specificity (58% and 42%, respectively) in a study by Fabergus, et al. [3]. It is quite obvious that leucocytosis may be induced by a variety of causes in a critically ill patient [3]. The high false positivity of purulent tracheal secretions is explained by the presence of purulent bronchitis in almost all patients particularly during prolonged MV. The negative results may be due to the peripheral location of pneumonia and impaired clearance of secretions [19] .

In the present study, chest X-ray was positive in all the patients at the time of enrolment. Similar observation was reported by Andrews, et al. [4] but only 57% had histological evidence of pneumonia. While Fabergus, et al. [3] found radiographic infiltrates as most sensitive (92%) clinical parameter for diagnosing VAP, the rate of false positive results was as high as 67%. This high false positivity may result from other causes mimicking VAP such as atelectasis, pulmonary infarction, alveolar hemorrhage and ARDS.

As per Pugin, et al. [6], a CPIS >6 was associated with high likelihood of pneumonia (sensitivity 93%, specificity 100%). The disadvantage with this score is dependence on tracheal aspirate gram stain and culture results and so the waiting period of 24 to 48 hours for the clinical diagnosis of VAP [9,20]. In the present study, simplified CPIS was used [9]. This score does not include any laboratory dependent variables. The maximum possible score is 10. Based on the previous studies patients with CPIS ≥6 were enrolled in the present study [3, 6-10, 21]. Pugin, et al. [6] did not find any positive bronchoscopic BAL in patients with CPIS<6. The likelihood ratio of detecting pneumonia was 1.46 and 0.68 in patients with CPIS ≥6 and <6, respectively [2]. In another study, CPIS >6 virtually ruled out other causes of pulmonary infiltrates in ICU patients [22].

The mean CPIS was significantly high in patients with definite VAP (P=0.007). CPIS has been studied in many adult studies as a diagnostic tool [3,6]. Results of these studies are conflicting. High sensitivity (93%) and high specificity (100%) of CPIS >6 were reported [6]. Papazian, et al. [7] used CPIS for selection of cases to test the diagnostic accuracy of bronchoscopic and nonbronchoscopic techniques. CPIS at the threshold of 6 achieved an accuracy of 79%, a sensitivity of 72% and a specificity of 85%.The individual variables of CPIS were not significantly different while the composite CPIS score was significantly higher in the VAP group than in the non-VAP group. Similar results were not reported in other adult studies [2,3]. The only pediatric study evaluating CPIS had enrolled 15 mechanically ventilated children. In 14 patients CPIS was more than 6 at the time of diagnosis of VAP with positive predictive value of 93% [10].

Serial CPIS and its components have been used previously to assess the treatment response in prospective studies. Progressively falling CPIS and rising PaO₂/FiO₂ ratio distinguished survivor from non-survivors [9]. CPIS had also identified patients requiring antibiotic therapy and so reduced the cost of therapy [22].