Abstract

Background: Nosocomial pneumonia (NP) is the second most common hospital acquired infection. Understanding the pattern of occurrence, risk factors and etiological agents of NP in a PICU, is essential for developing effective infection control measures. This prospective observational study was conducted in a PICU of a tertiary care teaching hospital, to determine the incidence, etiology and risk factors for NP.Materials and Methods: Patients admitted to the PICU, over a period of 1 year who had endotracheal (ET) intubation, were enrolled consecutively into the study. Demographic details were recorded at the time of inclusion. Diagnosis of NP was based on CDC criteria (1988).Semiquantiative assay of endotracheal aspirate (ETA) with a colony count of > 105 cfu/mL was taken as evidence of infection. Colonisation was defined as isolation of organism with <105 cfu/mL. Age, nutritional status, number and duration of intubations, duration of mechanical ventilation, sedation, nasogastric feeding were the risk factors studied for development of NP. Intubation attempts of more than one were defined as reintubation. Risk factors found significant on univariate analysis, were subjected to multiple regression analysis to determine the most important predictors of NP. Results: The study group comprised of 72 children with a median age of 3.7 years and boys: girls ratio of 1.9:1.Twenty two of 72 (30.5%) developed NP; the predominant isolates from ETA were Acinetobacter anitratus(12), Pseudomonas aeruginosa (5), Klebsiella sp(3) and Staphylococcus aureus and E.coli(1) each. Additionally 18(39%) had evidence of ET colonization, with Acinetobacter sp being the commonest 9(50%). Re-intubation, prolonged duration of intubation and mechanical ventilation were the significant risk factors on univariate analysis for development of NP.On multiple regression analysis, reintubation was the only independent risk factor for NP(OR 0.72, 95%CI 0.55-0.94).Overall mortality was 21%(15/72);7(47%)of these deaths were secondary to NP. Conclusions: NP developed in nearly one third of the intubated patients; Gram negative organisms were the predominant etiological agents and associated with high mortality. Re-intubation, prolonged duration of intubation and mechanical ventilation were the significant risk factors on univariate analysis for development of NP. On multiple regression analysis, reintubation was the only independent risk factor for NP

Key words: Endotracheal aspirate, Nosocomial pneumonia, Reintubation.

*p <0.05 by Student’s ‘t’ test.

Univariate analysis revealed that patients with NP had a significantly higher duration of mechanical ventilation and higher proportion and frequency of reintubation. There were, however, no statistically significant differences between the two groups with respect to age, sex, nutritional status, feeding and duration of sedation, paralysis and stress ulcer prophylaxis.

On subjecting the significant variables to multiple logistic regression, we found that reintubation was the only independent risk factor for development of NP (OR 0.72, 95% CI 0.55-0.94, p = 0.017).

The outcome of patients in terms of length of stay, survival and death are depicted in Table IV. The median duration of PICU stay was significantly longer in patients with NP as compared to those without [22 days Vs 14 days; p = 0.021].

* p<0.05 by Mann Whitney U test.

The percentage mortality in the group with NP and no NP was 31.8% and 16% respectively. However this difference was not statistically significant. All the deaths in the group with NP were secondary to Gram negative infections with Pseudomonas contributing to 4 (57.1%) deaths followed by Klebsiella, E. coli and Acinitobacter in one patient each.

The incidence of NP in our study was 30.5%, similar to the trend reported previously (7,9). Most studies on nosocomial pneumonia in intensive care units have been carried out in the developed countries, which reveals an incidence ranging from 7 to more than 40%.(10-13). Most studies on NP have used objective diagnostic criteria based on the combination of quantitative culture samples obtained with fiberoptic bronchoscopy using the protected specimen brush (PSB) or the bronchoalveolar lavage (BAL)(5,14-16). Both these methods have been shown in literature to have a specificity and sensitivity of greater than 95% in the diagnosis of NP(17,18). The quantitative bacteriologic methods have increased the reliability of sputum specimens for the diagnosis of lower respiratory tract infections when compared with conventional qualitative cultures because of more careful collection of sputum and the method of dilution which eliminates contaminating oropharyngeal secretions(13). However, Marquette, et al. (1993) in their study comparing ETA aspirate cultures with PSB found that at the cut off value of 106 cfu/mL, the former technique could be used as a reliable alternative to PSB(19). These findings compare favorably with our study where we had used the semiquantitative cultures of the ETA aspirates at a cut off point of 105 cfu/mL. Since the technique of PSB and BAL are not readily available to us in our ICU, the ET aspirate quantitative technique can be used as an alternative diagnostic tool for NP. It may be helpful in differentiating infection from colonisation and can thus reduce the overuse of antibiotics in PICU. Moreover, it is a simple, inexpensive and non-invasive technique and suited to our set up where cost becomes a major limiting factor.

We observed Gram-negative predominance as causative factor for NP in our study, a finding similar to that reported previously(7,20). Though it has been observed by some authors that the community acquired organisms such as Streptococcus pneumoniae, H. influenzae and MSSA were frequent cause of early onset NP (developing within 5 days of hospitalization) as against the resistant Enterobacter, Pseudomonas and Acinetobacter species encountered in late onset group (i.e., after 5 days of hospitalization)(21), Ibrahim, et al. in their study reported that this distinction did not hold true for an ICU setting and the organisms for both early and late NP were similar(20). Though we had not divided our NP into early and late infections, majority of them occurred within 3-10 days of PICU stay and the organisms encountered were mostly resistant Acinetobacter, Klebsiella, Pseudomonas and Enterobacter species.

The microbial flora associated with NP reflects the common organisms present in the gut, oropharynx and environment i.e., Gram-negative(22). Colonization in ICU patients has been recognized as an important source for Gram-negative infections(23). Frequent use of broad spectrum empiric antimicrobials in an ICU setting further enhances the risk of colonization with resistant organisms(22). This was shown in a study where the incidence of NP was 23% in colonised as against 3% in non-colonised patients(24). Our findings revealed that of the 18 patients having ET colonization, 14 were secondary to Gram-negative pathogens with Acinetobacter being the commonest (50%). Hence, the increased incidence of Gram-negative NP in our study population is a reflection of the pattern of colonization seen in our intubated patients.

Eleven (34.4%) of patients with NP in our study were less than one year. Though this number was not significantly different from the non-pneumonia group, it is well known that extremes of age are a high risk because of relative immaturity of the immunological system(1). Also the use of uncuffed ET enhances the risk of aspiration of upper GIT flora thus contributing to the development of NP(1).Other points borne out by the analysis of risk factors were that prolonged intubation and mechanical ventilation and reintubations were associated with a significantly high risk for NP.

Repeated intubations carried the maximum risk for development of NP on multivariate analysis, a finding similar to the one reported previously wherein 19 (47%) of the cases developed pneumonia following reintubation compared to 4 (10%) of the controls(9). The increased incidence of NP following reintubation was probably related to an enhanced risk of aspiration of the colonised oropharyngeal contents during each episode of intubation. The glottic dysfunction associated with prolonged intubation further aggravates this. Most patients in our study group were reintubated for falling SpO2, accidental extubations and rarely as an elective procedure. Our findings of increased risk of NP with reintubation imply that this procedure should not be included as a routine protocol in all mechanically ventilated patients. Also, it is safer to intubate patients (who have been extubated for some reason) as soon as possible to minimize the risk of aspiration during the interval between extubation and reintubation.

The relationship between prolonged duration of endotracheal intubation and ventilation and the development of NP has been examined by several authors(15,25-27).Our data suggests a similar trend for increasing lengths of endotracheal intubation and an interestingly high rate of acquisition of NP in the first 10 days of intubation. Though significant on univariate analysis both prolonged intubation and ventilation failed to show independent effect on multivariate analysis.

The overall mortality of NP in our study was 60% and is comparable to the mortality rates of 30-70% reported in several studies(7,14,16). A strong association between mortality rate and the infecting organism has been well documented. It has been shown that the attributable mortality in pneumonia due to Pseudomonas and Acinetobacter species exceeded 40%(10). Pseudomonas, Enterobacter, other Gram-negative bacilli and Staphylococcus aureus have been characterized as "high risk" since they are associated with high mortality(4). Our findings of organism specific mortality are in concordance with the previous study; with the highest mortality seen in Pseudomonas infection. NP not only increases incidence of mortality but also prolongs the length of ICU stay; thereby enhancing the cost of hospitalization(16). Just like organism specific mortality, increased length of ICU stay specially so with NP secondary to Pseudomonas and Acinetobacter species has also been observed(10). Our patients with NP had a significantly longer duration of ICU stay as compared to others. In a developing country like ours with limited resources this prolonged hospitalization imposes a significant economic burden.

NP in intubated patients was a significant problem encountered in our PICU. The predominant pathogens responsible for this infection were resistant Gram negatives with Acinetobacter being the commonest. Reintubation was the most significant risk factor identified for development of NP. Mortality was high; especially in patients with NP due to Pseudomonas species. The knowledge of common organisms and their antibiotic susceptibility is important for institution of appropriate antimicrobial therapy thus restricting the use of empiric broad spectrum antibiotics which pre-disposes to colonization.

Contributors: The study was conceived by PKP and MJ; PKP, MJ, SS, PR and AKS were involved in literature search, acquisition of data, analysis, interpretration of results and drafting the manuscript.

Funding: None.

Competing interests: None stated.