Predictive value of systemic immune inflammation index for infections caused by healthcare in pediatric patients hospitalized to the burn unit

Categories: Disease & Virus

April 26, 2025

In this study, a total of 26 pediatric patients hospitalized in the burn intensive care unit (ICU) of our tertiary care hospital were evaluated. The data of the uninfected and infected patient’s demographic and clinical data were presented in Table 1. The median age of infected patients was significantly higher than non-infected patients (p^*<0,05). The majority patients were boy, 73.9% were boy and 26.1% were female. The gender distribution of the infected and uninfected was similar (p^†>0,05). The proportion of uninfected patients receiving medical treatment and dressings was significantly higher than the proportion of infected patients receiving the same treatment (p^†<0,05). On the other hand, only the infected patients were grafting. When the type of burn was analyzed, scalds were the most common cause in both groups with 91.3% and 87%, respectively. Flame Burns ranked second. But, there was no statistically significant difference between the groups in terms of burn etiology (p^†<0,05). The proportion of infected patients with TBSA above 30% was statistically higher than the proportion of uninfected patients, while the proportion of uninfected patients with TBSA below 30% was statistically higher (p^†<0,05). There were statistically significant relationships between infection status and CVC use (p^†=0,007). The CVC usage proportion were higher in infected patients compared to uninfected patients. During the study period, the rate of ventilator use was recorded as 13% for only infected patients, but no case of ventilator-associated pneumonia was encountered. The mean body surface area of burn and median length of hospital and intensive care unit stays of infected patients were statistically significantly higher than uninfected patients (p^*<0,05).

Table 1 Comparison of demographic characteristics and clinical findings of noninfected and infected patients.

The distribution of infected patients according to the breeding site is given in detail in Table 2. There was a statistically significant relatinship between the burn degree and breeding site (p^†<0.05). In patients with second-degree burns, the proportions of growth in the blood (77,8%) was higher than the proportions of growth in the wound (12,5%). However, there was no statistically significant difference between the growth proportions in the catheter (50%) and the growth proportions in the wound and blood. In patients with third-degree burns, the growth proportions in the wound (87,5%) was higher than the growth proportions in the blood (22,2%). There was no statistically significant difference between the growth proportions in the catheter (50%) and the growth proportions in the wound and blood. (The proportions of burn degree were calculated from columns.) Distribution of type of burn, TBSA %, and the place of growth were homogeneous in the breeding site (p*>0,05).

Table 2 Distribution of infected patients according to site of reproduction.

The distribution of blood, wound, and central venous catheter culture samples obtained from our patients was homogenous, and, 39.1% of the blood, 34.8% of the wound, and 26.1% of the central venous catheter cultures were grown (p^*>0,05) in Table 3. The most frequently isolated microorganism was Pseudomonas aeruginosa (p^*<0,05). The microorganisms isolated according to the type of infection are presented in detail in Table 3.

Table 3 Microorganisms isolated according to the site of infection.

Table 4 shows the distribution of blood parameters and SII for patient groups. The results of infected patients represent samples taken in the middle of the patient’s ICU stay and when they did not show signs of infection. The Uninfected group includes the laboratory results of the day when samples were taken for microbiologic examination with suspicion of infection. Only hemoglobin mean, lymphocyte, CRP, and procalcitonin medians showed statistically significant differences between the infected and uninfected groups (p^*<0,05), while the other blood parameters, and SII did not show statistically significant differences in these groups (p*>0,05).

Table 4 Comparison of blood parameters of infected patients according to groups.

There was no statistically significant difference in the mean and/or median of blood parameters between burn degrees (p^*>0.05) (Table 5).

Table 5 Comparison of blood parameters of infected patients according to burn degree.

According to Table 1, CVC usage, body surface area and length of hospitalization are statistically significant risk factors for infection. CVC usage increases the risk of infection by 8.07 times (OR = 8,077; 95% CI 1,523 to 42,834). Each 1-unit increase in body surface area increases the risk of infection by 1,166 times (OR = 1,166; 95% CI 1,071 to 1,271) and each 1-day increase in length of hospitalization by 1.37 times (OR = 1,37; 95% CI 1,108 to 1,710). Each 1-unit increase in CRP increases the risk of infection by 1,492 times (OR = 1,492; 95% CI 1,128 to 1,972). Increases in Hb and lymphocytes decrease the risk of infection (respectively OR = 0,656, and 0,763; 95% CI 0,479 to 0,897 and 0,593 to 0,982). Factors that may affect the risk of infection are presented in detail in Table 6. According to Table 4, the SII is not a significant risk factor for infection. However, we evaluated and analyzed it as a clinical risk factor, and the result was not statistically significant (OR = 2,057;95% CI 0,489 to 8,657) (Table 6).

Table 6 Factors that May affect the risk of infection according to univarite logistic regression analysis.

Table 7 shows the diagnostic performances of blood parameters and SII for infecion. According to ROC analysis result, the SII was not successful in diagnosing infection. AUC value was 0,605 (p = 0,236) for the SII (Table 7; Fig. 1).

In Fig. 1, especially CRP was observed to be very successful in distinguishing between infected and uninfected patients (AUC = 0,877; p < 0,001). Those with a CRP value > 0,31 can be considered infected. The sensitivity value for CRP was quite high, 91,30% and the specificity was 73,91%. After CRP, the other blood parameter was Hb (AUC = 0,725; p = 0,009). The specificity of Hb was very high (Sp = 95.65%). Those with a Hb value ≤ 10,4 can be considered infected. Although the diagnostic performance of lymphocytes was not as good as the others, but the sensitivity value was high, 86.96% (AUC = 0.679; p = 0,02). Furthermore, when comparing the AUC values of these blood parameters in Table 7, we have observed that CRP was more successful than the SII in the diagnosis of infection (p*=0,001). But it was not a significant difference between the SII and the Hb and lymphocyte (respectively p^† =0,397; p^‡ =0,510) diagnostic performances’. Also, the difference between the AUC values of CRP and Hb is statistically significant (p^§=0,039). This indicates that the discrimination of CRP is more successful than Hb. When CRP is compared with lymphocyte, it can be said that the discrimination of CRP is more successful than lymphocyte (p^¶=0,033). When Hb is compared with lymphocyte, the difference between AUC values is not statistically significant (p^‖=0,663). Their diagnostic performances is similar.