[Clinical features and outcomes of cancer-related versus non-cancer-related sepsis in pediatric intensive care unit]

Objective: To compare the clinical features and outcomes of cancer-related and non-cancer-related sepsis in children who were admitted pediatric intensive care unit (PICU). Methods: The clinical history of patients with sepsis, who were admitted to PICU in Shanghai Children's Hospital, Shanghai Jiao Tong University from August 2016 to July 2019, were retrospectively reviewed. A total of 768 patients were divided into the cancer-related sepsis group (135 cases) and the non-cancer-related sepsis group (633 cases). The patients in the cancer-related group were further categorized into three subgroups including hematological malignancy (80 cases), solid tumor (43 cases) and hemophagocytic lymphohistiocytosis (HLH) (12 cases). The variables of clinical features, laboratory tests, pathogens, management strategies and in-hospital mortality were compared between the two groups by student t test, Mann-Whitney U test or Chi-square test. Results: The patients with cancer-related sepsis accounted for 17.6% of all patients (135/768). Regarding the site of initial infection, the incidence of gastrointestinal infection (43.0% (58/135) vs. 28.6% (181/633), χ(2)=10.718, P=0.001), blood stream infection (29.6% (40/135) vs. 17.1% (108/633), χ(2)=11.297, P=0.001) and skin and soft tissue infection (22.2% (30/135) vs. 4.1% (26/633), χ(2)=54.013, P<0.01) were higher in the patients with cancer-related sepsis than in those with non-cancer-related sepsis. On first PICU admission, the levels of hemoglobin (71 (61, 83) vs. 106 (92, 116) g/L, Z=13.594, P<0.01), white blood cell (1.4 (0.3, 5.2) vs. 9.8 (5.8, 15.1)×10(9)/L, Z=11.213, P<0.01), platelet count (51 (15, 121) vs. 286 (192, 384)×10(9)/L, Z=13.336, P<0.01), CD19(+)cells (0.106 (0.008, 0.274) vs. 0.325 (0.224, 0.454), Z=6.555, P<0.01), and neutrophil (0.449 (0.170, 0.730) vs. 0.683 (0.537, 0.800), Z=5.974, P<0.01) were significantly lower in patients with cancer-related sepsis; however, the levels of C-reactive protein (82 (25, 155) vs. 36 (11, 86) mg/L, Z=-5.257, P<0.01), procalcitonin (1.5 (0.3, 12.0) vs. 0.8 (0.2, 4.0) μg/L, Z=-2.767, P=0.006), CD8(+)cells (0.329 (0.253, 0.514) vs. 0.209 (0.156, 0.275), Z=-5.699, P<0.01), interleukin (IL) -6 (0.1 (0.1, 522.4) vs. 0.1 (0.1, 0.1) ng/L, Z=-2.747, P=0.006), IL-8 (0.1 (0.1, 177.0) vs. 0.1 (0.1, 4.5) ng/L, Z=-2.087, P=0.037), and IL-10 (0.1 (0.1, 42.7) vs. 0.1 (0.1, 6.6) ng/L, Z=-2.148, P=0.032) were significantly higher in patients with cancer-related sepsis. Similarly, the rate of continuous renal replacement therapy (CRRT) (34.8% (47/135) vs. 16.9% (107/633), χ(2)=26.267, P<0.01) and the use of intravenous immunoglobulin (IVIG) (83.0% (112/135) vs. 66.2% (419/633), χ(2)=14.667, P<0.01) were significantly higher in cancer-related sepsis group. Moreover, the incidence of co-infection with fungi was also higher in cancer-related sepsis group (14.1% (19/135) vs. 0.5%(3/633), χ(2)=73.965, P<0.01), and so was the number of multiple organ dysfunction (3 (2, 5) vs. 2 (1, 3), Z=-6.988, P<0.01). Finally, the in-hospital mortality rate of cancer-related sepsis and non-cancer-related sepsis were 36.3% (49/135) and 9.3% (59/633), respectively, also significantly different (χ(2)=67.000, P<0.01). There was no difference in the in-hospital mortality among children with hematologic tumors, solid tumors and HLH (35.0% (28/80) vs. 32.6% (14/43) vs. 7/12, χ(2)=2.838, P=0.242). Conclusions: The site of initial infection, inflammatory markers on PICU admission, and co-infection pathogen during hospitalization are different between patients with cancer-related sepsis and non-cancer-related sepsis. Besides, the in-hospital mortality of cancer-related sepsis is about 4-fold that of non-cancer-related sepsis. The monitoring of clinical features and organ dysfunction, and timely treatment are crucial for cancer-related sepsis.