A total of 1033 patients were included in our analysis based on the eligibility criteria. Among them, 195 (18.8%) of those patients were receiving aspirin. After propensity score matching (1:1 ratio), 352 patients were included based on predefined criteria. Most of the patients who received aspirin (69.2%) were using it for chronic conditions (e.g., stroke, IHD). Only 21 patients (12.2%) required discontinuation of aspirin during ICU stay secondary to thrombocytopenia (7 patients), bleeding (3 patients) and other reasons (11 patients).
Demographic and Clinical Characteristics
The majority of the patients included in both arms (69.1 %) were male, with a mean age of 61.7 (SD +14.77). Diabetes mellitus (59.2 %), hypertension (55.4%), and dyslipidemia (19.5%) were the most prevalent underlying comorbidities in our patients. There was a notable difference before propensity score matching, patients who did not take aspirin throughout their ICU stay were younger, had higher eGFR, and higher ferritin levels at baseline. On the other hand, patients who received aspirin had considerably higher SOFA scores and blood glucose levels at baseline. Most of these differences were comparable between the two groups after using propensity score matching (Table 1, Additional file 1).
30-day and in-hospital mortality
In crude analysis, patients who received aspirin showed lower in-hospital mortality (55.7 % vs. 61.9 %, p = 0.24) and 30-days mortality (54.0 % vs. 55.4 %, p = 0.78) as compared with patient who didn’t received aspirin, respectively. However, the difference was not statistically significant.
The multivariable Cox proportional hazards regression analyses showed a statistically significantly lower in-hospital mortality in patients who received aspirin (HR 0.73; 95% CI 0.56-0.97; P =0.03), but not the 30-day mortality (HR 0.86; 95% CI 0.65-1.14; P =0.30) (Table 2). Additionally, the overall survival probabilities were higher during hospital stay among patients who received aspirin after propensity score matched as presented in the survival curve (Figure 2).
Ventilator free days and Length of stay
Ventilator free days (VFD) during ICU stay were 9.5 days (±12.3) for patients who received aspirin compared to 8.8 days (±12.1) in the control group. However, did not reach to a statistical difference between the two groups with a beta coefficient (95%CI): 0.11 (-0.47, 0.69), p-value 0.71. Moreover, both ICU and hospital length of stay were not statistically significant between the two groups with a beta coefficient (95%CI): -0.06 (-0.30, 0.18), p-value 0.63, and beta coefficient (95%CI): 0.09 (-0.17, 0.35), p-value 0.51; respectively (Table 2).
ICU Complications during ICU stay
In crude analysis, the prevalence of major bleeding was non-significantly higher in the patients who received aspirin than patients who did not (6.3 % vs. 2.8 %; p-value= 0.12). Results from the multivariable logistic regression analysis; demonstrated a higher odd of major bleeding by 3-fold; however, it failed to reach the statistical significance (OR (95%CI): 2.92 (0.91, 9.36), p-value=0.07). Moreover, patients who received aspirin were shown to have a non-significantly higher odds of minor bleeding (OR (95%CI): 1.70 (0.68, 4.20), p-value=0.26). Furthermore, RBCs transfusion requirement was similar between the two groups (Table 3).
Interestingly, venous thromboembolism was 3-fold higher in patients who received aspirin (OR (95%CI): 3.14 (0.83, 11.81), p-value=0.09); but this difference did not reach a statistical significance. In terms of other secondary outcomes such as respiratory failure required MV (OR:1.14, p-value=0.75), acute kidney injury (OR:0.79, p-value=0.27) and liver injury (OR:1.20, p-value=0.59) did not differ substantially between the two groups (Table 3).
Follow-up inflammatory markers during ICU stay and RBC transfusion
Most of the follow-up inflammatory markers during stay (e.g. Ferritin, CRP, and CPK) were the same between the two groups, except that D-dimer was significantly lower and fibrinogen level was higher in patients who received aspirin in comparison to those who did not with a beta coefficient (95%CI): -2.84 (-3.32, -2.37), p-value <0.0001, and beta coefficient (95%CI): 0.62 (0.25, 0.99), p-value <0.001; respectively (Table 3).